Capsular Polysaccharide of Erysipelothrix rhusiopathiae, the Causative Agent of Swine Erysipelas, and Its Modification with Phosphorylcholine

Abstract

The capsule has been implicated in the virulence of the swine pathogen Erysipelothrix rhusiopathiae, a rod-shaped, intracellular Gram-positive bacterium that has a unique phylogenetic position in the phylum Firmicutes and is a close relative of Mollicutes (mycoplasma species). In this study, we analyzed the genetic locus and composition of the capsular polysaccharide (CPS) of the Fujisawa strain of E. rhusiopathiae. Genome analysis of the Fujisawa strain revealed that the genetic locus for capsular polysaccharide synthesis (cps) is located next to an lic operon, which is involved in the incorporation and expression of phosphorylcholine (PCho). Reverse transcription-PCR analysis showed that cps and lic are transcribed as a single mRNA, indicating that the loci form an operon. Using the cell surface antigen-specific monoclonal antibody (MAb) ER21 as a probe, the capsular materials were isolated from the Fujisawa strain by hot water extraction and treatment with DNase, RNase, pronase, and N-acetylmuramidase SG, followed by anion-exchange and gel filtration chromatography. The materials were then analyzed by high-performance liquid chromatography, mass spectrometry, and nuclear magnetic resonance (NMR) spectroscopy. The CPS of E. rhusiopathiae is heterogeneous and consists of the major monosaccharides galacturonic acid, galactose, mannose, glucose, arabinose, xylose, and N-acetylglucosamine and some minor monosaccharides containing ribose, rhamnose, and N-acetylgalactosamine. In addition, the capsule is modified by PCho, which comigrates with the capsular materials, as determined by Western immunoblotting, and colocalizes on the cell surface, as determined by immunogold electron microscopy. Virulence testing of PCho-defective mutants in mice demonstrated that PCho is critical for the virulence of this organism.

INTRODUCTION

Capsules are important virulence determinants on the cell surfaces of many bacterial pathogens, and they play a critical role in protection against innate host defenses during infection. Some pathogens have evolved several mechanisms to evade host immune responses. In mucosal pathogens, including Haemophilus influenzae (33), Pseudomonas aeruginosa (32), Neisseria species (5), and some serotypes of Streptococcus pneumoniae (12, 31), phosphorylcholine (PCho) is a surface component of lipopolysaccharide (LPS) or capsular polysaccharides (CPSs), and this surface modification plays an important role in the pathogenesis of infection.

Erysipelothrix rhusiopathiae is a Gram-positive bacterium that represents a new class, Erysipelotrichia, in the phylum Firmicutes and is a facultative intracellular pathogen that causes a variety of diseases in many species of birds and mammals, including humans. In pigs, E. rhusiopathiae can cause swine erysipelas, which may occur as acute septicemia or chronic endocarditis and polyarthritis (34). E. rhusiopathiae expresses a capsule (29), which plays an important role in the resistance of the organism to phagocytosis by polymorphonuclear leukocytes and intracellular killing by macrophages (24, 28). Whereas the importance of the capsule in virulence has long been recognized, the biochemical features and mechanisms by which the capsule prevents the recognition of E. rhusiopathiae by phagocytes or the immune system remain poorly understood.

Recently, we sequenced the genome of E. rhusiopathiae and reported that E. rhusiopathiae is phylogenetically related to Mollicutes (Mycoplasma species) (17). Based on genomic phylogenetic trees, we proposed that E. rhusiopathiae and other Erysipelotrichia strains may be separate from Firmicutes and classified as a distinct phylum (17). Furthermore, our genome analyses revealed that E. rhusiopathiae lacks many orthologous genes for the biosynthesis of wall teichoic acids (WTA) and lipoteichoic acids (LTA). In addition, E. rhusiopathiae lacks the dltABCD operon, which is responsible for d-alanine incorporation into WTA and LTA, suggesting that the organism may have an atypical cell wall (17). Thus, the phylogenetic position of E. rhusiopathiae is unique among Gram-positive pathogens, highlighting the need for detailed studies of E. rhusiopathiae cell surface antigens and cell wall structure. In the present study, we genetically and immunochemically analyzed the capsule of E. rhusiopathiae.

MATERIALS AND METHODS

Strain and growth conditions.

The wild-type E. rhusiopathiae strain Fujisawa (serotype 1a), which was originally isolated from a septicemic pig (29), a capsule-defective nonreverting mutant, YS-1, which was previously defined as an acapsular mutant of Fujisawa (26), and two PCho-defective insertional mutants of Fujisawa, clones 100 and 112, were cultured in brain heart infusion (BHI; Difco Laboratories, Detroit, MI) medium supplemented with 0.1% Tween 80 (pH 8.0) (BHI-T80) at 37°C.

MAbs.

The monoclonal antibodies (MAbs) ER21 and TEPC-15 (Sigma-Aldrich) were used to detect capsular material(s) and PCho, respectively. MAb ER21, which was prepared from a hybridoma generated from mice that had been immunized with irradiated E. rhusiopathiae bacteria, specifically agglutinates intact E. rhusiopathiae strains and recognizes the surface antigen(s) (25).

RT-PCR.

For reverse transcription-PCR (RT-PCR), total RNA from the Fujisawa strain was prepared with the RNeasy minikit (Qiagen, Valencia, CA) based on the protocol supplied. The RNA was further treated with RNase-free cloned DNase I (TaKaRa, Shiga, Japan) to remove DNA contaminants and quantified with a spectrophotometer (DU series 500; Beckman, Fullerton, CA). One microgram of total RNA in 20 μl reaction solution was converted to cDNA with Verso reverse transcriptase (Thermo Scientific Verso cDNA synthesis kit; ABgene, Surrey, United Kingdom) at 42°C for 55 min followed by 95°C for 2 min. cDNA was amplified with the primers shown in Table 1, and the amplicons were separated on a 1% agarose gel. The corresponding controls lacking reverse transcriptase or template RNA were included for each reaction to confirm the absence of genomic DNA.

Table 1.

Primers used in RT–PCR

Primer	Sequence	Product size (bp)
1F	5′-CAA CCT TGC GAA CTG AAA CA-3′	2,594
1R	5′-GGA GTT GGC TTC GTG CTT AG-3′
2F	5′-AGC ATG ATG ACC GAA TCA CA-3′	3,230
2R	5′-TTG CTT GAT ACC ATG CCT CA-3′
3F	5′-AGG GAT GCA TCG TCA CAG TA-3′	3,413
3R	5′-GGC GTA TCA GGA TCA AAT GG-3′
4F	5′-CTT GCG TAT GAC CGC AAT AA-3′	3,489
4R	5′-CTC TGC ACT GTC CGT TCG TA-3′
5F	5′-TGA TCT CGC AAA CGG ACT AA-3′	1,935
5R	5′-AGC CAG CAG TTC TCC CAG TA-3′
6F	5′-CGG AGG GTT TTA TGT TTT CG-3′	3,488
6R	5′-CGT GGA TTT TCG GAA GAA AG-3′
7F	5′-GG CAT CCA CGA TGT TCT TA-3′	2,702
7R	5′-GTG CAA GTG CTT TGT TTC CA-3′
8F	5′-ACG CTA CAC ACG CTC TAG CA-3′	2,807
8R	5′-AAC GGA AGG ATC AAC TGT GG-3′

Extraction and purification of CPS.

Bacteria were grown at 37°C in BHI-T80 and harvested at the end of logarithmic growth by centrifugation. The capsular polysaccharide (CPS) was then extracted with 65°C hot water for 20 min with constant stirring. The extracts were subjected to treatment with DNase (116 U/ml; Invitrogen, Carlsbad, CA), RNase (500 μg/ml; Invitrogen), pronase (1 mg/ml; Roche, Mannheim, Germany), and N-acetylmuramidase SG (66 U/ml; Seikagaku Biobusiness, Tokyo, Japan), followed by trichloroacetic acid (TCA) precipitation. The supernatant was neutralized with Trizma base, concentrated by ultrafiltration, and further subjected to methanol and chloroform extraction (1:1:0.5 methanol-chloroform-H₂O) (2) to remove possible lipophilic components of the cytoplasmic membrane. The crude CPS was exhaustively dialyzed against H₂O (cellulose tube; molecular weight cutoff [MWCO], 14,000). The CPS was further dialyzed against an equilibration buffer of 10 mM Tris-HCl (pH 7.5) containing 50 mM NaCl. The concentrated crude CPS was applied to a 2.5- by 12-cm DEAE-cellulose anion-exchange chromatography column (DE-52; Whatman), and fractions were eluted with a linear gradient of 50 mM to 500 mM NaCl in Tris-HCl (pH 7.5). Five-milliliter fractions were collected, and the CPS-containing fractions were pooled and further fractionated by gel filtration chromatography (GFC) on a 60- by 16-cm Sephacryl S-200 column (GE Healthcare Bio-Sciences AB, Uppsala, Sweden). The column was equilibrated and eluted with 150 mM NaCl in 10 mM Tris-HCl (pH 7.5). Each fraction was assayed with phenol sulfuric acid to determine the total amount of carbohydrates (10) and by dot immunoblotting to detect CPS with reactivity against MAb ER21. The approximate molecular weight of the CPS was estimated by comparing the elution volume of the CPS in GFC with the void volume, which was determined with dextran blue 2000 (GE Healthcare, Uppsala, Sweden).

Peroxidate oxidation of CPS.

CPS was oxidized with metaperiodate as previously described (4). Briefly, the CPS (400 ng) from each fraction eluted from GFC was dried in vacuo and resuspended in 10 μl 50 mM sodium acetate (pH 4.5) containing 20 mM metaperiodate. The CPS was then incubated in the dark with gentle shaking at room temperature for 1 h. The CPS was dissolved in 10 μl 50 mM sodium acetate (pH 4.5) without metaperiodate as a control.

Phospholipase C treatment of CPS.

The CPS was treated with phospholipase C, as described previously (16), with the following modifications. CPS (168 ng) dried in vacuo was dissolved in 0.5% deoxycholic acid (pH 8.0), followed by the addition of phospholipase C (1 U; Sigma-Aldrich, St. Louis, MO) dissolved in 10 mM 3,3-dimethylglutaric acid-NaOH with 0.1% bovine serum albumin (BSA). The solution was incubated at 37°C for 30 min, and the reaction was stopped with the addition of 1 M Tris-HCl containing 0.2% SDS. The corresponding control reactions were performed in the absence of CPS or phospholipase C. The reaction mixture was separated by SDS-PAGE and subjected to Western immunoblotting.

Dot and Western immunoblotting.

For dot immunoblotting, the CPS-containing solution or bacterial culture was spotted onto a nylon membrane (MagnaGraph; Funakoshi, Tokyo, Japan). For Western immunoblotting, the CPS-containing solution was separated by 15% SDS-PAGE and transferred to a polyvinylidene difluoride (PVDF) membrane (Immobilon-P; Millipore, Bedford, MA). The air-dried nylon membrane or PVDF membrane was blocked with 1% skim milk in phosphate-buffered saline (PBS) containing 0.05% Tween 20 and incubated with MAb ER21 or TEPC-15. The membranes were further treated with horseradish peroxidase conjugated with goat anti-mouse immunoglobulin antibody (IgG, IgM, and IgA; H+L) (Zymed Laboratories, San Francisco, CA). The blots were developed with 3,3-diaminobenzidine tetrahydrochloride–hydrogen peroxide (Wako, Tokyo, Japan).

Analytical procedures. (i) HPLC.

The monosaccharide components of the CPS from Fujisawa were released and labeled with p-aminobenzoic acid ethyl ester (ABEE) according to the manufacturer's instructions for the ABEE labeling kit Plus S (J-Oil Mills Co., Ltd., Tokyo, Japan). Briefly, the dried CPS (5 μg) was successively subjected to neuraminidase treatment, acid hydrolysis with 4 M trifluoroacetic acid (TFA), and N-acetylation. After the solvent was evaporated in vacuo, 10 μl distilled H₂O and 40 μl ABEE solution were added into the tube, and the mixture was incubated at 80°C for 1 h for labeling. The solution was cooled to room temperature, and 200 μl distilled H₂O and 200 μl chloroform were added and mixed vigorously. The aqueous phase was collected and used for high-performance liquid chromatography (HPLC) analysis. The ABEE-labeled monosaccharides were analyzed on a Honenpak C₁₈ column (75 by 4.6 mm) (Seikagaku Biobusiness) on a Prominence HPLC instrument (Shimadzu Corp., Kyoto, Japan) according to the method previously described (35). A 20-μl aliquot of ABEE-labeled monosaccharide was injected into the column and separated with 7% acetonitrile in 0.2 M potassium borate buffer (pH 8.9) over 50 min; the column was then washed with 50% acetonitrile in 0.02% TFA for 5 min at a flow rate of 1.0 ml/min at 30°C. The elution profile was monitored by fluorescence (excitation of 305 nm and emission of 360 nm) with an RF-10A detector (Shimadzu Corp.). The monosaccharide constituents of CPS were identified based on the retention time of each monosaccharide by comparison with the 11 standard monosaccharides supplied in the ABEE labeling kit, as well as d-galacturonic acid monohydrate (Fluka) and d-glucuronic acid (Sigma-Aldrich).

(ii) MALDI-TOF MS.

A 0.5-μl aliquot of native (45 μg/ml) or partially hydrolyzed CPS (2% acetic acid, 100°C, 2.5 h) was mixed with 0.5 μl α-cyano-4-hydroxycinnamic acid (CHCA) (Bruker Daltonics, Billerica, MA) saturated in a mixture of acetonitrile and 0.1% TFA (1:1). The solution was spotted onto a stainless steel matrix-assisted laser desorption ionization (MALDI) mass spectrometry (MS) target and air dried. Mass spectra were recorded on a 4800 plus MALDI-tandem time of flight (TOF/TOF) analyzer (AB SCIEX, Foster City, CA) in both the positive- and negative-ion modes. The mass spectrometer was tuned and calibrated with standard peptides or proteins (Bruker Daltonics) prior to measurements.

(iii) NMR spectroscopy.

Nuclear magnetic resonance (NMR) spectroscopy was performed as described previously (23). The lyophilized polysaccharide (200 μg) was dissolved in 0.25 ml deuterium oxide (99.96% D) and transferred into a 5-mm NMR tube (Shigemi, Tokyo, Japan). One-dimensional ¹H and distortionless enhancement by polarization transfer at 135° (DEPT-135) ¹³C, as well as two-dimensional double quantum filtered-correlated spectroscopy (DQF-COSY), triple quantum filtered (TQF)-COSY, total correlation spectroscopy (TOCSY), nuclear Overhauser effect spectroscopy (NOESY), ¹H-¹³C heteronuclear single quantum coherence (HSQC), ¹H-¹³C heteronuclear multiple bond correlation (HMBC), ¹H-¹⁵N HMBC, and ¹H-¹³C HSQC-TOCSY were measured on a Bruker Avance 800 spectrometer (Bruker Biospin, Rheinstetten, Germany) equipped with a TCI cryoprobe (¹H-¹³C/¹⁵N). For two-dimensional ¹H-³¹P HMBC, a Bruker DRX 600 spectrometer (Bruker Biospin) with a QXI probe (¹H-¹³C/¹⁵N/³¹P) was used. Both spectrometers were run with XWIN-NMR software. All NMR experiments were performed at a probe temperature of 25°C. The chemical shifts are expressed as δ-values relative to the internal standard sodium 2,2-dimethyl-2-silapentane-5-sulfonate (DSS).

Transposon mutagenesis.

PCho-defective transposon Fujisawa mutants were generated by a mariner-based transposition system, as previously described (3). Briefly, pMC plasmids were transferred into strain Fujisawa by electroporation according to the previously described method (27). Transformants were selected at 30°C on BHI-T80 plates supplemented with 1 μg/ml erythromycin. Individual colonies were grown in BHI-T80 with erythromycin at 30°C. The cultures were diluted and plated on BHI-T80 containing erythromycin at 30°C and then shifted to 40°C to eliminate the thermosensitive pMC plasmids. Erythromycin-resistant (Em^r) clones were picked and passaged on BHI-T80 plates supplemented with erythromycin. The transposon insertion site was determined by sequencing the transposon-flanking DNA regions of transformants that contained a single transposon insertion. The mutants of interest were analyzed for CPS and PCho production, and their virulence was further tested.

Effects of gene inactivation at the cps-lic locus.

Effects of gene inactivation at the cps-lic chromosomal region on CPS and PCho expression were examined. For electron microscope studies, strain Fujisawa, YS-1, clone 100, and clone 112 were cultured in BHI-T80 medium at 37°C overnight, harvested by centrifugation, and resuspended in a PBS solution containing 5% fetal calf serum (FCS). For metaperiodate treatment, Fujisawa cells were washed once with 50 mM sodium acetate (pH 4.5) and incubated with 50 mM sodium acetate (pH 4.5) containing 20 mM metaperiodate in the dark with gentle shaking at room temperature for 1 h. The bacteria were placed on Formvar-coated nickel grids (400 mesh; Nisshin, Tokyo, Japan) for 2 min. After being washed with a PBS solution containing 10% FCS, the grids were incubated with MAbs ER21 and TEPC-15, followed by incubation with 10-nm-diameter colloidal gold-conjugated goat anti-mouse immunoglobulin (IgG, IgM, and IgA) (BB International, Cardiff, United Kingdom). The grids were briefly stained with 1% ammonium molybdate and air dried after successive washing with PBS containing 10% FCS and 1% ammonium acetate solution. The dried grids were observed by a transmission electron microscopy (H-7500; Hitachi, Japan). Samples treated with secondary antibody alone were included as controls. For morphological analysis with light microscopy, Fujisawa, YS-1, and clones 100 and 112 were cultured at 37°C in BHI-T80 overnight, and aliquots of each culture were smeared on a glass slide. After air drying and heat fixation, the slides were Gram stained using the Favor G Gram stain kit (Nissui Pharmacological, Tokyo, Japan) and viewed under a 100× objective lens with a Leica DM1000 microscope. Images were captured with LAS EZ software.

The effect of gene inactivation at the cps-lic region on CPS and PCho expression was further examined by dot immunoblotting and Western blotting. The Fujisawa and YS-1 strains and clones 100 and 112 were cultured in 500 ml of BHI-T80 at 37°C overnight. For dot immunoblotting, 3 μl of the overnight cultures was spotted onto a nylon membrane. Subsequently, 500-ml cultures of the strains were pelleted by centrifugation, and the pellets were washed with ice-cold PBS and suspended in 15 ml of water. Water-soluble antigens were extracted by treatment with 65°C water, as described above, and the crude capsular antigens were analyzed by Western blotting.

Virulence testing in mice.

Five female 6- to 8-week-old ddY mice (purchased from Japan SLC, Inc., Hamamatsu, Japan) were subcutaneously (s.c.) inoculated with 2.0 × 10⁸ CFU (approximately 10⁷ times the 50% lethal dose [LD₅₀] of the Fujisawa strain) of either the PCho-defective mutant clone 100 or clone 112. Five mice were inoculated s.c. with 2.0 × 10⁸ CFU of strain Fujisawa as a control. Mice were observed for 14 days for clinical symptoms and death.

RESULTS

Organization of a cps-lic gene cluster.

We previously constructed a nonreverting mutant Fujisawa strain of E. rhusiopathiae, YS-1, in which Tn916 was inserted into a gene (ERH_0855) and excised from the chromosome, resulting in a permanent inactivation of the gene (26). Genome analysis of the Fujisawa strain revealed that this gene is located in a cluster of genes encoding seven proteins (ERH_0855 to ERH_0861) that appear to be involved in capsular polysaccharide synthesis (cps genes) (17). The putative lic operon involved in the production and incorporation of PCho includes four genes (ERH_0862 to ERH_0865: licC-licA of ERH_0862, licD, licB, and licC-licA of ERH_0865) and lies immediately downstream of the cps region. In E. rhusiopathiae, the choline kinase (licA) and phosphorylcholine cytidylyltransferase (licC) genes are fused, suggesting that they form bifunctional enzymes (Fig. 1A). RT-PCR analysis was performed with seven primer sets that were designed to create overlapping PCR products encompassing the cps and lic regions and the downstream gene (ERH_0866; ethanolamine utilization protein gene [eutH]). A single band of the expected length was amplified (Fig. 1B), indicating that these genes are transcribed as a polycistronic mRNA, forming an operon. Ethanolamine can replace choline in the cell wall and substitute for a growth requirement of S. pneumoniae (8), and it is intriguing that eutH is located in this region.

Fig 1.

(A) Schematic representation of the cps-lic region of E. rhusiopathiae. The arrows indicate the orientation and corresponding locations of the primer pairs used (Table 1) in the RT-PCR assay. The black arrowhead denotes the transposon insertion site. (B) RT-PCR analysis of the E. rhusiopathiae cps-lic cluster. The numbers above the lanes correspond to the Arabic numerals of the primer names. All lanes labeled with a represent the analysis of PCRs performed with cDNA. All lanes labeled with b represent the corresponding control reactions with RNA extracts not subjected to reverse transcription. All lanes labeled with c show control reactions with mixtures not containing template. Lane 1a represents a control reaction with primers 1F and 1R, which were both designed to amplify an intragenic region of the 23S rRNA gene. A molecular size marker (1-kb ladder; Gibco BRL) is shown to the left (M).

Transposon mutagenesis and the construction of PCho-defective mutants.

To obtain PCho-defective mutants, we constructed approximately 300 distinct transformants that contained single transposon insertions and determined the insertion sites by sequencing. Two mutants, clone 100 and clone 112, which contained transposons inserted into licC-licA (ERH_0862) and licD (ERH_0863), respectively (Fig. 1A), were successfully obtained and used for further analysis.

Preparation of CPS.

To determine the chemical composition, the crude CPS of Fujisawa was extracted with hot water, treated with nucleases, pronase, and N-acetylmuramidase, delipidated, and purified by DEAE-cellulose anion-exchange chromatography. As shown in Fig. 2A, a peak that contained a large amount of carbohydrate was eluted at salt concentrations of 0.12 M to 0.2 M NaCl; the fractions from the latter part of the peak showed immunoreactivity with MAb ER21. The fractions containing MAb ER21-reacting antigen(s) were pooled and further purified by Sephacryl S-200 GFC. As shown in Fig. 2B, the fractions containing MAb ER21-reacting antigen(s) were eluted near the void volume. After Western immunoblotting with MAb ER21, a broadly smeared band that reacted with MAb ER21 was observed in the 17.4- to 25.7-kDa region. The patterns of reactivity were not altered when checked after each step, including water extraction, DNase, RNase, pronase, and N-acetylmuramidase treatment, methanol and chloroform treatment, and DEAE-cellulose anion-exchange chromatography (Fig. 2B, inset). To investigate the nature of the epitopes recognized by MAb ER21 and the presence of phosphorylcholine in the fractions, the antigenic fractions from GFC were treated with sodium metaperiodate or phospholipase C and analyzed by dot and Western immunoblotting. As shown in Fig. 2C, treatment with sodium metaperiodate, which cleaves the carbon-carbon bond between vicinal diols, abolished the binding of MAb ER21, suggesting that the epitopes recognized by MAb ER21 are carbohydrates. The CPS-containing fractions (no. 10 to 13) were pooled and further treated with phospholipase C. Treatment with phospholipase C resulted in a loss of reactivity with the PCho-specific MAb TEPC-15 (Fig. 2D). Taken together, these results indicate that the CPS of E. rhusiopathiae is modified by PCho, as has been shown for other mucosal pathogens, including N. meningitidis, S. pneumoniae, and H. influenzae (5, 31, 33). The pooled CPS-containing fractions were further analyzed.

Fig 2.

Elution profiles of DEAE-cellulose anion-exchange chromatography (A) and GFC (B). All of the fractions were assayed for reactivity with MAb ER21 by dot immunoblotting. X, the fractions that reacted with MAb ER21; V₀, void volume. The inset in panel B shows the Western immunoblot with MAb ER21 after each procedure: water extraction (lane 1); DNase, RNase, pronase, and N-acetylmuramidase treatment (lane 2); methanol-chloroform treatment (lane 3); and DEAE-cellulose anion-exchange chromatography (lane 4). (C) Dot immunoblot of fractions of GFC with MAbs ER21 (left) and TEPC-15 (right) after treatment with metaperiodate. The tube numbers correspond to the fraction numbers from GFC. (D) Western immunoblot of the pooled fractions (tube no. 10 to 13) from GFC with MAbs ER21 (lane 2) and TEPC-15 (lane 4) after treatment with phospholipase C. Lanes 1 and 3 are the corresponding control solutions without phospholipase C treatment. The positions of the protein molecular mass standards (kDa) are shown to the left of the inset in panel B and panel D.

Analysis of CPS components.

To identify the monosaccharide components of the CPS of Fujisawa, the ABEE-labeled monosaccharides were analyzed by HPLC. The CPS comprises seven major monosaccharides, galacturonic acid (GalA), galactose (Gal), mannose (Man), glucose (Glc), arabinose (Ara), xylose (Xyl), and N-acetylglucosamine (GlcNAc), and some minor monosaccharides, including ribose (Rib), rhamnose (Rha), and N-acetylgalactosamine (GalNAc). The identities of other minor monosaccharides were not determined due to a lack of standards (Fig. 3).

Fig 3.

HPLC chromatograms of ABEE-labeled 13 standard monosaccharides (A) and component monosaccharides released from the CPS of E. rhusiopathiae (B). The CPS was hydrolyzed, N-acetylated, and converted to ABEE-monosaccharide. The HPLC analysis was carried out as described in Materials and Methods. Asterisks represent peaks that were not determined due to a lack of standards. Peaks are numbered as follows: 1, GlcA; 2, GalA; 3, Gal; 4, Man; 5, Glc; 6, Ara; 7, Rib; 8, ManNAc; 9, Xyl; 10, GlcNAc; 11, Fuc; 12, Rha; and 13, GalNAc.

The purified native CPS and partially hydrolyzed CPS were analyzed by MALDI-TOF mass spectrometry to confirm the monosaccharide components of the CPS that were identified by HPLC and the presence of PCho in the CPS and to detect other noncarbohydrate moieties. MALDI-TOF mass spectra of the purified native CPS included a broad peak in the range of m/z 5,000 to 10,000, and the partially hydrolyzed CPS exhibited no distinct peaks in the high-mass range (data not shown). However, several peaks were observed in the low-mass range of both samples (data not shown). Figure 4 shows the representative tandem MS (MS/MS) spectra generated from four precursor ions in the positive- or negative-ion modes. The precursor ion at m/z 1,547.9 gave several ions by losing m/z 146, 203, and 243, indicating that the precursor ion contains Rha, N-acetylhexosamine, and an unidentified component (Fig. 4A). Likewise, the precursor ion at m/z 1,289.4 generated a series of ions by losing m/z 132, 162, and 275, indicating that it contains a pentose (Ara, Xyl, or Rib), five hexoses, and probably N-acetylmuramic acid (MurNAc), the latter of which may have been derived from peptidoglycan while preparing the CPS (Fig. 4B). The precursor ion at m/z 604 (Fig. 4C) gave product ions with mass differences of m/z 203 and 43, indicating the ion contains N-acetylhexosamine. The precursor ion at m/z 620 (Fig. 4D) gave product ions with mass differences of m/z 97, 44, 59, and 183; the latter two ions were also detected in the precursor ion at m/z 604 (Fig. 4C), which corresponds to neutral loss of H₂PO₄, (CH₂)₂O, trimethyl amine, and PCho, respectively, indicating that the ions contain PCho. Thus, mass spectrometric analysis confirmed the modification of the CPS by PCho and the monosaccharide composition of the CPS determined by HPLC.

Fig 4.

MALDI-TOF MS/MS spectra of precursor ions at m/z 1,547.9 (A) and m/z 1,289.4 (B) in the positive-ion mode and precursor ions at m/z 604.2 (C) and m/z 620.2 (D) in the negative-ion mode. Spectra were obtained from native CPS (A and B) or CPS partially hydrolyzed with 2% acetic acid at 100°C for 2.5 h (C and D).

Structural analysis of the CPS.

To reveal the chemical structure of the CPS, lyophilized CPS was dissolved in deuterium oxide and analyzed by NMR. Figure 5 shows the one-dimensional spectra of ¹H and ¹³C DEPT-135 (1D-¹H and 1D-¹³C). 1D-¹H analysis revealed several characteristic spectral clusters for the anomeric region at 4.4 to 5.5 ppm, the nonanomeric region at 3.3 to 4.4 ppm, and three kinds of methyl residues at 3.21 ppm, 1.90 to 2.10 ppm, and 1.10 to 1.17 ppm (Fig. 5A). The 1D-¹³C spectra were in accordance with those of 1D-¹H, except that the spectra of the methylenes of C-6 and PCho are shown downward in the DEPT-135 measurement (Fig. 5B). Analysis of the anomeric region from HSQC revealed that the CPS contains at least nine monosaccharides; the H-2 sites of these monosaccharides could be traced unequivocally from the anomeric H-1 sites in DQF-COSY (data not shown). These monosaccharides were labeled A, B, C, D, E, F, G, H, and I. The integrals of the H-1 proton peaks of the nine monosaccharides were different (Fig. 6). These results confirmed the heterogeneity of the CPS with deoxyhexose, N-acetylhexosamine and other monosaccharides, and PCho. The coupling constants (³J_H1,H2) of vicinal protons in monosaccharides A to F are no larger than 4 Hz, while those of G to I are 7 to 8 Hz. The H-1 and C-1 sites of these monosaccharides were assigned unambiguously from HSQC (Fig. 6; Table 2). The existence of PCho was further confirmed by HSQC analysis. Figure 7 shows the assignment of the three types of choline residues. Two types of CH₂ residues were first assigned based on the downward peaks in the one-dimensional ¹³C DEPT-135. Then, the phosphorus-bonded CH₂ peaks were assigned based on the characteristic split peak in ¹³C DEPT-135. (CH₃)₃ residues were assigned to the characteristic singlet with a low chemical shift and high integral value due to the nine equivalent protons. The chemical shifts are shown in Table 2. In our analyses, although the H-2 and C-2 sites of the nine monosaccharides could be traced, the monosaccharides and the linkages between the monosaccharides could not be identified due to heavy overlap of the cross peaks of H-3 to H-6 in DQF-COSY and the lack of cross peaks in the anomeric region in ¹H-¹³C HMBC and ¹H-¹H NOESY (data not shown).

Fig 5.

One-dimensional spectra from ¹H and ¹³C DEPT-135 NMR of the CPS of E. rhusiopathiae. The preparation of the CPS sample and the method for the NMR analysis were described in Materials and Methods. DSS (sodium 2,2-dimethyl-2-silapentane-5-sulfonate) was used as an internal standard for chemical shifts. *, process residuals; #, spectra derived from DSS. dHex, deoxyhexose.

Fig 6.

HSQC of the anomeric region of the CPS of E. rhusiopathiae. The cross peaks of nine monosaccharides for which H-2 sites could be unequivocally traced from anomeric H-1 sites in DQF-COSY are labeled from A to I, and their chemical shifts are shown in Table 2. The integrals of the H-1 protons of these monosaccharides were calculated from the area of each peak. HDO represents the residual H₂O in the CPS sample.

Table 2.

Chemical shifts of each anomeric resonance of the nine monosaccharides and phosphorylcholine of strain Fujisawa CPS

Residue	Chemical shift (ppm)
	H–1	C–1
Monosaccharides
A	5.44	98.2
B	5.14	98.8
C	5.10	99.1
D	5.08	110.3
E	5.01	100.6
F	4.96	100.5
G	4.73	105.1
H	4.64	107.2
I	4.61	105.6
Phosphorylcholine
CH3N–	3.21	56.8
–NCH2–	3.66	68.8
–CH2P	4.29	62.2

Fig 7.

Cross peaks for the PCho moiety characterized by HSQC, in which one-dimensional ¹³C was measured with DEPT-135, indicating the peak of the CH₂ residue pointing downward. Arrows indicate the residues of choline, and their chemical shifts are shown in Table 2. HDO represents the residual H₂O in the CPS sample.

Effects of gene inactivation at the cps-lic locus. (i) Electron and light microscopy studies.

As described above, our genetic and immunochemical analyses demonstrated that the capsular materials were a complex of polysaccharides and PCho. To examine the effects of gene inactivation at the cps-lic locus on CPS and PCho expression, wild-type Fujisawa and mutant strains were analyzed. Immunogold electron microscopic analysis of Fujisawa showed that the PCho colocalizes with the CPS on the cell surface, indicating that the CPS of E. rhusiopathiae is modified by PCho (Fig. 8A and B). The treatment of Fujisawa with metaperiodate abolished its reactivity with MAb ER21 (Fig. 8I and J) but not MAb TEPC-15 (data not shown), confirming the results obtained in dot blots (Fig. 2C). Clones 100 and 112, which had transposon insertions in licC-licA and licD, respectively, lost the reactivity with MAb TEPC-15 (Fig. 8F and H), showing that the lic region is indeed responsible for the production of PCho on the cell surface. In the capsule-defective mutant YS-1, the CPS and PCho were unexpectedly detected on the cell surface by immunogold particle labeling (Fig. 8C and D).

Fig 8.

Electron and light microscopic analyses of E. rhusiopathiae strains. (A to J) Immunogold electron microscopic analysis of strain Fujisawa, YS-1, and the PCho-defective mutants clones 100 and 112. In panel J, the whole cells of Fujisawa were pretreated with metaperiodate. Bacterial cells were loaded onto grids, incubated with MAb ER21 (A, C, E, G, I, and J) or TEPC-15 (B, D, F, and H), and subsequently incubated with a 10-nm colloidal gold particle-conjugated secondary antibody. (K to N) Light microscopy images of strain Fujisawa (K), YS-1 (L), clone 100 (M), and clone 112 (N).

The effect of gene inactivation at the cps-lic region on bacterial cell morphology was also examined. Light microscopic examination of the bacterial cells showed no clear difference between strain Fujisawa and the PCho-defective clones 100 and 112 (Fig. 8K, M, and N). In contrast, YS-1 showed unambiguous morphological changes; the size of the mutant (width) was increased, and some cells grew in long chains (Fig. 8L), suggesting that the integrity of the cell wall might be lost in this strain.

(ii) Immunoblotting analysis for CPS and PCho expression.

To examine CPS and PCho expression in the mutant strains of E. rhusiopathiae, dot and Western blotting analyses were performed. The dot blot analysis showed that intact whole cells of the YS-1 strain reacted with the MAbs ER21 and TEPC-15 (Fig. 9A), which was also observed in the electron microscopy analysis. However, in the Western blot analysis, the band that reacted with MAb ER21 was faint in the YS-1 strain, and retarded electrophoretic mobility was observed (Fig. 9B, lane 2). Taken together, these results strongly suggest that the YS-1 strain had changes in the capsule structure that affected its solubility and charge density. Likewise, the ER21-reactive bands of clones 100 and 112 also migrated with retarded mobility (Fig. 9B, lanes 3 and 4), suggesting that both the CPS and PCho are required to maintain the molecular integrity of the capsule.

Fig 9.

CPS and PCho production by strain Fujisawa and mutant strains of E. rhusiopathiae. The Fujisawa and YS-1 strains, clone 100, and clone 112 were cultured in 500 ml of BHI-T80 overnight at 37°C. (A) Dot blot detection of CPS and PCho. Three-microliter samples of fresh cultures were spotted onto a nylon membrane and probed with MAbs ER21 or TEPC-15. (B and C) Western blot detection of the CPS and PCho. The 500-ml cultures were pelleted by centrifugation. The crude CPS was extracted with 15 ml of hot water, as described in Materials and Methods, and subjected to immunoblotting with MAb ER21 (B) or TEPC-15 (C). Lanes: 1, Fujisawa; 2, YS-1; 3, clone 100; 4, clone 112. The positions of the protein molecular mass standards (kDa) are shown to the left in panel B.

Virulence testing in mice.

To determine whether PCho plays an important role in the virulence of E. rhusiopathiae, we inoculated mice s.c. with 2.0 × 10⁸ CFU (approximately 10⁷ times the LD₅₀) of the Fujisawa strain or the PCho-defective mutants. Mice inoculated with the Fujisawa strain died within 2 days of inoculation, whereas all of the mice inoculated with the PCho-defective mutants survived for 14 days without clinical symptoms. This result indicates that PCho plays a critical role in the virulence of this organism.

DISCUSSION

Capsular polysaccharides are important virulence factors in a wide range of bacterial species. The presence of a capsule in E. rhusiopathiae and the role of the capsule in virulence have been determined (26, 29). However, the chemical and biological properties of the capsule have not been characterized.

Lachmann and Deicher (14) analyzed the surface components of E. rhusiopathiae by sulfate-polyacrylamide gel electrophoresis and immunoblotting. They found that the major antigen with a molecular weight of 14,000 to 22,000 had a positive anthrone reaction and that its antigenicity was lost after treatment with sodium metaperiodate but not with alkali or organic acids; therefore, they concluded that the antigen is the capsular polysaccharide antigen. Schubert and Fiedler (22) investigated the cell surface of E. rhusiopathiae and found that the carbohydrate moiety of the cell wall is very complicated. The authors concluded that the cell surface polysaccharide is mainly composed of N-acetylfucosamine and smaller amounts of Gal, GlcNAc, and an unidentified sugar. Thus, information on the cell wall components and structure of E. rhusiopathiae has been limited.

In this study, we demonstrated that E. rhusiopathiae has a complex CPS and further clarified that the CPS is modified by PCho. The purified CPS of E. rhusiopathiae exhibited a smeared pattern of 17.4 to 25.7 kDa by Western immunoblotting, suggesting that the CPS is heterogeneous. The heterogeneity of the CPS was also suggested by MS spectrometry, which revealed an irregular reduction pattern of mass units from the MS spectra (data not shown) and by NMR analysis, which showed different ratios of integrals for each anomeric proton of the nine monosaccharide residues. The heterogeneous nature of the CPS of E. rhusiopathiae complicated the NMR spectra, which are typically simple for homogenous bacterial polysaccharides with repeating structure (6). In GFC, the CPS was eluted near the void volume. In addition, mass spectrometry analysis of the purified native CPS revealed a broad peak in the range of m/z 5,000 to 10,000. Treatment of intact Thermoplasma acidophilum LPS, which is highly aggregated in water and exhibits an apparent molecular weight greater than 1,200,000, which with 0.5% sodium dodecyl sulfate results in the dissociation of the LPS, yielding material with a molecular weight of 67,000 (15). This suggests that the discrepancy in the MWs determined by Western immunoblotting and GFC in this study is most likely due to the aggregation of CPS in the aqueous solutions used in GFC and mass spectrometry.

The CPS of E. rhusiopathiae contains major seven monosaccharides, GalA, Gal, Man, Glc, Ara, Xyl, and GlcNAc, and some minor monosaccharides, including Rib, Rha, and GalNAc. GalA, Gal, Man, Glc, GlcNAc, Rha, and GalNAc have been frequently identified as monosaccharide components of the CPSs of other Gram-positive bacteria (11, 18, 19), including S. pneumoniae (12, 13). Rib, Xyl, and Ara have been recently identified as monosaccharide components of the CPS and extracellular polysaccharide of the fish pathogen Streptococcus iniae (7). The difference in ratios between the major and minor monosaccharide components in strain Fujisawa could be due to the heterogeneity of the CPS. Alternatively, the minor components (Rib, Rha, and GalNAc) may be artifacts arising during purification procedures. Further experiments should be performed to clarify this point. Thus, the overall composition of the CPS of E. rhusiopathiae is similar to that of other Gram-positive bacteria, although it is more heterogeneous.

The electrophoretic mobility in SDS-PAGE of the CPS of E. rhusiopathiae and its modification by PCho exhibit some similarity to the LTA or lipoglycan of other Gram-positive bacteria, especially S. pneumoniae, and may raise the question of whether the CPS purified in this study is the LTA or lipoglycan of E. rhusiopathiae. However, the CPS appears to differ from prototypical LTA or lipoglycan of Gram-positive bacteria (21) for the following reasons. First, the monosaccharide constituents of the CPS, which were extracted by hot water, exhibit greater heterogeneity. Second, the CPS does not include glycerophosphate (GroP) or ribitol phosphate. We did not confirm the reactivity of the CPS with a GroP-LTA-specific MAb (clone 55; HyCult Biotechnology, Uden, Netherlands) (9) or characteristic cross peaks of GroP in two-dimensional NMR analysis (data not shown). Third, as previously shown (17), the E. rhusiopathiae genome lacks many orthologous genes for the biosynthesis of LTA and WTA and the dltABCD operon, all of which are mostly conserved in other Gram-positive bacteria. Using HPLC, we found that the cell wall of E. rhusiopathiae does not contain d-alanine (data not shown). The E. rhusiopathiae genome also lacks tarIJL genes, which are responsible for CDP-ribitol generation (20), suggesting that the cell wall of the organism may be atypical in structure and/or composition. In other Gram-positive bacteria, LTA and lipoglycan are both essential for bacterial physiology and, therefore, stable mutants with defects in LTA or lipoglycan metabolism are not available (30). In contrast, gene knockout mutants of Fujisawa in which transposons are inserted into genes within the cps (26) or lic loci (this study) could be constructed. Thus, there appears to be clear distinction between the CPS of E. rhusiopathiae and the LTA or lipoglycan of other Gram-positive bacteria.

Intriguingly, the Fujisawa genome lacks clear orthologues of ltaS, which encodes the polyglycerophosphate-LTA synthase and is indispensable in Staphylococcus aureus (9, 21). Furthermore, it appears that tagB and tagG (tarF), both of which are involved in the biosynthesis of WTA and are essential for cell viability of Bacillus subtilis (1), are fused together, with their functions encoded by a single gene (ERH_0432) in E. rhusiopathiae (17). Analysis of the E. rhusiopathiae genome also suggests that tagO, which is indispensable for the growth of Gram-positive bacteria (1), is most likely fused with mraY (17), a gene essential for peptidoglycan biosynthesis. Given the unique phylogenetic position of E. rhusiopathiae, there may be some overlap between the pathways for cps and other cell wall components, including WTA and/or LTA, due to reductive genome evolution. Such an overlap could explain the production of an atypical CPS by E. rhusiopathiae and the morphological changes observed in YS-1, in which the integrity of the cell wall might be lost. The gene ERH_0855, which was inactivated in the YS-1 strain, encodes a protein containing a domain that is highly conserved in a large number of different bacterial sugar transferases (pfam02397) in a broad spectrum of Gram-negative and Gram-positive bacteria (17). It is therefore possible that inactivation of this gene might have pleiotropic effects on the structure of the CPS and cell wall biosynthesis, which are important for maintaining cell wall integrity.

In summary, E. rhusiopathiae has a complex CPS, which is modified by PCho. The complexity of the E. rhusiopathiae CPS may be related to the unique phylogenetic position of the organism. To fully understand the pathogenicity and physiology of E. rhusiopathiae, further clarification of the cell wall structure and composition is needed.