INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Intra-abdominal infections present serious clinical conditions that are difficult to treat, frequently require surgical intervention, and often result in short-term and long-term complications. Bacteroides fragilis accounts for only 0.5% of the normal human colonic flora but is the anaerobic bacterium most frequently isolated from clinical infections in the abdominal cavity (9, 25). The capsular polysaccharides of B. fragilis are the most important virulence factors responsible for the formation of intra-abdominal abscesses by this organism (17, 19).

The capsular polysaccharides of B. fragilis exist on the organism's surface as part of a capsular polysaccharide complex (CPC). For the B. fragilis type strain, NCTC 9343, two polysaccharides of the CPC, PS A1 and PS B1, have been elucidated structurally (2, 31). The gene loci for biosynthesis of PS A1, PS B1, and a third, structurally unknown, capsular polysaccharide, PS C1, have recently been sequenced (3; M. J. Coyne, A. O. Tzianabos, B. E. Mallory, D. L. Kasper, and L. E. Comstock, submitted for publication). The most striking feature of all of the B. fragilis polysaccharides characterized to date is the presence of both positively and negatively charged substituent groups on each repeating unit. This specific charge motif (zwitterionic) is essential for the unique biological activity of this class of molecules (13, 30), including promotion of the formation of intra-abdominal abscesses (30).

Studies with monoclonal antibodies (MAb) and polyclonal antiserum have demonstrated, however, that the capsular polysaccharides of B. fragilis are quite heterogeneous (15, 26, 28). The B. fragilis strain that has been best studied, 638R, is known to produce a CPC that is immunologically distinct from the CPC of the type strain, NCTC 9343 (3, 21, 23, 27). However, the polysaccharides composing the 638R CPC have not been characterized. The biosynthesis loci of PS C from strain NCTC 9343 (PS C1) and strain 638R (PS C2) were the first loci to be cloned and sequenced. These loci are flanked by genes common to both strains; however, the genes involved in polysaccharide biosynthesis are distinct (3-5).

In this study, we sought to elucidate structural features of the B. fragilis strain 638R CPC to develop a clearer picture of the diversity of polysaccharides in the species and of the basic structure-activity and structure-biosynthesis relationships of abscess-inducing polysaccharides. To our knowledge, this is the first report of the isolation of three different surface-expressed capsular polysaccharidesPS A2, PS C2, and PS 3from one bacterial strain. The immunochemical characteristics and abscess-inducing potentials of these capsular polysaccharides are described and brought together with information about their biosynthetic pathways.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial strain and isolation of capsular polysaccharides of B. fragilis strain 638R. B. fragilis strain 638R (27) was maintained in peptone-yeast broth at 80°C. The production of CPC was enhanced and the glycogen content in extracts was minimized by the serial passage of bacteria five times through rat spleens (male Wistar rats; 175 to 200 g; Charles River Laboratories, Wilmington, Mass.) (14). Initially, 10⁸ organisms in phosphate-buffered saline (PBS) were injected intraperitoneally into a rat and recovered 24 h later by dispersing the spleen. The organisms were cultured overnight anaerobically on brucella agar with 5% sheep blood (PML Microbiologicals, Mississauga, Ontario, Canada). The inoculation and recovery processes were repeated. Enhanced production of capsule was confirmed by electron microscopy studies with polyvalent antiserum to whole bacteria (data not shown). After the passages, the bacteria were grown anaerobically in a 16-liter batch culture as described previously (22, 31). The cells were harvested by centrifugation (average wet weight, 244 g) and suspended in water. CPC was extracted from the bacterial cells with hot phenol-water and digested extensively with RNase, DNase, and pronase. The dialyzed product (average dry weight, 2 g) was chromatographed on a column of Sephracyl S-400 HR (Amersham Pharmacia Biotech, Piscataway, N.J.) in a 3% deoxycholate acid-containing buffer at pH 9.8 to separate lipopolysaccharide (LPS) from CPC (22, 31). Chromatographic fractions were analyzed throughout the isolation procedure by measurements of UV absorbance at 280, 260, and 206 nm; refractive index; and protein by the bicinchoninic acid method (Pierce, Rockford, Ill.), silver staining of sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) gels (16), dot blot assay, immunoelectrophoresis (IEP), and immunodiffusion (20, 22). After gel filtration chromatography with S-400 HR, silver-stained SDS-PAGE showed fractions with a high-molecular-mass antigen (>160 kDa according to protein standards) representing CPC. Subsequently eluted fractions showed antigens smaller than 20 kDa at the gel front that represented LPS and digested protein and nucleic acid. Fractions eluted between CPC and LPS contained both materials. Fractions containing capsular polysaccharide exclusively and reacting with antiserum to the whole bacteria by double immunodiffusion were pooled, concentrated, alcohol precipitated, dialyzed, and lyophilized. The CPC (average dry weight, 0.2 g) was examined by IEP (in 50 mM Tris buffer at pH 7.3), which showed three dominant components, two positively and one negatively charged, and two weak, distinct components migrating towards the anode and cathode (Fig. 1).

View larger version (28K): [in this window] [in a new window]   FIG. 1.   IEP of the CPC and separated, purified polysaccharide components PS 1, PS 2, and PS 3 of the capsule of B. fragilis strain 638R at different pH values. Each antigen preparation was reacted with polyvalent rabbit anti-B. fragilis strain 638R antiserum.

View larger version (39K): [in this window] [in a new window]   FIG. 2.   Recorded profile of UV absorbance at 280 nm, refractive index (RI), and conductivity (cond) of ion-exchange chromatography (Q-Sepharose FF) of the capsular polysaccharide complex of B. fragilis strain 638R with 50 mM Tris (pH 7.3) (A) and during elution with an NaCl gradient (B). The polysaccharide-containing fractions were identified by dot blot assay, IEP, immunodiffusion, and NMR.

Chemical characterization of capsular polysaccharides. Polysaccharides were analyzed for monosaccharide constituents and structure by both gas chromatography-mass spectroscopy (GC-MS) and NMR, as described elsewhere (32, 34). For GC-MS analysis, the polysaccharides were hydrolyzed completely and converted to alditol acetate derivatives. Analyses were performed on an HP 6890/5973 GC-MS spectrometer (Hewlett Packard, Wilmington, Del.) with a capillary DB17 column (J&W Scientific, Folsom, Calif.). GC was started at 150°C and held for 2 min; after an increase to 260°C at a rate of 5°C/min, GC was held at the final temperature for 20 min. Mass detection was obtained by electron ionization at 70 eV, and ions were scanned from 45 to 550 m/z. NMR experiments were performed on a Unity 500 spectrometer (Varian, Palo Alto, Calif.) with a proton resonance frequency of 500 MHz. All ¹H spectra were recorded at 70°C in ²H₂O, and chemical shifts were referenced in relation to H²HO resonance at 4.36 ppm. The molecular-size distribution of the polysaccharides was determined by gel filtration chromatography on Superose 6 columns calibrated with dextran standards according to the fractionation ranges of the gel (Amersham Pharmacia Biotech) at a flow rate of 0.5 ml/min (13, 24, 33).

Preparation of polyclonal antiserum. Antiserum to B. fragilis strain 638R was prepared by intravenous immunization of New Zealand White rabbits three times a week for 3 weeks followed by a single booster injection during the fifth week and bleeding after 7 weeks. Each injection consisted of ~10⁹ bacteria that were grown in basal medium, killed with formalin, and resuspended in 1 ml of sterile saline.

Immunological assays. Double diffusion in agarose was performed by the method of Ouchterlony (20) with 1% agarose in 50 mM Tris (pH 7.3). IEP was performed in 1% agarose in 50 mM Tris (pH 7.3), 50 mM phosphate buffer (pH 6.2), and 0.2 M barbital buffer (pH 8.6). The polysaccharide antigens were used as 1-mg/ml solutions. Each antigen preparation was reacted with polyvalent rabbit anti-B. fragilis strain 638R antiserum (10). SDS-PAGE was performed according to the method of Laemmli (16) with 4 to 20% gradient gels. The gels were stained with silver stain (Bio-Rad, Hercules, Calif.). Dot blotting was performed by applying 2.5 µl of capsular polysaccharide from a 1-mg/ml solution to a nitrocellulose paper as described previously (22).

Isolation of capsular polysaccharide PS A1 from B. fragilis NCTC 9343. PS A1 from B. fragilis NCTC 9343 was prepared as described previously (2, 22, 30, 31). The polysaccharide was isolated by hot phenol-water extraction, enzyme digestion, gel filtration, and anion-exchange chromatography, and isoelectric focusing. PS A1 was prepared in sterile, pyrogen-free saline for administration to animals.

Animal model for abscess induction. In two independent experiments, 0.5 ml of injection volume containing 100, 10, 1, or 0.1 µg of capsular polysaccharide and sterile cecal contents (dilution, 1:4 in PBS) was administered intraperitoneally by injection through an 18-gauge needle to groups of six outbred male Wistar rats (150 to 175 g; Charles River Laboratories). As the control, PBS was substituted for the capsular polysaccharide. Six days after challenge, the rats were sacrificed and then examined for macroscopically visible intraperitoneal abscesses by observers unaware of the treatment status. Grossly visible abscesses were confirmed microscopically. The development of one or more abscesses was considered a positive result.

Statistical analysis. In vivo experiments were analyzed in a structured logistic regression model that permitted evaluation of separate dose-response relationships (dose in micrograms) and direct interference on the median abscess inducing dose (AD⁵⁰), which is the theoretical dose of polysaccharide required to induce abscesses in 50% of animals (6, 13, 30). Likelihood ratio tests were performed for hypotheses concerning commonality of dose-response slopes and AD⁵⁰s. Rats that died within 2 days after challenge were not included in the analysis because their deaths were due to anesthesia.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Isolation of polysaccharides. Attempts to separate the differently charged polymers by isoelectric focusing or application of ion-exchange or gel filtration chromatography were not successful, whereas mild acid hydrolysis and further processing by ion-exchange and gel filtration chromatography and isoelectric focusing led to the isolation of distinct capsular polysaccharides.

Immunochemical characterization of polysaccharides. IEP of the isolated polysaccharides at acid, neutral, and basic pHs showed distinct patterns reflecting different structural compositions and net charge motifs. The distinct patterns represented the main components of the CPC described above (Fig. 1).

View larger version (25K): [in this window] [in a new window]   FIG. 3.   1H NMR spectra of PS 1, PS 2, and PS 3.

View larger version (31K): [in this window] [in a new window]   FIG. 4.   Inhibition by PS A2 and PS C2 of binding of rabbit polyclonal antiserum raised to whole bacteria of B. fragilis strain 638R on microtiter wells coated with PS A2. The values are the means of duplicate determinations.

Biological activity of polysaccharides in experimental abscess formation. All the polysaccharides in the tested dose ranges induced intra-abdominal abscesses in a dose-dependent fashion (testing for no dose effect; PS A1, PS A2, and PS 3, P < 0.001; PS C2, P = 0.19) (Fig. 5). PS A1 of B. fragilis NCTC 9343 was the most potent abscess-inducing zwitterionic polysaccharide, with an AD⁵⁰ of 0.28 µg. The AD⁵⁰ of the zwitterionic polysaccharide PS A2 of strain 638R was 4.79 µg17 times higher. At a theoretical dose of 3.58 µg, the predominantly negatively charged polysaccharide PS 3 would have induced abscesses in 50% of the animals. The AD⁵⁰ of the predominantly positively charged polysaccharide PS C2 was 749.89 µg.

View larger version (64K): [in this window] [in a new window]   FIG. 5.   Induction of intra-abdominal abscesses in rats with different doses of capsular polysaccharides PS A2, PS C2, and PS 3 of B. fragilis strain 638R, PS A1 of B. fragilis NCTC 9343, and saline. The rat model of intra-abdominal-abscess formation was used in two independent experiments with six rats per group in each experiment.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Studies with monoclonal and polyclonal antibodies have demonstrated that capsular polysaccharides of abscess-inducing B. fragilis strains, including the type strain, NCTC 9343, and the clinical strain 638R, are heterogeneous (3, 15, 21, 23, 26-28). For strain NCTC 9343, two polysaccharides of a CPC have been fully characterized (2, 31). Only a few bacterial organisms are known to produce such a complicated array of surface polysaccharides (1, 8, 12, 18). Recently, three polysaccharide biosynthesis loci have been identified for the strain NCTC 9343 (7; Coyne et al., submitted). In this study, we were able to isolate three different immunogenic capsular polysaccharides of the strain 638R.

Although the biological activity of PS A2 in vivo is very high, it is nevertheless 17 times lower than that of PS A1. Both polysaccharides have a balanced positive and negative charge motif on each repeating unit, and they have comparable molecular masses. However, several aspects of their three-dimensional structures might be different. Conformational modeling of PS A2 suggested a right-handed helix with two repeating units per turn and a pitch of 20 Å (34). Exposure of positive and negative charges on the outer surface of the polymer in a regularly spaced pattern renders them accessible to other molecules. In contrast to PS A2, PS A1 consists of only a tetrasaccharide instead of a pentasaccharide repeating unit, a difference that might contribute to a smaller pitch (31). The genes wcfR and wcfS are involved in the synthesis of the unusual sugar AAT and are found in the PS A1 locus of strain 9343 and in the PS A2 locus of strain 638R (Coyne et al., submitted). Although PS A1 and PS A2 have the -D-AAT sugar in common, the amino group is at position 2 in PS A2 and at position 4 in PS A1. This and other compositional differences might lead to different charge accessibilities, to different configurations of large grooves, to a slightly more preferred configuration of PS A1 that enhances the interaction between polysaccharides and protein molecules required for biological activity (13), and to an altogether different host response. Another explanation for the lower abscess-inducing activity of PS A2 might be that, although PS A2 demonstrated an average balanced zwitterionic charge motif, isoelectric focusing showed that PS A2, in contrast to PS A1, had a heterogeneous charge distribution ranging from a negative to a positive net charged motif. This heterogeneity may indicate that PS A2 is being isolated in a form with molecules having variable degrees of acetylation.

The PS C2 biosynthesis locus has been sequenced and found to be composed of 22 open reading frames (3). Two genes encoding products similar to aminotransferases were identified, suggesting that PS C2 contains at least two monosaccharides with free amino groups. This finding was confirmed by the results of the GC-MS analysis, NMR, and IEP. Complementation analysis has determined that the products of two genes, mnaA and mnaB (formerly wcgC and wcgD), are involved in the synthesis of N-acetyl-mannosaminuronic acid, which could be the uronic acid of hexosamine identified by GC-MS analysis. Three clustered genes, wcgJ to wcgL, encode products likely to be involved in the formation of N-acetylfucosamine. A possibly N-acetylated form of a fucosamine was identified by GC-MS analysis. Homology-based analysis identified five putative glycosyltransferases, suggesting that PS C2 contains at least five monosaccharides (3). GC-MS analysis identified six composing sugars. The AD⁵⁰ of PS C2, at 749.89 µg, was unexpectedly high. C-substance, the group polysaccharide from Streptococcus pneumoniae, is another naturally occuring zwitterionic polysaccharide with a positive net charge. It has a tetrasaccharide repeating unit with a total of three positive and two negative charges, which give a highly charged density to the polysaccharide. C-substance was shown to be a potent abscess-inducing polymer with an AD⁵⁰ of 5 µg (30). We hypothesize that contaminating polysaccharides isolated with PS C2 may be masking critical charges and configurations, resulting in a loss of abscess-inducing potency.

PS 3 was shown by GC-MS analysis to contain three sugars carrying uronic acids, to possess a net negative zwitterionic charge motif demonstrated by IEP, and to be a potent abscess-inducing capsular polysaccharide. We were not able to determine by GC-MS analysis which of the hexosamines is N acetylated. PS 3 certainly has a high density of charged molecules, which renders the polysaccharide very capable of inducing intra-abdominal abscesses. Previously, the capsular polysaccharide of S. pneumoniae type 1 was shown to induce intra-abdominal abscesses in 50% of animals at a dose of 31 µg, and PS B1 of B. fragilis NCTC 9343 had an AD⁵⁰ of 25 µg. Both of these capsular polysaccharides carry two negatively changed groups and one positively charged group per repeating unit (30).

We were able to isolate three distinct surface-expressed capsular polysaccharides from a single abscess-modulating bacterial strain, B. fragilis 638R. Compositional and immunochemical studies demonstrated a common zwitterionic charge motif in all of the capsular polysaccharides, with net charges ranging from predominantly positive to balanced to predominantly negative. This study reflects the importance of the determination of the structural composition of the CPC in a species in yielding information about biological activity, biosynthesis, and their relationships.