Recently, we have shown that the capsular polysaccharide of Bacteroides fragilis NCTC 9343 is composed of an aggregate of two discrete large molecular weight polysaccharides (designated polysaccharides A and B). Following disaggregation of this capsular complex by very mild acid treatment, high resolution NMR spectroscopy demonstrated that polysaccharides A and B consist of highly charged repeating unit structures with unusual substituent groups (Baumann, H., Tzianabos, A. O., Brisson, J.-R., Kasper, D.L., and Jennings, H.J. (1992) Biochemistry 31, 4081-4089). Presently, we report that the capsular polysaccharide of B. fragilis represents a complex structure that is formed as a result of ionic interactions between polysaccharides A and B. Electron microscopy of immunogold-labeled organisms (with monoclonal antibodies specific for polysaccharides A and B) demonstrated that the two polysaccharides are co-expressed on the cell surface of B. fragilis. We have shown that the purified capsule complex is made up exclusively of polysaccharide A and polysaccharide B (no other macromolecular structure was detected) in a 1:3.3 ratio and that disaggregation of this complex into the native forms of the constituent polysaccharides could be accomplished by preparative isoelectric focusing. Structural analyses of the native polysaccharides A and B showed that they possessed the same repeating unit structures as the respective acid-derived polysaccharides. The ionic nature of the linkage between polysaccharides A and B was demonstrated by reassociation of the native polysaccharides to form an aggregated polymer comparable to the original complex. The distinctive composition of this macromolecule may provide a rationale for the unusual biologic properties associated with the B. fragilis capsular polysaccharide.

Sialylation of bacterial capsules has been proposed as an important virulence factor for several species of encapsulated pathogens, including group B Streptococcus. We have constructed a transposon mutant strain of type III group B Streptococcus that expresses a capsular polysaccharide differing from the wild type only in that the mutant strain's capsule lacks sialic acid. The mutant polysaccharide is antigenically identical to the capsular polysaccharide of type 14 Streptococcus pneumoniae, as predicted by the structures of the type III group B Streptococcus and S. pneumoniae polysaccharides. Loss of capsular sialic acid was associated with loss of virulence in the mutant strain in a neonatal rat model of lethal group B Streptococcus infection. These studies demonstrate directly that capsular sialic acid is a critical virulence determinant for type III group B Streptococcus and support the general hypothesis that surface sialylation aids pathogenic microorganisms in evading host defenses.

Oligosaccharides consisting of one or more tetrasaccharide repeating units were derived from the capsular polysaccharide of type 14 pneumococcus (Pn14) by endo-beta-galactosidase digestion. The relative affinity of anticapsular antibody binding to derivative oligosaccharides of different chain lengths was measured in a Pn 14 ELISA inhibition assay. The concentration of inhibiting antigen required to achieve 50% inhibition of IgG binding increased progressively from 5.6 x 10(-4) M to 7.0 x 10(-11) M as the inhibiting saccharide chain length increased from 1 tetrasaccharide repeating unit to 2,500 repeating units. These data indicate that antibodies directed against the Pn14 polysaccharide recognize a conformational epitope fully expressed only in high molecular weight forms of the antigen. Similar results were found for inhibition of Fab fragment binding, suggesting that recognition of the conformational epitope is largely dependent on the intrinsic affinity of the Fab combining region. Unlike previously reported polysaccharides for which conformational epitopes have been described, the Pn14 polysaccharide does not contain negatively charged residues, indicating that expression of conformational determinants is not limited to acidic polysaccharides. Antibody recognition of conformational epitopes may be a common mechanism by which the host immune response discriminates between bacterial polysaccharides and host oligosaccharides of similar structure.

Immunization of pregnant women with a polysaccharide vaccine of group B streptococcus is a promising strategy for the prevention of perinatal infections caused by group B streptococci. To explore the feasibility of this strategy, we vaccinated 40 pregnant women at a mean gestation of 31 weeks with a single 50-microgram dose of the Type III capsular polysaccharide of group B streptococcus. The only adverse effect detected was a mild local reaction in nine women (22 percent). Of the 35 women with low or unprotective antibody levels before immunization (less than 2 micrograms per milliliter), 20 (57 percent) responded to the vaccine. The geometric mean antibody level rose from 1.3 to 7.1 micrograms per milliliter four weeks after vaccination (P less than 0.02), and these levels persisted at delivery and three months post partum. Sixty-two percent of the vaccine-induced immunoglobulin in the mothers was IgG, which readily crosses the placenta. Infant antibody levels in cord serum correlated directly with maternal antibody levels at delivery (r = 0.913, P less than 0.001). Of the 25 infants born to women who responded to the vaccine, 80 percent continued to have protective levels of antibody at one month of age and 64 percent had protective levels at three months. Serum samples from infants with greater than or equal to 2 micrograms of antibody to Type III group B streptococcus per milliliter uniformly promoted efficient opsonization, phagocytosis, and bacterial killing in vitro of Type III strains. This effect could be mediated exclusively by the alternative complement pathway. Although this vaccine with an overall response rate of 63 percent is not optimally immunogenic, we conclude that maternal immunization is feasible and can provide passive immunity against systemic infection with Type III group B streptococcus in the majority of newborns. Larger trials with better vaccines will be required to evaluate the safety and clinical effectiveness of this strategy.

Several lines of evidence suggest that there might be immunologic cross-reactivity between the thyroid plasma membrane in humans and antigenic determinants in the enteric pathogen Yersinia enterocolitica. Studies were therefore performed to determine whether Y. enterocolitica, like the thyroid membrane, contains a thyrotropin binding site. A saturable binding site for bovine thyrotropin was indeed demonstrable, particularly in preparations of the organism that have been treated with ethylenediaminetetraacetate and lysozyme. Hormonal specificity of the binding site, as judged from the inhibition of binding of 125I-labeled bovine thyrotropin, was similar to that of the thyrotropin receptor in human thyroid tissue.

Immunization of rabbits with group B type III streptococcus organisms induces two distinct populations of antibodies with a specificity for determinants on the native capsular polysaccharide antigen of these organisms. Some of the structural and conformational features of the two determinants responsible for the formation of these antibodies were elucidated by (13)C NMR and serological studies on the native type III polysaccharide and some of its structurally modified analogues. The specificity of the determinant corresponding to the major population of antibodies is dependent of the presence of sialic acid residues on the native type III antigen, and although these residues are not an integral part of the determinant, they exert conformational control over it. The carboxylate groups of the sialic acid residues are an important factor in this control mechanism which could possibly involve intramolecular hydrogen bonding. The terminal sialic acid residues control the orientation of the penultimate beta-d-galactopyranose residues with respect to the backbone of the native antigen. The orientation of these residues is critical to the determinant because the determinant is probably small and is located precisely at the junction of the same beta-d-galactopyranose residues with the backbone of the native type III antigen. The determinant corresponding to the other population of antibodies is not sialic acid dependent. This determinant is located on the backbone of the native antigen in the vicinity of the other determinant but on the opposite side to the oligosaccharide branches. In this position, its conformation is unaffected even by the removal of the oligosaccharide branches from the native antigen.

The immunogenicity and safety of two polysaccharides isolated from type III, group B Streptococcus, were tested in adults selected for existing low concentrations of natural antibody to the capsular polysaccharide of this organism. Both vaccine preparations (trichloroacetic acid and EDTA) were found to lack pyrogenicity and toxicity for experimental animals. A single 50-microgram subcutaneous injection of either polysaccharide in human subjects elicited significant increase in antibody concentration in immunized compared with control individuals receiving phosphate-buffered saline. Antibody responses were maximal by 2 wk and remained at 21 wk after immunization. Vaccine-induced antibody was primarily of the IgG class. Of the two vaccines, the larger molecular size polysaccharide was significantly more immunogenic. Although no systemic reactions were recorded, mild transient local reactions occurred in 45% of vaccinees.

We investigated the role of maternal antibody in neonatal Group B streptococcal infection with a radioactive antigen-binding assay employing a purified polysaccharide antigen with both Type III and Group B determinants. Serums from seven women who gave birth to infants who had invasive Group B streptococcal infection with Type III strains were all deficient in antibody. In contrast, serums from 22 of 29 pregnant Type III vaginal carriers whose infants were healthy contained antibody with a prevalence significantly different from that in women delivering infants with Type III disease (P less than 0.01). Three healthy neonates born to women with antibody in serums had demonstrable antibody in umbilical-cord serum. These data suggest that transplacental transfer of maternal antibody protects infants from invasive Group B streptococcal infection with Type III strains.