Twenty-nine murine monoclonal antibodies (MAbs) were prepared against antigenic determinants in the core and lipid A regions of Escherichia coli and Salmonella minnesota lipopolysaccharide (LPS). At least eight distinct MAb specificities were identified. Epitopes recognized by MAbs bearing these specificities were localized in the hexose, heptose, and 2-keto-3-deoxy-D-manno-octulosonic acid regions of the core oligosaccharide and on lipid A. Two groups of MAbs exhibited multispecificity for similar but distinct core- and lipid A-related epitopes. Some core-reactive MAbs cross-reacted with corresponding E. coli and Salmonella rough mutant chemotypes; others were specific for E. coli J5 LPS. Lipid A-specific MAbs reacted with free lipid A from diverse sources. Few MAbs reacted with smooth LPS. Antibody cross-reactivity was restricted by inter- and intraspecies differences in covalent core structure and by epitope concealment by overlying O-side chain and core sugars. The putative cross-reactive and antiendotoxic properties of MAbs specific for the core-lipid A complex may be limited by the inability of such MAbs to recognize determinants on "native" LPS.

In vitro and in vivo antibacterial and protective properties of murine monoclonal antibodies (MAbs) to Escherichia coli O111:B4 lipopolysaccharide (LPS) were evaluated in relation to antibody isotype and complement-fixing activity. Six O side chain-specific MAbs, including two IgMs and one of each IgG subclass, were analyzed for quantitative binding and C3 deposition on intact bacteria, complement-mediated bactericidal and opsonophagocytic activity, and protection against intraperitoneal infections in mice. Although C3 was deposited on bacteria in the presence of normal human serum (NHS) alone, LPS-specific MAbs increased C3 attachment in a dose-dependent manner. Bacterial killing occurred only in the presence of both antibody and complement NHS and required an intact alternative pathway. The efficiency of bacterial killing varied by antibody isotype (IgM greater than IgG2a greater than other IgG subclasses) and correlated with C3-fixing capacity. Opsonophagocytic activity of MAbs exhibited a similar isotype-related rank order. Likewise, IgM was more active than IgG, and IgG2a was superior to other IgG subclasses, in MAb-mediated protection against intraperitoneal infection. These data document the interdependent antibacterial and complement-fixing properties of LPS-reactive MAbs and the degree to which both activities are determined by antibody class and isotype.

Sixteen murine monoclonal antibodies (MAbs) reactive with the O-side chain, core oligosaccharide, or lipid A of Escherichia coli O111:B4 and Salmonella minnesota lipopolysaccharide (LPS) were evaluated for binding activity against wild-type and rough mutant strains using a fluorescence-activated cell sorter (FACS) and fluorescein-conjugated antiimmunoglobulin probe. O-side-chain-reactive MAbs produced immunofluorescence against homologous, smooth strains up to 500-fold higher than controls. Many core- and lipid A-reactive MAbs exhibited limited reactivity with smooth bacteria. Some core- and lipid A-associated epitopes, however, were better recognized by MAbs on intact bacteria than on isolated LPS. FACS analysis of binding by the core-reactive MAb, J8-4C10, to E. coli O26:B6 smooth bacteria revealed staining and non-staining bacterial phenotypes that were sorted and stably expressed in subculture. FACS analysis thus documented differences in the whole-cell reactivity of MAbs specific for various LPS subcomponents, differences in MAb reactivity with isolated and cell-associated LPS, and spontaneous changes in the phenotypic expression of certain LPS-associated epitopes on intact bacteria.

SununaryAntibodies bearing the T15 idiotype dominate the murine primary immune response to phosphocholine (PC). Analysis of antigen binding of antibodies derived from VI:DFL16.1:J.1 (V.1) germline and N region-derived variant heavy (H) chains and g22, K24, and K8 light (L) chains demonstrates that the T15H:K22L (T15) antibody binds PC at least 20-40 times better than other antibodies derived from alternate germline forms of the V.1 H chain and g22, K24, or K8 L chains. To achieve af~nities in the same range as the T15 antibody, K24 and Ic8 L chain-containing antibodies must have H chains derived from variant N region or somatically mutated V.1 genes. Single amino acid differences at the VD junction of the various germline and N region variant V.1 H chains dictate the L chain that can associate with the H chain to produce a PC-specific antibody. Several H:L combinations give rise to T15 or M167 idiotypepositive antibodies that lack specificity for PC, and single amino acid substitutions or insertions at the V.I:D junction result in the loss of T15 or M167 idiotopes. Based on these observations, our data support a molecular model involving both preferential gene rearrangement and antigendriven B cell selection to explain T15 idiotype dominance in the immune response to PC. In the absence of N region diversification, large numbers of neonatal B cells bearing the T15H:K22L surface immunoglobulin M (slgM) receptors would be selected and expanded by autologous or environmental PC antigen into the long-lived peripheral B cell pool.T he primary antibody response to PC in normal inbred mice is dominated by antibodies expressing H and L chain products of the V.1 and V~22 germline genes (1, 2). These antibodies are identical to the antibody produced by the T15 plasmacytoma (3). The T15 clone of B cells is present in the mouse spleen at a higher frequency than any other individual clone of B cells and comprises >75% of all phosphocholine (PC)l-specific B cells in normal mice (4). The reasons for T15-Id dominance of the anti-PC immune response are not fully understood; however, idiotype network selection, binding-site affinity, and preferential V gene expression have I Abbreviation used in this paper: PC, phosphocholine. all been suggested. A. J. Feeney (5) has recently shown that over 50% of the V1:DFL16.1junctions produced in neonatal pre-B cells have prototypic T15 V-D junctional sequences, and has hypothesized that this preferential recombination of these germline genes would account for the dominance of T15 clones in the anti-PC repertoire. If the prototypic T15H chain is preferentially produced and this H chain can only form a PC-specific antibody with a K22 L chain, the subsequent receptor-driven selection and amplification of this clone into a long-lived B cell pool could account for T15-Id dominance. This hypothesis is supported by our recent observation (6, 7) that PC-specific B cells in M167 H chain transgenic mice are selected and amplified by a receptor-mediated, antigen-driven process after they emerge from ...