The UspA1 and UspA2 proteins of Moraxella catarrhalis are structurally related, are exposed on the bacterial cell surface, and migrate as very high-molecular-weight complexes in sodium dodecyl sulfate-polyacrylamide gel electrophoresis. Previous analysis of uspA1 and uspA2 mutants of M. catarrhalis strain 035E indicated that UspA1 was involved in adherence of this organism to Chang conjunctival epithelial cells in vitro and that expression of UspA2 was essential for resistance of this strain to killing by normal human serum (C. Aebi, E. R. Lafontaine, L. D. Cope, J. L. Latimer, S. R. Lumbley, G. H. McCracken, Jr., and E. J. Hansen, Infect. Immun. 66:3113-3119, 1998). In the present study, isogenic uspA1, uspA2, and uspA1 uspA2 mutations were constructed in three additional M. catarrhalis strains: 012E, TTA37, and 046E. The uspA1 mutant of strain 012E had a decreased ability to attach to Chang cells. However, inactivation of the uspA1 gene in both strain TTA37 and strain 046E did not cause a significant decrease in attachment ability. Inactivation of the uspA2 gene of strain TTA37 did result in a loss of attachment ability. Nucleotide sequence analysis revealed that the predicted protein encoded by the uspA2 genes of both strains TTA37 and 046E had a N-terminal half that resembled the N-terminal half of UspA1 proteins, whereas the C-terminal half of this protein was nearly identical to those of previously characterized UspA2 proteins. The gene encoding this "hybrid" protein was designated uspA2H. PCR-based analysis revealed that approximately 20% of M. catarrhalis strains apparently possess a uspA2H gene instead of a uspA2 gene. The M. catarrhalis uspA1, uspA2, and uspA2H genes were cloned and expressed in Haemophilus influenzae cells, which were used to prove that both the UspA1 and UspA2H proteins can function as adhesins in vitro.Moraxella catarrhalis, an unencapsulated, gram-negative bacterium, can cause disease in both the upper and lower respiratory tracts (32). It has been estimated that approximately 20% of cases of acute bacterial otitis media in infants and young children are caused by this organism (6). M. catarrhalis is also associated with nearly one-third of infectious exacerbations of chronic obstructive pulmonary disease in adults (16). The ability of this organism to cause significant morbidity has resulted in increased efforts to develop an efficacious M. catarrhalis vaccine (35).Outer membrane proteins have received the most attention as possible M. catarrhalis vaccine candidates (9,19,20,31,33,43), and even M. catarrhalis lipooligosaccharide may contain potential vaccine components (15). A few of these outer membrane proteins, especially CopB (OMP B2) (4, 38), OMP CD (24), TbpA and TbpB (28), LbpA and LbpB (12), and UspA (ubiquitous surface protein A or HMW-OMP) (20, 26), which consists of two related proteins, UspA1 and UspA2 (2, 3), have been characterized in some detail. Furthermore, changes in expression of M. catarrhalis outer membrane proteins have been shown to affect the ability of thi...

Little is known about the virulence mechanisms employed by Haemophilus ducreyi in the production of genital ulcers. This Gram-negative bacterium previously has been shown to produce a soluble cytotoxic activity that kills HeLa and HEp-2 cells. We have now identified a cluster of three H. ducreyi genes that encode this cytotoxic activity. The predicted proteins encoded by these genes are most similar to the products of the Escherichia coli cdtABC genes that comprise the cytolethal distending toxin (CDT) of this enteric pathogen. Eleven of 12 H. ducreyi strains were shown to possess this gene cluster and culture supernatants from these strains readily killed HeLa cells. The culture supernatant from a single strain of H. ducreyi that lacked these genes was unable to kill HeLa cells. When the H. ducreyi cdtABC gene cluster was cloned into E. coli, culture supernatant from the recombinant E. coli clone killed HeLa cells. A monoclonal antibody that neutralized this soluble cytotoxic activity of H. ducreyi was shown to bind to the H. ducreyi cdtC gene product. This soluble H. ducreyi cytotoxin may play a role in the development or persistence of the ulcerative lesions characteristic of chancroid.

Cells of an attenuated live vaccine strain (LVS) of F. tularensis grown under iron-restricted conditions were found to contain increased quantities of several proteins relative to cells of this same strain grown under iron-replete conditions. Mass spectrometric analysis identified two of these proteins as IglC and PdpB, both of which are encoded by genes located in a previously identified pathogenicity island in F. tularensis LVS. Regions with homology to the consensus Fur box sequence were located immediately in front of the iglC and pdpB open reading frames (ORFs), and in silico analysis of the F. tularensis Schu4 genome detected a number of predicted 5 untranslated regions that contained putative Fur boxes. The putative Fur box preceding Francisella iron-regulated gene A (figA) had the highest degree of identity with the consensus Fur box sequence. DNA microarray analysis showed that nearly 80 of the genes in the F. tularensis LVS genome were up-or down-regulated at least twofold under iron-restricted growth conditions. When tested for possible siderophore production by means of the Chrome Azurol S assay, a wild-type F. novicida strain produced a large reaction zone whereas its figA mutant produced very little reactivity in this assay. In addition, a cross-feeding experiment demonstrated that this siderophore-like activity produced by the wild-type F. novicida strain could enhance the ability of the F. novicida figA mutant to grow under iron-restricted conditions. This study provides the first identification of iron-regulated genes in F. tularensis LVS and evidence for the production of a siderophore-like molecule by F. novicida.

Haemophilus influenzae type b (Hib) strains N0100 and COL10 were found to produce bacteremia in infant rats at a much lower frequency than other Hib strains previously tested. These relatively avirulent strains were the only Hib strains among 200 clinical isolates examined to date which failed to react with two Hib lipopolysaccharide (LPS)-specific monoclonal antibodies (MAbs). LPS analysis by sodium dodecyl sulfatepolyacrylamide gel electrophoresis showed that strains NOIOO and COL10 possessed LPS which migrated faster than the LPS of Hib strains that reacted with one of the two or with both of these MAbs. These observations suggested that the relative lack of virulence of strains NO100 and COL10 might be related to their unusual LPS phenotype. To determine whether alteration of LPS structure would affect the virulence of these strains, we identified and isolated isogenic LPS antigenic variants of strains NO100 and COL10 using the LPS-specific MAbs 4C4 and 5G8 in a colony blot radioimmunoassay. Antigenic variation of LPS was found to occur spontaneously in these two strains at a relatively high frequency in terms of both acquisition and loss of MAb reactivity (ca. 0.2 to 16.7%). LPS antigenic variants of strains N0100 and COL10 reactive with both MAbs 4C4 and 5G8 (4C4+ 5G8+) were more virulent in the infant rat model than their respective 4C4-5G8parental strains (P < 0.01). An antigenic variant of COL10 reactive with only MAb 4C4 (4C4+ 5G8-) was also significantly more virulent than its 4C4-5G8parent. These LPS antigenic variants with increased virulence synthesized altered LPS molecules which possessed apparent molecular weights higher than those of the LPS of the parental strains. Increased resistance of strain N0100 to the bactericidal activity of normal infant rat serum was associated with changes in LPS structure, while strain COL10 and its LPS variants were all uniformly resistant to serum bactericidal activity. Our results demonstrate that (i) spontaneous antigenic and phenotypic variation of LPS occurs at a relatively high frequency in some strains of Hib; (ii) the highermolecular-weight type of LPS is associated with the full expression of Hib virulence; (iii) LPS phenotype may not correlate with Hib serum resistance; and (iv) serum resistance of Hib is not an accurate indicator of virulence.

The utilization of heme bound to the serum glycoprotein hemopexin by Haemophilus influenzae type b (Hib) strain DL42 requires the presence of the 100-kDa heme:hemopexin-binding protein encoded by the hxuA gene (M. S. Hanson, S. E. Pelzel, J. Latimer, U. Muller-Eberhard, and E. J. Hansen, Proc. Natl. Acad. Sci. USA 89:1973-1977, 1992). Nucleotide sequence analysis of a 5-kb region immediately upstream from the hxuA gene revealed the presence of two genes, designated hxuC and hxuB, which encoded outer membrane proteins. The 78-kDa HxuC protein had similarity to TonB-dependent outer membrane proteins of other organisms, whereas the 60-kDa HxuB molecule most closely resembled the ShlB protein of Serratia marcescens. A set of three isogenic Hib mutants with cat cartridges inserted individually into their hxuA, hxuB, and hxuC genes was constructed. None of these mutants could utilize heme:hemopexin. The hxuC mutant was also unable to utilize low levels of free heme, whereas both the hxuA and hxuB mutants could utilize free heme. When the wild-type hxuC gene was present in trans, the hxuC mutant regained its ability to utilize low levels of free heme but still could not utilize heme:hemopexin. The hxuA mutant could utilize heme:hemopexin when a functional hxuA gene from a nontypeable H. influenzae strain was present in trans. Complementation analysis using this cloned nontypeable H. influenzae hxuA gene also indicated that the HxuB protein likely functions in the release of soluble HxuA from the Hib cell. These studies indicate that at least two and possibly three gene products are required for utilization of heme bound to hemopexin by Hib strain DL42.

Moraxella (Branhamella) catarrhalis is an important cause of disease in both the upper and lower respiratory tracts (35,48). This unencapsulated gram-negative coccobacillus has been shown to express a number of different outer membrane proteins on its cell surface, some of which are antigenically conserved (47, 49). At present, information about the M. catarrhalis gene products that are involved in the ability of this organism to colonize the mucosa of the nasopharynx and survive in this hostile environment is limited at best. Much effort has been expended recently on documenting the human immune response to selected M. catarrhalis surface-exposed proteins (6,12,25,53,65), providing evidence that these particular gene products are expressed in vivo during otitis media or infections of the bronchial tree. A few of these outer membrane proteins now have a function ascribed to them, mainly with respect to iron acquisition (7,9,10,15,42,43).In contrast, there is relatively little known about other surface proteins of M. catarrhalis that might be involved in the ability of this organism to colonize and survive in the nasopharynx (35). The CD outer membrane protein (33) has been shown to bind middle ear mucin in vitro (51), a function that could be involved in the colonization process or in the development of otitis media. The UspA1 protein has been shown to be an adhesin, at least in vitro (38), whereas both the UspA2 protein (38) and outer membrane protein E (50) have been implicated in serum resistance. Both UspA1 and UspA2, consistent with their functional activities, have been localized to the surface of M. catarrhalis, where they are accessible to antibodies (2, 45).Scott and colleagues (16, 17) correlated both hemagglutination activity and the expression of a 200-kDa protein by some M. catarrhalis isolates with the presence of a fibrillar surface array. In addition, Sasaki and colleagues reported that the 200-kDa protein expressed by M. catarrhalis was subject to phase variation in vitro (K. Sasaki, L. Myers, S. M. Loosmore, and M. H. Klein, Abstr. 99th Gen. Meet. Am. Soc. Microbiol., abstr. B/D-306, 1999) and determined the nucleotide sequence of the gene encoding this protein (54). In the present study, we used analysis of mutants to show that this protein, designated Hag (hemagglutinin), is involved not only in hemagglutination but also in autoagglutination and the binding of human immunoglobulin D (IgD) by M. catarrhalis strain O35E. In addition, we determined that the Hag protein, together with the UspA1 and UspA2 proteins (3), all form fibrillar projections on the M. catarrhalis cell surface.

SummaryAlthough previous studies using human cytokines in rabbits and rats have provided evidence ofthe participation oftumor necrosis factor a (TNF-(x) and interleukin lß (IL1ß) in the meningeal inflammatory cascade, the results obtained by several groups ofinvestigators have been discordant or, at times, contradictory. In the present study, homologous cytokines were applied to the rabbit meningitis model. Intracisternal administration of 102-10 1 IU of purified rabbit TNF-a (RaTNF-a) produced significant cerebrospinal fluid (CSF) inflammation. A similar response was observed after intracisternal inoculation of 5-200 ng of rabbit recombinant ILlß (rrIL1ß) . Preincubation of these two mediators with their specific antibodies resulted in an almost complete suppression of the CSF inflammatory response. In animals with Haemophilus influenzae type b lipooligosaccharide-induced meningitis, intracisternal administration of anti-rrlIrlß, anti-RaTNF-a, or both resulted in a significant modulation of meningeal inflammation. Simultaneous administration of 103 IU of RaTNF-a and 5 ng of rrILlß resulted in a synergistic inflammatory response manifested by a more rapid and significantly increased influx of white blood cells into the CSF compared with results after each cytokine given alone. These data provide evidence for a seminal role of TNF-a and IL1ß in the initial events of meningeal inflammation.

A monoclonal antibody (MAb) to Moraxella catarrhalis O35E bound to a surface-exposed epitope of a proteinaceous antigen of this organism. The antigen, designated UspA, was present in every strain of the pathogen tested in a colony blot RIA. UspA had a molecular mass on SDS-PAGE that varied between 300 and 400 kDa, depending on the individual M. catarrhalis strain. Passive immunization of mice with the UspA-reactive Mab enhanced pulmonary clearance of M. catarrhalis. Use of this Mab to screen a M. catarrhalis genomic DNA library permitted identification of a recombinant bacteriophage expressing the M. catarrhalis UspA protein. The recombinant UspA protein was used in Western blot analysis with sera from patients with M. catarrhalis pneumonia. Convalescent-phase sera but not acute-phase sera from these patients contained antibodies to this M. catarrhalis surface protein, indicating that M. catarrhalis strains growing in vivo express this molecule.