A 69-kDa protein has been identified on the surface of the Gram-negative pathogenBordetela peilussis that can elicit a protective immune response in animal models. This protein is associated with virulent strains ofB. pertussis but its function has remained unclear. In this report we demonstrate that purified preparations of the 69-kDa outer membrane protein can promote the attachment of Chinese hamster ovary (CHO) cells. The interaction between the mammalian cells and this protein can be specifically inhibited by an Arg-Gly-Asp (RGD)-containing synthetic peptide that is homologous with a region found in the 69-kDa protein sequence. These studies indicate that a specific cell binding site containing an Arg-GlyAsp sequence may be involved in the interaction of this bacterial protein with mammalian cell surfaces. To further investigate the role of this protein as a bacterial adhesin, a mutant of B. pertussis W28 that does not express the 69-kDa protein was constructed using the plasmid vector pRTP1. This mutant was 30-40% less efficient at adhering to CHO cells and to human HeLa cells than was the parent strain. These data support a role for this 69-kDa outer membrane protein in the attachment ofB. pertussis to mammalian cells. We propose the name "pertactin" for this protein.

Protein P.69 is localized on the outer membrane of Bordetella pertussis and is one of the virulence factors believed to contribute to the disease state of whooping cough. We demonstrate that protein synthesis of P.69 is under genetic control of the vir locus. Using oligonucleotide probes derived from the protein sequence of a cyanogen bromide fragment, we have cloned the gene for P.69 from B. pertussis CN2992. Analysis of the DNA sequence reveals a G+C-rich gene capable of encoding a protein of 910 amino acids with a Mr of 93,478, suggesting that P.69 is a processed form of a larger precursor. In common with some of the genes in the pertussis toxin operon, the sequence CCTGG was found 5' to the ATG initiation codon. At the 3' end, 29 bases after the TAA stop codon, the sequence GTTTTTCCT was found and may have some function in transcription termination. A full-length clone ofthe gene for P.69 carried by the cosmid pBPI69 was unable to direct the expression of P.69 protein in an Escherichia coli host. The generation of P.69-fusion products allowed the detection of P.69-specific protein products synthesized in E. coli.Bordetella pertussis, Bordetella parapertussis, and Bordetella bronchiseptica are closely related pathogenic organisms. In humans, whooping cough is caused by B. pertussis or B. parapertussis. B. bronchiseptica is an animal pathogen, but this organism has also been isolated from children with whooping cough-like disease (1). There are many similarities between the diseases caused by these three species. All species undergo phenotypic changes that are genetically modulated by the vir locus, which regulates the expression of many proteins, some of which are apparently correlated with bacterial virulence and immunogenicity (2). These include filamentous hemagglutinin, adenylate cyclase (AdeCase), hemolysin, pili, and in B. pertussis, pertussis toxin (PTX

FOR the development of an effective gonococcal vaccine it is likely that knowledge of virulence factors is required so that antigens of possible protective value can be defined. It is then necessary to determine whether such protective antigens are produced by gonococci cultured in vitro.Much of the work being carried out at the present time on gonococci was stimulated by the discovery by Kellogg et al. (1963) of four different colonial types designated Tl-T4. Although all the colony types derived from a single strain were virulent for human volunteers for up to 38 selective passages, only T1 and T2 colonies were able to produce the disease after several hundred passages (Kellogg et al., 1968).Work on the gonococcus has been hampered by the lack of an animal model, since the only susceptible species appear to be man and chimpanzees (Lucas et a!., 1971). Recently, however, there have been reports that claim to measure virulence by simple in-vivo procedures. Ark0 (1972Ark0 ( , 1973 used implanted capsules in laboratory animals to cultivate gonococci in vivo and showed that the avirulent types when injected into these were cleared much faster than virulent types. The model was also useful in the study of active immunisation (Arko, 1974). Buchanan and Gotschlich (1973) demonstrated that T1 and T2 gonococci survived for longer periods than their T3 and T4 counterparts when inoculated on the chorioallantoic membrane of embryonated eggs. Bumgarner and Finkelstein (1 973), although failing to support this finding, showed remarkable differences in virulence between the types on intravenous injection into 1 1-day-old chick embryos. In addition, experiments with human polymorphs have shown that T1 cells are retained to a greater extent and are more resistant to digestion than T4 cells (Thomas, Hill and Tyeryar, 1973) and that virulent cells have greater resistance to phagocytosis (Ofek, Beachey and Bisno, 1974). Human bactericidal serum (Thomas et al., 1973) or rabbit antiserum to T2 gonococci enhanced phagocytosis of virulent but not avirulent cells (Ofek et al., 1974).Although it is possible by laboratory culture to distinguish colonies of Neisseria gonurrhoeae that are believed to be virulent, the factors responsible for virulence are far from clear. Jephcott, Reyn and Birch-Andersen (1971) and independently Swanson, Kraus and Gotschlich (1971) demonstrated that gonococci of colonial types T1 and T2 were piliate whereas those of T3 and T4 were not. As this appeared to be the most significant morphological difference between the virulent and avirulent colony types it was thought that piliation might be the factor that conferred virulence. Although in describing these appendages the term fimbriae has a prior claim over that of pili (Brinton, 1965 ;Duguid, Anderson and Campbell, 1966), the latter is in more general use as applied to gonococci and is used throughout the

The gene prn encoding the outer-membrane protein P.70 (pertactin) from Bordetella parapertussis has been cloned in Escherichia coli and its DNA sequence determined. Analysis of the DNA sequence reveals that the gene has an open reading frame comprising 922 amino acids capable of encoding a protein with a molecular weight of 95,177 (P.95). In vivo processing of this precursor yields a protein with an estimated Mr of 70 kDa (P.70) which is located on the surface of B. parapertussis. Homology between the prn gene from B. parapertussis and that from Bordetella pertussis is 91.3%. The homology is 93% when the protein sequence of P.95 is aligned with that of P.93 from B. pertussis. The major differences between the P.70 pertactin from B. parapertussis and the P.69 pertactin from B. pertussis occur in the number of reiterated units within the repeat motifs found in both proteins; the sequence Gly-Gly-Xaa-Xaa-Pro is repeated four times in the P.70 pertactin, and five times in the P.69 pertactin, while the sequence Pro-Gln-Pro occurs nine times in P.70 pertactin and five times in P.69 pertactin. Cloning of the gene for P.95 in an E. coli expression vector results in the synthesis of a protein that mimics native gene expression in B. parapertussis, i.e. the P.95 protein is synthesized and subsequently processed to yield the P.70 form of the protein on the surface of the cell.

No abstract