FHA is encoded by the fhaB gene and synthesized as a precursor of 367 kDa, the C-terminal portion of which is cleaved to yield the mature 220kDa FHA protein28. Mature FHA is exported and expressed on the bacterial surface, and is also released from the cell surface into the environment. FHA has several different binding activities that implicate it as an important adhesin. An Arg–Gly–Asp (RGD) amino acid motif promotes adherence to monocytes by the leukocyte-response integrin/integrin-associated protein complex and complement receptor type 3 (CR3). A carbo-hydrate-recognition domain allows adherence to cilia, and a lectin-like binding domain promotes adherence to sulphated carbohydrates and heparin, which are found at the surface of the respiratory tract. Despite these binding activities, in vivo studies have failed to define clearly the role of FHA in Bordetella pathogenesis. Some have failed to observe a difference between wild-type and an FHA-mutant B. pertussis in a mouse model of infection, whereas others have observed B. pertussis FHA mutants to be deficient in tracheal colonization. A protease, SphB1, was recently shown to be required for the secretion of FHA. B. pertussis SphB1 mutants were impaired in their ability to colonize the mouse respiratory tract. Instillation of purified FHA into the respiratory tract before inoculation with the SphB1 mutants, or co-infection with wild-type bacteria alleviated the attenuation of the mutants. This indicates that release of FHA from the bacterial surface contributes to the colonization process. Such contradictions might derive from the use of the human-specific pathogen B. pertussis in a mouse model. Studies using B. bronchiseptica indicated that FHA is required, but is not sufficient, for colonization of the rat trachea.
It has been proposed that FHA forms a hairpin-like rod. Its amino acid sequence contains two regions of different, imperfect direct repeats of 19 residues, known as repeat regions 1 and 2 (R1 and R2), which are proposed to form β-sheet structures which comprise much of the hairpin shaft. Both the RGD and carbohydrate recognition domains are located at the tip of the hairpin shaft, and might be ideally placed for interaction with host structures. Study of the amino-terminal segment of FHA led to the proposal of an alternative structure in which the overall shape of the molecule is a rod, but in which the shaft comprises β-helical segments that are derived mainly from the repeat regions. In this model, the CR3-binding domain, the RGD motif and the carbohydrate-recognition domain are located in the middle of the shaft.
Genome sequence information revealed that the fhaB sequences of the three bordetellae are highly conserved. The B. bronchiseptica FhaB protein is predicted to contain copies of R1, whereas the B. pertussis and B. parapertussis FhaB proteins are predicted to each contain only 38 copies. The FhaB protein of B. bronchiseptica might therefore be expected to form a longer shaft than the FhaB protein of B. pertussis or B. parapertussis. Previous studies have also determined that B. bronchiseptica FhaB contains more copies of R1 than the FhaB of B. pertussis; however, the B. bronchiseptica strain that was studied (strain GP1) contained only 40 copies of R1. So, the repeat number of R1 could vary not only between species, but also between strains of the same species. The consequences of this variability should be further investigated. An FHA homologue was recently characterized in Bordetella avium. Although the amino acid sequence of this FhaB is clearly different to those of the three mammalian-adapted species, B. avium FHA seems to have a similar role in the infection biology of this species. The completion of the B. avium genome sequence project will enable a full evaluation of B. avium FHA-like genes and their similarity to those described here. Two other genes, fhaS and fhaL, which are predicted to encode FHA-like proteins, have also been identified in the Bordetella genome sequences.
The B. bronchiseptica genome contains a locus that seems to encode the bio-synthesis of Type IV secretion system protein ptlF homolog (ptlf processs) Tfg and which is absent from the B. pertussis and B. parapertussis genomes. Tfg are distinguished from other pili by their polar location. They have several functions, including adherence of pathogens to host cells, twitching and social motility and DNA uptake in Neisseria gonorrhoeae and Bacillus subtilis. The Tfp secretion machinery has homology to the type II secretion apparatus, which forms the general secretory pathway. Tfp fibres are formed from multimers of the major pilin subunit, although minor subunits might also be present. The identity of the major pilin subunit gene in the putative B. bronchiseptica tfp locus is not obvious. However, there are several genes that could encode either the major or minor pilin subunits on the basis of a highly conserved amino-terminal leader sequence that is characteristic of Tfp pilins and is cleaved by the prepilin peptidase PilD (BB0792). After cleavage, the N-terminal residue is often phenylalanine, which subsequently becomes N-methylated. Many of the predicted proteins of the tfp locus are also homologous to type II secretion apparatus proteins. We propose that this locus encodes Tfp biosynthesis on the basis that the predicted protein product of BB0791 is approximately 70% identical to PilT proteins of other Tfp secretion systems throughout the entire protein. PilT is specific to tfp loci and is responsible for pilus retraction, which is required for twitching motility56. There are no reports of Tfp secretion systems in the bordetellae. Future work is likely to investigate the ability of this locus to direct expression of the Tfp secretion apparatus and characterize its role in the adherence of B. bronchiseptica to host structures.
B. bronchiseptica encodes a type III secretion system (TTSS). A TTSS is an export apparatus that delivers specific effector proteins into host cells with subsequent alterations in host-cell behaviour. The B. bronchiseptica TTSS is involved in the alteration of host immune cell function, which reduces the effectiveness of the host response to B. bronchiseptica infection. Previously, the locus had been shown to be present in B. parapertussis and B. pertussis. An ovine isolate of B. parapertussis was shown to transcribe some TTSS genes, but information regarding the expression of the TTSS locus in B. pertussis is contradictory. One study failed to demonstrate expression of some genes of the locus, whereas another study demonstrated the transcription of TTSS genes. The genome sequence information showed that, in B. parapertussis, two of the gene— one regulatory (BPP2241) and one structural (BPP2262)— are pseudogenes, which might indicate that the locus is non-functional in B. parapertussis, at least in this human isolate. However the B. pertussis locus is intact, which might indicate that there are mutations in unlinked regulatory genes that are required for TTSS expression. Some of this confusion might stem from the fact that studies of the expression of the B. pertussis TTSS locus are incomplete, and have only examined a small number of the genes in the locus. Future work is likely to investigate the expression of the locus as a whole, at both the translational and transcriptional levels.
 Bishop R E. The lipid A palmitoyltransferase PagP: molecular mechanisms and role in bacterial pathogenesis [J]. Molecular Microbiology, 2005, 57(4):900-912.
 The bordetellae: lessons from genomics. Nat Rev Microbiol. 2004 May; 2(5):379-90.
 Cherry, J. D. Pertussis — the trials and tribulations of old and new pertussis vaccines. Vaccine 10, 1033–1038 (1992). 2.
 Cherry, J. D. Historical review of pertussis and the classical vaccine. J. Infect. Dis. 174 (Suppl.), S259–S263 (1996).
 Heininger, U. et al. Clinical characteristics of illness caused by Bordetella parapertussis compared with illness caused by Bordetella pertussis. Pediatr. Infect. Dis. J. 13, 306–309 (1994).