Next: -defensins Up: Other Cyclic Peptides Previous: Cyclic Bacteriocins
The largest of the known circular proteins, and the only ones to have their biosynthetic pathway determined are a class of transfer proteins acting in bacterial conjugation and type IV secretion systems. These proteins are from 73 to 78 residues in length and constitute the major subunit of conjugative pili in some type IV secretion systems. The proteins are plasmid-expressed from the so called broad host range plasmids. An example of this class of plasmid is RP4 which was originally discovered in the common Gram-negative bacteria Pseudomonas aeruginosa [150,151]. The RP4 plasmid endows resistance to a number of antibiotics including ampicillin, kanamycin and tetracycline  and codes for at least 74 genes . RP4 containing bacteria produce a tube-like extracellular filamentous structure, the ``sex pilus'', which establishes bacterial conjugation and facilitates the transfer of single stranded DNA. These filamentous structures consist of multiple subunits of a circular protein called the pilin. Although no structural information has been obtained for the pilin it is known that, like AS-48, they are susceptible to proteolysis by trypsin and chymotrypsin  and that they appear to be cyclised in a two step reaction facilitated by a signal peptidase-like protease.
The pilin subunit is expressed as a 145 residue prepro-protein from the trbC gene [153,154]. Figure 1.19 sets out the steps in the production of the mature cyclic pilin. Firstly, the 27 residue C-terminal pre-protein is cleaved by an unknown protease and subsequently the pro-TrbC has its 36 residue signal peptide removed by the host's leader peptidase LepB, which thereby attaches the intermediate TrbC to the inner membrane. Two transmembrane helices within TrbC enable the intermediate to fold within the membrane, hence positioning the N and C-termini in close proximity. It has been proposed that the serine protease TraF then cleaves a C-terminal tetrapeptide from TrbC and the proximate N-terminal aminolyzes (Figure 1.20) the TrbC-acyl-TraA complex to yield the circular pilin .
The mature pilin is cyclised between Ser37 and Gly114 and mutagenesis experiments have shown that these two residues appear to be essential for the reaction. Cyclisation efficiency was significantly reduced with S37T and G114S mutants and activity was completely abolished with a S37C mutant . Interestingly, in no mutant was cleavage of the tetrapeptide observed without subsequent cyclisation, suggesting that the cyclisation reaction uses conserved energy from the cleavage reaction and that the two events cannot be uncoupled. The cyclising enzyme, TraF, which is similar in sequence to known signal peptidases , appears to contain a catalytic dyad consisting of S37 and K89. Mutagenesis experiments on this enzyme showed loss of activity when these residues were removed . Additionally, D155 was found to be essential for activity and it is possible that this residue acts as a proton acceptor during the aminolysis reaction. As the only cyclisation mechanism yet characterised it is intriguing that a serine protease, with similarities to typical signal peptidases, catalyses both the cleavage and ligation of the precursor protein. As we have seen proteases are suspects in the cyclisation reactions of the cyclotides, and SFTI-1 and it is possible that similar reactions occur in those systems.
Further strengthening this idea, the circular pilus characteristic of a number of other conjugative systems. These include the pertussis toxin operon of Bordetella pertussis, which encodes the pilin like protein Pt1A , and two related virulence operons from Brucella suis  and Brucella abortus  which contain the virB2 genes. Each of these proteins bears sequence similarity with TrbC , however, in the case of VirB2 marked differences do exist in the processing mechanism. The C-terminal pre-peptide of VirB2 is not cleaved before the cyclising reaction and, significantly, the plasmid containing virB2 contains no homologue to TraF and it has been shown that no plasmid borne gene is necessary for the cyclisation of the gene product. This would appear to suggest that a host-chromosome encoded signal petidase or other factor performs the cyclising reaction. The possible processing of VirB2 by endogenous proteases once again highlights the possibility of precursor proteins providing the driving force for backbone cyclisation; if the cyclising reaction in this case is mediated by a signal peptidase with a generic function then it suggests that the important factor in the cyclising of these proteins is the sequence of the cyclised protein. In this scenario existing cellular machinery, such as proteases, are the unspecialised mediators of cyclisation and evolution of cyclisation occurs in the processed proteins themselves. This resolves difficulties that could arise by contemplating the simultaneous evolution of cyclised proteins and a specialised mechanism to cyclise them.
Next: -defensins Up: Other Cyclic Peptides Previous: Cyclic Bacteriocins Jason Mulvenna