25–1 h. The OD600 nm of cultures were normalized to allow comparison of secreted protein levels, pelleted as before, and the supernatants were then filtered through 0.22-μm pore-size filters (Millipore). A 10% (w/v) final concentration of trichloroacetic acid (BDH Laboratory Supplies, UK) was used to precipitate the proteins as described previously (Leyton et al., 2003). Supernatant proteins were separated by sodium dodecyl sulfate-polyacrylamide electrophoresis (SDS-PAGE) (Sambrook et al., 1989) and detected by staining with Coomassie brilliant blue R250 (BDH Laboratory Supplies). Overnight E. coli HB101 cultures transformed with pPetssPet, pMBPssPet, pDsbAssPet and pPhoAssPet were diluted

1 : 100 into a fresh medium, grown Akt inhibitor to an OD600 nm=0.2 and induced by the addition of isopropylthiogalactoside (Sigma-Aldrich) at a concentration of 0.5 mM until an OD600 nm=0.8 MDV3100 was reached. The bacteria were pelleted and supernatant fractions were prepared as described above. Pet was localized by SDS-PAGE followed by Western immunoblotting (Sambrook et al., 1989). For cytotoxicity assays, Pet was expressed as described above and clarified supernatants were then concentrated 100-fold using 50 000 MW cut-off Vivaspin 20 columns (Sartorius Stedim Biotech, France) at 4 °C. Assays for cytotoxicity

were carried out as described previously (Guyer et al., 2000) using cultured HEp-2 cells. Cells were stained with Giemsa (Sigma-Aldrich) and observed for morphological changes by bright-field microscopy using a Leica DM-RB HC fluorescence phase-contrast microscope. Eslava et al. (1998) showed that the Pet signal peptide comprises 52 amino acids spanning residues M1–A52 (Fig. 1). Szabady et al. (2005) demonstrated that although the EspP ESPR is not required for inner membrane translocation, deletion of the ESPR inhibited the translocation of the protein across the outer membrane only due to incomplete folding of the EspP passenger domain in the periplasm. However, we previously demonstrated that site-directed mutagenesis of

the conserved residues within the Pet ESPR had only mild effects on Pet biogenesis (Desvaux et al., 2007), suggesting that deletion of the ESPR from the native Pet signal peptide would not abolish the ability of this construct to secrete Pet into the extracellular space. To investigate this hypothesis, an ESPR deletion construct was created in pBADPet in which expression was controlled by an arabinose-inducible pBAD promoter (Fig. 1) and monitored by SDS-PAGE analysis of supernatant fractions for the ability of cells containing this construct to secrete Pet. In contrast to the work on EspP carried out by Szabady et al. (2005), we demonstrated that the passenger domain of Pet (108 kDa) was released into the culture supernatant by cells containing the ESPR deletion mutant, pBADPetΔN1H1, with the level of secretion equaling that of the wild type at all stages of growth (Fig. 2).