During infection of the bladder mucosa by uropathogenic Escherichia coli (UPEC), uroepithelial cells will often exfoliate, a process that benefits the host by helping to clear the infection but can also benefit the bacterium by promoting dissemination. However, the host and bacterial factors that drive uroepithelial cell exfoliation have remained poorly understood. Writing in Cell Host & Microbe, Dhakal and Mulvey now show that the UPEC pore-forming toxin α-haemolysin (HlyA) stimulates the proteolysis of host proteins involved in cell adhesion.

These data ... add to the growing list of secondary functions that can be attributed to bacterial pore-forming toxins.

Exfoliation of host cells generally requires the degradation of cell-to-cell adhesion factors, leading the authors to look at the effects of UPEC infection on host adhesion components. They found that UPEC infection of bladder epithelial cells (BECs) resulted in the complete loss of paxillin, a scaffold protein that can modulate the dynamics of cytoskeletal rearrangements. This effect was specific to uropathogenic strains of E. coli, and in a panel of UPEC isolates 60% were capable of triggering paxillin degradation in BECs. Interestingly, all the isolates that triggered paxillin loss were haemolytic on blood agar plates, suggesting a role for the pore-forming toxin HlyA. Accordingly, the authors found that a UPEC strain in which hlyA had been deleted was unable to stimulate paxillin degradation in BECs. The hly operon also encodes HlyC, an acyltransferase that acylates two conserved lysine residues in HlyA, allowing the toxin to oligomerize and form functional pores in the host cell membrane. As for hlyA, deletion of hlyC rendered UPEC unable to degrade paxillin in BECs. In addition, depletion of cholesterol from the host cell membrane or chelation of extracellular Ca²+ (cholesterol and Ca²+ are required for HlyA insertion into the membrane) blocked paxillin degradation. Importantly, degradation of paxillin did not require lytic concentrations of HlyA, as BECs remained intact and viable.

To investigate the proteolytic pathway by which paxillin is degraded, the authors used a range of pharmacological inhibitors to target different protease types. They found that on infection with UPEC there was a rapid activation of the caspases CASP3 and CASP7, which have important roles in regulating the inflammatory response to infection; however, inhibition of these proteases did not block paxillin degradation. Conversely, addition of the serine protease inhibitor tosyl-L-lysine chloromethyl ketone completely inhibited paxillin degradation. Furthermore, using a fluorescent probe specific to serine proteases, the authors observed a significant activation of serine proteases within 2 hours of infection by wild-type UPEC but not the hlyA mutant. Immunoprecipitation and mass spectrometry identified mesotrypsin, a serine protease that is generated through processing of three zymogens known as trypsinogens 3, 4 and 5. Silencing the expression of trypsinogen 4 using small interfering RNA led to a significant delay in the degradation of paxillin in UPEC-infected BECs. In addition, the authors observed that trypsinogen 4 localization in infected BECs changed from a diffuse cytosolic pattern to a more reticulated and punctate distribution at the cell periphery along microtubules, which is consistent with a role for mesotrypsin in modulating cell adherence in response to HlyA.

Finally, the authors looked for proteins other than paxillin that might be degraded by mesotrypsin. Using a panel of antibodies against 24 cellular proteins, 12 proteins were found to be degraded on infection of BECs by wild-type UPEC but not by an hlyA mutant, including proteins that function in cell adherence and signal transduction. One such protein was β-catenin, a negative regulator of the inflammatory transcription factor nuclear factor-κB.

Taken together, these data shed light on the processes underlying exfoliation of the bladder mucosa during UPEC infection and add to the growing list of secondary functions that can be attributed to bacterial pore-forming toxins.