To the editor:

We are writing in response to the letter of Phillips et al.1, which addresses the identity of the cellular tyrosine kinase(s) required for the formation of actin-filled membrane protrusions (pedestals) induced by enteropathogenic Escherichia coli (EPEC). Previous reports have established that the EPEC virulence factor Tir is translocated into the host plasma membrane and binds to the EPEC outer membrane protein intimin2, thereby allowing tight adhesion of the bacterium to the host cell. Tyrosine phosphorylation of Tir on Y474 is required for recruitment of Nck, N-WASP and the Arp2/3 complex, as well as for actin polymerization, which constructs and braces the pedestal3,4. An outstanding question has been the identity of the host tyrosine kinase(s) responsible for phosphorylating Tir. Phillips et al. provide evidence that the Src family kinase c-Fyn is both necessary and sufficient for Tir phosphorylation and actin pedestal formation induced by EPEC. Importantly, Phillips et al. form actin pedestals by using a dual bacterial system in which the haemagglutinin (HA)-tagged Tir was introduced into the host cell by an EPEC strain lacking wild-type Tir and intimin (“priming”); a K12 strain expressing intimin in the outer membrane was then used to cluster Tir and induce pedestal formation (“challenge”).

Our published work5 indicates that the situation after infection with a wild-type strain of EPEC (WT-EPEC, strain 2348/69) using a standard infection protocol is more complex. In particular, we demonstrate that cells lacking the three kinases Src, Fyn and Yes (the same cells used by Phillips et al.), are still capable of forming actin pedestals when infected with WT-EPEC, suggesting that Fyn is not necessary for pedestal formation5. Using WT-EPEC, we also show that several tyrosine kinases, including Abl and Arg, localize and are persistently activated beneath attached EPEC, and that Abl and Arg are each sufficient for pedestal formation, but neither alone is necessary. Indeed our results suggest that several functionally redundant kinases can suffice, perhaps allowing the bacterium to increase its host range. We cannot rule out that Fyn, similar to Abl and Arg, functions in a redundant fashion; however, our data indicate that phosphorylation and dephosphorylation of Tir are dynamic and require sustained local kinase activity; furthermore, Fyn is not evident beneath attached EPEC in our experiments, nor beneath K12-intimin strains in the experiments of Phillips et al. Thus, the dual bacteria system used by Phillips et al. has different requirements for host effector molecules than wild-type EPEC strains. The apparent incongruities between data obtained with the dual bacteria approach of Phillips et al. and that obtained with WT-EPEC raise important questions relevant to treating disease caused by pathogenic E. coli. Kinases phosphorylating Tir represent targets for drug development, and our data suggest that effective drugs will probably have to target multiple kinases.