Salmonella translocates across an in vitro M cell model independently of SPI-1 and SPI-2

We have used an in vitro model of intestinal M cells to examine the mechanisms by which Salmonella enterica translocates across these specialized cells, which constitute a primary site of infection of the mammalian host. S. enterica can invade cultured cells by deploying a type III secretion system (TTSS) encoded within Salmonella pathogenicity island 1 (SPI-1) to translocate effector proteins into the host cell cytoplasm that trigger cellular responses, including prominent cytoskeletal rearrangements. After Salmonella enters the host cell, a second TTSS encoded in SPI-2 modulates intracellular trafficking and enables the bacteria to replicate within a modified vacuolar compartment. Within the host intestine, specialized antigen-sampling M cells, which reside in the epithelium overlying lymphoid tissues in the gut, are a preferential site of Salmonella invasion. The mechanisms of infection of M cells remain poorly defined and it is not known whether either SPI-1 or SPI-2 is required for infection of these cells. To address these questions we have employed an in vitro M cell model involving co-culture of polarized Caco-2 intestinal epithelial cells with Raji B cells. S. enterica serovar Typhimurium translocated across Caco-2/Raji co-cultures to a much greater extent than they cross native Caco-2 cell monolayers. Salmonella translocation was greatly reduced by heat treatment or fixation, suggesting that processes distinct from the sampling of inert particles are the main determinants of bacterial translocation. Translocation across both mono-cultured and co-cultured Caco-2 cells was partially inhibited by treatment with the dynamin inhibitor dynasore, but resistant to EIPA, an inhibitor of macropinocytosis. There was no difference between the abilities of wild-type Salmonella Typhimurium and mutants lacking multiple SPI-1 effectors to translocate across the M cell model, although the SPI-1 effector mutants were somewhat attenuated for translocation across native Caco-2 layers. There was also no difference between wild-type and SPI-2 mutants in M cell translocation. Together these data suggest that that SPI-1 and SPI-2 are dispensable for rapid M cell translocation and that infection at these specialized epithelial sites involves distinctive mechanisms that are not reliably modelled using conventional cell culture infection models.