Molecular mechanism of suppression of T cell responses by Salmonella effector SteD

Abstract

Salmonella enterica is an intracellular pathogen that resides within specialised vacuoles and creates a niche that enables its survival and replication. Salmonella achieves this by translocation of virulence proteins, effectors, into the host cell by a type 3 secretion system. One such effector, SteD, depletes surface levels of three immune receptors from infected dendritic cells - antigen-loaded MHC class II, CD97 and CD86 – by inducing their ubiquitination and subsequent lysosomal degradation. A genome-wide mutant screen identified two host proteins are required for this process: the HECT E3 ubiquitin ligase WWP2 and its adaptor the transmembrane protein TMEM127(Alix et al., 2020). Here I show that although TMEM127 has a PPxY motif it is unable to activate WWP2. In contrast, SteD is an Ndfip and can activate WWP2 in the presence of TMEM127. Therefore, SteD and TMEM127 are co-activators. However, SteD is a non-canonical Ndfip, as it requires the co- activating protein TMEM127, it does not have a canonical activating PPxY motif and interacts with the C2 domain of WWP2 and not the WW domains. Remarkably, SteD and TMEM127 together mimic how the mammalian Ndfip Dishevelled-2 activates WWP2. I also show that the canonical Ndfip NDFIP2 interacts with multiple WW domains of WWP2 and that interaction via multiple L/PPxY motifs of NDFIP2 is required for interaction with and activation of WWP2. Furthermore, I show that SUSD6 is an Ndfip with a single LPxY motif and although SUSD6 can interact with multiple WW domains of WWP2, it is a weak activator of WWP2. This work has expanded our knowledge of how Ndfips activate WWP2. To our knowledge, SteD is the first description of a bacterial protein that activates an E3 ubiquitin ligase and is another example of how a bacterial protein co-opts the host machinery to interfere with the host immunity.