Crystal structure of the N-terminal domain of the effector protein SidI of Legionella pneumophila reveals a glucosyl transferase domain.

The Gram-negative bacterium Legionella pneumophila is an accidental human pathogen that can cause a life-threatening respiratory infection called Legionellosis. In the course of infection, L. pneumophila injects more than 300 effector proteins into the host cell. The effector proteins modify the intracellular environment in order to create a stable compartment for proliferation within the host cell. The effector protein SidI has been shown to potently inhibit host translation upon translocation. SidI is able to interact with the translation elongation factor eEF1A, which has been hypothesized to be a target of SidI. A postulated glycosyltransferase domain in the C-terminal half may be responsible for the toxic effect of SidI. Here, we present the crystal structure of an N-terminal fragment of SidI containing residues 37-573. The structure is divided into three subdomains, two of which display a novel fold. The third subdomain shows close structural homology to GT-B fold glycosyltransferases. Based on structural analysis we predict that the two previously identified residues R453 and E482 assume roles in the catalytic activity of SidI. Furthermore, we show that the N-terminal fragment of SidI is able to directly interact with a postulated target, the translation elongation factor eEF1A.

Crystal structure of the metaeffector MesI (Lpg2505) from Legionella pneumophila.

Persistence and replication of the gram-negative bacterium Legionella pneumophila in the human host cell depend on so-called effector proteins that target diverse cellular functions and modulate them in favor of the pathogen. We solved the crystal structure of the L. pneumophila effector protein MesI de novo to a resolution of 2.2 Å. The 34 kDa polypeptide chain folds into two distinct α-helical domains. The larger C-terminal domain shows similarity to tetratricopeptide repeat proteins. Using size-exclusion chromatography, we confirmed that MesI binds tightly to full-length SidI and that deletion of either the N- or the C-terminus weakens the interaction. Based on the three-dimensional structure of MesI we suggest a possible binding mode for SidI and identified two homologs of MesI within the proteome of L. pneumophila that do not bind to SidI, but may act as specific inhibitors of other yet to be identified effectors.