|dc.description.abstract||Legionella pneumophila is a Gram-negative intracellular pathogen that causes Legionnaires’ disease and Pontiac fever in elderly or immunocompromised humans. The ability of Legionella to thrive within host cells depends on the Legionella-containing vacuole (LCV) which, in turn, relies on the activity of secreted effector proteins for its formation. Effectors are highly variable in structure and function, and functional redundancy is prevalent among them. Consequently, relating structural data to function provides an attractive avenue of research into molecules which are unlikely to exhibit a phenotype upon gene deletion. Our lab relies on X-ray crystallography for macromolecular structure determination. Structural data may point to a function for the protein of interest, which can be verified using mutagenesis, biochemical assays or some combination thereof. This dissertation explores the structure and putative function of effectors LpnE(lpg2222), MavE(lpg2344) and MavL(lpg2526).
LpnE (Legionella pneumophila Entry) is a Sel1-like repeat (SLR) protein implicated in host cell invasion. During infection, a eukaryotic polyphosphate 5-phosphatase, called Oculocerebrorenal syndrome of Lowe protein 1 (OCRL1), is recruited to the LCV by an interaction with LpnE and restricts bacterial replication by an unknown mechanism. The crystal structure of His-LpnE(73-375) reveals a typical SLR super-helix with a concave surface implicated in protein-protein interactions. Herein, critical residues promoting the LpnE-OCRL interaction are uncovered using size exclusion chromatography with multi-angle light scattering (SEC-MALS). In addition, we show that LpnE localizes to cis¬-Golgi using its signal peptide. These findings are compiled into a mechanistic hypothesis where: (1) LpnE localizes to the LCV by its predicted signal peptide. (2) OCRL binding prevents liberation of LpnE from the LCV and (3) renders LpnE unable to promote infection by mediating protein-protein interactions in the cytosol.
MavE is one of many proteins identified as a secreted effector based on its ability to rescue LCV localization of a translocation deficient SidC (SidCΔ100). Our collaborator, Dr. Yousef Abu-Kwaik, has obtained a unique phenotype for Δlpg2344 (MavE) mutants, in which the LCV fuses with lysosomes (unpublished data). He suggests that MavE interacts with proteins harbouring phosphotyrosine-binding domains (PTBs) using its NPxY motif. The recruitment of these binding partners may impede autophagic trafficking. The crystal structure of MavE(39-172) presented in this dissertation has an overall structure reminiscent of the grass pollen allergen, Phlp 5b, with the NPxY motif located on a loop of poorly defined electron density. This loop has no counterpart in Phlp 5b and has flexibility that may accommodate protein-protein interactions. These structural data corroborate the proposed role of the NPxY motif while revealing a scaffold domain previously seen in the grass pollen allergen, Phlp 5b.
MavL is another secreted effector identified in the same manner as MavE. Presently, there is little published data available on the function of MavL. Elizabeth Hartland et al. have found by yeast two-hybrid that an E2 ubiquitin conjugating enzyme called Ube2q1 interacts directly with MavL, although we were unable to reproduce this interaction in vitro. The crystal structure of MavL(42-435) reveals an ADP-ribose binding macrodomain with homology to those that recognize mono-ADP-ribosylated targets. We confirmed the interaction of MavL and ADP-ribose by isothermal titration calorimetry (ITC), giving a dissociation constant of 13µM. Intriguingly, MavL contains a pair of neighboring aspartate residues in the same location as the catalytic glutamates of poly-ADP-ribose glycohydrolase (PARG) enzymes. We propose that MavL exhibits either ADP-ribose reader or eraser activity. Further studies are needed to investigate the role of ADP-ribosylation in MavL functionality.||