Legionella pneumophila LegC7 effector protein drives aberrant endoplasmic reticulum:endosome contacts in yeast

Abstract

Legionella pneumophila is a facultative intracellular bacterial pathogen, causing the severe form of pneumonia known as Legionnaires' disease. Legionella actively alters host organelle trafficking through the activities of "effector" proteins secreted via a type-IVB secretion system, in order to construct the bacteria-laden Legionella-containing vacuole (LCV) and prevent lysosomal degradation. The LCV is created with membrane derived from host endoplasmic reticulum (ER), secretory vesicles and phagosomes, although the precise molecular mechanisms that drive its synthesis remain poorly understood. In an effort to characterize the in vivo activity of the LegC7/YlfA SNARE-like effector protein from Legionella in the context of eukaryotic membrane trafficking in yeast, we find that LegC7 interacts with the Emp46p/Emp47p ER-to-Golgi glycoprotein cargo adapter complex, alters ER morphology and induces aberrant ER:endosome interactions, as measured by visualization of ER cargo degradation, reconstitution of split-GFP proteins and enhanced oxidation of the ER lumen. LegC7-dependent toxicity, disruption of ER morphology and ER:endosome fusion events were dependent upon endosomal VPS class C tethering complexes and the endosomal t-SNARE, Pep12p. This work establishes a model in which LegC7 functions to recruit host ER material to the bacterial phagosome during infection by driving ER:endosome contacts, potentially through interaction with host membrane tethering complexes and/or cargo adapters.

Keywords: LegC7; Legionella pneumophila; SNARE proteins; Saccharomyces cerevisiae; VPS class C tethering complexes; cargo adapters; membrane fusion.

Figure 1.. LegC7 binds to Emp47 but does not colocalize with COPII machinery.

(A) α-myc immunoprecipitations from yeast strains containing LegC7, myc-Emp47, or both were separated via SDS-PAGE and immunoblotted for Emp47 and LegC7. Strains were grown to saturation in CSM medium at 30°C, harvested via centrifugation, resuspended in an equal volume of fresh CSM containing 1% raffinose and 1% galactose, and grown for an additional 4 h at 30°C before immunoprecipitation. The arrow indicates Emp47 (wild type or myc-tagged). The asterisk indicates a protein that cross reacts with the α-LegC7 antibody and serves as a loading control. (B) Yeast strains harboring LegC7-mRuby2 and GFP-tagged COPII coatamer subunit Sec13 were visualized to discern potential colocalization of LegC7 and COPII machiney. Strains were grown to saturation in CSM medium at 30°C, harvested via centrifugation, resuspended in an equal volume of fresh CSM containing 1% raffinose and 1% galactose, and grown for an additional 18 h at 30°C before imaging. Scale bars = 5 μ. (C) Colocalization of GFP and mRuby2 was quantified in the form of Pearson correlation coefficients (PCCs) for 100 cells per strain using the Coloc2 plugin in Fiji (ImageJ). Error bars represent 95% confidence intervals of the average PCC.

Figure 2.. An emp46Δemp47Δ double mutation and LegC7 expression both alter localization of Carboxypeptidase S.

Yeast strains harboring GFP-CPS were visualized to determine vacuolar localization and number of punctae. All strains were grown to saturation in CSM medium at 30°C, harvested via centrifugation, resuspended in an equal volume of fresh CSM containing 1% raffinose and 1% galactose, and grown for an additional 18 h at 30°C before imaging. (A-C) are representative images of cells for each strain screened. The white arrow indicates vacuolar localization and the white asterisks indication cells without vacuolar localization. Scale bars = 5 μ. Individual cells were analyzed to determine vacuolar localization of GFP (D) and number of punctae (E). ≥ 100 cells each per three independent experiments were analyzed; error bars represent ± the standard deviation between experiments; (*) = P < 0.05; (**) = P < 0.005; (***) = P < 0.0005.

Figure 3.. LegC7 alters ER morphology but does not affect Golgi morphology in a discernable manner.

Yeast strains expressing LegC7-mRuby2 and GFP fusions of ER translocon subunit Sec63p (A) or cis-Golgi GDP-mannose transporter Vrg4p (B) were visualized. All strains were grown to saturation in CSM medium at 30°C, harvested via centrifugation, resuspended in an equal volume of fresh CSM containing 1% raffinose and 1% galactose, and grown for an additional 18 h at 30°C before imaging. Arrows indicate patches where LegC7-mRuby2 and Sec63-GFP colocalize. Scale bars = 5 μ.

Figure 4.. Deletion of Class C tethering complex subunits suppresses LegC7-mediated growth inhibition.

(A) Yeast deletion mutants harboring either pYES2NT C or pYES2-LEGC7⁺ were grown to saturation in CSM at 30°C. For each strain, 1 OD₆₀₀ unit was harvested via centrifugation, resuspended in 1 mL of 0.9% NaCl, and serially diluted 1:10 four times. 5 μL of each dilution was plated onto CSM containing 2% glucose and CSM containing 1% galactose and 1% raffinose to induce LegC7 expression. (B) Models of the class C tethering complexes are displayed, with the core complex in gray and the Rab-specific subunits in blue and orange.

Figure 5.. LegC7 colocalizes with early endosomal fusion machinery more than that of late endosomes.

Yeast strains expressing LegC7-mRuby2 and GFP fusions CORVET-specific subunit Vps8p (A), Class C core complex subunit Vps33p (B), HOPS-specific subunit Vps41p (C), early endosomal Rab5 homolog Vps21p (D), and late endosomal Rab7 homolog Ypt7p (E) were visualized. All strains were grown to saturation in CSM medium at 30°C, harvested via centrifugation, resuspended in an equal volume of fresh CSM containing 1% raffinose and 1% galactose, and grown for an additional 18 h at 30°C before imaging. Scale bars = 5 μ. (F) Colocalization of GFP and mRuby2 was quantified in the form of Pearson correlation coefficients (PCCs) for 100 cells per strain using the Coloc2 plugin in Fiji (ImageJ). Error bars represent 95% confidence intervals of the average PCC.

Figure 6.. CORVET subunit Vps8p localizes to the ER during LegC7 expression, and endosomal tethering machinery is required for LegC7-induced alteration of ER morphology.

(A)Yeast strains containing Vps8-GFP, Sec63-RFP, and pYES2NT C or pYES2-LEGC7⁺ were visualized. (B) Colocalization of GFP and RFP was quantified in the form of Pearson correlation coefficients (PCCs) for 100 cells per strain in (A) using the Coloc2 plugin in Fiji (ImageJ). Error bars represent 95% confidence intervals of the average PCC. (C) Yeast strains containing Sec63-GFP and expressing LegC7-mRuby2 were visualized to observe the effect of a vps33Δ deletion on ER morphology during LegC7 expression. All strains were grown to saturation in CSM medium at 30°C, harvested via centrifugation, resuspended in an equal volume of fresh CSM containing 1% raffinose and 1% galactose, and grown for an additional 18 h at 30°C before imaging. Scale bars = 5 μ.

Figure 7.. LegC7 causes the degradation of ER luminal ATPase Kar2p, dependent upon endosomal fusion machinery and vacuolar proteases.

(A) Yeast strains containing pYES2NT C or pYES2-LEGC7⁺ were grown to saturation in CSM medium at 30°C, harvested via centrifugation, resuspended in an equal volume of fresh CSM containing 1% raffinose and 1% galactose, and grown for an additional 18 h at 30°C. Equal amounts of each strain were harvested via centrifugation, dounced, fractionated into 10 fractions, and fractions were TCA precipitated (Materials and Methods). Equal volumes of each fraction were separated via SDS-PAGE and immunoblotted for the ER luminal ATPase Kar2p, the cis-Golgi mannosyltransferase Och1p, and LegC7. (B) Strains were treated as in (A) but not fractionated. Equal volumes of whole cell extracts were separated via SDS-PAGE and immunoblotted for Kar2 and LegC7. The black arrows in (A) and (B) indicate Kar2 degradation product caused by LegC7. (C) Yeast deletion mutants harboring either pYES2NT C or pYES2-LEGC7⁺ were grown to saturation in CSM at 30°C. For each strain, 1 OD₆₀₀ unit was harvested via centrifugation, resuspended in 0.9% NaCl, and serially diluted 1:10 four times. 5 μL of each dilution was plated onto CSM containing 2% glucose and CSM containing 1% galactose and 1% raffinose to induce LegC7 expression.

Figure 8.. LegC7 expression oxidizes the ER lumen, but not in a vps33Δ mutant.

Yeast strains containing ER-targeted redox-sensitive eroGFP were grown to saturation in CSM medium at 30°C, harvested via centrifugation, resuspended in an equal volume of fresh CSM containing 1% raffinose and 1% galactose, and grown for an additional 6 h at 30°C. Cells were then analyzed via flow cytometry (n = 10⁵). Gates for cells without eroGFP (A) and heat-killed cells containing eroGFP (B) were used to eliminate low fluorescence and dead cell populations, respectively, from strains in panels (C – F) to generate panels (G – J). (K) Ratios of GFP fluorescence (Ex₄₀₅/Ex₄₈₈) were calculated, and ratios for (G) and (H) were normalized by the same factor such that the ratio for (G) = 1. ratios for (I) and (J) were normalized such that the ratio for (I) = 1. Three independent experiments were performed to generate ratios displayed in (K); error bars represent ± the standard deviation between experiments; (*) = P < 0.05. Panels (A – J) are representative plots, taken from one experiment.

Figure 9.. LegC7 may cause fusion of ER-derived and endosomal compartments.

(A) Model of the split-GFP system constructed for this study. Strains containing individual constructs were grown to saturation in CSM medium at 30°C for imaging. Scale bars = 5 μ. (B) Strains containing ER-targeted mRuby2-ScsΔTM-GFP₁₁ and endosome-targeted CPYss-GFP_1-10 and either pYES2NT C (left) or pYES2-LEGC7⁺ (right) were grown to saturation in CSM medium at 30°C, harvested via centrifugation, resuspended in an equal volume of fresh CSM containing 1% raffinose and 1% galactose, and grown for an additional 18 h at 30°C. Equal amounts of each strain were fractionated into 25 fractions (Materials and Methods), and 20 μL volumes of each fraction were measured in triplicate for GFP fluorescence and averaged. Error bars represent ± the standard deviation across three independent experiments. The remainder of each fraction was TCA precipitated, and equal volumes were separated via SDS-PAGE and immunoblotted for the ER luminal ATPase Kar2p, the cis-Golgi mannosyltransferase Och1p, the early endosomal t-SNARE Pep12p, the vacuolar ATPase subunit Vph1p, and LegC7. The black arrow indicates Kar2 degradation product and the bracket indicates fractions 19 and 20 where we see Kar2p degradation and a shifted fluorescence peak. (C) GFP fluorescence plots were combined for comparison; (*) = P < 0.05; (**) = P < 0.005.