Fig 1.
Infection with R. rickettsii Sheila Smith causes dispersal of the trans-Golgi network.
A) Representative images showing the trans-Golgi network is dispersed in cells infected with the highly virulent Sheila Smith strain of R. rickettsii but not in cells infected with the avirulent Iowa strain or uninfected control cells. The cis-Golgi apparatus remains largely intact in cells infected with either strain. Vero cells were infected at an MOI of 1 and fixed at 24 hpi. The cis-Golgi protein GM130 (red) and the trans-Golgi protein TGN46 (red) were stained with specific antibodies. Rickettsia are labeled with anti-rOmpB specific monoclonal antibodies (green). Nucleic acids were stained with DAPI (blue). Bar = 10 μm. B) Immunoelectron microscopy confirms dispersal of the trans-Golgi network in cells infected with R. rickettsii Sheila Smith but not Iowa. The cis-Golgi apparatus appears vacuolated in cells infected with R. rickettsii Sheila Smith in contrast to cells infected with R. rickettsii Iowa or uninfected cells. Vero cells were infected at an MOI of 1 and fixed at 48 hpi. Primary antibodies targeted GM130 or TGN46, followed by horseradish peroxidase conjugated secondary antibodies and diaminobenzidine-based detection. Bar = 1 μm.
Fig 2.
A) Lectin labeling of the Golgi apparatus in uninfected Vero cells (Uninf) or infected with R. rickettsii (green) strains Sheila Smith or Iowa for 24 hr. To confirm dispersal of the TGN using an antibody-independent labeling method, we employed the lectins wheat germ agglutinin (WGA, red) and Helix pomatia agglutinin (HPA; red). HPA selectively binds to terminal α-N-acetylgalactosaminyl residues–intermediate sugars added to serine and threonine residues in cis-Golgi cisternae. WGA binds sialic acid and N-acetylglucosaminyl residues and labels predominantly mature glycoproteins in the TGN [6,7]. Labeling with fluorescently labeled HPA resulted in condensed signals resembling those observed for GM130 while labeling for WGA was dispersed when cells were infected with R.rickettsii Sheila Smith. Bar = 10 um. B) Labeling of the Golgi apparatus with fluorescent C6-NBD-ceramide showed dispersed signal in cells infected with R. rickettsii Sheila Smith, but not for uninfected or R. rickettsii Iowa infected cells. Bar = 10 um.
Fig 3.
The effector protein RARP2 of R. rickettsii Sheila Smith causes dispersal of the trans-Golgi network.
A) Schematic of RARP2 variants showing R. rickettsii Sheila Smith (SS-RARP2) contains 10 ankyrin repeats units relative to RARP2 of R. rickettsii Iowa (Iowa-RARP2) which contains only 3 Ankyrin repeats. The cleavage site of the type IV secretion signal (green triangle), a predicted catalytic cysteine (red asterisk) and non-synonymous SNPs (white bar in the Iowa protein schematic) are identified. B) Expression of RARP2 of R. rickettsii Sheila Smith (Iowa:SS-RARP2), but not RARP2 of the Iowa strain (Iowa:Io-RARP2) in the avirulent Iowa strain causes dispersal of the trans-Golgi network. Genes were cloned into the vector pRAMF2, transformed into R. rickettsii Iowa and used to infect Vero cells. Cells were fixed 24 hpi and the trans-Golgi protein TGN46 (red) and rickettsiae (green) were detected by specific antibodies. Nucleic acids were stained with DAPI (blue). Bar = 10μm. C) Quantitation of the TGN46 observed in cells infected with R. rickettsii Sheila Smith or Iowa or infected with R. rickettsii Iowa expressing SS-RARP2. GM130 and TGN46 localization was determined for at least 350 host cells per strain for each of three biological replicates and normalized to uninfected control cells. Shown is the mean +/- the S.E.M. Statistics were performed using One-way ANOVA and post-hoc Tukey test. Significant differences to the uninfected control are indicated (**** p<0.0001 and * p < 0.05).
Fig 4.
Mutation of a putative catalytic cysteine residue or deletion of the ankyrin-repeat domain abolishes TGN dispersal by SS-RARP2.
R. rickettsii Iowa expressing SS-RARP2 carrying a mutation of the putative catalytic cysteine at position 109 to an alanine (SS-RARP2-C109A); R. rickettsii Iowa expressing SS-RARP2 with the ankyrin-repeat domain deleted (SS-RARP2ΔAnk); R. rickettsii Iowa expressing SS-RARP2 (SS-RARP2) were used to infect Vero cells. Cultures were fixed at 48 hpi and stained for TGN46 (red) and R. rickettsii (green). Bar = 10 μm.
Fig 5.
Ectopic expression of SS-RARP2 causes dispersal of the trans-Golgi network.
Sheila Smith RARP2 (SS), Iowa RARP2 (Iowa), or the SS-RARP2-C109A mutant (C109A) were expressed in Vero cells as EGFP fusions from pEGFP-C1 and probed for TGN46 (red). Parental pEGFP-C1 transfected cells served as a negative control. The TGN is dispersed in cells expressing SS-RARP2 (green), but not in cells expressing Iowa-RARP2, SS-RARP2-C109A, or negative control cells expressing EGFP (green). Ectopically expressed SS-RARP2-GFP and SS-RARP2-C109A are enriched in vesicular structures as previously observed [4]. Outlines of transfected cells expressing SS-RARP2 are shown in white. Bar = 10 μm. Higher magnification images are shown in S6 Fig.
Fig 6.
Temporal analysis of TGN dispersal.
A) Time course of trans-Golgi dispersal in R. rickettsii Sheila Smith infected cells. Infected cells were fixed and labeled at the indicated times. TGN disruption is evident by 4 hpi. TGN46 (red), rickettsiae (pink), and DAPI (blue). Bar = 10 um. B) Effects of inhibitors of rickettsial transcription (rifampicin) or translation (chloramphenicol) when added at various times pre- or post-infection. Inhibitors were added at 30 min pre-infection (30 min pre), 5 min post-infection (5 min pi), or 2 hr post-infection (2 hpi). Cultures were fixed and stained at 24 hpi. Uninfected cells are shown in the upper right of panel B. The remainder of the cells are infected with R. rickettsii Sheila Smith (green). TGN46 is shown in red and cells couterstained with DAPI to identify the nucleus. Bar = 10 μm.
Fig 7.
TGN46 glycosylation defects in Sheila Smith infected cells.
A) TGN46 in R. rickettsii Sheila Smith (SS) infected cells is unglycosylated and therefore shows no effect of deglycosylation treatment. R. rickettsii Iowa (Io) or uninfected Vero cell lysates (Un) migrate at a lower apparent mol. wgt. following sialidase treatment. Red represents TGN46; tubulin (green) was used as a loading control. Left arrowhead indicates the major deglysylated form of TGN46; right arrowhead the glycosylated form. B) Temporal analysis of TGN46 deglycosylation. Uninfected Vero cells (Un) or cells infected with R. rickettsii Sheila Smith (SS) or Iowa (Io) were harvested for SDS-PAGE at various times post infection and immunoblotted for TGN46, which identifies an immature glycosylated form of TGN46 (arrowhead) apparent by 4 hpi. A lesser amount of unglycosylated TGN46 is apparent even in the Iowa infected cells at the later time points at 48 and 72 hpi. We believe this represents activity of RARP2 which even though inappropriately targeted, may be in sufficient quantity by that time to show an effect. Tubulin (Tub) was used as a loading control.
Fig 8.
SS-RARP2 mediated dispersal of the trans-Golgi interferes with protein transport to the host cell surface.
A) Infection of Vero cells with R. rickettsii Sheila Smith (SS) or Iowa expressing SS-RARP2 (Iowa:SS-RARP2) inhibits trafficking of GFP-tagged TNFalpha to the cell surface. Vero cells were transfected with RUSH-constructs [9] encoding GFP-tagged TNFalpha and infected with rickettsiae 24 hours after transfection. Biotin was added 24 hpi to release reporter proteins from the endoplasmic reticulum and fixed 80 min after initiation of trafficking. GFP-tagged reporter proteins on the host cell surface were detected with anti-GFP antibody (red) without permeabilization, while intracellular GFP-tagged reporter protein is shown in green. A red signal thus indicates normal trafficking to the cells surface such that the reporter is detected with the anti-GFP antibody and red secondary antibody. Nucleic acids stained with DAPI are shown in blue, rickettsiae in white. B) Quantification of cells showing surface-labeling for GFP. Signals were counted for at least 100 cells for each of three biological replicates, and numbers were normalized to uninfected control cells. Shown is the mean +/- the S.E.M. Statistics were performed using One-way ANOVA and post-hoc Tukey test. Significant differences to the uninfected control are indicated with *** (p < 0.0005). C) Vero cells were infected with rickettsiae, and pulse-labeled for 15 min. with 35S-methionine, incubated for two more hours, surface biotinylated and then surface proteins pulled-down with streptavidin coated beads for SDS-PAGE, autoradiography, and immunoblotting. D) Radioactively-labeled surface proteins are pulled down from uninfected cells (Un) and cells infected with the Iowa strain (Io), but surface exposure of newly synthesized protein is greatly inhibited in cells infected with R. rickettsii Sheila Smith (SS). Autoradiographs of surface-associated proteins and total cell lysates are shown in the left hand panel and demonstrate a substantial decrease in surface exposure of de novo synthesized proteins in R. rickettsii Sheila Smith (SS) infected cells. The Western blot on the right demonstrates that MHC-1 was pulled down under these conditions and indicates the position of MHC-I (green) in surface protein and total protein lysates. GAPDH (red) is used as a loading control and seen primarily in the total cell lysates. Molecular weight (MW) is indicated on the left. E) A large fraction of newly synthesized MHCI is sensitive to Endoglycosidase H (EndoH) treatment in cells infected with R. rickettsii Sheila Smith (SS) or Iowa expressing SS-RARP2, indicating incomplete processing and maturation of the glycoprotein in the Golgi apparatus. Cells were pulsed-labeled with 35S, and MHC-I was pulled down with specific antibodies. Western blots (α-MHCI) and autoradiographs (35S) of pulled-down proteins treated with EndoH (+ EndoH) or untreated (- EndoH) are shown. Iowa expressing Io-RARP2 and Iowa expressing a methyltransferase of the Sheila Smith strain (Iowa:SS-MTase) were used as controls. Arrowheads indicate the position of the EndoH-sensitive, immature forms of MHC-I.