Bacterial pathogens hijack host cell peroxisomes for replication vacuole expansion and integrity

Abstract

Pathogens manipulate host cell organelles to establish infection. There is extensive evidence of pathogen modulation of the endoplasmic reticulum, Golgi apparatus, mitochondria, endosomes, lysosomes, and nucleus. However, one organelle that has been largely overlooked in connection with bacterial pathogenesis is peroxisomes. Here, we demonstrate that Legionella actively recruits peroxisomes to its replication vacuole using a secreted bacterial effector protein. Defects in peroxisome metabolic function restrict expansion of the Legionella vacuole membrane and cause rupture of this compartment, inhibiting bacterial replication and leading to bacterial degradation. Similarly, peroxisome dysfunction causes Salmonella replication vacuole destabilization and reduced bacterial burden within host cells. Thus, these two intracellular bacterial pathogens exploit host cell peroxisomes to maintain their replication compartments, establishing a critical role for this organelle in disease.

Fig. 1.. MavP interacts with the host peroxisomal membrane protein PMP70.

(A) MavP biotinylates the host protein PMP70. Affinity-captured biotinylated proteins from HEK293T cells transiently expressing BirA or BirA:MavP were subjected to Western analysis. T, total protein; UB, unbound; B, bound (biotinylated) protein. Data are representative of three biological replicates. (B) MavP interacts with PMP70. Lysates of HEK293T cells transiently expressing 3xFLAG epitope–tagged PMP70 (3xFLAG:PMP70) and BirA:MavP or BirA were subject to immunoprecipitation (IP) and examined by Western analysis. Data are representative of two biological replicates. (C) PMP70 redistributes to sites of MavP puncta. Left panel: HEK293T cells were mock transfected or transfected with plasmid encoding BirA or BirA:MavP and their subcellular distribution was compared to endogenous PMP70 by fluorescence microscopy. Right panels: Quantification of punctate BirA:MavP distribution pattern and colocalization of PMP70 signal with BirA:MavP puncta. Data are the mean ± SD of three biological replicates, consisting of three technical replicates each, scoring 100 cells per technical replicate. An asterisk indicates a Student’s t test P < 0.05. (D) MavP interacts with PMP70 during infection. Human monocyte–derived macrophages (THP-1 cells) were infected with the indicated L. pneumophila strains and at 6 hours cells were harvested. 3xFLAG:MavP was affinity captured from cell lysates and the presence of PMP70 was examined by Western analysis, loading 22.5× the amount of bound sample relative to total and unbound samples. Data are representative of three biological replicates.

Fig. 2.. MavP localizes to the Legionella replication vacuole at the onset of bacterial replication.

(A) MavP accumulation at the LCV temporally correlates with bacterial growth. Primary bone marrow–derived murine macrophages were infected with L. pneumophila (Lp) expressing 3xFLAG:MavP for the indicated times then fixed and visualized by fluorescence microscopy. (B) The number of MavP-positive LCVs in (A) were scored. Data are the mean ± SD of three biological replicates consisting of three technical replicates, scoring 100 vacuoles per technical replicate. An asterisk indicates a Student’s t test P < 0.05 compared to WT at 1 hour.

Fig. 3.. MavP recruits peroxisomes to the Legionella replication vacuole at the onset of bacterial replication.

(A) MavP-dependent recruitment of PMP70 to the LCV. Primary bone marrow–derived murine macrophages were infected with wild type (WT), ΔmavP mutant, or dot⁻ bacteria, fixed, and examined by fluorescence microscopy. Representative images at 6 hours are shown. (B) The number of PMP70 puncta colocalizing with the LCV at 1, 3, and 6 hpi based on fluorescence microscopy as in (A) were scored. (C) Disruption of Pmp70 in murine macrophages. Western analysis of whole-cell lysates of WT and Pmp70^−/− RAW264.7 cell lines. (D) Peroxisome recruitment to the LCV depends on PMP70. WT or Pmp70^−/− RAW264.7 cells were infected with WT or ΔmavP mutant bacteria for 6 hours, fixed, and examined by fluorescence microscopy for peroxisome colocalization with the LCV based on the peroxisomal membrane protein PEX14. (E) The number of PEX14 puncta at the LCV in (A) were scored. (F) Loss of PMP70 impairs growth of the L. pneumophila ΔlidA and ΔwipB deletion mutants. WT RAW264.7 cells or cells depleted of PMP70 were infected with the indicated L. pneumophila strains. Bacteria growth was quantified on the basis of recovered colony-forming units (cfus) from host cell lysates at 24 hours, equivalent to a single round of infection, and normalized to the WT bacteria in WT host cells by the number of intracellular bacteria at 1 hpi. [(B), (E), and (F)] Data are the mean ± SD of three biological replicates consisting of three technical replicates. For microscopy, 100 vacuoles (B) or 50 vacuoles (E) per technical replicate were scored. An asterisk indicates a Student’s t test P < 0.05 compared to WT at 1 hpi (B) or WT bacteria in the corresponding cell line [(E) and (F)], unless otherwise indicated.

Fig. 4.. Peroxisomes are important for L. pneumophila intracellular growth.

(A) RAW264.7 macrophage knockout (KO) cell lines depleted of PEX19 or PEX5 based on Western analysis of whole-cell lysates. (B) L. pneumophila growth is impaired in peroxisome-deficient cells. Cells in (A) were infected with the indicated L. pneumophila strains. Bacterial growth was quantified on the basis of recovered cfus from host cell lysates at 24 hours and normalized to the WT bacteria in WT host cells by the number of intracellular bacteria at 1 hpi. Data are the mean ± SD of three to four biological replicates consisting of three technical replications each. An asterisk indicates a Student’s t test P < 0.05 compared to WT bacteria in WT host cells.

Fig. 5.. Alterations in peroxisome lipid metabolism restricts L. pneumophila replication.

(A) Peroxisomes at the LCV are larger than those in the rest of the cell. Primary murine macrophages were challenged with WT bacteria, fixed, and visualized by fluorescence microscopy (as in Fig. 3A). Images were deconvoluted in three dimensions and the size of PMP70 puncta, based on the volume occupied in three-dimensional space were quantified (see Materials and Methods). (B) The relative abundance of peroxisomes at the LCV is similar to the rest of the cell. The number of PMP70 puncta from experiments in (A) was quantified, normalizing to the volume of the respective compartments. [(A) and (B)] Data are the combined measurements of two biological replicates, scoring 75 cells each. An asterisk indicates a two-way analysis of variance (ANOVA) with a post hoc two-tailed, nonparametric t test with Welch correction P < 0.0001. (C) L. pneumophila infection alters the abundance of peroxisomal proteins. THP-1 cells were infected for the indicated times and the amount of ACOX1, GNPAT, and AGPS in infected (I) cells was compared to uninfected (UI) cells by Western analysis (upper panel), and quantified (lower panel), normalizing to total protein (fig. S11B) and reported as fold change relative to 1 hour in uninfected cells. (D) Disruption of peroxisome lipid biosynthetic genes impairs L. pneumophila intracellular growth. THP-1 cells depleted of GNPAT or AGPS, based on Western analysis of whole-cell lysates (left panel), were infected with WT L. pneumophila and bacterial replication was quantified on the basis of recovered cfus from host cell lysates at 24 hours and normalized to the WT bacteria in WT host cells by the number of bacteria at 1 hpi. [(C) and (D)] Data are the mean ± SD of three biological replicates consisting of three technical replicates each. An asterisk indicates a Student’s t test P < 0.05 compared to uninfected cells at the same time point (C) or WT bacteria in WT host cells (D).

Fig. 6.. Defects in peroxisome function limit LCV expansion and integrity.

(A) LCVs in macrophages lacking functional peroxisomes exhibit compact architecture. Wild-type (WT) immortalized murine macrophages (RAW264.7 cells) lacking either PEX19 or PEX5 (Fig. 3A) or cells lacking PEX5 that were exogenously supplemented with palmitate (PA) were infected with L. pneumophila for 12 hours, fixed, and visualized by fluorescence microscopy. (B) LCVs in (A) were scored as spread (the majority of bacteria aligned end to end or at various angles) or compact (the majority of bacteria bundled together with their sidewalls juxtaposed to one another, with no bacteria extending beyond the boundary of a spherical shape) based on the organization of the bacteria in three-dimensional space (23). (C) Peroxisome defects lead to increased numbers of ruptured LCVs. Infected cells as in (A) were stained for L. pneumophila (Lp) and Galectin-3 (Gal3). (D) LCVs in (C) were scored for colocalization with Gal3. (E) Disruption of peroxisome lipid metabolic pathways limits LCV expansion and causes LCV destabilization. WT human monocyte–derived macrophages (THP-1 cells) lacking PEX5, GNPAT, or AGPS (Fig. 5B) were infected with L. pneumophila for 12 hours, fixed, and visualized by fluorescence microscopy. (F) LCVs in (E) were scored as spread or compact, as described in (B). (G) Gal3 colocalization with LCVs in (E) was quantified. [(B), (C), (F), (G)] Data are the mean ± SD of three biological replicates consisting of three technical replications each, scoring 50 to 100 vacuoles per technical replicate. An asterisk indicates a Student’s t test P < 0.05 compared to WT, unless otherwise indicated. ns, not significant.

Fig. 7.. Peroxisomes promote S. Typhimurium replication and vacuole integrity.

(A) S. Typhimurium growth is reduced in peroxisome-deficient cells. Wild-type (WT), PEX19-, and PEX5-deficient murine macrophages (RAW264.7 cells) (Fig. 4) were infected with WT S. Typhimurium (St) or an avirulent phoP^- mutant strain. Bacterial growth was measured on the basis of recovered cfus from host cell lysates at 21 hpi, normalized to the WT strain in WT host cells by the number of intracellular bacteria at 2 hpi. Data are the mean ± SD of three biological replicates consisting of three technical replications each. An asterisk indicates a Student’s t test P < 0.05 compared to WT bacteria in WT host cells. (B) Defects in peroxisome function lead to more compact S. Typhimurium vacuoles. WT RAW264.7 cells or cells lacking functional peroxisomes (Pex5^−/−) were infected for 20 hours with S. Typhimurium constitutively expressing mCherry fluorescent protein, fixed, stained and visualized by fluorescence microscopy (left panel). The area of the vacuoles, based on the contour of the bacteria, and the number of bacteria within the vacuole, based on fluorescence intensity, were measured and the concentration of bacteria per vacuole (area/bacteria) was determined (right panels). Data are the combined measurements of three biological replicates, scoring 100 vacuoles per replicate. An asterisk indicates a two-tailed, nonparametric unpaired t test with Mann-Whitney correction P < 0.0001 relative to WT cells. (C) Peroxisome dysfunction leads to the destabilization of Salmonella-containing vacuoles (SCVs). Cells in (B) were stained for Galectin-3 (Gal3) (left panel) and the percentage of SCVs colocalizing with Galectin-3 were quantified (right panel). Data are the mean ± SD of three biological replicates consisting of three technical replications each, scoring 100 vacuoles per technical replicate. An asterisk indicates a Student’s t test P < 0.05 compared to WT cells.