The intracellular bacteria Chlamydia hijack peroxisomes and utilize their enzymatic capacity to produce bacteria-specific phospholipids

Abstract

Chlamydia trachomatis is an obligate intracellular pathogen responsible for loss of eyesight through trachoma and for millions of cases annually of sexually transmitted diseases. The bacteria develop within a membrane-bounded inclusion. They lack enzymes for several biosynthetic pathways, including those to make some phospholipids, and exploit their host to compensate. Three-dimensional fluorescence microscopy demonstrates that small organelles of the host, peroxisomes, are translocated into the Chlamydia inclusion and are found adjacent to the bacteria. In cells deficient for peroxisome biogenesis the bacteria are able to multiply and give rise to infectious progeny, demonstrating that peroxisomes are not essential for bacterial development in vitro. Mass spectrometry-based lipidomics reveal the presence in C. trachomatis of plasmalogens, ether phospholipids whose synthesis begins in peroxisomes and have never been described in aerobic bacteria before. Some of the bacterial plasmalogens are novel structures containing bacteria-specific odd-chain fatty acids; they are not made in uninfected cells nor in peroxisome-deficient cells. Their biosynthesis is thus accomplished by the metabolic collaboration of peroxisomes and bacteria.

Figure 1. Peroxisomes are translocated into the Chlamydia inclusion during infection.

A- HeLa cells were infected with C. trachomatis L2 for 20 h. The inclusion membrane was labeled with an anti-CT813 antibody (green), peroxisomes with an anti-ALDP antibody (red) and bacterial and nuclear DNA with Hoechst (Blue). A single ApoTome x-y section is shown in the central image. The z-x projection on the top shows the peroxisome indicated by a white arrowhead in the x-y image. Scale bar: 5 µm. B- HeLa cells were transfected with cytosolic GFP (shown in blue) to illuminate the entire cell except for the Chlamydia inclusion and were infected with C. trachomatis L2 for 20 h. Bacteria were labeled with an anti-Hsp60 antibody (green) and peroxisomes with an anti-ALDP (red). A single ApoTome x-y section is shown in the central image. z-x and z-y projections on the top and on the right side, respectively, are centered on the peroxisome indicated by a white arrowhead. Scale bar: 5 µm. C- One optical section from the stack of images shown in Movie S1. Cells were prepared as in Figure 1 B; the colors are different: bacteria are in blue, peroxisomes in red, GFP in green.

Figure 2. Peroxisomes are close to bacteria.

A- Quantitative image analysis. A green polygon representing the Region Of Interest (ROI) was drawn over an optical section from the stack of images shown in Movie S1 (left image): bacteria are in blue, peroxisomes in red, GFP in green. Peroxisomes and bacteria detected within the ROI are circled in the middle and right images, respectively. Scale bar: 2 µm. B- Quantification of the distances between intra-inclusion peroxisomes and bacteria. The minimal distances between peroxisomes and bacteria within the ROI were calculated (from three different cells with respectively 6, 13 and 14 peroxisomes each, n = 33 peroxisomes in total) and the distribution of these distances is shown in the histogram. We calculated (see Methods) that a random distribution of bacteria and peroxisomes within the ROI would result in an average distance of 1.35 µm (p = 0.05, dotted line). The observed distribution is strongly shifted to the left and supports the hypothesis of a contact, or close proximity, between intra-inclusion peroxisomes and bacteria.

Figure 3. Peroxisomes are not essential for C. trachomatis infection and development.

A- Control or PEX19-deficient fibroblasts were infected with C. trachomatis L2 for 24 h. Infected cells were fixed and labeled with an antibody against bacterial lipopolysaccharide (LPS) coupled to FITC (green) and DNA was stained with Hoechst (blue). Scale bar: 20 µm. B- The number of infection-forming units (IFU) was determined by titrating cell lysates at 48 hrs post infection on fresh HeLa cell monolayers. The average of three experiments is shown, error bars indicate standard deviation. C- Areas of inclusions. Results for two experiments are shown (Number of inclusions measured: experiment 1: control n = 111, PEX19-deficient n = 111; experiment 2: control n = 65, PEX19-deficient n = 76).

Figure 4. Many plasmalogens decrease upon infection.

A- Structure of plasmalogens. Phosphatidylethanolamine (left) and the plasmalogen version of phosphatidylethanolamine (right) are depicted. The plasmalogen variant contains a long chain alcohol in place of a fatty acid and has a double bond between the first two carbons of the alcohol (1-O-alkenyl-2-acyl-sn-glycero-3-phosphoethanolamine; plasmenylethanolamine). B- Example of plasmalogens abundant in the non-infected control fibroblasts. Putative plasmalogens were identified based on their predicted formula as C_xH_2x–2yNO₇P (where y = the number of double bonds in the fatty acid) after ICR-FT/MS analysis (mass error <100 parts/billion, MassTRIX [46]), and on their sensitivity to acid hydrolysis. Lipid species fulfilling these criteria were either missing or severely reduced in peroxisome-deficient cells (“PEX19”) compared to control fibroblasts (CTL), confirming that they are plasmalogens. Note that, globally, plasmalogen content decreased during infection, most notably at the later time point. Identification of plasmalogens was performed on samples from a single experiment. C- Abundance of control phospholipids. Phospholipids that gave the highest signal intensity in the ICR-FT/MS (corresponding to C42H82NO8P) were arbitrarily chosen as control. D- Total phospholipids detected with ICR-FT/MS. The intensities of all annotated phospholipids (mass error <1 ppm) were summed and normalized to the total sum of intensities in the sample. The data in C and D are representative of two independent experiments.

Figure 5. C. trachomatis contain bacteria-specific plasmalogens.

A- Plasmalogens that increase in infected cells. Extracted ion chromatograms of plasmalogens of control and peroxisome-deficient (“PEX19”) fibroblasts, analyzed with RP LC-MS. The peaks of plasmalogens that increased upon 24 and 48 h of infection in control cells are shown. Notation, e.g. for PE(P-32:0): PE(P = plasmalogen version of phosphatidylethanolamine; 32:0 = 32 carbons in the combined side chains and 0 double bonds in the fatty acid. B- Plasmalogens in C. trachomatis. Phospholipids extracted from bacteria purified on a density gradient were analyzed as in B, with or without prior acid hydrolysis. The peaks with the same masses as in Panel B are shown. C- High resolution analysis and fragmentation of the plasmalogens of Panel C by SORI ICR/FT-MS². The fragmentation pattern of PE(P-33:0) is shown as an example. D- Plasmalogen structures of C. trachomatis. Detected fatty acids and inferred structures for the Chlamydia-derived plasmalogens. S/N: signal to noise ratio.