SEARCH

SEARCH BY CITATION

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Coxiella burnetii directs the synthesis of a large parasitophorous vacuole (PV) required for replication. While some lysosomal characteristics of the PV have been described, the origin and composition of the PV membrane is largely undefined. Cholesterol is an essential component of mammalian cell membranes where it plays important regulatory and structural roles. Here we investigated the role of host cholesterol in biogenesis and maintenance of the C. burnetii PV in Vero cells. The C. burnetii PV membrane stained with filipin and was positive for the lipid raft protein flotillin-1, suggesting PV membranes are enriched in cholesterol and contain lipid raft microdomains. C. burnetii infection increased host cell cholesterol content by 1.75-fold with a coincident upregulation of host genes involved in cholesterol metabolism. Treatment with U18666A, lovastatin, or 25-hydroxycholesterol, pharmacological agents that inhibit cholesterol uptake and/or biosynthesis, altered PV morphology and partially inhibited C. burnetii replication. Complete inhibition of C. burnetii PV development and replication was observed when infected cells were treated with imipramine or ketoconazole, inhibitors of cholesterol uptake and biosynthesis respectively. We conclude that C. burnetii infection perturbs host cell cholesterol metabolism and that free access to host cholesterol stores is required for optimal C. burnetii replication.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Coxiella burnetii is an obligate intracellular bacterium and the causative agent of human acute and chronic Q fever. Acute disease normally manifests as a flu-like illness and is effectively treated with antibiotics. Chronic disease, usually exhibited as endocarditis, is rare and often refractory to antibiotic treatment. Humans are primarily infected by inhaling contaminated aerosols generated during domestic livestock operations (reviewed in Maurin and Raoult, 1999). Unlike other obligate intracellular bacteria, C. burnetii is highly resistant to environmental stress, a characteristic attributed to a small cell variant form that is part of a biphasic developmental cycle that includes a more metabolically and replicatively active large cell variant (LCV) form (Heinzen et al., 1999). C. burnetii replicates to high numbers within an acidic (pH ∼5) parasitophorous vacuole (PV) with characteristics of a phagolysosome (Heinzen et al., 1996). Within this vacuole, acid pH activates C. burnetii metabolism and initiates replication of the organism (Hackstadt and Williams, 1981).

A growing body of evidence indicates that biogenesis of the C. burnetii PV deviates from typical phagosome to lysosome maturation. Fusion between lysosomes and nascent phagosomes harbouring viable C. burnetiiTakes about 4 h, while fusion between lysosomes and phagosomes containing latex beads or inactivated C. burnetii occurs within 30 min (Howe and Mallavia, 2000). C. burnetii PV demonstrate promiscuous homotypic and heterotypic fusion with endolysosomal vacuoles, a process that presumably provides membrane for the expanding PV and nutrients for pathogen replication (Veras et al., 1994; 1995; Howe et al., 2003). C. burnetii PV further deviate from typical endosomal maturation by interacting with the autophagy pathway to acquire some characteristics of autophagolysosomes (Beron et al., 2002; Gutierrez et al., 2005).

Coxiella burnetii protein expression is required for maturation of the large and spacious PV that supports pathogen replication (Howe et al., 2003). In infected cells treated with chloramphenicol to block Coxiella protein synthesis, nascent phagosomes mature into lysosomes (e.g. acidic and LAMP-1-positive); however, these vacuoles do not homotypically fuse or become spacious (Howe et al., 2003). Although C. burnetii effectors of PV maturation have not been identified, the C. burnetii genome contains homologues of nearly all 26 Legionella pneumophila dot/icm genes that encode a type IV secretory apparatus (Sexton and Vogel, 2002). Thus, it is reasonable to assume that C. burnetii assembles a functional type IV secretory apparatus that secretes protein effectors of PV maturation.

Mature C. burnetii PV can occupy the majority of the host cell volume in both epithelial cells and macrophages (Baca et al., 1985; Roman et al., 1986; Coleman et al., 2004). These large vacuoles are relatively insensitive to mechanical disruption, a characteristic observed during microinjection experiments (Heinzen and Hackstadt, 1997). Moreover, they display a low buoyant density when isolated by sucrose density gradient centrifugation (D. Howe, unpublished). These features suggested that the C. burnetii PV membrane is rich in cholesterol as membranes with a high cholesterol content show increased buoyancy and resistance to physical disruption (Needham et al., 1988).

Cholesterol is a critical component of mammalian cell membranes where it helps define membrane structure and function. Cholesterol increases the mechanical strength and decreases the ionic permeability of membranes (Needham et al., 1988; Corvera et al., 1992). The curvature of lipid bilayers is affected by the ratio of cholesterol to phospholipid, and changes in sterol composition can induce tubulation of spherical vesicles (Bacia et al., 2005). Cholesterol and sphingolipid-rich microdomains of membranes known as lipid rafts are thought to cluster various regulatory proteins that can mediate a variety of events such as signal transduction and membrane fusion (Munro, 2003). The cholesterol concentration of organelle membranes is maintained at different levels by homeostatic mechanisms. For example, lysosomal membranes have a cholesterol: phospholipid molar ratio of 0.38, which is between that of the plasma membrane (0.65) and that of the endoplasmic reticulum (0.16) (Schoer et al., 2000). Cholesterol is extremely hydrophobic and moves very slowly by passive diffusion. Therefore, most intracellular transport of cholesterol is mediated by vesicle membranes or carrier proteins (Soccio and Breslow, 2004). However, despite the importance of cholesterol to cellular function, mechanisms controlling cholesterol transport and membrane cholesterol content are poorly understood.

The importance of cholesterol in the biogenesis of pathogen PV is becoming increasingly evident. Chlamydia trachomatis (Carabeo et al., 2003) and Salmonella enterica serovar Typhimurium (Catron et al., 2002; Garner et al., 2002) are both bound by vacuolar membranes rich in cholesterol. The chlamydial PV (inclusion) obtains cholesterol directly from the Golgi apparatus by a pathogen-directed process (Carabeo et al., 2003). Late in infection of epithelial cells and macrophages by S. enterica serovar Typhimurium, there is striking redistribution of total cellular cholesterol to the Salmonella-containing vacuole (SCV) (Catron et al., 2002; Garner et al., 2002). This accumulation, which can account for 30% of the cellular cholesterol, is dependent on bacterial replication (Catron et al., 2002). The high cholesterol content of the late SCV is unusual as these vacuoles have characteristics of late endosomes, and membrane cholesterol typically decreases as early endosomes mature into lysosomes (reviewed in Holthuis et al., 2003)

Here we examine the role of cholesterol in development of the lysosome-like PV of C. burnetii. Our results indicate that the PV membrane is cholesterol-rich and contains lipid raft proteins. Moreover, trafficking of cholesterol to the PV is required for optimal C. burnetii replication.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

The C. burnetii PV membrane is cholesterol-rich

To test for the presence of cholesterol in the C. burnetii PV membrane, infected Vero cells were stained with filipin, a fluorescent polyene antifungal agent that binds cholesterol (Kruth et al., 1986). At 2 days post infection, intense staining of the PV membrane was observed (Fig. 1A). The PV membrane generally labelled as intensely as the plasma membrane where >  75% of cellular cholesterol normally resides (Prinz, 2002). Staining of some PV membranes was uneven, and small cholesterol-rich vesicles were often observed in close association with PV. The PV lumen also appeared to contain cholesterol-rich particles. These may be multilamellar bodies that are trafficked to the C. burnetii PV by autophagic vacuoles (Punnonen et al., 1988; Coleman et al., 2004; Gutierrez et al., 2005; Lajoie et al., 2005). To determine whether PV cholesterol correlated with the presence of lipid raft proteins, infected cells were labelled by indirect immunofluorescence for flotillin-1 and flotillin-2 (Fig. 1B and data not shown). Both proteins localized to the C. burnetii PV membrane at 2 days post infection, with smaller C. burnetii PV labelling more intensely than larger PV. Intense staining of the plasma membrane was also observed as previously described (Underwood et al., 1998). Collectively, these data indicate that the C. burnetii PV membrane is cholesterol-rich with potential concentrations of this sterol in lipid rafts.

image

Figure 1. The C. burnetii PV membrane contains cholesterol and the lipid raft protein flotillin-1. Infected Vero cells were fixed at 2 days post infection and C. burnetii stained red by indirect immunofluoresence. Cells were then counterstained for cholesterol (blue) (A) or flotillin-1 (green) (B). A. Epifluoresence image shows filipin labelling of the C. burnetii PV membrane (arrows) and the plasma membrane (arrowhead), indicating the presence of cholesterol. The upper and lower grey scale images are 2× magnifications of the boxed region and show filipin and C. burnetii staining, respectively. B. Indirect immunofluoresence shows flotillin-1 associated with the C. burnetii PV membrane. Small PV (arrowhead) labelled more intensely than the plasma membrane or larger PV (arrow). The upper and lower grey scale images are 2× magnifications of the boxed region and show flotillin-1 and C. burnetii staining respectively. Samples were viewed by laser scanning confocal microscopy and a single optical section (0.2 µm) is shown. Bars, 5 µm.

Download figure to PowerPoint

Coxiella burnetii infection increases total cellular cholesterol and upregulates genes involved in cholesterol metabolism

To investigate whether cholesterol trafficking to the C. burnetii PV affects cellular cholesterol levels, we measured the cholesterol content of Vero cells infected for 6 days using a fluorometric enzyme-based assay. At this time point C. burnetii is in the early stationary phase of its growth cycle and the PV occupies the majority of the cell volume (Coleman et al., 2004). There was 73% more cellular cholesterol in infected cell monolayers than in mock infected cell monolayers; [3.33 ± 0.15 (mean ± SD) versus 1.93 ± 0.49 µg of cholesterol per 106 cells in three independent experiments, P < 0.01].

To assess whether the increase in cellular cholesterol accompanying C. burnetii infection correlated with the upregulation of genes involved in cholesterol metabolism, the expression of genes involved in both cholesterol uptake and biosynthesis was examined. Specifically, the gene encoding the low-density lipoprotein (LDL) receptor (LDLR), responsible for uptake of exogenous cholesterol, was evaluated as were three genes encoding enzymes of the biosynthetic pathway; namely, 3-hydroxy-3-methylglutaryl coenzyme A (HMGCoA) reductase (HMGCR), which catalyses the rate-limiting step in the cholesterol biosynthetic pathway, lanosterol synthase (LSS), which catalyses the conversion of squalene to lanosterol, and C-24 sterol reductase (DHCR24), which catalyses the final step in cholesterol biosynthesis (Fig. 2). Gene expression was measured by quantitative polymerase chain reaction (qPCR) at 0 h, 2 days, 4 days and 6 days post infection, time points where C. burnetii growth is in early lag, early log, mid-log and early stationary phase respectively (Coleman et al., 2004). At 0 h post infection, the relative expression of genes encoding HMGCR, LSS and DHCR24 was higher in infected cells versus mock infected cells (Fig. 3). This response may reflect plasma membrane remodelling that occurs after internalization of multiple C. burnetii. At 2 days post infection, C. burnetii containing phagosomes have coalesced to form small PV that are visible by phase contrast microscopy and the organism has begun to replicate. Here the relative expression of genes encoding HMGCR, LSS and DHCR24 was lower in infected cells compared with mock infected cells. There was no difference in relative LDLR gene expression between infected and mock infected cells at 0 h and 2 days post infection. At 4 days post infection, relative expression of all genes was higher in infected cells compared with mock infected cells. This response is likely required to fulfil the cholesterol requirements of the maturing PV, which has dramatically expanded at this time point to accommodate exponential growth of C. burnetii. At 6 days post infection, the relative expression of all genes was roughly the same in infected and mock infected cells, and had dropped below the level observed for mock infected cells at 0 h post infection. At this time point C. burnetii has entered stationary phase and PV expansion has ceased. The overall expression profile of mock infected cells, with expression of three of four genes increasing from 0 h to 4 days, then decreasing below the 0 h level at 6 days, may reflect Vero cell division within the confluent monolayer. However, the effect of cell division on gene expression would be the same in infected and mock infected cell cultures as infection by C. burnetii does not alter the generation time or cell cycle of host cells (Baca et al., 1985).

image

Figure 2. Steps in cholesterol biosynthesis and uptake and inhibitors of these pathways. Pharmacological inhibitors of these pathways and the steps they inhibit are depicted in red and green respectively. In the cholesterol uptake pathway, LDL particles are internalized by LDL receptor-mediated endocytosis and delivered to lysosomes. Here acid lipase hydrolyses LDL-cholesterol esters whereupon free cholesterol associates with the lysosomal membrane or is complexed to carrier proteins and transported via the Golgi apparatus to the endoplasmic reticulum or plasma membrane. U18666A and imipramine reversibly inhibit cholesterol transport from lysosomes, causing accumulation of the molecule in this compartment. In the biosynthetic pathway, cholesterol is synthesized in the endoplasmic reticulum via a multistep enzymatic pathway. Lovastatin inhibits the activity of 3-hydroxy-3-methylglutaryl coenzyme A (HMGCoA) reductase, the rate-limiting step in cholesterol biosynthesis, and thus inhibits biosynthesis of isoprenoids and sterols. Ketoconazole, an inhibitor of ergosterol biosynthesis in fungi, also inhibits lanosterol 14α-demethylase in mammalian cells, an enzyme catalysing one step in the 19-step conversion of lanosterol to cholesterol. C-24 sterol reductase catalyses one of the final steps in the conversion of lanosterol to cholesterol. The oxysterol, 25-hydroxycholesterol, regulates cholesterol homeostasis. It signals the presence of excess cholesterol and induces the downregulation of genes involved in both cholesterol biosynthesis and uptake.

Download figure to PowerPoint

image

Figure 3. Coxiella burnetii infection alters expression of host cell genes involved in cholesterol metabolism. Vero cell monolayers were incubated with C. burnetii for 2 h to allow adherence and internalization. Extracellular organisms were then washed from cell monolayers and fresh medium was added. This time point was designated as 0 h post infection. Total RNA was extracted at the indicated times. A temporal analysis of gene transcription was performed using quantitative reverse transcription PCR with TaqMan primers/probe specific for each gene. Transcript copy number was normalized to β-actin transcript in each sample. Results are expressed as fold-change in gene expression relative to mock infected Vero cells at 0 h post infection and are representative of two independent experiments.

Download figure to PowerPoint

Pharmacological inhibitors of cholesterol biosynthesis and uptake disrupt PV morphology and inhibit C. burnetii replication

To assess the relative importance of exogenously and endogenously acquired cholesterol to C. burnetii replication and PV development, infected cells were treated with pharmacological agents that inhibit various steps in cholesterol metabolism (Fig. 2). Lovastatin, an inhibitor of HMGCR (Catron et al., 2004) and ketoconazole, an inhibitor of lanosterol 14α-demethylase (Iglesias and Gibbons, 1989), were used to block steps in cholesterol biosynthesis. Imipramine and 3-β-[2-(diethylamino) ethoxy]androst-5-en-17-one (U18666A) were used to inhibit transport of LDL-derived cholesterol from lysosomes (Liscum and Faust, 1989; Lange et al., 1997). The oxysterol, 25-hydroxycholesterol (25-OHC), was used to inhibit both pathways. It signals excess cellular cholesterol and acts as a negative feedback regulator of cholesterol metabolism (Lange and Steck, 1994; King et al., 2004).

We first investigated the effect of inhibitors on PV development. Infected cells were cultured in the presence of inhibitors for 3 days, then fixed and stained for C. burnetii and LAMP-1. A continuum of effects on the gross morphology of PV was observed following inhibitor treatments (Fig. 4). Treatment with lovastatin did not dramatically alter the general appearance of PV although these vacuoles were smaller than those of untreated cells. Treatment with 25-OHC resulted in unusually intense labelling of LAMP-1 on the PV membrane. These vacuoles were often lobed with LAMP-1-positive extensions (sometimes harbouring organisms) that wrapped around the nucleus and extended to the cell periphery. In cells treated with U18666A, homotypic fusion of C. burnetii PV appeared inhibited. Many small LAMP-1-positive PV were observed, which were arranged in tight clusters and contained a single organism. Occasionally, an unusually swollen PV that contained multiple organisms with weak LAMP-1 staining was also observed. In cells treated with ketoconazole or imipramine, complete inhibition of homotypic fusion occurred as small LAMP-1-positive vacuoles harbouring single organisms were observed scattered throughout the cytoplasm that had failed to coalesce or enlarge. Lysosomes of cells treated with these inhibitors also appeared much larger than those of untreated cells. Collectively, these observations indicate that host cell cholesterol, derived from either pathway, is required for proper PV biogenesis and maintenance.

image

Figure 4. Inhibitors of cholesterol metabolism alter C. burnetii PV morphology. Infected Vero cells were incubated for 3 days in the presence of inhibitors, then stained by indirect immunofluoresence using monoclonal antibodies directed against LAMP-1 (green) and polyclonal guinea pig antiserum against C. burnetii (red). DRAQ5 was used to visualize nuclear DNA (blue). Samples were viewed by laser scanning confocal microscopy and a single optical section (0.2 µm) is shown. Bar, 5 µm.

Download figure to PowerPoint

The alteration of PV morphology associated with inhibition of cholesterol biosynthesis and egress from lysosomes appeared, in most cases, to correlate with inhibition of C. burnetii replication. Therefore, we quantified C. burnetii genomes by qPCR following a 6-day treatment of infected monolayers with inhibitors. C. burnetii replication was significantly inhibited following treatments (Fig. 5). Lovastatin, U18666A and 25-OHC inhibited C. burnetii growth by approximately 30%. Treatment with imipramine or ketoconazole, which had the most striking effect on PV morphology, completely inhibited C. burnetii replication. Indeed, there was a net decrease in C. burnetii genomes over the 6-day incubation period, indicating degradation of some organisms. Taken together, these data indicate that cholesterol acquisition by the PV is not only required for proper development of the vacuole, but also for optimal replication of C. burnetii.

image

Figure 5. Coxiella burnetii replication is inhibited by pharmacological inhibitors of cholesterol metabolism. Vero cells were infected with C. burnetii and incubated in the presence of inhibitors for 6 days. At 0 h and 6 days post infection, total DNA was extracted and C. burnetii genome equivalents were quantified by qPCR. Results are expressed as fold-change in genomes at 6 days post infection normalized to genome equivalents extracted at 0 h post infection, and are the mean of three independent experiments with error bars representing standard deviation. C. burnetii replication was significantly inhibited by all cholesterol metabolism inhibitors (P-values <  0.05) as calculated using one-way analysis of variance followed by a Dunnett's post hoc test.

Download figure to PowerPoint

The bactericidal effect of ketoconazole, but not imipramine, is determined by the maturation state of the C. burnetii PV

Because ketoconazole and imipramine most profoundly inhibited PV maturation and C. burnetii replication, we examined the effects of these drugs in greater detail. We specifically asked two questions: first, is their activity affected by the maturation state of the PV and/or the growth phase of the organism and second, is inhibition of C. burnetii replication reversible? To address the first question, C. burnetii-infected cells were treated with either inhibitor beginning immediately after internalization (0 h post infection), or at 2 days post infection when PV are established and C. burnetii is entering its exponential growth phase. Bacterial genomes were then quantified at 6 days post infection (Fig. 6). In cells treated with ketoconazole or imipramine at 0 h post infection, C. burnetii genomes decreased 3.1-fold and 4.5-fold respectively. In cells treated at 2 days post infection, C. burnetii genomes decreased 7.7-fold in imipramine treated cells. Conversely, a 1.5-fold increase in genomes was observed in ketoconazole-treated cells. This is much lower than the roughly 10-fold increase in C. burnetii genomes observed in untreated cells at 6 days post infection but approximates the slight increase in genomes typically observed at 2 days post infection in Vero cells (Coleman et al., 2004). Thus, imipramine has a bactericidal effect on both non-replicating and replicating C. burnetii harboured in immature and mature PV respectively. In contrast, ketoconazole is bactericidal to lag-phase bacteria occupying nascent phagosomes and only bacteristatic to replicating C. burnetii within mature PV. Therefore, the maturation state of the PV and the replication status of C. burnetii correlate with the bactericidal effect of ketoconazole, but not imipramine.

image

Figure 6. Inhibition of C. burnetii replication by ketoconazole, but not imipramine, correlates with the maturation state of the PV. Infected Vero cells were incubated in the presence of ketoconazole or imipramine beginning at 0 h or 2 days post infection (PI). Total DNA was extracted at 0 h and 6 days post infection and C. burnetii genome equivalents were determined using qPCR. The results are expressed as fold-change in genomes normalized to genome equivalents extracted at 0 h post infection, and are the mean of three independent experiments with error bars representing standard deviation. At 6 days post infection, genome numbers in all treated samples were significantly different from those in untreated samples (P-values <  0.02) as calculated using one-way analysis of variance followed by a Dunnett's post hoc test.

Download figure to PowerPoint

Inhibition of C. burnetii replication by ketoconazole, but not imipramine, is reversible

To determine if inhibition of C. burnetii replication by ketoconazole or imipramine is reversible, infected cells were incubated in the presence of inhibitors for 6 days, followed by their removal and continued incubation for an additional 6 days. At 0 h, 6 days and 12 days post infection, C. burnetii genomes were quantified by qPCR. After a 6-day treatment with ketoconazole or imipramine, C. burnetii genomes decreased 1.7-fold and 5.6-fold respectively, relative to genomes extracted from cells at 0 h post infection (Fig. 7). At the same time point there was an approximately ninefold increase in genomes in untreated cells. When ketoconazole was removed at 6 days post infection, C. burnetii replicated over the next 6 days at a rate similar to that observed for C. burnetii in untreated cells between 0 and 6 days post infection. In contrast, the number of C. burnetii genomes did not significantly increase after removal of imipramine. Thus, although treatment with ketoconazole or imipramine results in a similar immature PV phenotype, the mechanisms by which these agents inhibit C. burnetii replication are different, with the inhibitory effect of imipramine being irreversible.

image

Figure 7. Inhibition of C. burnetii replication by ketoconazole, but not imipramine, is reversible. Infected Vero cells were treated with ketoconazole or imipramine beginning at 0 h post infection. At 6 days post infection, inhibitors were removed (washout) and cells incubated for an additional 6 days. Total DNA was extracted at 0 h, 6 days and 12 days post infection and C. burnetii genome equivalents were quantified by qPCR. The results are expressed as fold-change in bacterial genome equivalents at 6 and 12 days post infection normalized to genome equivalents extracted at 0 h post infection, and are the mean of three independent experiments with error bars representing standard deviation. Genome numbers that were significantly different from those of untreated samples at the same time point (P-values < 0.02) are designated with an asterisk. Significance was determined using a one-way analysis of variance followed by a Dunnett's post hoc test.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Coxiella burnetii is absolutely dependent on the biogenesis of a PV that provides a lysosome-like environment permissive for replication (Heinzen et al., 1996). Unusual characteristics of this compartment include its promiscuous fusogenicity with endolysosomal vacuoles (Howe et al., 2003) and the ability to maintain a low pH (∼5) for many weeks (Maurin et al., 1992a). Late in the infectious cycle, the PV enlarges to occupy the majority of the cell volume and accommodates a large number of replicating bacteria (Coleman et al., 2004). Presumably this process imposes a tremendous metabolic burden on the host cell. Yet little cytopathic effect is observed during C. burnetii infection, indicating the host cell has a remarkable ability to adjust nutrient and lipid fluxes to accommodate the parasite. Despite the necessity of PV development for C. burnetii pathogenicity, little is known about the structure of the PV membrane or the nature and source of its constituent lipids. Here we describe a C. burnetii PV that is highly enriched in cholesterol. Indeed, the C. burnetii PV membrane stains for cholesterol at roughly the same intensity as the plasma membrane where 75–95% of cellular cholesterol resides (Lange et al., 1989). The high cholesterol content of mature C. burnetii PV is unusual as these vacuoles have characteristics of lysosomes, which typically contain approximately half the cholesterol of the plasma membrane (Schoer et al., 2000).

Trafficking of cholesterol to the C. burnetii PV is accompanied by a 73% increase in the cholesterol content of infected cell cultures relative to uninfected cultures at 6 days post infection. We employed an infection procedure that results in initial infection of approximately 80% of host cells. However, this percentage decreases somewhat during the 6-day incubation period as some Vero cell division occurs within the confluent monolayer and PV segregate to only one daughter cell. Thus, we predict that individual infected cells actually contain closer to 100% more cholesterol than uninfected cells.

The accumulation of cholesterol induced by C. burnetii infection is striking as cellular cholesterol metabolism is a tightly regulated process controlled by multiple feedback mechanisms at work in the endoplasmic reticulum (ER) (Brown and Goldstein, 1999; Feng et al., 2003). Genes involved in both cholesterol uptake and biosynthesis are upregulated during the mid-exponential growth phase of C. burnetii when the PV is approaching its maximum volume (Coleman et al., 2004). In a recent study, Catron et al. (Catron et al., 2002) demonstrated that mouse macrophage cultures infected by S. enterica serovar Typhimurium accumulate 30% more cholesterol than uninfected cultures, and that sequestration of cholesterol by the SCV is dependent on bacterial replication and possibly effectors secreted by a Salmonella pathogenicity island 2 (SPI-2) type III secretion system (Catron et al., 2002). Dramatic expansion of the C. burnetii PV is co-ordinated with bacterial replication and possibly peak expression of a type IV secretion apparatus (Coleman et al., 2004). Thus, it is intriguing to speculate that secreted effector proteins mediate recruitment of cholesterol to the C. burnetii PV.

Cholesterol-containing ‘vesicles’ are also observed in the PV lumen by filipin staining. These vesicles may be cholesterol-rich multilamellar bodies (MLBs) that have been previously observed within C. burnetii PV (Coleman et al., 2004). Swollen lysosomes that contain MLBs are also associated with Neimann-Pick Type C (NPC) lysosomal storage disease (Blanchette-Mackie, 2000). This disease occurs because of deficient expression of the NPC1 protein, a protein that facilitates egress of LDL-derived cholesterol from lysosomes (Blanchette-Mackie, 2000). C. burnetii may also inhibit this process to result in accumulation of cholesterol-rich MLBs in its PV. Alternatively, MLBs may be trafficked to the PV by fusion with autophagic vacuoles (Gutierrez et al., 2005; Lajoie et al., 2005).

Expression of genes associated with cholesterol metabolism is responsive to the growth phase of C. burnetti and the associated PV phenotype. Specifically, when compared with uninfected cells, three cholesterol biosynthesis genes (encoding HMGCR, LSS and DHCR24) are downregulated at 2 days post infection. These genes, along with the gene encoding the LDLR, are then upregulated at 4 days post infection, with expression returning to the same level as uninfected cells at 6 days post infection. This behaviour contrasts to cholesterol gene expression patterns associated with development of large and spacious, phase-lucent lysosomes (sucrosomes) that result from internalization of indigestible sucrose by mammalian cells. Genes involved in cholesterol metabolism are upregulated approximately twofold in human fibroblasts after 1 day of sucrose treatment, with expression remaining relatively unchanged through 5 days of treatment (Helip-Wooley and Thoene, 2004). Thus, in contrast to sucrosomes, the C. burnetii PV is a bioactive organelle that is integrated into the cellular cholesterol distribution machinery and cellular cholesterol homeostatic mechanisms respond to the progression of C. burnetii infection.

Both exogenously acquired and endogenously synthesized cholesterol are important for normal PV structure and robust C. burnetii replication as pharmacological agents that block specific steps in either pathway inhibit growth and disrupt PV formation. A continuum of PV structural changes are observed following inhibitor treatments, ranging from smaller but otherwise normal appearing PV in lovastatin-treated cells, to non-fusogenic PV in ketoconazole and imipramine-treated cells that are the same size as slightly swollen resident lysosomes. Intermediate PV phenotypes are induced by 25-OHC and U18666A, with elongated PV induced by 25-OHC, a phenotype previously observed for lysosomes of treated murine astrocytes (Patel et al., 1994), and convoluted PV induced by U18666A. With the exception of PV induced by ketoconazole and imipramine, aberrant PV support C. burnetii replication, albeit at approximately 70% the level of untreated cells. If the exogenous pathway of cholesterol acquisition is inhibited, the endogenous pathway is upregulated (and vice versa) to maintain cellular cholesterol levels. Thus, based on inhibitor studies, the C. burnetii PV is adaptable in acquiring cholesterol from multiple sources. Whether the PV can intercept de novo ER-derived cholesterol before or after its transit through the Golgi apparatus en route to the plasma membrane, or if the PV is capable of directly utilizing cholesterol esters hydrolized from LDL particles trafficked to the PV, is unknown and requires additional study.

Maturation of the C. burnetii PV and replication of the organism are exquisitely sensitive to ketoconazole and imipramine, inhibitors of cholesterol biosynthesis and LDL transport respectively. Both drugs completely inhibit C. burnetii replication when added at 0 h post infection. Moreover, treatments result in a net loss of C. burnetii genomes, indicating degradation of the organism. This result is somewhat surprising as antibiotics are generally bacteriostatic to C. burnetii in cell culture, and treatments do not typically result in a net loss of viable organisms relative to the original inoculum (Maurin et al., 1992b; Brennan and Samuel, 2003). Inhibition by ketoconazole is reversible with surviving organisms replicating at a normal rate after its removal. Conversely, inhibition by imipramine is irreversible and furthermore, the drug is more bactericidal when added at 2 days post infection when intracellular C. burnetii are in the replicatively active but environmentally fragile LCV form (Coleman et al., 2004). This is not the case with the effect of ketoconazole, which changes from cidal to static when added at 2 days post infection. Thus, although both ketoconazole and imipramine completely inhibit formation of mature C. burnetii PV, the bactericidal effects of imipramine occur regardless of the phase of C. burnetii morphological development or the state of PV maturation. Residence within a mature PV and/or morphogenesis to the LCV appear to somehow protect C. burnetii from killing by ketoconazole. Collectively, these results suggest the anti-C. burnetii activities of ketoconazole and imipramine occur through different cholesterol-related mechanisms.

What are potential functions of cholesterol in the C. burnetii PV membrane? Cholesterol affects the biophysical characteristics of membranes and consequently, the structure and function of organelles. Cholesterol increases the mechanical strength of a phospholipid bilayer (Needham et al., 1988) and decreases its ionic permeability (Corvera et al., 1992). Thus, cholesterol may contribute to the mechanical stability of the PV and maintenance of its luminal acidic pH (Maurin et al., 1992a; Heinzen and Hackstadt, 1997). Decoration of the PV by flotillin-1 and flotillin-2 suggests cholesterol, in conjunction with sphingolipids, may form lipid raft microdomains in the PV membrane. Lipid rafts are enriched in the plasma membrane and are implicated in a number of important cellular functions including signal transduction, cytoskeletal interactions and membrane fusion (Simons and Ikonen, 2000). These functions are largely mediated by proteins that specifically partition to raft domains such as glycosylphosphatidylinositol-anchored proteins (Li et al., 2003). Although it has been proposed that lipid rafts are excluded from degradative lysosomes (Simons and Gruenberg, 2000), there is increasing evidence for their presence in late phagosomes where they presumably have functional roles (Dermine et al., 2001; Li et al., 2003). Therefore, lipid raft-based signalling may control, among other things, the promiscuous fusogenicity of the C. burnetii PV. Lipid rafts are also associated with the cholesterol-rich vacuole of S. enterica serovar Typhimurium (Catron et al., 2002; Knodler et al., 2003). Interestingly, the accumulation of lipid rafts in lysosomes has been implicated as a cause of some lysosomal storage diseases (Simons and Gruenberg, 2000). Proposed pathological consequences of this process include the accumulation of MLBs that may alter normal trafficking of the organelle and the disruption of lysosome biogenesis (Simons and Gruenberg, 2000). As discussed earlier, MLBs are a hallmark of the C. burnetii PV and infected cells contain a paucity of lysosomes relative to uninfected cells (Howe et al., 2003).

The importance of cholesterol and its precursors to C. burnetii infectivity raises some interesting questions related to the pathophysiology of Q fever. First, are individuals with elevated plasma cholesterol more likely to experience symptomatic acute or chronic disease following infection? Epidemiologic evidence suggests acute infection by C. burnetii can result in long-term arteriovascular disease such as cardiac ischaemia (Lovey et al., 1999). Cholesterol ester-laden macrophages known as foam cells are thought to contribute to the inflammatory damage observed in these conditions (Kalayoglu and Byrne, 1998; Linton and Fazio, 2003). Second, what role do sterols play in the sex-related manifestation of Q fever? Men are more likely to experience symptomatic Q fever than women and pregnancy, with its associated changes in hormone levels, predisposes women to evolution of chronic disease following acute Q fever (Tissot Dupont et al., 1992; Raoult et al., 2002). Sterol-derived sex hormones affect the outcome of infection by a variety of microbes including C. burnetii (Leone et al. 2004; Roberts et al., 2001; Kaushic et al., 2000) and some, such as estrogen and progesterone, directly reduce cellular cholesterol levels (McCrohon et al., 1999). Finally, would cholesterol-lowering drugs augment antibiotic treatment of acute or chronic Q fever? Lovastatin was recently demonstrated to lower S. enterica serovar Typhimurium replication in a murine model of infection (Catron et al., 2004).

Experimental procedures

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Coxiella burnetii, cell culture and inhibitor treatments

Coxiella burnetii (Nine Mile strain in phase II) was propagated in African green monkey kidney cells (Vero) fibroblasts (CCL-81; American Type Culture Collection, Manassas, VA) and purified as previously described (Hackstadt et al., 1992). Vero cells were maintained in RPMI supplemented with 10% fetal bovine serum (FBS) (Invitrogen, Carlsbad, CA) and incubated at 37°C in 5% CO2. Confluent Vero cells in six-well plates (inhibitor studies) or 24-well plates containing 12 mm coverslips (fluorescence microscopy) were incubated with C. burnetii at a multiplicity of infection of 10 for 2 h at room temperature to allow for attachment and internalization. Extracellular organisms were then washed from cell monolayers, fresh media was added, and incubation continued at 37°C. (This time point was considered 0 h post infection.) Inhibitors were added at the indicated times and replaced every 3 days. Inhibitor treatments did not affect the integrity of Vero cell monolayers as assessed by light microscopy. Inhibitors were used at the following final concentrations: lovastatin (1 µM), ketoconazole (20 µM), imipramine (100 µM), U18666A (5 µM) (Sigma-Aldrich, St Louis, MO) and 25 hydroxycholesterol (25 µM) (25-OHC) (Steraloids, Newport, RI).

Fluorescent staining

All fixation and staining procedures were conducted at room temperature. For immunofluorescent labelling, infected Vero cells on 12 mm coverslips were fixed and permeabilized by incubation for 2 min in 100% cold methanol. For filipin staining, cells were fixed with 4% paraformaldehyde in phosphate buffered saline (PBS; 150 mM NaCl, 10 mM NaPO4, pH 7.2) for 10 min, followed by permeabilization for 10 min with 0.1% saponin in PBS containing 10% goat serum. C. burnetii were stained with antiserum from a convalescent guinea pig followed by Alexa Fluor 594 goat anti-guinea pig IgG (H&L) (Invitrogen). The C. burnetii PV membrane was stained with monoclonal antibodies directed against human LAMP-1 (Clone H4A3; Developmental Studies Hybridoma Bank, University of Iowa, Iowa City, IA), or human flotillin-1 or flotillin-2 (BD Pharmingen, San Diego, CA) followed by AlexaFluor 488 goat anti-mouse IgG (Invitrogen). Cellular cholesterol was stained with filipin (50 µg ml−1) (Sigma-Aldrich) as previously described (Kruth et al., 1986). Nuclear DNA was stained by incubating fixed cells for 5 min with DRAQ5 (AXXORA, San Diego, CA) diluted 1:1000 in PBS. Filipin and DRAQ5 staining was conducted after immunostaining. Coverslips were mounted onto glass slides with Mowiol (Sigma-Aldrich).

Microscopy

Epifluorescent images were obtained using a Nikon TE-2000 E inverted microscope equipped with a CoolSNAP HQ digital camera (Roper Scientific, Tuscon, AZ). Confocal fluorescence microscopy was conducted with a modified Perkin Elmer UltraView spinning disc confocal system connected to a Nikon Eclipse TE-2000 S microscope. Images were viewed with a 60× Plan-Apo oil immersion objective (Nikon, NA 1.4) and a Photometrics Cascade:512F digital camera (Roper Scientific). Epifluorescent and confocal images (0.2 µm sections) were acquired using Metamorph software (Universal Imaging, Dowingtown, PA). All images were processed using ImageJ software (written by Wayne Rasband at the US National Institutes of Health and available by anonymous FTP from http://zippy.nimh.nih.gov) and Adobe Photoshop (Adobe Systems, San Jose, CA).

Quantification of cellular cholesterol

The cholesterol content of infected and uninfected Vero cells was determined using an enzyme-based assay. Vero cell monolayers in 25 cm2 flasks were washed with PBS and removed with trypsin (Invitrogen). Cellular lipids were extracted in hexane : isopropanol (Sigma-Aldrich) (3:2 v/v) and total cholesterol was quantified using an Amplex Red Cholesterol Assay Kit (Molecular Probes, Eugene, OR) according to the manufacturer's instructions.

Quantitative PCR of C. burnetii genome equivalents

Quantification of C. burnetii genome equivalents was conducted using TaqMan qPCR. Infected or mock infected Vero cells in individual wells of a six-well tissue culture plate were harvested by scraping. Total DNA was extracted using an UltraClean microbial DNA Isolation Kit (MoBio laboratories, Carlsbad, CA) including an optional heat treatment (85°C for 30 min) before physical disruption of the samples. The forward primer (5′-GCG CAATACGCTCAATCACA-3′), reverse primer (5′-CCATGGCCCC AATTCTCTT-3′) and probe (5′-CCGGAGATACCGGCGGTGGG-3′) were designed with PrimerExpress software (Applied Biosystems, Foster City, CA) and are specific for the C. burnetii dotA gene. Dilutions from 103To 107 copies of the C. burnetii dotA sequence cloned into pCR2.1 TOPO (Invitrogen) were used to generate a standard curve. Amplification efficiencies of plasmid and genomic DNA were the same. Amplification was conducted using a Prism 7000 sequence detection system (Applied Biosystems).

Quantitative reverse transcription PCR of cholesterol metabolism genes

Reverse-transcriptase qPCR was used to quantify the expression of Vero cell genes involved in cholesterol metabolism. Total RNA was harvested from infected or mock infected Vero cells cultivated in six-well plates by lysing in situ with 1 ml of RNA Wiz (Ambion, Austin, TX). RNA was purified using a Qiashredder and an RNeasy Mini kit (QIAGEN, Valencia, CA) following the manufacturers’ protocol. A High Capacity cDNA Archive kit (Applied Biosystems) was used to synthesize cDNA from RNA. Primers and probe specific for genes encoding LDLR, HMGCR, LSS and DHCR24 were purchased from Applied Biosystems as ready-to-use kits (Assay-on-Demand assay numbers: Hs99999903_ml, Hs00158906_ml, Hs00181192_ml, Hs00168352_ml and Hs00207388). LSS, LDLR, HMGCR and DHCR24 expression was normalized to expression of the gene encoding β-actin.

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

We thank Harlan Caldwell, Ted Hackstadt, Olivia Steele-Mortimer, Leigh Knodler and Shelly Robertson for review of this manuscript, and Anita Mora for assistance with graphics. This research was supported by the Intramural Research Program of the National Institutes of Health, National Institute of Allergy and Infectious Diseases.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References