ABSTRACT
Coxiella burnetii is an obligate intracellular bacterial pathogen and a causative agent of culture-negative endocarditis. While C. burnetii initially infects alveolar macrophages, it was also found in lipid droplet (LD)-containing foamy macrophages in the cardiac valves of endocarditis patients. In addition, transcriptional studies of C. burnetii-infected macrophages reported differential regulation of the LD coat protein-encoding gene perilipin 2 (plin-2). To further investigate the relationship between LDs and C. burnetii, we compared LD numbers in mock-infected and C. burnetii-infected alveolar macrophages using fluorescence microscopy. Compared to only 10% of mock-infected cells, 50% of C. burnetii-infected cells had more than 50 LDs/cell as early as 24 hours post-infection, indicating a significant increase in LDs in infected cells. Increased LDs required the C. burnetii Type 4B Secretion System (T4BSS), a major virulence factor that manipulates host cellular processes by secreting bacterial effector proteins into the host cell cytoplasm. To determine the importance of LDs during C. burnetii infection, we assessed the effect of manipulating LD homeostasis on C. burnetii intracellular growth. Surprisingly, blocking LD formation with the pharmacological inhibitors triascin C or T863, or knocking out acyl-CoA transferase-1 (acat-1) in alveolar macrophages, increased C. burnetii growth at least 2-fold. Conversely, preventing LD lipolysis by inhibiting adipose triglyceride lipase (ATGL) with atglistatin almost completely blocked bacterial growth, suggesting LD breakdown is essential for C. burnetii. Together these data suggest that LDs are detrimental to C. burnetii and maintenance of LD homeostasis, possibly via the T4BSS, is critical for bacterial growth.
IMPORTANCE Host neutral lipid storage organelles known as lipid droplets (LDs) serve as a source of energy, nutrients, or signaling lipids during infection with intracellular bacteria, such as Mycobacterium spp., and Chlamydia spp. LDs have also been associated with infection of the intracellular bacterial pathogen Coxiella burnetii, a significant cause of culture-negative infectious endocarditis. Although C. burnetii was found in LD-rich foam macrophages in endocarditis patients, little is known about the host LD-C. burnetii relationship. We demonstrated a C. burnetii Type 4B Secretion System (T4BSS)-dependent LD accumulation in macrophages, suggesting that the T4BSS plays a key role in regulating host cell LD formation or breakdown. Further, manipulation of LD homeostasis significantly affected C. burnetii intracellular growth, indicating LDs play an important role during C. burnetii infection. Since C. burnetii endocarditis has a 19% mortality rate even in treated patients, exploring the LD-C. burnetii association might identify novel therapeutic targets.
INTRODUCTION
Lipid droplets (LDs) are dynamic cytoplasmic organelles which store cellular lipids in eukaryotic cells. LDs are uniquely comprised of a phospholipid monolayer surrounding a hydrophobic core of neutral lipids, primarily sterol esters and triacylglycerols. LD assembly begins with neutral lipid synthesis, where fatty acyl CoA synthetases generate long chain fatty acids which are converted to sterol esters and triacyglycerols by acyl-CoA:cholesterol acyltransferase (ACAT) and acyl-CoA:diacylglycerol acyltransferase (DGAT), respectively. Progressive accumulation of neutral lipids in the ER leads to budding of the lipid ester globule, surrounded by the ER membrane cytoplasmic leaflet (1, 2). Hormone sensitive lipase (HSL) (3) and adipose triglyceride lipase (ATGL) (4) mediate LD breakdown and release of free cholesterol and fatty acids. Functionally, LDs serve as intracellular lipid reservoirs for membrane synthesis or energy metabolism. In addition, LDs are linked to a range of cellular functions including protein storage, protein degradation and signaling (2, 5).
LDs are emerging as important players during host-pathogen interactions. During infection of host cells, Hepatitis C virus (HCV) (6) and Dengue virus (7) co-opt LDs as platforms for viral assembly and replication. Even though pharmacological manipulation of LD content reduced viral numbers, the importance of LDs during viral infection still remains elusive (8). Increased LD numbers in host cells is observed upon infection with several pathogens including HCV (6) and Dengue virus (7), as well as the protozoan parasites Trypanosoma cruzi (9), Plasmodium berghei (10), Toxoplasma gondii (11), Leishmania amazonensis (12) and Leishmania major (13). In addition, the intracellular bacterial pathogens Chlamydia spp. (14), Mycobacterium spp. (15-18), Orientia tsutugamushi (19), and Salmonella typhimurium (20) also increase LD numbers in infected cells. C. trachomatis (14, 21) and M. tuberculosis (15) are thought to use triacylglycerol and cholesterol esters stored in LDs as a major source of energy and nutrients. Furthermore, in cells infected with M. leprae (22), M. bovis (16), T. cruzi (23), and Leishmania infantum chagasi (24), LDs serve as a source of prostaglandin and leukotriene eicosanoids, important signaling lipids which modulate inflammation and the immune response. These LD-derived eicosanoids potentially favor intracellular pathogen survival by downregulating the immune response (25).
LDs have been implicated during infection by Coxiella burnetii, a gram-negative intracellular bacterium and the causative agent of human Q fever. Primarily spread through aerosols, C. burnetii acute infection is characterized by a debilitating flu-like illness, while chronic disease results in endocarditis. Although in vitro and in vivo C. burnetii can infect a wide range of cells including epithelial cells and fibroblasts, the bacterium first infects alveolar macrophages during natural infection. Inside the host cell, C. burnetii directs formation of a specialized lysosome-like compartment called the parasitophorous vacuole (PV) which is essential for C. burnetii survival. PV biogenesis requires the C. burnetii type 4B secretion system (T4BSS), which secretes effector proteins into the host cell cytoplasm where they manipulate a wide range of cellular processes. While not established to be T4BSS-dependent, C. burnetii is thought to manipulate LDs and other components of host cell lipid metabolism (26-29). C. burnetii-containing LD-filled foam cells were found in heart valves of an infected patient (30), and LDs were observed in the C. burnetii PV lumen of infected human alveolar macrophages (31). Further, two separate microarray analyses reported differential regulation of the LD coat protein plin-2 in C. burnetii-infected human macrophage-like cells (THP-1) (28, 29), suggesting C. burnetii induced changes in host cell LDs. Intriguingly, siRNA depletion of the phospholipase involved in LD breakdown, PNPLA2 (also known as ATGL), increased the number of C. burnetii PVs in HeLa epithelial cells (32). In addition, treatment of monkey kidney epithelial cells (Vero cells) with a broad spectrum antiviral molecule ST699, which localizes to the host cell LDs, inhibited C. burnetii intracellular growth (33). Despite these observations, the importance of LDs during C. burnetii infection is not known. In this study, we further examined the relationship between host LDs and C. burnetii. We observed a T4BSS-dependent increase in LD numbers in infected alveolar and monocyte-derived macrophages. Furthermore, manipulation of LD homeostasis significantly altered C. burnetii intracellular growth, thus strongly indicating that LDs play an important role during C. burnetii infection.
RESULTS AND DISCUSSION
C. burnetii infection results in host cell LD accumulation
To examine the role of LDs in C. burnetii pathogenesis, we first quantitated LDs in infected cells. Previously, two separate microarray studies reported differential regulation of the LD coat protein-encoding gene plin-2 in C. burnetii-infected cells (28, 29), generally indicating changes in LD numbers. As C. burnetii preferentially infects alveolar macrophages during natural infection, we utilized a mouse alveolar macrophage cell line (MH-S) previously shown as a model for C. burnetii infection (34). Cells were stained by immunofluorescence for the LD coat protein PLIN2, and LD number per cell determine by fluorescence microscopy. During a 4-day mock infection, the majority of macrophages (~80%) had less than 50 LDs per cell, irrespective of the time point (Figure 1). In contrast, we observed a significant increase in LDs per cell at 1, 2 and 4 days after C. burnetii infection, with >60% of infected cells having more than 50 LDs. Notably, LD accumulation occurred as early as 1 day post-infection, when the PV has not expanded and the bacteria are not in log growth. This suggests that LD accumulation is not a host response to a large PV, but could be a result of C. burnetii directly manipulating host LDs.
Host cell LD accumulation is dependent on the C. burnetii Type 4B Secretion System (T4BSS)
Previously, based on microarray analysis of C. burnetii-infected cells where bacterial protein synthesis was blocked, Mahapatra et al. identified 36 host cell genes specifically regulated by C. burnetii during early stages of infection. These genes were predominantly involved in the innate immune response, cell death and proliferation, vesicular trafficking, cytoskeletal organization and lipid homeostasis. Interestingly, changes in plin-2 expression level in infected cells was dependent on C. burnetii protein synthesis (28). Together with our data, this suggests that C. burnetii actively manipulates host LDs as early as day 1 post-infection, possibly through C. burnetii T4BSS effector proteins secreted into the host cytoplasm. The C. burnetii T4BSS detectably secretes effector proteins beginning 1 hour post-infection in bone marrow-derived macrophages and 8 hours post-infection in HeLa cells (35). To test if the C. burnetii T4BSS was responsible for LD accumulation in murine alveolar macrophages, LD numbers were analyzed at 1, 2, and 4 days after infection with a T4BSS dotA mutant. While at least 75% of the wild-type C. burnetii-infected cells had >50 LDs at all time points, only 40% cells infected with T4BSS mutant had >50 LDs per cell, similar to mock-infected cells (Figure 2A and B). This suggests the T4BSS is involved in increased LDs in C. burnetii-infected alveolar macrophages.
To confirm this finding, we analyzed LD numbers in human macrophage-like cells (THP-1). Similar to murine alveolar macrophages, when compared to mock-or T4SS mutant-infected cells, wild-type C. burnetii-infected THP-1 cells had increased LD numbers at 1 and 4 days post-infection (Figure 2C). Interestingly, we did not observe T4BSS-dependent LD accumulation at 2 days post-infection. However, these results demonstrate that the C. burnetii-induced increase in LDs in both human and mouse macrophages is species-independent and dependent on the C. burnetii T4BSS.
The requirement for the C. burnetii T4BSS suggests that one or more C. burnetii T4BSS effector proteins may actively manipulate LDs. Other bacteria have been shown to target host LDs via effector proteins. For example, the C. trachomatis secreted protein Lda3 localizes to the LD surface and is involved in LD translocation into the Chlamydia-containing inclusion (21). The Salmonella Typhimurium type 3 secretion system (T3SS) effector protein SseJ esterifies cholesterol and increases LD numbers when expressed in epithelial and macrophage cells (20). Thus far, none of the identified C. burnetii T4BSS secreted effector proteins localize to host LDs. While C. burnetii effectors might directly target proteins involved in LD formation or LDs themselves, it is also possible that fission of preexisting LDs, and not de novo formation, is responsible for increased numbers of LDs (36). As observed in Figure 2A, the LD size in C. burnetii-infected macrophages appeared smaller than the mock-or T4BSS mutant-infected macrophages. Smaller, more numerous LDs might result from C. burnetii T4BSS-mediated fission of the existing LDs.
While C. burnetii T4SS effector proteins might directly target LDs or LD pathways, LD accumulation may also be a host innate immune response. In other diseases, LD accumulation occurs during the inflammatory response in macrophages in atherosclerotic lesions (37), leukocytes from joints of patients with inflammatory arthritis (38), and eosinophils in allergic inflammation (39). Thus, an innate immune response to the T4BSS apparatus or T4BSS effector proteins may increase LD numbers in C. burnetii-infected macrophages. While our data demonstrate that the C. burnetii T4BSS is involved in LD accumulation in both mouse and human macrophages, the bacterial effector proteins and the specific LD processes involved remain unknown.
Blocking LD formation increases C. burnetii growth
Given our finding that the C. burnetii T4BSS may manipulate host LD accumulation, we next assessed the importance of LDs during C. burnetii infection. We first blocked LD formation using triascin C, a long chain fatty acyl CoA synthetase inhibitor (40). Compared to vehicle control, triascin C significantly reduced macrophage LDs, with <5 LDs per cell (Figure 3A). We next treated macrophages with triascin C during C. burnetii infection, and quantitated bacterial growth using a fluorescent infectious focus-forming unit (FFU) assay. At various times post-infection, we recovered bacteria from MH-S cells, replated onto a monolayer of Vero cells, and incubated for 5 days. After staining for C. burnetii, we counted the number of fluorescent foci, with 1 focus equivalent to 1 viable bacterium. Surprisingly, compared to vehicle-treated cells, triascin C treatment increased C. burnetii growth 5-fold at 4 days post-infection (Figure 3B).
To further validate these results, we used CRISPR/Cas-9 to knockout acat-1 (Figure 3C), the enzyme responsible for sterol esterification. While acat-1-/- LDs lack sterol esters, fluorescence microscopy revealed similar LD numbers in wild-type and acat-1-/- cells (Figure 3D). Compared to wild-type cells, C. burnetii growth in acat-1-/- cells increased 2-fold at 4 days post-infection (Figure 3E), indicating that blocking sterol esterification favors C. burnetii growth. To further deplete both triacylglycerol-and sterol ester-containing LDs, we treated acat-1-/- cells with the DGAT1 inhibitor T863, which specifically blocks formation of triacylglycerols (41). T863 treatment significantly reduced LDs in acat-1-/- macrophages, compared to untreated wild-type or acat-1-/- macrophages (Figure 3D). C. burnetii growth increased 2-fold in T863-treated acat-1-/- cells compared to vehicle-treated acat-1-/- cells (Figure 3F), demonstrating that blocking both sterol ester- and triacylglycerol-containing LDs improves C. burnetii growth.
These studies demonstrate that both pharmaceutical and genetic approaches to blocking LD formation increases C. burnetii fitness in macrophages. Interestingly, Mahapatra et al. observed an increase in plin-2 transcript levels after transiently blocking C. burnetii protein synthesis, suggesting that wild-type C. burnetii downregulates LD formation (28). It is not clear why inhibiting LD formation is advantageous to C. burnetii. While it is not known if C. burnetii uses host fatty acids, unesterified free fatty acids or sterols in LD-deficient cells may support C. burnetii growth.
Inhibiting LD breakdown blocks C. burnetii growth
Because blocking LD formation appeared to benefit C. burnetii, we next examined C. burnetii growth after inhibiting LD breakdown and increasing LD numbers. When cells or tissues need fatty acids or cholesterol, cytosolic lipases such as hormone sensitive lipase (HSL) (2, 3) and adipose triglyceride lipase (ATGL) (4) hydrolyze triacylglycerols and sterol esters stored in LDs. To block LD breakdown in murine macrophages, we inhibited ATGL with the selective and competitive inhibitor atglistatin, which binds the ATGL patatin-like phospholipase domain (42). To eliminate the possibility of ATGL inhibiting a C. burnetii phospholipase, we first measured viability of axenic C. burnetii cultures in the presence or absence of atglistatin (Figure 4A). Treatment for 4 days had no effect on axenic bacterial growth, indicating atglistatin does not directly affect C. burnetii.
We next tested the effect of atglistatin on intracellular bacteria. After atglistatin treatment of wild-type MH-S cells, we observed larger LDs by immunofluorescence microscopy, although the number did not significantly increase (Figure 3A). Interestingly, C. burnetii intracellular growth in atglistatin-treated wild-type MH-S cells decreased 5-fold, with essentially no growth (Figure 4B). Further, atglistatin-treated acat-1-/- cells, which contain triacylglycerol-containing LDs, also showed reduced bacterial growth (Figure 4C). Together, these data demonstrate that blocking LD breakdown significantly inhibits intracellular C. burnetii growth, regardless of LD composition.
Our data suggest that LD breakdown is essential for C. burnetii intracellular growth in macrophages. Previously, siRNA knockdown of ATGL in HeLa cells increased the number of C. burnetii PVs, although the effect on C. burnetii growth was not determined (32). LDs are less abundant in HeLa cells compared to macrophages, and LDs and LD breakdown may play a larger role during C. burnetii infection of macrophages. LD breakdown liberates fatty acids, which can be reesterified or serve as signaling cofactors, building blocks for membranes, or substrates forβ-oxidation to generate energy (2). Several intracellular bacteria use LD-derived fatty acids as a source of energy and carbon. M. tuberculosis, which can make its own LDs, converts host-derived fatty acids into triacylglycerol, which is then deposited in bacterial LDs (15). The M. tuberculosis lipase is hypothesized to release stored bacterial fatty acids, but can also degrade host LD-derived triacylglycerol (43). Host LDs translocated into the C. trachomatis inclusion may be broken down to provide lipids for bacterial growth (14). We did not observe LDs in the C. burnetii PV lumen in murine macrophages or THP-1 macrophage-like cells, in contrast to reports in human alveolar macrophages (31). It is not known if free fatty acids or sterols liberated from LDs, either in the cytosol or possibly the PV lumen, support C. burnetii growth.
In addition to serving as a source of free fatty acids and sterols, macrophage LDs are rich in substrates and enzymes that generate prostaglandins and leukotrienes, which are arachidonic acid-derived inflammatory lipid mediators (44, 45). In M. leprae-infected Schwann cells and M. bovis BCG-infected macrophages, increased LD biogenesis correlates with increased production of prostaglandin E2 (PGE2), linking LDs to the production of innate immunity modulators (17, 46). Thus, LDs can serve as a source of inflammatory mediators in response to pathogen infection. Interestingly, elevated levels of PGE2 were observed in C. burnetii endocarditis patients and linked to C. burnetii-mediated immunosuppression. Koster et al. reported lymphocytes from chronic Q fever patients being unresponsive to C. burnetii antigens, an effect reversed by PGE2 suppression with indomethacin (47). In addition, after stimulation with C. burnetii antigens, monocytes from Q fever patients produced PGE2, which in turn downregulated T lymphocyte-mediated IL-2 and IFNγ production. Interestingly, PGE2 synthesis inhibitor Piroxicam reversed this downregulation of pro-inflammatory cytokine production (48). Thus, while PGE2 appears to play a role in Q fever patients, the relationship between C. burnetii-induced LDs and PGE2 production is not known. Considering that LD breakdown can serve multiple functions, C. burnetii could use LDs either as a source of nutrients or for production of lipid immune mediators like PGE2, which could then modulate the host cell response to promote C. burnetii intracellular growth.
In summary, our data demonstrate that LD homeostasis is important for C. burnetii intracellular survival. Because the C. burnetii T4BSS is involved in LD accumulation, characterizing bacterial T4BSS effector proteins that target host LD homeostasis will help further understand the role of LDs in C. burnetii pathogenesis.
MATERIALS AND METHODS
Bacteria and mammalian cells
C. burnetii Nine Mile Phase II (NMII; clone 4, RSA439) were purified from Vero cells (African green monkey kidney epithelial cells, ATCC CCL-81; American Type Culture Collection, Manassas, VA) and stored as previously described (49). For experiments examining T4BSS-dependent accumulation of LDs, NMII and the dotA mutant (50) were grown for 4 days in ACCM-2, washed twice with phosphate buffered saline (PBS) and stored as previously described (51). Vero, mouse alveolar macrophages (MH-S; ATCC CRL-2019) and human monocytes (THP-1; ATCC TIB-202) were maintained in RPMI (Roswell Park Memorial Institute) 1640 medium (Corning, New York, NY, USA) containing 10% fetal bovine serum (Atlanta Biologicals, Norcross, GA, USA) at 37°C and 5% CO2 and human embryonic kidney 293 (HEK293T; ATCC CRL-3216) in DMEM (Dulbecco’s Modified Eagle Medium) (Corning, New York, NY, USA) containing 10% fetal bovine serum at 37°C and 5% CO2. THP-1 cells were differentiated with 200 nM of phorbol 12-myristate 13-acetate (PMA) for 24 hours. PMA was removed, and the cells rested for 48 hours prior to infection. The multiplicity of infection (MOI) was optimized for each bacterial stock, cell type and infection condition for a final infection of ~1 internalized bacterium/cell at 37oC and 5% CO2.
Generating acat-1-/- MH-S cell line
The guide RNA sequence 5′TCGCGTCTCCATGGCTGCCC3′ to mouse acat-1 was selected using the optimized CRISPR design site crispr.mit.edu. Oligonucleotides were synthesized (IDT, Coralville, IA, USA), annealed, and cloned into the lentiCRISPRv2 plasmid (a gift from Feng Zhang, Addgene # 52961, Cambridge, MA, USA) (52), at the BsmBI restriction site to generate plentiCRISPRv2-acat-1. To generate lentivirus, HEK293T cells were co-transfected with plentiCRISPRv2-acat-1 and packaging plasmids pVSVg (Addgene # 8454), pRSV-Rev (Addgene # 12253), and pMDLg/pRRE (Addgene # 12251) using FuGENE6 reagent (Promega, Madison, WI, USA). At 48 hours post-transfection, supernatant was collected and centrifuged at 3000xg, and then filtered with 0.45μm filter to remove cells and debris. Supernatant was concentrated using the Lenti-X concentrator (Catalog # PT4421-2, Clontech, USA) and viral RNA isolated using Viral RNA isolation kit (Catalog # 740956, Macherey-Nagel, Germany) to determine viral titer using Lenti-X qRT-PCR titration kit (Catalog # PT4006-2, Clontech). Viral titers were optimized for transduction of MH-S cells to generate stable acat-1-/- cells.
2×105 MH-S cells were plated in a 6 well plate and transduced with 5.8 × 106 viral particles/ml. 1 μg/ml puromycin was used for selection 48 hours post-transduction and continued for 24 hours. The puromycin was then removed and the cells allowed to recover before isolating individual clones by limiting dilution.
To confirm disruption of acat-1, clones were lysed in 2% SDS (Sigma-Aldrich, St. Louis, MO, USA) for SDS-PAGE and immunoblotting with 1:1000 rabbit anti-mouse ACAT1-specific antibody (Catalog # NBP189285, Novus Biologicals, Littleton, CO, USA) and 1:4000 GAPDH loading control monoclonal antibody (Catalog # MA5-15738, ThermoFisher Scientific, Waltham, MA, USA). The multiplicity of infection (MOI) was optimized for each bacterial stock for a final infection of ~1 internalized bacterium/cell.
LD quantitation
1×105 MH-S cells were plated onto ibidi-treated channel μslide VI0.4 (3×103 cells per channel; Ibidi, Verona, WI) and allowed to adhere overnight. After infecting with C. burnetii for 1 hour, cells were gently washed with phosphate buffered saline (PBS) to remove extracellular bacteria, and incubated in 10% FBS-RPMI. At different times post-infection, infected cells were fixed with 2.5% paraformaldehyde on ice for 15 min, then permeabilized/blocked for 15 min with 0.1% saponin and 1% bovine serum albumin (BSA) in PBS (saponin-BSA-PBS) and stained with 1:1000 rabbit anti-mouse PLIN2 primary antibody (Catalog # PA1-16972, ThermoFisher Scientific), 1:2000 guinea-pig anti-C. burnetii primary antibody (53) and 1:1000 rat anti-LAMP (Catalog # 553792, BD Biosciences) primary antibody in saponin-BSA-PBS for 1 hour. THP-1 cells were stained with 1:500 guinea-pig anti-human PLIN2 primary antibody (Catalog # 20R-AP002, Fitzgerald Industries International, Acton, MA), 1:2000 rabbit anti-C. burnetii primary antibody and 1:1000 rat anti-LAMP primary antibody in saponin-BSA-PBS for 1 hour. After three washes with PBS, cells were stained with 1:2000 AlexaFluor 488 anti-rabbit, AlexaFluor 594 anti-guinea pig and AlexaFluor 647 anti-rat secondary antibodies (Invitrogen) for 1 hour. ProLong Gold mount (Invitrogen) was added to the wells after washing with PBS and slides visualized on a Leica inverted DMI6000B microscope (100X oil). The number of LDs per cell was quantitated for 50 cells per condition in three individual experiments, with only bacteria-containing cells counted for C. burnetii-infected cells. Each experiment was done in duplicate.
Inhibitors
Each LD homeostasis inhibitor used was diluted in DMSO based on manufacturer’s instructions and optimum inhibitor concentration was determined based on 100% host cell viability determined by trypan blue staining, and changes in LD numbers per cell. The optimum concentrations determined for each inhibitor was: Triascin C (Enzo Life Sciences, Farmingdale, NY, USA) – 10 μM, T863 (Sigma-Aldrich) – 10 μM, Atglistatin (Cayman Chemicals, Ann Arbor, MI, USA) – 20 μM.
C. burnetii growth by fluorescent infectious focus-forming unit (FFU) assay
To measure growth of C. burnetii in wild-type and acat-1-/- MH-S cells, 5×104 cells/well were infected for 1 hour in a 48 well plate, washed with PBS, and then incubated with media containing respective vehicle and inhibitors. At the indicated time points, the media was removed and cells were incubated with sterile water for 5 min, pipetted up and down and the lysate diluted 1:5 in 2% FBS-RPMI. Serial dilutions were added to 24 well plate containing confluent monolayers of Vero cells, incubated for 5 days, fixed with methanol and stained with rabbit anti-C. burnetii antibody as well as DAPI to confirm monolayer integrity. Four fields per well were captured on an Evos automated microscope (ThermoFisher) with 4X objective and fluorescent foci units were quantitated using ImageJ. Each experiment was done in duplicate.
Atglistatin-treatment of C. burnetii axenic cultures
To test bacterial sensitivity to atglistatin, ACCM-2 was inoculated at approximately 1x105 bacteria/ml with C. burnetii NMII and grown for 3 days as previously described (51). Bacteria (500 μl) were then incubated with DMSO or atglistatin in 24 well plates under normal C. burnetii culture conditions. Media was replenished every 24 hours by centrifuging the supernatant at 20000×g for 10 min, and bacterial pellet resuspended in new media containing inhibitor. After 4 days, bacteria were diluted 1:10 in 2% FBS-RPMI and serial dilutions were added to confluent Vero cell monolayers in a 96 well ibidi-treated μplate. At 5 days post-infection, the plate was stained and fluorescent foci were determined as above. Each experiment was done in duplicate.
Statistical analysis
Statistical analyses were performed using ordinary one-way ANOVA or two-way ANOVA with Tukey’s or Bonferroni’s multiple comparisons test in Prism (GraphPad Software, Inc., La Jolla, CA).
ACKNOWLEDGMENTS
This research was supported by National Institutes of Health (5R21AI21786; S.D.G.), American Heart Association (16POST27250157; M.M.) and National Institutes of Health (5R25GM07592; B.Z.). We thank Anna Justis, Tatiana Clemente and Rajshekar Gaji for critical reading of the manuscript and members of the IU Biology of Intracellular Pathogens Group for helpful suggestions.
We have no conflicts of interest to declare.