Francisella tularensis Catalase Restricts Immune Function by Impairing TRPM2 Channel Activity*

As an innate defense mechanism, macrophages produce reactive oxygen species that weaken pathogens and serve as secondary messengers involved in immune function. The Gram-negative bacterium Francisella tularensis utilizes its antioxidant armature to limit the host immune response, but the mechanism behind this suppression is not defined. Here we establish that F. tularensis limits Ca2+ entry in macrophages, thereby limiting actin reorganization and IL-6 production in a redox-dependent fashion. Wild type (live vaccine strain) or catalase-deficient F. tularensis (ΔkatG) show distinct profiles in their H2O2 scavenging rates, 1 and 0.015 pm/s, respectively. Murine alveolar macrophages infected with ΔkatG display abnormally high basal intracellular Ca2+ concentration that did not increase further in response to H2O2. Additionally, ΔkatG-infected macrophages displayed limited Ca2+ influx in response to ionomycin, as a result of ionophore H2O2 sensitivity. Exogenously added H2O2 or H2O2 generated by ΔkatG likely oxidizes ionomycin and alters its ability to transport Ca2+. Basal increases in cytosolic Ca2+ and insensitivity to H2O2-mediated Ca2+ entry in ΔkatG-infected cells are reversed by the Ca2+ channel inhibitors 2-aminoethyl diphenylborinate and SKF-96365. 2-Aminoethyl diphenylborinate but not SKF-96365 abrogated ΔkatG-dependent increases in macrophage actin remodeling and IL-6 secretion, suggesting a role for H2O2-mediated Ca2+ entry through the transient receptor potential melastatin 2 (TRPM2) channel in macrophages. Indeed, increases in basal Ca2+, actin polymerization, and IL-6 production are reversed in TRPM2-null macrophages infected with ΔkatG. Together, our findings provide compelling evidence that F. tularensis catalase restricts reactive oxygen species to temper macrophage TRPM2-mediated Ca2+ signaling and limit host immune function.

Francisella tularensis is a Gram-negative intracellular bacterium currently classified as a category A biological warfare agent. F. tularensis is transmissible through virtually all routes (airborne, contact, and ingestion), and inhalation of the bacteria is associated with a 50% lethality rate in the absence of antibiotic intervention (1,2). These attributes of F. tularensis necessitate research into the mechanisms of bacterial pathogenicity and/or host response to make any advances in prevention or treatment of infection.
Upon infection, host macrophages take up Francisella via phagocytosis. Francisella has been shown to prevent the formation of the NADPH oxidase complex at the phagosomal membrane, which in turn allows its initial survival within the phagosome (3). Francisella subsequently escapes from the phagosome and proliferates within the cytoplasm prior to exiting the macrophage (4 -6). As a defense mechanism, macrophages produce ROS including, but not limited to, superoxide (O 2 . ) and H 2 O 2 . All F. tularensis subspecies contain multiple mechanisms for ROS detoxification. Francisella superoxide dismutases (SodB and SodC) are responsible for converting superoxide to H 2 O 2 and oxygen. H 2 O 2 is further detoxified by catalase (KatG) and/or alkyl hydroperoxide reductase family members (AhpC) (7,8). Although all subspecies of F. tularensis contain a similar repertoire of antioxidant proteins, their ability to evade destruction by host ROS is variable. We and other groups have previously demonstrated that the highly virulent SchuS4 strain is extremely resistant to ROS when compared with less virulent subspecies (7,9,10). Francisella has been shown to subvert numerous host immune processes following infection to enhance survival. A pathway that is central to Francisella manipulation is the PI3K/ Akt signal cascade (9,(11)(12)(13)(14). Signal propagation through this path is highly reliant on phosphatase and kinase signaling, which have been shown to be regulated in part by the redox state of the cell (15). Our group has previously shown that there is a relationship between the scavenging capacity of Francisella and its ability to restrict redox-mediated protein modifications, thereby altering downstream signaling events. Specifically, the increased ability of Francisella to remove H 2 O 2 from its environment allows for the preservation of host PTEN phosphatase activity, an important negative regulator of AKT (9). The rapid replication of this bacterium coupled with its ability to manipulate the host immune response allows Francisella to remain virtually undetected early after infection, contributing to its ability to widely disseminate.
Parallel to the ability of H 2 O 2 to act on numerous pathways, calcium is also a small molecule mediator that facilitates crosstalk throughout the cell (15)(16)(17)(18)(19). Immune cell function is highly impacted by intracellular Ca 2ϩ concentrations because many processes rely on local alterations of Ca 2ϩ . Subsequent to a rise in intracellular Ca 2ϩ necessary for phagocytosis, Ca 2ϩ signaling is also required for later fusion events, as well as the secretion of inflammatory cytokines (20 -25). Several microarray studies have shown altered regulation of host genes important for Ca 2ϩ homeostasis after Francisella infection, but no further studies into the contribution of Ca 2ϩ signaling to the immune response have been conducted (26,27). Several other intracellular bacteria including Mycobacterium spp., Salmonella spp., and Listeria spp. all have been shown to manipulate host Ca 2ϩ signaling through various mechanisms, suggesting that manipulation by Francisella is also possible (17, 28 -32).
Cytosolic Ca 2ϩ homeostasis is maintained through mobilization of Ca 2ϩ from both intracellular stores and the extracellular space. The interplay between intracellular and extracellular Ca 2ϩ is regulated by diverse signals that control cell function and fate. Receptor-mediated Ca 2ϩ signaling involves release of internal stores mainly from the endoplasmic reticulum and subsequent activation of Ca 2ϩ influx channels at the plasma membrane (33). Other Ca 2ϩ entry pathways exist that are not dependent on the state of internal Ca 2ϩ stores. These pathways are mediated in part by members of the transient receptor potential (TRP) 2 channel family, of which there are 28 channel members (34). Although most members of the TRP family are permeable to Ca 2ϩ , the selectivity varies greatly among members because of differences in both pore structure and regulation (35). Ca 2ϩ mobilization is also intimately associated with the production of ROS. The transient receptor potential melastatin 2 (TRPM2) channels have recently been shown not only to be involved in regulation of immune function as it relates to bacterial infection but also to be gated by H 2 O 2 through cysteine oxidation (36 -38). We have identified a H 2 O 2 -sensitive mobilization of Ca 2ϩ through TRPM2 in Francisella-infected macrophages that appears to be integral for the increased host inflammatory response upon infection with catalase-deficient F. tularensis, suggesting that H 2 O 2 is responsible for TRPM2 activation in response to live vaccine strain (LVS) infection. This study identifies the critical components controlling redox-dependent Ca 2ϩ mobilization that are crucial to the regulation of the inflammatory response to F. tularensis infection.
We propose that in addition to being a means of protection for itself, antioxidants of Francisella, and likely those of other intracellular pathogens, function to modulate Ca 2ϩ -dependent host cell signaling in a redox-dependent manner. We establish that Francisella is capable of scavenging host-derived ROS and that this scavenging leads to early modulation of immune Ca 2ϩ signaling.

Experimental Procedures
Bacterial Strains, Cell Lines, and Media-F. tularensis subsp. Holarctica LVS (ATCC 29684; American Type Culture Collec-tion, Manassas, VA) was used in this study. The catalase (katG) mutants were kindly provided by Dr. Anders Sjostedt. Dr. Karsten R. O. Hazlett generously provided wbtA, fopA, mglA, and iglC mutants. All bacteria were grown on chocolate agar plates supplemented with IsoVitaleX (Becton Dickinson) at 37°C with 5% CO 2 . Bacteria were grown in Mueller Hinton broth supplemented with IsoVitaleX in a shaking incubator (160 rpm) at 37°C where indicated. GFP-expressing LVS were created by transforming electrocompetent Francisella with the plasmid pkk::GFP. Selection of GFP-positive colonies was achieved with kanamycin-containing chocolate agar plates and confirmation via microscopy.
Bacterial Scavenging Assay-Bacteria were grown in MHB broth to an A 600 of 0.200 four times consecutively prior to being spun and washed with PBS. The bacteria were resuspended to an A 600 of 0.200, and 1 ϫ 10 9 bacteria were incubated with 50 M H 2 O 2 at 37°C. At the indicated time points, the aliquots were removed, and H 2 O 2 concentration was determined via the Amplex Red method (41).
ImageStream Analysis-MHS cells were grown to 90% confluence prior to infection. Cells were infected with an MOI of 100 for time indicated. 20,000 events were collected for each sample after gating out debris. Sample data were collected utilizing an ImageStream imaging flow cytometer from Amnis Corp. Data files were analyzed using IDEAS statistical analysis software (Amnis).
Phalloidin-Prior to analysis, cells were thoroughly washed with PBS and lifted with trypsin EDTA at the indicated time point. The cells were fixed and stained with Invitrogen Phalloidin Alexa 488 per the manufacturer's instructions. All cells were spun and resuspended at a concentration of 1 ϫ 10 6 cells/60 l for ImageStream analysis. Immediately prior to analysis, cells were vortexed to disrupt cell clumps. Utilizing a 488 laser, brightfield and GFP channels were collected.
Fura Red-Macrophages were loaded with Fura Red (Invitrogen) for 15 min at 4°C and 30 min at 37°C. The cells were then washed and infected with indicated bacteria for 1 h at an MOI of 100. The wells were washed three times with HBSS containing 100 mM CaCl 2 and 1 mM dextrose and lifted with trypsin at the indicated time point. The cells were concentrated as described above. The samples were excited with a 488 laser, and the brightfield, GFP, and RFP channels were collected on the ImageStream.
An additional panel of LVS mutants was assayed for Fura Red fluorescence utilizing a high throughput plate reader method. Briefly, MHS cells were seeded in a 96-well plate 24 h prior to the experiment. The cells were then loaded and infected as described above. Following three HBSS washes, fluorescence was measured on a FlexStation 3 Molecular Devices plate reader with the software SoftMaxPro 5.4.1 at an excitation wavelength of 420 nm and emission at 650 nm. Calcium Imaging-Macrophages were seeded on #30 glass coverslips 24 h prior to the experiment to a confluence of 80%. The cells were washed several times and loaded with 4 M Fura-2-AM (Invitrogen) for 15 min at 4°C and 30 min at 37°C. The cells were then washed and infected with indicated bacteria for 1 h at an MOI of 100. Coverslips were washed three times with HBSS containing 2 mM CaCl 2 and 1 mM dextrose. Fluorescence experiments were recorded and analyzed using a digital fluorescence imaging system (Intracellular Imaging). At the indicated times, the solutions on the coverslips were exchanged to include 300 M hydrogen peroxide Ϯ 2 mM calcium (42,43).
Intracellular calcium levels were determined real time by ratiometric analysis of the calcium binding dye Fura-2-AM. This cell-permeable dye excites at 340 nm when bound to calcium and 380 nm when unbound. Fluorescence intensities were then measured at the emission wavelength of 510 nm. These intensities were represented as a ratio of 340/380.
Microplate-based Fura-2-AM Analysis-MHS cells were seeded on 96-well plates 24 h before performing assay. At the time of assay, the medium was removed, and the cells were loaded with 4 M Fura-2-AM for 15 min at 4°C followed by incubation at 37°C for 30 min. Then the cells were washed with 1ϫ HBSS with dextrose, three times. The 96-well plate was placed in a FlexStation 3 Molecular Devices plate reader, the treatments were added at specific time points using the Flex mode, and fluorescence measurements were obtained. For monitoring effects of H 2 O 2 on ionomycin Ca 2ϩ transport activity, the medium was removed, and cells were loaded with 4 M Fura-2-AM for 15 min at 4°C followed by incubation at 37°C for 30 min. The cells were washed with 1ϫ HBSS with dextrose and incubated with indicated doses of H 2 O 2 at 37°C for 10 min with no loss in cell viability. 10 M ionomycin with 2 mM Ca 2ϩ was added at 30 s during the read, and fluorescence traces were monitored for 240 s.
Cytokine Analysis-Secretion of proinflammatory cytokine IL-6 was measured 24 h postinfection. Mouse alveolar macrophage cells were plated in 6-well plates and incubated for 24 h at 37°C. The cells were infected with MOI 100 or 50 ng of LPS and incubated for 24 h at 37°C. Supernatants were collected and analyzed using a MSD Mouse Proinflammatory 7-plex ultrasensitive kit according to the manufacturer's instructions. Samples and standards were run in duplicate, and the data were collected utilizing a Sector 2400 imager (Meso Scale Discovery, Gaithersburg, MD).
Bone marrow-derived macrophages were assessed for increased transcription of IL-6 as follows. Total RNA was extracted from cells using an RNeasy isolation kit (Qiagen) per the manufacturer's instructions and stored at Ϫ80°C until use. cDNA was synthesized utilizing a Maxima Universal First Strand cDNA synthesis kit (Fisher) with random hexamers. Real time PCR was performed with an Applied Biosystems 7500 real time thermocycler with SYBR green. Specific primers were used for each target: IL-6 sense, 5Ј-TGG AGT CAC AGA AGG AGT GGC TAA G-3Ј; and IL-6 antisense, 5Ј-TCT GAC CAC AGT GAG GAA TGT CCA C-3Ј. PCR conditions were as follows: 95°C for 10 min, followed by 40 cycles of melting at 95°C for 15 s, and annealing and elongation at 60°C for 1 min. Product specificity was determined by melt curve CT values were normalized to a control housekeeping gene, GAPDH (sense, 5Ј-TTC ACC ACC ATG GAG AAG GC-3Ј; and antisense, 5Ј-GGC ATG GAC TGT GGT CAT GA-3Ј), and the fold change was found using the ⌬⌬CT method of analysis.
Statistical Analysis-All statistical analyses were performed with the GraphPad 5.04 statistical software package. Statistical significance was assessed using the two-tailed Student's t test or one-way analysis of variance with a Tukey-Kramer post-test. p values of Ͻ0.05 were considered significant.

Infection with Catalase-deficient F. tularensis LVS Alters
Macrophage Ca 2ϩ Response-Previous studies from our lab and others have shown that F. tularensis antioxidants are capable of altering immune signaling postinfection through a disturbance of the AKT, MAPK, and NF-B pathways (9, 44 -46). Although ROS on their own can act as secondary messengers in the course of infection (15,16,47), the impact that ROS may have on the secondary messenger, Ca 2ϩ , has yet to be investigated in the context of Francisella. Other groups have shown that a number of Ca 2ϩ channels are impacted by the intracellular redox state of a cell (48 -51). Given our knowledge that infection with a catalase-deficient mutant of Francisella increases the intracellular steady state concentration of H 2 O 2 within 30 min (1), we hypothesize that intracellular Ca 2ϩ signaling may also be altered.
Several TRP channels have been implicated in the response to oxidative stress including members of the TRPC, TRPM, and TRPV families (52). Zou et al. (43) recently published a study illustrating the Ca 2ϩ response of several different types of macrophages (both immortal cell lines and primary cells) to an exogenous H 2 O 2 stimulus. Utilizing their experimental approach, we aimed to determine whether infection with Francisella will impact the ability of mouse alveolar macrophages to respond to a secondary stimulus shown to elicit a calcium response.
MHS cells were infected with wild type F. tularensis subspp. holartica LVS and catalase-deficient LVS ⌬katG), respectively, and the relative concentration of intracellular Ca 2ϩ was assessed with the ratiometric dye Fura-2. One hour postinfection, baseline Ca 2ϩ levels within ⌬katG-infected macrophages was significantly higher than LVS-infected cells (Fig. 1, A and  B). Cells stimulated with 300 M H 2 O 2 in the absence of extracellular Ca 2ϩ displayed no response when either strain was used for infection. However, when extracellular Ca 2ϩ (2 mM) was reintroduced to the bath solution, both uninfected and LVS-infected MHS cells showed Ca 2ϩ influx in response to H 2 O 2 , whereas the ⌬katG failed to respond (Fig. 1C). Additionally, macrophages infected with ⌬katG showed a reduced Ca 2ϩ signal upon ionomycin treatment (Fig. 1A). Together, these data indicate that Francisella LVS catalase modulates macrophage Ca 2ϩ homeostasis following infection.
Francisella Catalase Is Necessary to Maintain Ca 2ϩ Homeostasis Following Infection-To determine whether this disruption in Ca 2ϩ homeostasis is exclusive to infected cells, relative intracellular Ca 2ϩ levels were also measured with Fura Red, the fluorescence of which is inhibited by the presence of calcium (53). MHS cells were infected with GFP-expressing LVS, and imaging flow cytometry was used to monitor Fura Red intensity 1 h postinfection. Similar, to Fura-2, Fura Red intensity profiling revealed that intracellular Ca 2ϩ is increased postinfection with ⌬katG as shown by the decrease in Fura Red fluorescence ( Fig. 2A). Fig. 2B depicts images from the mean bins of each sample, illustrating the vast decrease in fluorescence after infection with ⌬katG.
To ensure that the alterations in intracellular Ca 2ϩ are not attributed to the hypercytotoxicity that often accompanies engineered mutations in Francisella, we tested a panel of mutants that have been reported to display aberrant bacterial lysis and are inherently proinflammatory (54). Monack and coworkers (54) previously demonstrated that several Francisella mutants in the literature induce proinflammatory responses caused by decreased bacterial integrity and increased the release of pathogen-associated molecular patterns. The integrity of ⌬katG was assessed by scanning electron microscopy and vital staining and was found to be structurally sound (data not shown). To determine whether Ca 2ϩ modulation was specific to the ⌬katG mutant, we chose a panel of four LVS mutants including ⌬wbtA, ⌬fopA, ⌬iglc, and ⌬mglA (Table 1) to examine their impact on host cell Ca 2ϩ modulation. Utilizing a high throughput assay, we monitored Fura Red fluorescence 1 h postinfection with the indicated bacterial strains in Fig. 2C. ⌬katG is the only mutant that exhibited a significant increase in intracellular Ca 2ϩ as illustrated by a decrease in Fura Red fluorescence. Overall, these findings establish that Francisella catalase is likely the key participant in the regulation of host cell Ca 2ϩ homeostasis.

Calcium Channel Inhibitors 2APB and SKF-96365 Prevent
Increased Ca 2ϩ in ⌬katG-infected MHS Cells-To determine whether a redox-sensitive plasma membrane ion channel controls the rise in intracellular Ca 2ϩ postinfection, we employed two broad pharmacological inhibitors of Ca 2ϩ influx channels: 2APB and SKF-96365. 2APB was first described as a membrane-permeable inositol 1,4,5-trisphosphate receptor antagonist that was shown later to inhibit store-operated Ca 2ϩ entry channels mediated by STIM/Orai proteins, as well as several TRP channels (55)(56)(57)(58)(59). SKF-96365 was first introduced as a store-operated Ca 2ϩ entry inhibitor but was later shown to inhibit several members of the TRPC channel family (60). To confirm that the rise in cytosolic Ca 2ϩ is mediated by a plasma membrane channel, we treated MHS cells with the aforementioned modulators during infection and measured intracellular Ca 2ϩ .
Treatment with the broad spectrum inhibitor 2APB had no effect on the baseline level of intracellular Ca 2ϩ of uninfected cells; however, it did significantly decrease the basal intracellular Ca 2ϩ of cells infected with either LVS or ⌬katG (Fig. 3, A-C  and G). The ability of uninfected and LVS-infected macrophages to mobilize Ca 2ϩ in response to H 2 O 2 was impaired by 2APB treatment, while rescuing the H 2 O 2 response of ⌬katGinfected MHS cells (Fig. 3, B, C, and H). Together, these data suggests that Francisella catalase regulates Ca 2ϩ influx through a channel sensitive to 2APB.
The Ca 2ϩ channel modulator SKF-96365 has previously been used to illustrate the dependence of macrophage innate immune survival signaling on Ca 2ϩ fluxes (61)(62)(63). Uninfected and LVS-infected cells exhibited no statistical difference in baseline intracellular Ca 2ϩ following SKF-96365 treatment At 300 s, medium was exchanged to include 300 M H 2 O 2 ; at 600 s, medium was exchanged to include 2 mM Ca 2ϩ ; and at 2400 s, medium was exchanged to include 10 M ionomycin. B depicts the baseline Fura ratio as measured from time 150 -250. C illustrates the total change in intracellular calcium after stimulus. *, p Ͻ 0.01; **, p Ͻ 0.001; ***, p Ͻ 0.0001. (Fig. 3, D, E, and I). However, cells pretreated with SKF-96365 and subsequently infected with ⌬katG display basal intracellular Ca 2ϩ levels similar to those of LVS-infected cells (Fig. 3, E, F,  and I). SKF-96365 treatment also rescues the ability of macrophages infected with ⌬katG to increase cytosolic Ca 2ϩ in response to H 2 O 2 (Fig. 3, F and J). Taken together, these data suggest that the failure of catalase-deficient Francisella to detoxify H 2 O 2 dramatically alters the macrophages ability to mobilize Ca 2ϩ in response to infection.
Francisella Requires Catalase in Order to Inhibit Macrophage Function-Ca 2ϩ is required for many steps of phagocytosis, including the depolymerization of actin that encases phagosomes, phagolysosomal fusion, and the assembly and activation of the ROS producing complex NADPH oxidase (21). Other intracellular pathogens such as Mycobacteria and Listeria have been shown to abrogate the necessary rises in Ca 2ϩ for proper phagosome maturation, therefore enhancing their survival (17). To determine whether the ability of Francisella to modulate Ca 2ϩ also impacts the cytoskeleton, global cytoskeletal rearrangements after infection were analyzed with the use of imaging flow cytometry. Within 30 min postinfection, macrophages infected with catalase-deficient Francisella exhibit a significant rise in F-actin as monitored by mean phalloidin intensity (Fig. 4A). Fig. 4B depicts images from the mean bin of each data set represented in Fig. 4A, clearly illustrating the increased fluorescence following infection with GFP-expressing ⌬katG relative to control GFP-LVS. Restriction of Ca 2ϩ influx with 2APB prevented the increase in actin polymerization following infection with ⌬katG (Fig. 4C). Overall, these data indicate that the H 2 O 2 scavenging activity Francisella limits actin polymerization and may impair efficient phagosome formation in a Ca 2ϩ -dependent manner.
Our lab has gathered compelling evidence that Francisella antioxidants contribute to virulence and actively participate in restricting host innate immune function (9,64). We have previously published that infection with antioxidant-deficient Francisella potentiates macrophage proinflammatory cytokine production relative to LVS infection. The studies illustrated that loss of catalase impairs the cytokine-suppressing capacity of F. tularensis by enhancing NF-B activation. Kinase signaling cascades upstream of NF-B can also be influenced by the intracellular Ca 2ϩ concentration (63). To elucidate whether modulation by Francisella of intracellular Ca 2ϩ postinfection contributes to the cytokine suppression, MHS cells were treated with Ca 2ϩ channel blockers prior to infection, and IL-6 secretion was measured. Pretreatment of alveolar macrophages with either 2APB or SKF-96365 reduced the secretion of IL-6 after wild type infection or following TLR engagement by LPS. However, IL-6 production was only significantly decreased with 2APB and not SKF-96365 following infection with ⌬katG, suggesting that the Ca 2ϩ channel involved in this response is likely more efficiently blocked by 2APB than by SKF-96365 (Fig.  4D). Based on this pharmacology, we therefore reasoned that this Ca 2ϩ -dependent response is likely mediated by a TRP channel family member rather than the store-operated Ca 2ϩ entry pathway mediated by Orai channels. Additionally, this pattern was only present in the absence of catalase, suggesting a redox-dependent mechanism in the Ca 2ϩ -dependent cytokine response.
Catalase-dependent Restriction of Intracellular Ca 2ϩ Increase Appears to be TRPM2-dependent-We next sought to define the precise channel responsible for the increase in basal Ca 2ϩ in response to infection with catalase-deficient Francisella. Among all receptor-activated Ca 2ϩ -conducting channels, TRPM2 stands out for being unambiguously characterized as  Francisella Infection and TRPM2 channels FEBRUARY 19, 2016 • VOLUME 291 • NUMBER 8 redox-sensitive (48,(65)(66)(67). Recent work demonstrating that macrophage TRPM2 is highly redox-responsive (36,43,51) and the block of both actin polymerization and cytokine production by 2APB in ⌬katG-infected MHS cells led us to explore using molecular tools the contribution of TRPM2 to the Ca 2ϩ entry pathway involved post-Francisella infection. BMDM from TRPM2 KO mice, which lack transmembrane domains 5 and 6, as well as the TRPM2 C terminus including the putative calcium pore, rendering the TRPM2 channel inactive (39), were infected with both LVS and ⌬katG, and intracellular Ca 2ϩ was monitored. In the absence of TRPM2, ⌬katG no longer elicits a significant increase of intracellular Ca 2ϩ 1 h postinfection, whereas WT BMDM behave similarly to MHS cells (Fig. 5D). WT and TRPM2 KO BMDM were also assessed for their ability to respond to H 2 O 2 following Francisella infection. Similar to the MHS cell line (Fig. 1), WT BMDM that were either uninfected or infected with LVS increased cytosolic Ca 2ϩ in response to 300 M H 2 O 2 . WT BMDM infected with ⌬katG were in this case capable of responding to H 2 O 2 although to a lesser extent than that of LVS-infected cells (Fig. 5E, dark bars).
In accordance with other published studies, TRPM2 KO BMDM were defective in their response to H 2 O 2 (Fig. 5E, light   bars). Together, these data suggest that Francisella modulation of intracellular Ca 2ϩ is mediated by TRPM2. We next determined whether ⌬katG-dependent increase in macrophage actin polymerization was TRPM2-dependent. Infection of WT BMDM with LVS caused a small decrease in actin polymerization, whereas infection with ⌬katG significantly increased actin polymerization as measured by phalloidin staining. This increase is completely abrogated in TRPM2null BMDM (Fig. 6A), suggesting that Ca 2ϩ influx through TRPM2 is responsible for increased actin polymerization following infection with ⌬katG infection.
We set out to determine whether TRPM2 plays a role in the cytokine response to Francisella infection. Following 24 h of infection, mRNA was collected from both WT and TRPM2 KO BMDM and assayed for IL-6 expression. Consistent with our previous findings (9), WT BMDM infected with ⌬katG displayed a significant elevation in IL-6 expression as compared with LVS infection. This pattern was not observed in the TRPM2 KO BMDM, further supporting the importance of TRPM2-mediated Ca 2ϩ signaling in enhancing cytokine production in response to ⌬katG infection (Fig.  6B).    FEBRUARY 19, 2016 • VOLUME 291 • NUMBER 8

JOURNAL OF BIOLOGICAL CHEMISTRY 3877
Finally, in an effort to determine the mechanistic basis for the impaired ionomycin response in response to ⌬katG infection (Figs. 1A, 3C, and 5C), we first evaluated whether thapsigargin induced changes in cytosolic calcium levels. Although treatment of uninfected, LVS-and ⌬katG-infected MHS cells with thapsigargin did not increase Fura-2 ratio in the absence of extracellular Ca 2ϩ , addition of Ca 2ϩ caused an equivalent increase in the levels of cytosolic Ca 2ϩ that was not affected by infection with either LVS or ⌬katG (Fig. 7, A and B). Thus, the ⌬katG-infected macrophages do not display an intrinsic defect in extracellular Ca 2ϩ entry.
We have previously reported that upon infection with the ⌬katG, there occurs a more than 2-fold increase in the steady state levels of H 2 O 2 relative to LVS-infected macrophages (9). Ionomycin has a complex chemical structure that coordinates Ca 2ϩ hexavalently, and it is possible that its Ca 2ϩ binding activity may be impaired by H 2 O 2 . We first validated the impairment in ionomycin-dependent Ca 2ϩ entry by ⌬katG infection using a microplate-based assay (Fig. 7C). We next evaluated whether H 2 O 2 alters the capacity of ionomycin to transport Ca 2ϩ into the cytosol. H 2 O 2 displayed a dose-dependent inhibition of ionomycin Ca 2ϩ transport capacity that is exacerbated when the drug is not freshly prepared (Fig. 7, D and E). Together, these data suggest that Francisella catalase restricts Ca 2ϩ mobilization by impairing TRPM2 activity and suppressing macrophage function (Fig. 8).

Discussion
Overall, elucidating how Francisella survives in and restricts macrophage function is fundamental to understanding the pathogenic mechanisms of this highly virulent bacterium. To further our understanding of how Francisella antioxidants impede host immune signaling, we have begun investigating the role of intracellular Ca 2ϩ in this process. Here, we establish that F. tularensis limits Ca 2ϩ entry in a redox-dependent fashion. Utilizing two separate fluorescence detection methods, the Fura Red results (Fig. 2) that suggested an effect on basal Ca 2ϩ were confirmed/supported by Fura-2 experiments (Fig. 3 (9) activating TRPM2 leading to an influx of Ca 2ϩ . We propose that TRPM2 is already "on" in the case of ⌬katG infection, and as such it can no longer respond to the addition of exogenous H 2 O 2 as seen in Fig. 1A. Together, the dysregulation of intracellular Ca 2ϩ can severely influence downstream kinase signaling, which in the case of Francisella infection is already impaired (9, 45, 68 -71).
Additionally, macrophages infected with ⌬katG fail to respond to ionomycin treatment. We demonstrate that ionomycin Ca 2ϩ transport capacity is sensitive to direct inhibition by H 2 O 2 . Thus, it is very likely that the increases in steady state H 2 O 2 production resulting from the ⌬katG infection may limit the capacity of ionomycin to transport Ca 2ϩ . The chemistry of H 2 O 2 is complex because it can work as a reductant, oxidant, and a radical initiator; thus, the exact mechanism behind the H 2 O 2 -dependent inhibition is unclear. We must emphasize that agents that alter the cellular redox state, such as infection with ⌬katG, may impair the Ca 2ϩ binding capacity of ionomycin. Moreover, air exposure (oxidation) sensitizes ionomycin to H 2 O 2 -dependent inhibition (Fig. 8D). Although we are actively exploring the redox responsiveness of ionomycin, biochemically, this work is beyond the scope of the current report.
In conjunction with their Ca 2ϩ profiles, macrophages infected with ⌬katG also displayed distinct actin polymerization patterns as compared with macrophages infected with LVS.WeestablishedthatFrancisellacatalaserestrictsactinpolymerization in a Ca 2ϩ -dependent manner. Actin polymerization is essential for several key steps in the clearance of infection including phagocytosis, as well as phagolysosome fusion. It has already been established that Francisella inhibits the phagolysosome fusion and goes on to escape and replicate within the cytoplasm (72,73). Constraints on actin polymerization caused by the restriction of Ca 2ϩ mobilization may represent a redundant pathway to achieve the same goal of bacterial survival.
Pharmacological inhibitors 2APB and SKF-96365 not only rescue the ability of LVS⌬katG-infected macrophages to respond to H 2 O 2 but also differentially alter proinflammatory cytokine secretion, suggesting that there is interplay between intracellular Ca 2ϩ release and the activation of membrane channels that are redox-regulated. The use of these inhibitors, as well as recent publications, led us to hypothesize that the channel involved was TRPM2.
We established that increases in intracellular Ca 2ϩ following infection with ⌬katG, as well as the failure to respond to exogenous H 2 O 2 , is mediated by TRPM2. Additionally, we demonstrate that the catalase-dependent restriction of actin mobilization postinfection is mediated by Ca 2ϩ influx through TRPM2. Finally, we present evidence that the increased cytokine output observed following infection with ⌬katG is dependent upon intracellular Ca 2ϩ mediated by TRPM2. Recently several studies have illustrated the importance of TRPM2 for the clearance of the intracellular bacteria Listeria monocytogenes (36,37,74,75). Knock-out of TRPM2 led to increased morbidity due in part to decreased cytokine output. A hallmark of Francisella infection is the delayed secretion of cytokines followed by a cytokine storm. A variety of mechanisms have been proposed for how Francisella accomplishes this feat (69). Intracellular bacteria are known to have redundant mechanisms to accomplish the same task, thereby increasing their odds of survival, and we propose that catalase restriction of Ca 2ϩ signaling via TRPM2 represents an additional mechanism that Francisella utilizes to thwart the hosts immune response.
Together, our findings provide compelling evidence that bacterial antioxidants harness H 2 O 2 to limit Ca 2ϩ signaling and restrict host immune function. Continued work will further the understanding of how bacterial antioxidant enzymes impair macrophage antimicrobial mechanisms and will contribute to design and implementation of effective vaccination strategies against Francisella and potentially other infectious pathogens.
Author Contributions-N. L. S. and J. A. M. designed the study and wrote the paper. A. C. preformed and analyzed the experiments shown in Fig. 7. M. T. and B. A. M. provided technical assistance and contributed to the preparation of the manuscript. All authors reviewed the results and approved the final version of the manuscript.