The Sphingosine-1-Phosphate Lyase (LegS2) Contributes to the Restriction of Legionella pneumophila in Murine Macrophages

L. pneumophila is the causative agent of Legionnaires’ disease, a human illness characterized by severe pneumonia. In contrast to those derived from humans, macrophages derived from most mouse strains restrict L. pneumophila replication. The restriction of L. pneumophila replication has been shown to require bacterial flagellin, a component of the type IV secretion system as well as the cytosolic NOD-like receptor (NLR) Nlrc4/ Ipaf. These events lead to caspase-1 activation which, in turn, activates caspase-7. Following caspase-7 activation, the phagosome-containing L. pneumophila fuses with the lysosome, resulting in the restriction of L. pneumophila growth. The LegS2 effector is injected by the type IV secretion system and functions as a sphingosine 1-phosphate lyase. It is homologous to the eukaryotic sphingosine lyase (SPL), an enzyme required in the terminal steps of sphingolipid metabolism. Herein, we show that mice Bone Marrow-Derived Macrophages (BMDMs) and human Monocyte-Derived Macrophages (hMDMs) are more permissive to L. pneumophila legS2 mutants than wild-type (WT) strains. This permissiveness to L. pneumophila legS2 is neither attributed to abolished caspase-1, caspase-7 or caspase-3 activation, nor due to the impairment of phagosome-lysosome fusion. Instead, an infection with the legS2 mutant resulted in the reduction of some inflammatory cytokines and their corresponding mRNA; this effect is mediated by the inhibition of the nuclear transcription factor kappa-B (NF-κB). Moreover, BMDMs infected with L. pneumophila legS2 mutant showed elongated mitochondria that resembles mitochondrial fusion. Therefore, the absence of LegS2 effector is associated with reduced NF-κB activation and atypical morphology of mitochondria.


Introduction
The facultative intracellular pathogen L. pneumophila multiplies within human alveolar macrophages and modulates host cell signaling. Following internalization by macrophages or amoeba, L. pneumophila form a unique compartment called the Legionella containing vacuole Donna McCarthy Beckett, Ph.D., R.N., F.A.A.N. Human cells were obtained from anonymous samples from the Red Cross, http://www.redcrossblood.org/centralohio. Blood from the red cross does not have any information or link to the donor and therefore does not require IRB or consent. Primary macrophages derived from mice bone marrow were used.

Bacterial strains
L. pneumophila JR32 [33] and the legS2 mutant [30] were kindly provided by Dr. Amal Amer, at The Ohio State University. L. pneumophila strains were grown on buffered charcoal yeast extract solid media (BCYE) plates at 37°C, or in BYE broth containing L-cysteine, thymidine and ferric nitrate supplements at 37°C with shaking [2,34].

Mice, cell culture and infections
Wild-type mice were purchased from Jackson laboratory. Mice were housed in AALACaccredited facilities and kept in ventilated cages with automatic water valves (lixit). Mice were fed Harlan 7912 as needed and kept in 12hr light/dark cycles 6am-6pm light, 6pm-6am-dark for their circadian rhythms. Plastic houses/huts were provided for enrichment if one mouse per cage or mice needed to feel more secure. Mice were euthanized by CO 2 then cervical dislocation. Anesthesias for in vivo experiments were done by exposing the mice to isoflurane gas for 30 seconds until they are asleep. Mice were monitored daily for signs of pain or distress. All of the animals received humane care according to the criteria outlined in the Guide for the Care and Use of Laboratory Animals published by National Institutes of Health and The Ohio State University. C57BL/6 BMDMs were prepared from the femurs of five to eight-week-old mice as previously described [2,35] and grown in IMDM (Iscove's modified Dulbecco's medium supplemented with the L929 fibroblast cell line (ATCC)-cultured supernatant and 10% heat inactivated fetal bovine serum (HIFBS) (Gibco). Infection was done in IMDM supplemented with 10% HIFBS. The isolation and preparation of the hMDMs from peripheral blood (obtained from the Red Cross) was carried out as previously described [34,36,37].

In vitro and in vivo infection
The in vitro BMDM infections were carried with a multiplicity of infection (MOI) of 0.2-0.5 unless otherwise specified. For in vivo experiments, C57BL/6 mice were infected with 1x10 6 bacteria intratracheally. The lungs were harvested 4 or 48 h post-infection, homogenized, and plated on BCYE with streptomycin. Colony forming units (CFUs) were then quantified.

Western blot
Cell extracts were prepared and immunoblotted with antibodies that recognize caspase-1, caspase-7 and caspase-3 (Cell Signaling), followed by the appropriate secondary antibody. Equal concentration of protein was loaded on a 12% SDS polyacrylamide gel.

Macrophage cytotoxicity assay
In vitro quantification of cytoplasmic histone-associated-DNA-fragments (apoptosis) was performed using the Cell Death Detection ELISA and the photometric enzyme immunoassay (Roche Applied Science) according to the manufacturer's specifications. The fold change of macrophage necrosis was determined by measuring the release of host cell cytoplasmic lactate dehydrogenase (LDH) using the cytotoxicity detection kit (Roche Applied Science) according to the manufacturer's specifications. BMDMs (0.5 x 10 6 ) were plated in 12-well plates and infected with either JR32 or the legS2 mutant at an MOI of 0.5, 1, or 5. The supernatants were then collected 24 h post-infection and assayed.

NF-κB DNA binding activity assay
Nuclear extracts of WT BMDMs untreated or treated with the JR32 or legS2 mutant were prepared as described [2]. EMSA was used to measure NF-κB DNA binding activity as noted [38].

Lyso tracker red colocalization
Wild-type mouse macrophages were plated at a density of 0.5 x 10 6 cells per well in a 24 well plate containing sterile coverslips for 48 h. The lyso tracker red was added to the monolayer for 30 min at 37°C. The monolayer was then washed with the infection medium 3 times and kept for 30 min at 37°C in infection media alone. Next, it was infected with JR32, dotA, or legS2 at an MOI of 0.2 for 1 h. Then, the monolayer was washed with the infection medium, fixed with paraformaldehyde, and finally stained with the anti-L. pneumophila antibody, followed by the secondary antibody Alexa flour 488 goat anti-mouse. The coverslips were mounted and viewed by the Olympus FV1000 spectral Confocal Laser Scanning Microscopy (CLSM).

Enzyme-linked immuno sorbent assay (ELISA)
Macrophages were infected with L. pneumophila JR32 or the legS2 mutant for 24 h. Then, culture supernatants were collected, centrifuged and stored at -20°C until assayed for cytokine content. The amount of IL-1β, IL-6, IL10 and TNFα in the supernatants was determined by specific sandwich ELISA following the manufacturer's protocol (R&D system Inc.,

Quantitative PCR
Total RNA was extracted from mouse macrophages using Trizol (Invitrogen Life Technologies). Then, 1-2 μg of the RNA were converted to cDNA by ThermoScript RNase H-Reverse Transcriptase (Invitrogen, Life Technologies). 20-60 ng of the converted cDNA was used for quantitative PCR with SYBR Green I PCR Master Mix in the Step One Plus Real Time PCR System (Applied Biosystems). The target gene Ct values were normalized to the Ct values of two housekeeping genes (human GAPDH and CAP-1, accordingly to the cell origin) and expressed as relative copy number (RCN) (34).

Transmission Electron Microscopy
BMDMs were plated at a density of 4x10 6 per well on coverslips in Permanox chamber slides, with 2 chambers and 8 units per tray (lab tech) for 48 h; they were then infected with an MOI of 1 for 4 h or MOI of 0.5 for 24 h. The coverslips were fixed with 2.5% glutaraldehyde in 0.1M phosphate buffer and 0.1M sucrose at a pH of 7.4. The cells were post-fixed with 1% osmium tetroxide in the phosphate buffer; then, they were en bloc stained with 2% uranyl acetate in 10% ethanol, dehydrated in a graded series of ethanol, and embedded in Eponate 12 epoxy resin (Ted Pella Inc., 18012). Ultrathin sections were cut on a Leica EM UC6 ultramicrotome (Leica microsystems, EM FC7), collected on copper grids and stained. Images were acquired with an FEI Technai G 2 Spirit transmission electron microscope (FEI), a Macrofire (Optronics) digital camera, and AMT image capture software.

Statistical analysis
The comparisons of groups for statistical difference were done using a Student's two-tailed ttest. P values less than 0.05 were considered significant.

L. pneumophila legS2 strain replicates in WT BMDMs and hMDMs
To delineate the role of LegS2 in L. pneumophila infection, we tested the replication of the JR32 and legS2 mutant in WT BMDMs. The BMDMs restricted the growth of the JR32 strain, while the legS2 mutant showed a 10-fold increase in the colony forming units (CFUs) over time  h) compared to the JR32 strain ( Fig 1A). Since humans are permissive to L. pneumophila replication (39), we also examined the intracellular growth of legS2 mutant in hMDMs ( Fig  1B). Both JR32 and legS2 mutant replicated in hMDMs; however, the legS2 mutant replicated more than the JR32 strain.
Legionnaires' disease stems from the successful replication of L. pneumophila in human alveolar macrophages [39,40]. Therefore, we tested legS2 mutant replication in vivo by examining the bacterial CFUs within the lungs of infected mice. C57BL/6 mice were infected with either 1x10 6 JR32 or legS2 mutant intratracheally. The initial bacterial load was not significantly different 4 h post-infection ( Fig 1C). However, 48 h post-infection, the legS2 mutant-infected mice harbored 4 fold more CFUs compared to the JR32 strain ( Fig  1D). Our data showed that the legS2 mutant replicates in WT mice and their derived macrophages. The growth of legs2 mutant is independent of caspase-1 or caspase-7 activation Caspase-1 activates caspase-7, resulting in the restriction of L. pneumophila growth by mediating the fusion of the LCV to the lysosome [41][42][43]. Casp-1 -/-BMDMs are permissive to L. pneumophila replication [2]. Therefore, impaired caspase-1 activation could result in the increased replication of the legS2 mutant. Therefore, we examined the replication of legS2 mutant bacteria in Casp-1 -/-BMDMs. The legS2 mutant showed higher replication than JR32 24, 48, and 72 h post-infection in casp-1 -/-BMDMs (Fig 2A). Since the cleavage and maturation of IL-1β is mediated by caspase-1 activation, we examined the level of IL-1β 24 h post-infection with the JR32 or the legS2 mutant by ELISA. No significant differences were seen in the level of IL-1β 24 h post-infection with either the JR32 or the legS2 mutant ( Fig 2B). Moreover, we tested the possibility of differential activation of caspase-1 by the WT JR32 or legS2 mutant. Indeed, both JR32 and legS2 infected BMDMs showed capsase-1 cleavage as detected by the mature subunits of casp-1 (Fig 2C). Taken together, the replication of the legS2 mutant in WT macrophages is independent of caspase-1 activation.
The growth restriction of L. pneumophila is also attributed to caspase-7 mediated bacterial delivery to the lysosome and the early death of murine macrophages [2,44]. Hence, impaired caspase-7 activation could result in increased replication of the legS2 mutant. To that end, we assessed the activation of caspase-7 by Western blot. Both the JR32 and the legS2 mutant cleaved caspase-7 in WT macrophages ( Fig 2D). Thus, the replication of the legS2 mutant is independent of caspase-1 or caspase-7 activation.

The growth of legs2 mutant is independent of host cell death
Interference with the host's survival is a key determinant in the L. pneumophila replication cycle. Early cell death aborts L. pneumophila replication and terminates the infection [45]. It is possible that the legS2 mutant modulates the macrophage's life span to sustain its replication. To dissect the role of LegS2 in mediating host cell death, we quantified the level of LDH released in the culture supernatants of macrophages infected with either JR32 or the legS2 mutant. Comparable LDH release was observed in the culture supernatants of both the JR32 and the legS2 mutantinfected macrophages at the two different MOIs (Fig 3A). Moreover, we measured apoptosis in the overall population of macrophages by determining the cytoplasmic apoptosis histone-associated-DNA fragments at MOIs of 0.5, 1 and 5. We found that infection by JR32 and the legS2 mutant leads to comparable cell death at both low and high MOIs (Fig 3B). Overall, our data demonstrates that the replication of the legS2 mutant in macrophages is independent of cell death. Further, the activation of caspase-3 is comparable in JR32 and legS2 infected macrophages ( Fig 3C). Consequently, the replication of the legS2 mutant is independent of host cell death.
The replication of the legS2 mutant in WT BMDM is not attributed to impaired phagosome-lysosome maturation The restriction of L. pneumophila in WT BMDMs is dependent upon the delivery of the LCV to the lysosome, which ultimately results in bacterial degradation. To determine the mechanism by which the legS2 mutant bacteria replicate in macrophages, we examined the colocalization of the JR32, dotA, and legS2 mutants with lyso-tracker Red, an acidic dye used to track the lysosome. (Fig 4A) shows images obtained by the Confocal Laser Scanning Microscopy (CLSM), and (Fig 4B) shows the score of the percentage of bacteria colocalized within the lysotracker positive compartment. Approximately 60% of the translocation-defective dotA mutant was delivered to the lyso tracker-labeled compartment within 1 h of infection. In addition, 52% of the JR32 were destined to the lyso tracker labeled compartment. This is due to the fact that WT cells are restrictive to L. pneumophila replication. The trafficking of the legS2 mutant was comparable to that of the JR32 with~57% lyso tracker labeled compartment. Therefore, the legS2 mutant's substantial replication in WT macrophages is not attributed to enhanced evasion of LCV-lysosome fusion.

The increased replication of the legS2 mutant in BMDMs is associated with reduced cytokine production
The production and secretion of cytokines and chemokines is concomitant with L. pneumophila intracellular growth [46]. Cytokines are chemical mediators that play fundamental roles in eliciting the necessary immune response for controlling lung infection. We examined the level of several pro-inflammatory and anti-inflammatory cytokines following infection with either the JR32 or the legS2 mutant by ELISA. After 24 h, the levels of the cytokines IL-10, IL-6, and TNFα were significantly less in macrophages infected with the legS2 mutant as compared to those infected with JR32. The level of IL-10 in JR32 infected cells was double the amount in legS2 infected cells, (Fig 5A), the levels of IL-6 and TNFα in JR32 infected cells were triple the amount in legS2 infected cells, (Fig 5B and 5C). Differences in the cytokine levels were also evident in the Casp1 -/-BMDMs 24 h post-infection. Our data suggested that cytokine production is associated with increased replication of the legS2 mutant in WT BMDMs.
The legS2 mutant infected macrophages exhibited reduced cytokine mRNAs accompanied by inhibition of the NF-κB pathway To decipher the mechanism by which legS2 affects cytokine level, we used quantitative RT-PCR to examine the levels of the cytokine mRNA following infection with either the JR32 or the legS2 mutant. At 4 h post-infection, the transcription levels of IL-10, IL-6 and TNFα were reduced in the legS2 infected monolayers as compared to the JR32 infected monolayers (Fig 6A-6C). Our results indicate that the decrease in cytokine level from the legS2 infected cells was due to lower levels of cytokine mRNA.
NF-κB is a transcription factor that regulates the mammalian innate immune system by modulating the transcription of pro-inflammatory cytokine and chemokine genes [47]. Indeed, infection with L. pneumophila activates the host NF-κB pathway [1,48]. Therefore, we examined NF-κB activation by an electrophoretic mobility shift assay (EMSA) (Fig 7). The EMSA data showed that, 4-8 h post-infection, NF-κB was inhibited in the legS2 mutant and not in the JR32 strain. This suggests that the reduced cytokine production is associated with the inhibition of the NF-κB pathway.

Infection with the legS2 mutant is accompanied by change in mitochondrial morphology
The dynamic nature of mitochondrial morphology is shown by its continuously fusing and interconnecting network [49]. The fusion process is critical for the maintenance of mitochondrial function [50]. However, mitochondrial fission is an early event during apoptosis, either occurring before or coinciding with cytochrome c release [49,51,52]. It is upstream of caspase activation and characterized by mitochondrial fragmentation into multiple small units [49,51,52]. We employed transmission electron microscopy to examine cells infected with either the JR32 or the legS2 mutant at 4 and 24 h post-infection (Fig 8). Our data showed that, 24 h post-infection, BMDMs infected with the legS2 mutant exhibited elongated mitochondria that seemed to be fused with each other. In BMDMs infected with the JR32 strain, however, we observed fragmented mitochondria, indicating fission. These data suggest that the legS2 mutant bacteria manipulate mitochondrial fusion and fission, respectively.

Discussion
In summary, we show that the legS2 mutant bacteria replicate in WT mice and their derived macrophages. The permissiveness of BMDMs to infections with the legS2 mutant, is accompanied by the dampening of pro-inflammatory and anti-inflammatory cytokines, inhibiting the NF-κB pathway resulting in changes in mitochondrial morphology.
Studies with the A/J-derived mouse macrophages, which are permissive to L. pneumophila infection, showed a biphasic activation of NF-κB during infection [1,48,53,54]. The early phase is characterized by strong but transient activation. Even though this phase is beneficial for the host cell, it must be dampened quickly to overcome the toxic effect of inflammatory cytokines [54]. IL-6 and TNFα are major proinflammatory cytokines that are produced in response to NF-κB activation. They regulate inflammatory responses locally and systemically. Moreover, they recruit leukocytes and mediate more cytokine production and increased bactericidal activity [55,56]. The limitation of the above cytokines following infection with the legS2 mutant is likely responsible for the enhanced L. pneumophila growth in vivo. Indeed, patients with Legionnaires' disease show increased level of IL-6 and TNFα in their lungs [57][58][59][60]. Furthermore, the addition of TNFα renders macrophages less permissive for L. pneumophila and renders neutrophils more bactericidal [61] [62]. Also, mice with a mutation that abrogates Tolllike receptors (TLRs) exhibit limited cytokines, which in turn results in increased susceptibility to L. pneumophila lung infection [63][64][65]. Therefore, the reduction in the levels of IL-10, IL-6 and TNFα could result in the increased replication of the legS2 mutant.
We also observed that infecting BMDMs with legS2 mutant results in the reduction of IL-10 level, and other studies have shown that IL-10 is induced in vivo upon L. pneumophila infection [57,59,[66][67][68]. IL-10 is an anti-inflammatory cytokine that represses the expression of Th1 cytokines [69][70][71]; therefore, it seems that, by dampening the IL-10, the legS2 mutant reduces L. pneumophila replication. However, it is quite likely that limiting IL-6 and TNFα outweighs the effect of IL-10. Thus, the overall impact of a legS2 mutant infection is promoting growth in BMDMs. Even though we focused our study on IL-6, IL-10, TNFα and IL-1β, it is likely that legS2 affects other cytokines that are produced by macrophages as well.
Based upon our RT-PCR analysis of IL-6, IL-10 and TNFα in macrophages infected with legS2 mutants, we can conclude that infection with L. pneumophila legS2 dampens the mRNA  Role of LegS2 in Legionella expression of cytokines in infected cells. The reduction in the above cytokines is likely a consequence of the inhibition of the NF-κB pathway seen by EMSA.
Our TEM data show that infection with legS2 mutant bacteria results in changes in mitochondrial fusion or fission. We observed mitochondrial morphologies that resemble fused mitochondria in legS2 infected cells. Mitochondrial fusion has been linked to an increase in the total pool of ATP during starvation [51]. Consistent with our results, Degtyar et. al showed that legS2 colocalizes to the mitochondria in U937 infected cells [31].
Listeria monocytogenes is another pathogen that has been shown to manipulate mitochondrial fusion/fission. The link between mitochondrial dynamics and the efficiency of the L. monocytogenes infection has been demonstrated. Indeed, the inhibition of the mitochondrial fusion has been shown to decrease the efficiency of L. monocytogenes infection [72]. Similarly, inducing mitochondrial fusion the legS2 mutant resulted in a more efficient infection than in the WT. Therefore, modulating mitochondrial fusion/fission by these two pathogens could impact the efficiency of macrophage infection.
We show that legS2 mutant replicated more than the JR32 strain. It seems contradictory that L. pneumophila would maintain legS2 in the genome, even though it is dispensable for replication in Acanthamoeba castellanii [31]. It is possible that retaining redundant genes in the genome of L. pneumophila is a factor utilized during evolution to prevent chromosomal lesions. Such lesions could affect the ability of the bacteria to multiply intracellularly [30,73]. Moreover, the survival of L. pneumophila in several protozoan hosts could improve the hosts' tropism. However, it could prevent host specialization by retaining pathways that are essential for growth in macrophages and other hosts. Therefore, growing L. pneumophila in laboratories or in murine macrophages, which are not the natural host for this bacteria, could greatly affect the fitness and host range of L. pneumophila [74]. Furthermore, the restriction of L. pneumophila growth in murine macrophages by the acquisition of legS2 could select for bacteria that has increased adaptability to other hosts.