A Toll-Like Receptor 2 Pathway Regulates the Ppargc1a/b Metabolic Co-Activators in Mice with Staphylococcal aureus Sepsis

Activation of the host antibacterial defenses by the toll-like receptors (TLR) also selectively activates energy-sensing and metabolic pathways, but the mechanisms are poorly understood. This includes the metabolic and mitochondrial biogenesis master co-activators, Ppargc1a (PGC-1α) and Ppargc1b (PGC-1β) in Staphylococcus aureus (S. aureus) sepsis. The expression of these genes in the liver is markedly attenuated inTLR2−/− mice and markedly accentuated in TLR4−/− mice compared with wild type (WT) mice. We sought to explain this difference by using specific TLR-pathway knockout mice to test the hypothesis that these co-activator genes are directly regulated through TLR2 signaling. By comparing their responses to S. aureus with WT mice, we found that MyD88-deficient and MAL-deficient mice expressed hepatic Ppargc1a and Ppargc1b normally, but that neither gene was activated in TRAM-deficient mice. Ppargc1a/b activation did not require NF-kβ, but did require an interferon response factor (IRF), because neither gene was activated in IRF-3/7 double-knockout mice in sepsis, but both were activated normally in Unc93b1-deficient (3d) mice. Nuclear IRF-7 levels in TLR2−/− and TLR4−/− mice decreased and increased respectively post-inoculation and IRF-7 DNA-binding at the Ppargc1a promoter was demonstrated by chromatin immunoprecipitation. Also, a TLR2-TLR4-TRAM native hepatic protein complex was detected by immunoprecipitation within 6 h of S. aureus inoculation that could support MyD88-independent signaling to Ppargc1a/b. Overall, these findings disclose a novel MyD88-independent pathway in S. aureus sepsis that links TLR2 and TLR4 signaling in innate immunity to Ppargc1a/b gene regulation in a critical metabolic organ, the liver, by means of TRAM, TRIF, and IRF-7.

Immune hyper-activation in sepsis produces metabolic stress, e.g. from cytokine synthesis, fever, catecholamine release, NO production, and changes in carbon substrate and oxygen utilization [17]. In this setting, several energy-producing metabolic and catabolic pathways are activated in response to the increased cellular ATP and substrate requirements, but this also generates, excessive reactive oxygen and nitrogen species, and this set of conditions may promote mitochondrial damage and metabolic dysregulation [18,19,20]. The energy-protective responses of the cell also include mitochondrial biogenesis, which is initiated through nuclear gene activation [21,22] controlled by ''master'' co-activator genes, e.g. the peroxisome proliferator-activated receptor gamma co-activators, Ppargc1a, Ppargc1b, and Pprc [23,24,25], whose protein products (PGC-1a PGC-1b and PRC) partner with transcription factors that regulate and enhance mitochondrial quality control [26]. PGC-1 is also critically involved in lipid homeostasis and glucose metabolism [27,28], especially in the liver, wherein heterozygosity of PGC-1a reduces the level of gene expression, leading to impaired fatty acid oxidation, steatosis, and insulin resistance [28]-the metabolic hallmarks of sepsis.
Under the metabolic stress of S. aureus sepsis, Ppargc1a and Ppargc1b are up-regulated synchronously, but independently of Pprc. In peritonitis, Ppargc1a/Ppargc1b mRNA levels increase ,5fold in the liver in WT mice, but neither mRNA increases in TLR2 2/2 mice, and both increase by 10-15-fold in TLR4 2/2 mice, in part through suppression of microRNA-mediated mRNA degradation [29]. Of further interest, both Ppargc1 genes are upregulated in sepsis through an unknown cascade involving the TLR2 and TLR4 signaling pathways. These findings led us to postulate that S. aureus infected mice up-regulate Ppargc1a/Ppargc1b through a unique arrangement of TLR2/TLR4 and adaptor proteins that links innate immunity to cell metabolism and mitochondrial biogenesis in the liver, a crucial metabolic and immune organ.
Our findings indicate that hepatic Ppargc1a/Ppargc1b upregulation in S. aureus sepsis is independent of MyD88 and MAL and does not require NF-kB, but relies instead on a novel TLR2 pathway involving TRAM, TRIF, and IRF-3/7. Studies of Ppargc1 regulation in Unc93b1 2/2 (3d) mice also indicate a lack of involvement of nucleic acid sensing TLRs (TLR3, 7-9), and we identify a post-inoculation interaction of TRAM with TLR2 and TLR4 that may represent a platform for TLR2 signaling to TRAM and IRF-3/7.

Murine Model
S. aureus sepsis in mice produced by fibrin-clot implantation is characterized by hepatic TLR2 and TLR4 up-regulation without involvement of exogenous LPS [22,29]. The liver also demonstrates an early up-regulation of the PGC-1 co-activator family of genes, but Ppargc1a and Ppargc1b are not up-regulated in TLR2 2/2 mice and are amplified in TLR4 2/2 mice [29].
Liver cytokine expression in WT, TLR2 2/2 , and TLR4 2/2 mice In order to check for appropriate cytokine responses to S. aureus, we measured Tnf, Il6, and Il10 levels by Q-PCR in the liver in the peritonitis model (Fig. 1). All three cytokines were up-regulated in WT mice by 6 h PI, and declined towards baseline by 24 h. TLR2 2/2 mice showed greater increases in all three cytokines than WT mice at 6 h PI, but statistically only Tnf levels were higher (WT Tnf 6 h PI: 8.0462.32; TLR2 2/2 Tnf 6 h PI: 27.51610.29; P,0.05). In contrast, TLR4 2/2 mice had depressed cytokine up-regulation compared with WT, but between the two strains only Tnf was statistically different at 6 h PI (TLR4 2/2 Tnf 6 h PI: 0.6960.28; P,0.01 vs. WT). Since Tnf production after S. aureus required TLR4, we checked LPS levels by the Limulus assay and detected only 0.04 ng LPS per clot. These abdominal clots undergo lysis over several days, so the mice absorbed less than 0.04 ng of exogenous LPS each day.

NF-kB activation
The unexpected increase in NF-kB-related cytokine production exhibited by TLR2 2/2 mice in response to S. aureus was evaluated further in liver homogenates and nuclei from healthy control (HC), WT, TLR2 2/2 , and TLR4 2/2 mice. We checked NF-kB activation by probing whole cell extracts for phospho-ser276-p65, and found p65 phosphorylation in WT and TLR4 2/2 mice, but not in TLR2 2/2 mice ( Fig. 2A). Nuclear p65 protein in WT mice was comparable among HC mice and stable at 6 h PI, while HC TLR2 2/2 and TLR4 2/2 mice had variable nuclear p65 levels preinfection (intra-experiment variability) and between-strain similarity in nuclear p65 levels at 6 h PI. Thus, the p65 phosphorylation and p65 nuclear patterns did not correspond. Nuclear p50 was detected in similar amounts in the HC mice of the three strains and did not increase 6 h PI. Nuclear cRel levels were stable at 6 h PI in WT mice, but increased in TLR4 2/2 and TLR2 2/2 mice. Thus, NF-kB activation in the liver after S. aureus inoculation was variable in the TLR-deficient strains and no pattern found that was consistent with Pparg1a/b mRNA expression.
The role of NF-kB on Ppargc1a activation was examined after S. aureus sepsis in two ways. WT mice were injected with an inhibitor of IkB-a phosphorylation, BAY-11-7082 [30] at 20 mg/kg [31,32], and then inoculated with S. aureus. IkB-a binds preferentially to the p65 homodimer or to the p50-p65 heterodimer [33]; thus, BAY-treated mice showed no nuclear translocation of p50/p65. NF-kB activity in S. aureus sepsis was also evaluated in p50 2/2 mice (the p65 knockout is lethal) by Q-PCR for Tnf mRNA compared with Ppargc1a mRNA. BAY-treated mice had no increase in Tnf expression at 6 h PI (WT: 8.0-fold PI vs. HC; BAY: 1.1-fold PI vs. HC; WT vs. BAY, P,0.01; Fig. 2B), thus Tnf induction depended on p50/p65 activation. The p50 2/2 mice showed more variability in Tnf activity at 6 h PI (P = NS compared to WT), but Tnf mRNA was still induced. Ppargc1a mRNA was measured in BAY-treated WT and in p50 2/2 mice, and neither experiment produced significantly different Ppargc1a mRNA levels compared with controls (P = NS at 6 h PI). Thus, Ppargc1a induction after S. aureus did not track TNF-a production and was independent of classical NF-kB activation.
Since only four TLR adaptors are known, and the phenotype of MAL 2/2 and MyD88 2/2 mice did not match the TLR2 2/2 mice, we considered the possibility that TLR2 could signal through TRAM and/or TRIF to induce Ppargc1a/b expression. We therefore exposed TRAM 2/2 mice and TRIF 2/2 mice to S. aureus and found that they did not up-regulate Ppargc1a at 6 h PI (TRAM 2/2 : 1.1-fold and TRIF 2/2 : 2.3-fold vs. HC, P = NS, and P,0.05 compared to WT at 6 h for both) (Fig. 4A). TRAM 2/2 Figure 2. Nuclear p65, p50, and c-rel, and whole-cell phospho-p65. Immunoblots are shown for NF-kB family members in nuclear extracts and in whole-liver extracts from WT, TLR2 2/2 , and TLR4 2/2 mice in HC and at 6 h PI (A). Ppargc1a and Tnf mRNA levels in S. aureus sepsis (B). Ppargc1a and Tnf mRNA levels at 6 h PI (compared to HC) were measured in WT, p50 2/2 , and BAY-11-7082-treated mice (n = 3 mice of each strain); * P,0.01 compared with WT Tnf levels at 6 h PI. Vertical bars are SD. doi:10.1371/journal.pone.0025249.g002 mice failed to up-regulate Ppargc1b, while TRIF 2/2 mice showed some Ppargc1b activation at 6 h PI, but this was much less than for WT mice (TRAM 2/2 : 0.7-fold vs. HC, P,0.05 compared to WT at 6 h; TRIF 2/2 : 2.8-fold vs. HC, P = NS compared to WT at 6 h, P,0.05 compared to HC) (Fig. 4B). This indicated that the gene induction was dependent on TRAM and partly dependent on TRIF. Neither TRAM 2/2 nor TRIF 2/2 mice showed a significant difference in Tnf production compared with WT mice (Fig. 4C). Thus, TLR2 signaling for Ppargc1a/b gene induction operates through TRAM and TRIF because the absence of either interferes with the response.

IRF-3 and IRF-7 activation and Ppargc1a transcription
The IRF-3 and IRF-7 transcription factors are the major known effectors of TRAM and TRIF activity, and these were assayed in WT and TLR2 2/2 mice. IRF-3 and IRF-7 are constitutive and translocate to the nucleus upon activation [36,37]; however,  immunoblots did not suggest differences in nuclear IRF-3 protein levels between HC mice and WT and TLR2 2/2 mice, but there was a small increase in TLR4 2/2 mice at 6 h PI (Fig. 5A). In WT mice, nuclear IRF-7 showed little change at 6 h PI followed by a decline at 24 h PI, but TLR2 2/2 mice showed a markedly low baseline level of nuclear IRF-7 and a further decrease at 6 h, whereas TLR4 2/2 mice showed a marked increase in nuclear IRF-7 at 6 h post-inoculation (Fig. 5B). Thus, nuclear IRF-7 levels and nuclear IRF-7 translocation were deficient in TLR2 2/2 mice and fit the pattern of Ppargc1a and Ppargc1b mRNA expression in TLR2 2/2 [38] and in TRAM 2/2 and TRIF 2/2 mice.
The translation of Ppargc1a mRNA was checked by comparing the expression levels of total PGC-1a protein in WT, TLR2 2/2 , and TLR4 2/2 mice after S. aureus inoculation (Fig. 5C). PGC-1a was up-regulated in WT and TLR4 2/2 mice, but not inTLR2 2/2 mice. We also monitored mitochondrial levels of the fatty acid oxidation enzyme, very long-chain specific acyl-CoA dehydrogenase (VLCAD), which is strongly regulated by PGC-1a. Hepatic VLCAD levels decreased in WT and especially in TLR2 2/2 mice in sepsis, but increased in TLR4 2/2 mice relative to the outer membrane reference protein porin (Fig. 5D).
The Ppargc1a and Ppargc1b promoter regions were examined for interferon-sensitive response elements (GAAANNGAAANN) where IRF-3 and IRF-7 binding occurs [39] and sites were found in both with close homology to the IRF-7 consensus. One Ppargc1a site around 2289 Bp from the transcription start site (TSS) had a conserved ISRE in mouse and human genes (Table S1). For this site, we performed chromatin immunoprecipitation assays for IRF-7 and found that it was active in WT mice, but not in TLR2 2/2 mice (Fig. 5E). Positive (RNA Polymerase II and transcription factor EF1a) and negative (negative IgG) controls confirmed specificity for IRF-7 occupancy of the Ppargc1a promoter.

TLR2-TLR4 signaling
Since TLR2 ligands act mainly through MyD88, MyD88independent effects have drawn little further attention since the original studies [7,43]. Macrophage and dendritic cells stimulated with TLR2 ligands show no ISRE-binding activity or interferon-b (IFN-b) up-regulation or IRF-3 translocation [6,44,45,46]. However, viral particles do activate TRIF and IRF3/7 by TLR2 dimerization with TLR4, leading to TLR2-dependent TLR4 activation and signaling through TRIF and TRAM [9]. TLRintegrin constructs also form TLR2/4 dimers [47], while complementation assays demonstrate cytoplasmic TLR2-TLR4 binding [48]. In macrophages, damage-associated ligands, e.g. biglycan, signal jointly through TLR2 and TLR4 [16,49]; reviewed in [50]. Using immunohistochemistry, we found that S. aureus inoculation simultaneously and widely up-regulates both TLR2 and TLR4 by 6 h in WT mouse liver (Fig. 7A). To explore non-canonical TLR2 and TLR4 interactions in this setting, we compared the native (complexed) and reduced states by Blue native PAGE [51,52] in pre-and post-inoculation WT liver extracts prepared at 6 h PI in 0.5% n-dodecyl b-D-maltoside (DDM) without denaturing (no DTT or heating), or in 4% sodium dodecyl sulfate (SDS) plus 100 mM DTT with denaturing. The membranes were independently blotted for TLR2, TLR4, and TRAM (Fig. 7B). In nondenaturing conditions, each of the three antibodies independently identified the same complex at ,300 kD. We also exposed TLR2 2/2 6TLR4 2/2 mice to S. aureus, but in three trials of paired mice, only two survived for 6 h and were moribund, indicating that the inoculation stress was overly severe in the absence of either TLR2 or TLR4. Since a TLR2/4 independent contribution to Ppargc1a/Ppargc1b induction had not been excluded, we exposed Unc93b1-mutant (3d) mice (deficient in TLR3, 7, 8 and 9 signaling) to S. aureus because the unc93b1 protein functions in ER trafficking and mediates translocation of nucleotide-sensing TLRs from endoplasmic reticulum to endolysosomes, allowing for their activation by microbial nucleic acids [53,54]. This 3d mouse lacks endosome-dependent TLR signaling and its responses signify a role for nucleotide-sensing TLRs in gene activation. We measured Ppargc1a, Ppargc1b, Tnf, and Il10 mRNA levels in Unc93b1 2/2 mice and found that all four genes responded similarly to WT mice (Fig. 8), but there were trends towards more Tnf and less Il10 activation at 6 h PI. Thus, TLR 3 or 7-9 do not regulate Ppargc1a and Ppargc1b gene expression in S. aureus sepsis, suggesting the TLR2/TLR4 balance is specifically involved in the regulation of these genes.

Discussion
The key new finding is the existence of previously unsuspected NF-kb-independent transcriptional cross-talk between hepatic TLR2 and TLR4 and the Ppargc1a and Ppargc1b metabolic coactivator genes. Metabolic dysfunction and organ failure are common and potentially lethal problems in sepsis where prosurvival energy-sensing pathways must be activated in concert with the innate host defenses. Information on the regulation of energysensing functions in this setting is limited, but the response is controlled by an integrated transcriptional network that includes NF-kb [55] and the mitochondrial damage response [56]. Mitochondrial DNA copy number falls in several organs in sepsis, which puts oxidative phosphorylation at risk. The restoration of mitochondrial density is delayed in TLR2 or TLR4 knockout mice relative to WT controls [29].
As crucial co-activators of mitochondrial biogenesis [24,25], the loss of Ppargc1a and Ppargc1b function results in a decline in mitochondrial DNA copy number and ultimately in mitochondrial dysfunction [57]. Specifically, in S. aureus infection, WT mice upregulate Ppargc1a and Ppargc1b, butTLR2 2/2 mice do not, while TLR4 2/2 mice display much greater increases in these mRNA levels than do WT mice. Our findings also indicate that these metabolic co-activator genes are regulated by a novel MyD88independent mechanism.
TLR2 ligands rapidly activate NF-kB, so we first checked for NF-kB regulation of Ppargc1a and Ppargc1b and found no evidence for involvement of the classical pro-inflammatory NF-kB pathway. TLR2 2/2 mice surprisingly showed higher and TLR4 2/2 mice lower early-phase cytokine levels after S. aureus compared with WT mice. Moreover, p50 2/2 and BAY-11-7082-treated mice exhibited Ppargc1a up-regulation that was comparable to WT controls, implying that NF-kB activation is not required. Moreover, Ppargc1a/b is induced in MAL 2/2 and MyD88 2/2 mice after S. aureus, even though both lacked NF-kB activation demonstrated by weak Tnf expression.
Since Ppargc1a/b induction was not impaired in either MAL 2/2 or MyD88 2/2 mice, we tested TRAM 2/2 and TRIF 2/2 mice and found, like the TLR2 2/2 mice, that neither strain induced these genes [38]. The main downstream signal of TRAM/TRIF is the phosphorylation of IRF-3 and IRF-7, and our data indicated that nuclear IRF-7 increases in TLR4 2/2 and decreases in TLR2 2/2 mice compared with WT mice, reflecting the levels of Ppargc1 mRNA and the mitochondrial fatty acid oxidation enzyme VLCAD.
The proximal promoter regions of Ppargc1a/b in the mouse and human contain multiple partially-conserved ISRE sites, and by ChIP assay the one that spans 2289 bp from the Ppargc1a transcription start site (TSS) was activated after S aureus infection. Given IRF-7 binding to Ppargc1a, we exposed IRF-3 2/2 6IRF-7 2/2 double-knockout mice to S. aureus sepsis and saw impaired Ppargc1a/ b up-regulation, documenting a role for IRF-3/7. Using TLR3 agonist PolyI:C to induce IRF-7 in TLR2 2/2 mice, we found no increase in basal Ppargc1a mRNA levels, but we did rescue the Ppargc1a response in sepsis. Thus, other factors are also involved in IRF-7 induction of Ppargc1a, e.g. similar to the type I interferon response that follows TLR2 translocation to endolysosomes after ligand engagement [58], though this response is not typical of S. aureus sepsis. In any case, Ppargc1a induction in mice in response to S. aureus infection clearly involves IRF-7, and the TRAM/TRI-FRIRF-7RPpargc1a/b pathway represents a broadening of the scope of TLR2 functionality to encompass a rapid metabolic response. Some intriguing differences in cytokine regulation were also observed in WT, TLR2 2/2 , and TLR4 2/2 mice, but these were not pursued due to insufficient information on the membrane proteins involved and the known discrepancies in vivo and in vitro in response to live S. aureus and to Gram-positive cell wall constituents, e.g. in TLR4 2/2 mice [10,11,12,13,14,15,16,29]. Because TLR4 does not bind Gram-positive ligands, a requirement for TLR4 in this in vivo study suggests the possibility that endogenous ligands are involved in the induction of TLR2/TLR4 interactions that are not found in cell systems. For example, fibrin breakdown products and physiological factors absent in cell systems are present in peritonitis models, such as altered intestinal epithelial barrier function, and generate additional DAMPs (e.g. extracellular matrix products) or PAMPs (e.g. LPS translocation).
In this respect, the clot model mimics clinical peritonitis where physical damage and deposition of hemoglobin and fibrin form a bed for infection, and endogenous cell-surface or damage-receptor ligands may contribute to the TLR2/4 interaction. Whatever factors are responsible -microbial or host-TLR4 participates in the defense against S. aureus. TLR4 2/2 mice show decreases in S. aureus clearance and increases in mortality similar to TLR2 2/2 mice [11,29], and although Ppargc1a/b is essential for metabolic gene expression and for mitochondrial biogenesis, these are not the sole survival genes in the intact animal.
The formation of a TLR2-TLR4-TRAM complex may have important implications for the host response to sepsis, but this aspect is preliminary and there is insufficient data to propose a definitive role for it the initiation of the hepatic response to S. aureus inoculation. S. aureus rapidly up-regulates TLR2 and TLR4 in the mouse liver, and the use of weak non-ionizing detergent and non-reducing conditions allows the detection of a native complex that appears at a molecular weight ,60 kd higher than the predicted triplex. This implies that one or more other factors, such as post-translational modifications or adapter or chaperone molecules, are involved. The establishment of a functional role for a TLR2-TLR4 -TRAM complex would require cell studies beyond the scope of this in vivo paper, and any such complex could be unique to the liver [2][3][4][5][6][7][8]. There are multiple potential TLR2 interactions that might explain both the known mitochondrial protective effects of TLR4 and the pronounced up-regulation of the Ppargc1 metabolic co-activator genes observed here when TLR4 is genetically deleted. Based on our findings, some possibilities for TLR2-TLR4-TRAM interactions leading to IRF7 activation are illustrated in Figure 9. The diagram puts our findings into the context of the well-known TLR2 and TLR4 signaling pathways and outlines testable possibilities for TLR2dependent MyD88-independent IRF7 activation.
In any case, the TLR2-dependent regulation of Ppargc1a and Ppargc1b through a MyD88-independent pathway has been established, and this finding not only extends TLR2 signaling to encompass key metabolic genes, but identifies distinct regulation of the NF-kB-dependent pro-inflammatory genes and the hepatic metabolic genes that maintain energy production and initiate mitochondrial biogenesis after S. aureus infection. Although the receptor signaling pathways will require detailed molecular and cellular studies, the in vivo biology does suggest novel therapeutic approaches. If Ppargc1a/b activation by IRF-7 translates to the clinical setting, it should be possible to establish whether this Figure 8. Ppargc1a, Ppargc1b, Il10, and Tnf mRNA levels in Unc93b1 2/2 mice. Hepatic mRNA levels of Ppargc1a, Ppargc1b, Il10, and Tnf were measured in healthy controls (HC) and in S. aureus sepsis at 6 h PI in WT and Unc93b1 2/2 mice. There was no significant difference between induction levels in WT and Unc93b1 2/2 mice for the four genes (n$3 mice at each point for each strain). Vertical bars are SD. doi:10.1371/journal.pone.0025249.g008 pathway protects metabolic and organ function during sepsis. Ppargc1a and Ppargc1b expression increase functional mitochondrial mass [24,25] and preserved mitochondrial function, e.g. in skeletal muscle, is associated with better outcomes in sepsis [59]. Thus, timed interventions to manipulate IRF-7 may improve cell protection and hasten the resolution of multiple organ failure in patients with sepsis and mitochondrial dysfunction.
Mice were anesthetized with intraperitoneal xylazine and ketamine, the abdomen shaved and cleaned with povidone-iodine, and a midline laparotomy was performed. The peritoneum was inoculated with a fibrin clot containing S. aureus and the incision closed in two layers. Mice were resuscitated with 1 ml of subcutaneous 0.9% NaCl. Healthy control (HC) mice of each strain were also used. Mice were humanely killed at 6, 24, 48, or 72 h post-injury (PI) and the livers immediately harvested to isolate mitochondria or snap-frozen and stored at 280uC.
For the clots, Staphylococcus aureus ssp aureus (ATCC) was reconstituted and suspended in bovine fibrin [22]. The bacteria were inoculated sterilely onto agar slants for 18 h and then resuspended to a concentration of 10 10 CFU/ml based on optical Figure 9. Potential TLR signaling pathways for Ppargc1 metabolic co-activator gene activation after S. aureus infection. Pathway 1 shows the canonical TLR2 MyD88-dependent signaling pathway that activates NF-kB after S. aureus. Pathway 2 shows TLR4 MyD88-dependent signaling to NF-kB and MyD88-independent signaling to TRIF/TRAM. Both MyD88 pathways have been excluded as causes of the Ppargc1a gene expression. Pathway 3 shows a putative TLR2-TLR4 heterodimer interacting with TRIF/TRAM. Pathway 4 indicates TLR2 in the TLR4 null state, as a homodimer or a heterodimer involving a non-TLR4 partner such as TLR1 or 6, interacting with TRIF/TRAM and unmasking the innate immune regulation of Ppargc1a expression. Pathway 5 shows canonical TLR3 endosome signaling also excluded in Ppargc1 gene regulation after S. aureus; however, independent TLR3 activation partially rescues the Ppargc1 phenotype in mice. TIRAP is Toll/interleukin-1 receptor domain-containing adapter protein (MAL); IRAK4 is Interleukin-1 receptor-associated kinase 4; TRAF3 and TRAF6 are TNF receptor-associated factor 3 and 6; TAK1 is TGFbeta-activated kinase 1 and TBK1 is NF-kappa-B-activating kinase. doi:10.1371/journal.pone.0025249.g009 density at 550 nm. Doses of 10 5 , 10 6 , or 10 7 CFU were then suspended in 500 ml fibrin clots (500 ml of 10 mg/ml bovine fibrinogen, fraction 1, plus 10 ml of bovine plasma thrombin) (Sigma, St Louis, MO). Pour plates were used to confirm microbial counts. The Limulus Amebocyte Lysate (LAL) assay was performed with a GenScript Chromogenic LAL endotoxin assay kit (Gen-Script, Piscataway, NJ). Thrombin and fibrinogen were prepared in the standard fashion and tested in duplicate for endotoxin.

Real-Time RT-PCR
RNA was extracted with TRIzol reagent (Invitrogen, Oslo, Norway) and reverse transcribed with the ImProm-II Reverse Transcription System (Promega, Madison, WI). Mouse-specific primers were designed or purchased from Applied Biosystems ( Table 1) and real-time PCR carried out in triplicate, using 18 s primers as internal controls [56]. Real-time PCR output for HC mice of each strain was set to one, and relative quotients obtained at the later time points.

Western Blots
Whole cell extracts or nuclei were sonicated and standardized for protein using bicinchoninic acid. Proteins were resolved by sodium dodecyl sulfate-PAGE on 4-20% gels and transferred to PDVF membranes. Membranes were probed with affinity-purified primary antibodies (Table 1) and exposed to the appropriate secondary antibody (Santa Cruz). Membranes were developed with ECL (Santa Cruz) and imaged on X-ray film in the middynamic range. Membranes were stained with Coomassie blue as a loading control. The blots were quantified on a BioRad G-710 densitometer.

Tissue Immunofluorescence
Livers were fixed in 4% paraformaldehyde, dehydrated, paraffin-embedded, and cut into 4-5 micron sections. After antigen retrieval, the slides were stained with primary TLR2 or TLR4 antibody (SC-52735, mouse monoclonal, and SC-10741, rabbit polyclonal, Santa Cruz), a fluorescently-labeled secondary, and counterstained with DAPI. Confocal images were collected in fluorescence mode followed by electronic image merging.

Chromatin Immunoprecipitation
Nuclear extracts were exposed to 1% formaldehyde for 15 min at 24uC, and the reaction quenched in 0.125 M glycine for 5 min. DNA was sheared with a sonicator into ,200-800 bp fragments. ChIP was carried out using the ChIP-IT Express Kit (Active Motif, Carlsbad, CA) and the manufacturer's instructions using mouse monoclonal anti-IRF-7 and rabbit polyclonal anti-Pol-II (Santa Cruz Biotechnology). Primers were designed for the ISRE sequence for IRF-7 in the Ppargc1a promoter and the promoter of EF1a (for Pol-II). Conventional PCR was carried out to 40 cycles at 60uC.

Blue Native PAGE
Snap-frozen liver from WT mice was homogenized in Native PAGE bis-tris buffer (Invitrogen) with either 0.5% DDM or 4% SDS. The lysates were centrifuged at 14,0006 g for 20 min, and supernatant protein content measured. DTT was added to a final concentration of 100 mM, and the samples boiled for 5 min at 95uC. The samples were mixed 1:1 with Blue Native running buffer from the NativePAGE TM NovexH Bis-TrisGel System (Invitrogen), and run on 3-12% bis-tris polyacrylamide gels with NativeMark unstained molecular weight standards (Invitrogen). Gels were transferred to PVDF and washed twice in methanol to remove excess Coomassie blue before immunoblotting.

Statistics
Grouped data are presented as means 6 SD. The n values in the experiments are for the total number of mice of each strain. Each point in the real-time PCR experiments was compared to the healthy control (HC) of its own strain using the Student's t-test. The 6 h between strain points were compared with Student's ttests with adjustment for multiple comparisons where necessary. The statistical significance levels (P) are provided with the Results.

Supporting Information
Table S1 Mouse (Mm) Ppargc1a and human (Hs) PPARGC1A promoter alignment. ChIP primer sites and the IRF7 consensus sequence for the mouse are indicated. TSS = transcription start site. Note the presence of expanded ISREs in the Hs promoter around the same site. (DOCX)