MTFMT deficiency correlates with reduced mitochondrial integrity and enhanced host susceptibility to intracellular infection

Mitochondria behave as functional and structural hubs for innate defense against intracellular infection. While the mitochondrial membrane serves as a platform for the assembly of signaling complexes activated by intracellular infection, various danger molecules derived from impaired mitochondria activate innate signaling pathways. Using methionyl-tRNA formyl transferase (MTFMT)-deficient cells, which exhibit impaired mitochondrial activity, we examined the role of mitochondrial integrity in regulating innate defense against infection. Since MTFMT functions at the early steps of mitochondrial translation, its loss was expected to cause defects in mitochondrial activity. Under transient MTFMT gene silencing conditions, we observed shortened mitochondria along with reduced activity. MTFMT-silenced cells were more susceptible to intracellular infection, as examined by infection with RNA viruses and the intracellular bacterium Shigella flexneri. In support of this observation, MTFMT-silenced cells possessed lowered basal NF-κB activity, which remained low after S. flexneri infection. In addition, the mitochondrial accumulation of evolutionarily conserved signaling intermediate in Toll pathway (ECSIT), an adaptor protein for NF-κB activation, was significantly decreased in MTFMT-silenced cells, explaining the reduced NF-κB activity observed in these cells. Since impaired mitochondria likely release mitochondrial molecules, we evaluated the contribution of mitochondrial N-formyl peptides to the regulation of bacterial infection. Transient transfection of mitochondrial-derived N-formyl peptides favored S. flexneri infection, which was accompanied by enhanced bacterial survival, but did not affect host cell viability. However, transient transfection of mitochondrial-derived N-formyl peptides did not affect basal NF-κB activity. Altogether, these data suggest that the integrity of mitochondria is essential to their proper function in protecting against infection, as intact mitochondria not only block the release of danger molecules but also serve as signaling hubs for the downstream NF-κB pathway.

. MTFMT deficiency disrupts the integrity of mitochondria. (a) HeLa cells were transfected with control or MTFMT siRNA and stained with MitoTracker Deep Red (400 nM) and DAPI. Four representative images of each group are shown (left). To statistically analyze mitochondrial length and mass, mitochondrial network analysis (MiNA), a macro for the measurement of mitochondrial morphology, was used to analyze > 400 cells collected from three independent experiments (right). The error bars indicate SDs (n = 459 for siCON and n = 424 for siMTFMT), and significance was assessed by an unpaired t-test. ****, P < 0.0001. (b) HeLa cells were transfected with control or MTFMT siRNA and stained with JC-1 or MitoSOX to measure mitochondrial membrane potential and mitochondrial ROS, respectively. The fluorescence intensity of JC-1 aggregates, JC-1 monomers and MitoSOX was measured by FACS (left). The error bars indicate SDs, and significance was assessed by an unpaired t-test. **, P < 0.01; ****, P < 0.0001. The MTFMT knockdown efficiency was assessed by RT-PCR, and representative data are shown after the inversion of an agarose gel image (right). For the original gel image, see Supplementary Fig. S4. (c,d) The cellular OCR in the control or MTFMT siRNA-transfected HeLa cells was measured using a Seahorse XFe96 analyzer. The parameters of basal respiration (d, left) and ATP-linked respiration (d, right) were automatically calculated by the Seahorse XF Mito Stress Test Report Generator based on the data in (c). The error bars indicate SDs (n = 3 for siCON and n = 4 for siMTFMT), and significance was assessed by an unpaired t-test. *, P < 0.05; **, P < 0.01. www.nature.com/scientificreports/ tary Fig. S2). We also measured the functionality of mitochondria based on their membrane potential, which can be directly assessed by determining the ratio of an aggregated form of JC-1 dye trapped in healthy mitochondria relative to the monomeric form of JC-1 outside mitochondria. The ratio value was reduced by 30% when MTFMT expression was silenced (Fig. 1b), indicating that the integrity of the mitochondria was impaired when normal levels of MTFMT expression were not maintained. FCCP is a potent mitochondrial uncoupler that dissipates mitochondrial membrane potential by augmenting proton permeability across the mitochondrial inner membrane. Although the addition of FCCP reduced the mitochondrial membrane potential, MTFMT silencing did not lead to a further decrease in mitochondrial membrane potential ( Supplementary Fig. S3a). In addition, levels of mitochondrial ROS, products of an abnormal operation of the mitochondrial electron transport chain (ETC) system, were similarly elevated in MTFMT-silenced cells (Fig. 1b). As a control, we observed the levels of mitochondrial ROS increased by the addition of FCCP in dose-dependent manner ( Supplementary Fig. S3b). Next, the functionality of mitochondria was assessed by measuring the cellular oxygen consumption rate (OCR). Following the sequential addition of oligomycin, FCCP, and rotenone/antimycin A, which target mitochondrial electron transport chains, the OCR values of control-or MTFMT-silenced cells were measured using Seahorse XFe Analyzers. In every step with added chemical, MTFMT-silenced cells behaved similarly to control cells, but with relatively lower OCR value (Fig. 1c). The results showed that in MTFMT-silenced cells, the basal and the maximal respiratory capacities, representing OCR values taken before oligomycin treatment and after FCCP treatment, respectively, were significantly decreased (Fig. 1d). In addition, the ATP production capacity of MTFMT-silenced cells, measured by ATP-linked respiration (OCR Oligomycin − OCR FCCP ), was also significantly reduced (Fig. 1d). The cellular concentrations of ATP were also separately measured to support the above results (Supplementary Fig. S3c and d). These results suggest both structural and functional alterations of mitochondria occurred in MTFMT-silenced cells. Since MTFMT functions during mitochondrial translation processes, we suspected that the general mitochondrial translation efficiency may have been altered. To test this hypothesis, the steady-state levels of mitochondrial proteins were examined (Fig. 1e). We selected voltage-dependent anion-selective channel 1 (VDAC1) as a chromosomal DNA-encoded mitochondrial protein and MT-CO2 as a mitochondrial DNA-encoded mitochondrial protein. However, the levels of neither protein differed between the control-and MTFMT-silenced cells, indicating that although MTFMT functions during mitochondrial translation, its transient deficiency does not markedly affect translational efficiency.
MTFMT deficiency weakens the cellular response against pathogenic infections. As a primary target organelle and central hub for innate immunity, mitochondria are known to mediate cellular host responses upon both virus and bacterial infection 21 . Since we observed altered mitochondrial structure and function in MTFMT-deficient cells, we wondered whether the cellular response to intracellular infection was similarly altered. To this end, we infected HeLa cells with RNA viruses, including influenza A virus, Sendai virus, or Newcastle disease virus, the host response to which depends on cytosolic RIG-I-MAVS signaling pathways in mitochondria. MTFMT-silenced cells exhibited a weakened cellular response against viral infections, as judged by lower RNA expression levels of type I IFN (Fig. 2a-c) and decreased protein secretion (Fig. 2d-f) compared to that observed in control cells. IRF-3, a key regulator of type I IFN gene expression, was also less activated upon viral infections in MTFMT-silenced cells (Fig. 2g). In agreement with the cellular response against infection, the amounts of intracellular virus, measured by viral proteins, were significantly higher in MTFMT-silenced cells at  transfected with control or MTFMT siRNA  were infected with Sendai virus and harvested at the indicated time points for real-time PCR analysis of IFNβ.  (c) HeLa cells transfected with control or MTFMT siRNA were infected with Newcastle disease virus and  harvested at the indicated time points for real-time PCR analysis of IFNβ. (d-f) Forty-eight hours after siRNA transfection, HeLa cells were infected with influenza A virus (d), Sendai virus (e), or Newcastle disease virus (f) for 12 h. Cell medium was harvested for quantifying the human IFNβ concentration as described in the Materials and Methods. The error bars of graphs from a-f indicate SDs, and significances were assessed by an unpaired t-test. *, P < 0.05; **, P < 0.01; ***, P < 0.001. (g) HeLa cells transfected with control or MTFMT siRNA were infected with influenza A virus (left), Sendai virus (middle) or Newcastle disease virus (right) and harvested for immunoblotting assays at the indicated time points. To detect phosphorylated IRF3, membranes were incubated with Tris-buffered saline (TBS) containing 0.05% Tween-20 during immunoblotting.   www.nature.com/scientificreports/ 24 h post infections ( Fig. 2h and i). However, in contrast to IFNβ, the production of IL-6 upon viral infection was not significantly different between the control-and MTFMT-silenced cells ( Supplementary Fig. S7).
In addition to cellular responses against intracellular viral infection, increasing evidence also suggests that mitochondria play crucial roles during the cellular response against bacterial infection. Therefore, we also tested the effect of MTFMT deficiency during bacterial infection, using a Shigella flexneri infection system. S. flexneri is known to infect and replicate within the cytoplasm of epithelial cells 22 , and the number of replicating intracellular S. flexneri can be quantified using a gentamicin protection assay (Fig. 3a). S. flexneri-infected host cells were chased with gentamicin-containing media for the indicated times, and cell lysates were then plated on agar plates containing Congo red. The obtained colony numbers indicate the number of bacteria in proliferative and healthy states within the infected cells. To assess the initial bacterial invasion efficiency, bacteria-infected HeLa cells were chased with gentamicin-containing media for 1 h. An approximately two-fold increase in the number of intracellular bacteria during the very early stage of infection was detected in MTFMT-silenced cells compared to that observed in the control cells (Fig. 3b, left). Since cellular toxicity was not significantly altered by MTFMT   Mock   RNF114  CD83  SLC1A3  OASL  ALDH1A3  GDF15  ALPP  SMAD3  ICAM1  ANXA11  PTGES  NCOA7  KRT17  MKNK2  TNFAIP2  IRF1  VTN  LAMB3  LOXL2  KRT80  SLC12A3  PDGFB  APLP1  MTFMT  RARRES3  BIRC3  TIFA  SLC25A45  EFHD1  C1QTNF1  PBXIP1  IGSF3  AIF1L  CCL5  IFT27  CYP11A1  STYXL1  HIST1H2BJ  HIST1H2BD  CA9  HIST2H2BE  MESDC2 HSDL2 HIST1H1C  HIST1H2BK  CGA  GANAB  TPM4  ISG15  CPA4   IGSF3  KRT80  SLC12A3  C1QTNF1  CRIP2  IFIT3  IFIT2  CCL5  EBAG9  EFHD1  ST6GAL1  ABCC3  ABTB1  TRAF1  RAB28  RARRES3  HIST1H2AE  HIST1H2AC  HIST1H2BJ  TCTA  CYP11A1 ALDH1A3  www.nature.com/scientificreports/ silencing during the first hour of infection (Fig. 3b, right), we concluded that MTFMT deficiency increases the susceptibility of cells to S. flexneri infection. Next, we measured the kinetics of intracellular S. flexneri proliferation under normal and MTFMT-deficient conditions. To this end, S. flexneri-infected cells were cultured in gentamicin-containing medium for an additional 4 to 24 h, and proliferating intracellular bacteria were counted using the gentamicin protection assay. In the control siRNA-transfected cells, the number of intracellular S. flexneri increased consistently over time, and proliferating intracellular bacteria were detected even after 24 h of infection (Fig. 3c, left). In contrast, MTFMTsilenced cells showed a different kinetics pattern of intracellular S. flexneri proliferation. During the early stage of infection, more proliferating bacteria were detected in MTFMT-silenced cells than in the control cells. However, the number of proliferating intracellular S. flexneri plateaued approximately 8 h after infection and then decreased (Fig. 3c, left). Consequently, substantially fewer intracellular S. flexneri were detected in MTFMT-silenced cells than in control cells at the later stage of infection (Fig. 3c, right). To determine whether MTFMT deficiency along with mitochondrial alterations leads to massive cell death at the later stage of infection, which may explain the peculiar kinetics of intracellular bacteria within, we next examined the occurrence of cell death. Under the noninfection conditions, programmed cell death accompanied by cytochrome C release was not observed in either  Fig. S8a). Even throughout the entire period of infection, live cell numbers were not significantly different between control-and MTFMT-silenced cells ( Supplementary  Fig. S8b). These results indicate that MTFMT deficiency has a dual function in which pathogens and the host are favored during the early and later stages of infection, respectively.

MTFMT deficiency results in decreased basal NF-κB activity. To gain insight into how MTFMT
deficiency attenuates cellular immune responses during the early phase of infection, we performed a genomewide transcriptome profiling assay. HeLa cells transfected with control or MTFMT siRNA were subsequently infected with S. flexneri for 4 h, after which RNA sequencing was performed for both infected and non-infected cells (Fig. 4a). By pairwise comparison of control and MTFMT siRNA-transfected and uninfected cells, 85 genes were identified as differentially expressed genes (DEGs) (fold-change ≥ 2, p-value ≤ 0.05) (Fig. 4b). Among these genes, the expression of approximately 71.8% (61 genes) was downregulated in MTFMT-silenced cells, with a maximum reduction of tenfold observed compared to the expression measured in control cells (Fig. 4b). The expression of these genes was similarly reduced in bacteria-infected MTFMT-silenced cells (Fig. 4c). The most interesting finding regarding these downregulated genes was their biological relationships with TNF signaling, NF-κB signaling, papillomavirus infection, and RIG-I-like receptor signaling pathways according to KEGG pathway enrichment analysis of DEGs (Fig. 4d) 23 . The expression of groups of genes involved in TNF-mediated NF-κB signaling (CCL5, TRAF1, BIRC3, ICAM1 and JAG1) and the RIG-I-like receptor signaling pathway (DDX58 and ISG15) were separately validated by real-time quantitative PCR (Fig. 4e). Another subset of genes (IL-6 and TNFα) that were not included among the DEGs shown in Fig. 4a, despite being known as NF-κBresponsive genes, were also examined for their mRNA expression level ( Supplementary Fig. S9). Among the identified DEGs, the reduced expression of NF-κB target genes was of particular interest. Since the basal and infection-induced expression levels of those NF-κB target genes were low in MTFMT-silenced cells, and the NF-κB signaling pathway is a well-characterized central player that controls the host response against intracellular infection 24 , we hypothesized that MTFMT-silenced cells have intrinsically lower basal NF-κB activity. The low levels of basal NF-κB activity may explain the reduced viral and bacterial responses, especially during an early phase of infections, in MTFMT-silenced cells. To test this hypothesis, we assessed NF-κB activity by measuring the active portion of total p65, an active NF-κB component, in the nuclear fraction. Under basal conditions, slightly reduced amounts of p65 were observed in the nuclear fraction of MTFMT-silenced cells. In control cells, S. flexneri infection induced the translocation of p65 from the cytoplasm to the nucleus. In contrast, in MTFMT-deficient cells, the nuclear translocation of p65 was significantly inhibited until 90 min after infection (Fig. 4f and Supplementary Fig. S11). However, IκBα degradation was not clearly observed upon S. flexneri infection in our experimental setting (Supplementary Fig. S12). The activity of the NF-κB signaling pathway was separately tested using an NF-κB-responsive luciferase reporter construct, the activity of which was verified in control experiments ( Supplementary Fig. S13). Using a luciferase reporter assay, reduced NF-κB activity in MTFMT-silenced cells was separately confirmed (Fig. 4g). These results suggest that MTFMT-silenced cells with defective mitochondria possess decreased NF-κB activity under basal conditions, which may explain their attenuated cellular responses against intracellular infection during the early stage of infection.

Recruitment of the NF-κB-activating complex ECSIT to the mitochondrial membrane under basal conditions is reduced in MTFMT-deficient cells.
To understand the molecular mechanisms associated with the reduced basal NF-κB activity in untreated MTFMT-silenced cells, we hypothesized that signaling complex formation on the mitochondrial membrane may be altered in cells lacking MTFMT expression. The detection of basal levels of NF-κB activity in control cells (Fig. 4f) suggests that spontaneous NF-κB signaling complex formation may occur in unstimulated cells. As a possible candidate molecule that activates NF-κB under basal conditions, we evaluated ECSIT, which interacts with TRAF6 on mitochondria 10 . When ECSIT expression was silenced in unstimulated cells, basal levels of NF-κB activity were reduced, confirming that ECSIT plays a role in the activation of basal NF-κB (Fig. 5a). In agreement with the observed basal NF-κB activity, the basal expression level of CCL5, a target gene of NF-κB, was similarly reduced in ECSIT-silenced cells (Fig. 5b), suggesting that ECSIT plays a role in the activation of basal NF-κB-mediated signaling pathways. Using confocal microscopy, we then observed the endogenous localization of ECSIT protein (Supplementary Fig. S14).  www.nature.com/scientificreports/ When mitochondria were co-stained with MitoTracker, substantial amounts of ECSIT protein overlapped with areas harboring mitochondria in control cells (Fig. 5c). Although ECSIT has been reported to be recruited to mitochondria upon TLR activation in macrophages 10 , we observed that substantial amounts of ECSIT were already located at mitochondria under basal conditions (Supplementary Fig. S15). In contrast, those areas showing overlap between ECSIT and mitochondria decreased in MTFMT-silenced cells under basal conditions (Fig. 5c,d). Statistical analysis of the differences in ECSIT-and mitochondrial staining intensity demonstrated that the correlation score was significantly reduced in MTFMT-silenced cells. Based on these observations, we postulate that mitochondrial defects in MTFMT-deficient cells may prevent the stable recruitment of ECSIT to mitochondria, which inhibits the efficient activation of the NF-κB signaling pathway under basal conditions.
Cytosolic N-formyl peptides promote infection independent of the NF-κB pathway. In addition to reduced NF-κB/ECSIT formation on mitochondria, we explored additional effects caused by altered mitochondrial integrity. One hallmark of impaired mitochondria is a leakage of their contents, including mitochondrial DNA and peptides. Since we observed mitochondrial alterations in MTFMT-deficient cells, we wondered whether mitochondrial-derived DNA or peptides could be released from the impaired mitochondria. However, the amounts of detected cytosolic mtDNA were not significantly different (Supplementary Fig. S16).
To measure mitochondrial peptides, we used a double-sandwich ELISA kit designed to detect N-formylmethionine (fMet) in serum. In this assay, a fMet-specific antibody, which is prebound to a reaction plate, can capture fMet-containing products (N-formyl peptide) in the sample up to 1,000 pg/ml (Fig. 6a). First, the mitochondrial and cytosolic fractions were separately harvested by differential centrifugation from HeLa cell lysates and then assayed for the presence of N-formyl peptides. In a test using a mitochondrial lysate (from 7.5 to 30 μg total), a www.nature.com/scientificreports/ positive correlation between the quantity of the lysates and fMet-reactive signal was observed (Fig. 6b). On average, 54.9 ± 23.3 pg of fMet-reactive signal per μg of mitochondrial lysate was detected. We next used four-fold excess amounts of cytosolic lysates (from 30 to 120 μg total) to detect any fMet-reactive signal. Although much less than that observed for mitochondrial lysates, significant levels of fMet-reactive signal (6.43 ± 2.4 pg/μg) were detected in the cytosolic fraction (Fig. 6b). Since this finding indicates that N-formyl peptides may be present in the cytoplasm, we next examined whether the fMet-reactive signal could be altered when mitochondria were disrupted by UV irradiation. Upon UV irradiation, where a clear cytosolic release of mitochondrial cytochrome C was observed, increased levels of fMet-reactive signal were detected (Fig. 6c). Therefore, using this assay, we measured fMet-reactive signal from the cytosolic fraction of MTFMT-silenced cells (Fig. 6d). From the repeated experiments, we clearly observed significantly higher levels of fMet-reactive signal in MTFMA-silenced cells compared to that observed in control HeLa cells. These results suggest that mitochondrial formyl peptide may be released from mitochondria in MTFMT-silenced cells, although the mechanism associated with this release remains unclear. Next, we assessed whether such mitochondrial-derived N-formyl peptide plays any roles during the host defense response against infection inside cells. To directly address this question, cells were transiently transfected with synthesized N-formyl peptides, and the effect of cytosolic N-formyl peptides on the defense response was assessed after S. flexneri infection. In this assay, through peptide transfection, we attempted to mimic MTFMTsilenced basal cell conditions. We used four different N-formyl peptides, where each sequence was derived from one of the mitochondrial proteins ND6, ND4, COXI, or Cyto b (Fig. 7a). Each control peptide had the same amino acid sequence, except for the absence of a formyl group on the first methionine residue. The transfection efficiency and stability of the peptides inside cells was examined using R-phycoerythrin (R-PE) transfection (Fig. 7b). The transfection of R-PE alone did not dramatically alter cell viability. However, the bacterial infection of control-or formyl peptide-transfected cells caused cell death at a later stage of infection (data not shown). Since peptide transfection along with bacterial infection was not observed to have a cytotoxic effect during the early stage of infection (Fig. 7c), we examined the regulatory effects of cytosolic N-formyl peptides during this early period in infection.
When individual N-formyl peptides were transfected, their effects on intracellular bacterial survival were marginal, with the exception of the COXI-derived N-formyl peptide (Fig. 7d). However, the effect on bacterial survival was most significant when a mixture of COXI-and ND6-derived N-formyl peptides was used, following which an average bacterial survival rate of more than 60% was observed compared to that of control peptidetransfected cells (Fig. 7e and f). Finally, we examined whether cytosolic N-formyl peptides directly control basal NF-κB activity. However, basal NF-κB activity was not significantly altered by the transfection of either control or formyl peptides alone (Supplementary Fig. S19). Furthermore, when S. flexneri cells were pre-incubated with N-formyl peptides before infection, their invasive capacity and virulence were much stronger ( Fig. 8a and  b). These data indicate that mitochondrial-derived N-formyl peptides directly favor bacteria but not through regulating the host defense response mediated by the NF-κB signaling pathway.

Discussion
In this study, we investigated the role of mitochondrial integrity in the defense against infection. Taking advantage of MTFMT-deficient cells that have both structurally and functionally impaired mitochondria but relatively intact mitochondria translation potential, we propose that mitochondrial integrity has dual functions: it is required for the proper assembly of signaling complexes, observed in this study by the NF-κB-activating ECSIT on mitochondria, and the blockage of mitochondrial DAMPs, observed in this study as mitochondrial N-formyl peptides, likely released from impaired mitochondria (Fig. 8c).
On the mitochondrial membrane, a signaling complex activates NF-κB or IRF assembly upon intracellular infection. One of the best-studied adaptor molecules on mitochondria is MAVS. RNA virus-infection triggers RIG-I activation through a conformational change, after which activated RIG-I moves into mitochondria to bind its adaptor protein MAVS. Since MAVS is a transmembrane protein, it forms a signaling complex on mitochondria and consequentially transmits the activating signal to downstream components. In contrast, ECSIT which functions as an adaptor protein in the Toll/IL-1 receptor-mediated signaling pathway, is primarily located in the cytoplasm 25 . However, upon ligand binding, TRAF6 ubiquitinates ECSIT for signal transmission, and their interaction can be detected on mitochondria 10 . The NF-κB pathway is known to function downstream of ECSIT. In this study, we observed that (1) MTFMT deficiency attenuates immunity-related signaling pathways, such as the NF-κB pathway; and (2) the mitochondrial accumulation of ECIST diminishes when mitochondrial structure and function is altered by MTFMT deficiency. Indeed, ECSIT has recently been shown to mediate a physical connection between RIG-I and MAVS on the mitochondria upon viral infection 11 . Since RIG-I-like receptor-signaling pathways are one of the affected KEGG pathways in our transcriptome analysis (Fig. 4d), we speculate that the interaction between RIG-I and MAVS could also be attenuated by MTFMT deficiency. In this case, MTFMT deficiency appears to inhibit upstream of MAVS, because "basal" expression and accumulation of MAVS on mitochondria was not significantly altered in MTFMT-deficient cells (Supplementary Fig. S20).
In this study, we tested the effects of cytosolic formyl peptides in response to infection. The mitochondrion is an intrinsically non-self organelle 26 harboring many danger molecules that function during defense responses against infection by directly or indirectly stimulating innate immune responses 5,27 . Mitochondrial N-formyl peptides are structurally and functionally similar to bacterial formyl peptides. The binding of N-formyl peptides to FPRs in the plasma membrane leads to changes in calcium flux and MAPK pathway activation 28 . Consequent gene expression leads to the recruitment of immune cells around the site of infection or injury, which explains how N-formyl peptides act as chemoattractants 29,30 . The release of mitochondrial-derived molecules, which is typically accompanied by the breakdown of mitochondrial integrity, is a warning sign of danger 5 www.nature.com/scientificreports/ of cytochrome c protein from mitochondria into the cytoplasm indicates a specific type of cell death, apoptosis. Furthermore, the release of mitochondrial peptides during the mitochondrial unfolded protein response (mtUPR) turns on the cellular stress response 31 . The mtUPR induces the accumulation of mitochondrial ROS and the cleavage and subsequent release of mitochondrial proteins. In C. elegans, the mitochondrial proteases LONP-1 and CLPP-1 participate in the cleavage process during the mtUPR 32,33 , and the resultant cleaved peptides are released into the cytoplasm through the peptide exporter HAF-1 34 . Consequently, ATFS-1, a major transcription factor, translocates to the nucleus, where it regulates the expression of stress-responsive genes 33 . The presence of mitochondrial-derived peptides has also been reported in the cytosol of S. cerevisiae 35 . In this study, we mimicked the cytosolic formyl peptide by transfection and assessed its role in the regulation of bacterial infection. Currently, the mechanism of this process remains unknown. However, we demonstrated that pre-existing mitochondrial N-formyl peptide favors bacterial survival.
In this study, we used four different types of synthetic peptides with methyl or formyl groups attached to their N-terminus. Each amino acid sequence was derived from mitochondrial proteins, which are translated inside of the mitochondrial matrix and not by the cytosolic translation machinery. Although 13 proteins can be synthesized in the mitochondria, only a subset of mitochondrial N-formyl peptides were effective. We chose ND6, ND4, and COXI because they were previously reported to show high agonistic activity against FPR1 and FPR2 based on their capacity to increase intracellular calcium flux 36 . We also tested the Cyto b peptide based on its ability to induce β-hexosaminidase release and chemotaxis in HL-60 cells 37 . Although all the peptides used in this study possessed an N-formyl methionine residue at their N-terminus, each peptide had a different effect (Fig. 5), suggesting that the formyl group itself is not sufficient to enhance bacterial virulence, although it is required. It will be of further interest to establish effective peptide parameters, such as peptide length, amino acid composition or order to tailor this effect.
In summary, mitochondrial integrity is crucial in mediating the defense response against infection. Healthy mitochondria exhibit dual functions by providing a physical platform for the stable assembly of signaling complexes and preventing the leakage of unwanted danger molecules from the mitochondria to the cytosol. The results of our study confirm that mitochondria are central organelles that regulate the immune response to infection and demonstrates the importance of maintaining mitochondrial health.

Measurement of mitochondrial length and mass. For measurements mitochondrial morphological
parameters, fluorescence images of mitochondria were obtained by confocal microscopy (FLUOVIEW FV3000; Olympus, Tokyo, Japan). Each image that contained only one cell was turned into a binary image (black/white) and subsequently analyzed using a ImageJ macro called MiNA. This plug-in automatically measures the volume and length of mitochondria based on white pixels.
Mitochondrial stress assay. siRNA-transfected cells were trypsinized and recounted for reseeding in an XF96 cell culture microplate (Agilent, cat. #102416-100, Santa Clara, CA, USA) 36 h before the assay. XF base medium (Agilent, cat. #102353-100) was supplemented with 1 mM pyruvate, 2 mM glutamine, and 10 mM glucose, and the pH was adjusted to 7.4 immediately before the assay. Each mitochondrial inhibitor (Agilent, cat. #103015-100) was dissolved in assay medium (at the following concentrations: 100 µM oligomycin, 100 µM FCCP, and 50 µM rotenone/antimycin A) and loaded onto an XFe96 sensor cartridge (Agilent, cat. #102416-100). After incubating in a non-CO 2 incubator for 1 h, the microplate and sensor cartridge were loaded onto a Seahorse XFe96 analyzer, and the assay was performed. Each parameter was automatically calculated by the Seahorse XF Mito Stress Test Report Generator.
Cell fractionation. For nuclear/cytosolic fractionation, cells were lysed with sucrose buffer [10 mM Tris-HCl (pH 8.0), 320 mM sucrose, 3 mM CaCl 2 , 2 mM MgOAc, 0.1 mM EDTA, and 0.5% NP-40] and divided into cytosolic (supernatant) and nuclear (pellet) fractions by centrifugation at 2,400 rpm for 5 min at 4 °C. For mitochondrial/cytosolic fractionation, cells were lysed with mitochondrial isolation buffer [10 mM Tris-HCl (pH 7.5), 320 mM sucrose, and 1 mM EDTA], and the supernatant was obtained by centrifugation at 1,000 × g for 10 min at 4 °C. After one more round of centrifugation at 11,000 rpm for 15 min at 4 °C, the cytosolic (supernatant) and mitochondrial (pellet) fractions were obtained. www.nature.com/scientificreports/ Wilmington, DE, USA) to collect 4 × 10 5 cells. A cell pellet was obtained by centrifugation, and the cells were subsequently resuspended in 1 ml of PBS. Then, the cells were collected by centrifugation and lysed with 1 × Passive lysis buffer (Promega, cat. #E1941, Madison, WI, USA) in a vertical microtube rotator for 15 min at 4 °C. The supernatant was separated from cell debris by centrifugation at 13,000 rpm for 5 min and then assayed to determine the cellular ATP concentration using a Luminescent ATP detection assay kit (Abcam, cat. # ab113849, Cambridge, UK).
Bacterial strain and gentamicin protection assay. Shigella flexneri (2457 T) was a kind gift from Dr.
Jae-Ouk Kim (International Vaccine Institute, Seoul, Korea). One day before infection, colonies were inoculated and incubated in 5 ml of tryptic soy broth (TSB) overnight. On the day of infection, the bacterial broth was diluted 1:100 in fresh TSB and incubated at 37 °C in a shaking incubator at 2,400 rpm until the OD value reached 0.6-0.8. After determining the bacterial dose, the bacterial pellet was obtained by centrifugation at 3,400 rpm for 10 min and then resuspended in 500 µl of PBS. The bacterial suspension was then added to a dish in which the cell medium had been replaced with fresh complete medium before infection. For the bacterial invasion assays, cells with the added bacterial suspension were centrifuged at room temperature at 1,800 rpm for 10 min and then incubated at 37 °C. After 30 min of incubation, the cells were washed three times with cold PBS and cultured in fresh complete medium containing 50 µg/ml gentamicin for 1 h. Cell pellets were then obtained by trypsinization, after which the pelleted cells were lysed, diluted in 0.1% Triton X-100 and spread onto TSB agar plates containing Congo red (Serva, cat. #27215.01, Heidelberg, Germany) to reconstitute intracellular bacteria. For the bacterial survival assay, infected cells were incubated in gentamicin-containing medium for 24 h.

Viral infection.
For viral infection assays, the cell medium was replaced with serum-free medium containing 10 TCID50/ml of influenza A virus (A/PR8/1934(H1N1)), 300 TCID50/ml of Sendai virus, or 15 TCID50/ ml of Newcastle disease virus (NDV, La Sota). After 3 h of incubation at 37 °C, the medium was replaced with fresh DMEM containing 10% FBS, and cells were harvested for RNA extraction at various time points.
Human IFNβ and IL-6 ELISA. HeLa cells were seeded at 2 × 10 5 cells/well in six well tissue culture plate.
Twelve hours after cell seeding, siRNA was transfected into cells with Lipofectamine 2000 reagent (Invitrogen, cat. #11668019). Forty-eight hours after siRNA transfection, the cell medium was replaced with serum-free medium containing each virus. After 3 h of incubation, the medium was replaced with fresh DMEM containing 10% FBS. The supernatants of virus-infected cells were collected at 12 h post infection and used to quantify the concentrations of human IFNβ and IL-6 using a Human IFNβ Quantikine ELISA kit (R&D Systems, cat. #DIFNB0, Minneapolis, MN, USA) and a Human IL-6 Quantikine ELISA kit (R&D Systems, cat. #D6050), respectively.
fMet ELISA and CLIA. For measurements of intracellular levels of formyl peptide, two ELISA kits were used: fMet ELISA (MyBioSource, cat. # MBS265039, San Diego, CA, USA) and fMet CLIA (Elabscience, cat. # E-CL-H0950, Houston, TX, USA). Cytosolic or mitochondrial fractions were quantified by Bradford assay. The measurement of fMet in the cellular lysates was conducted according to the manufacturer's instructions. Briefly, appropriate amounts of sample were loaded into fMet-antibody-precoated 96-well plates and incubated for 90 min at 37 °C. After washing with wash buffer supplied by the manufacturer, biotinylated a fMet-antibody was added into the plate and incubated for 60 min at 37 °C. After washing, an avidin-conjugated solution was added and incubated for 30 min at 37 °C. After the final round of washing, color reagent was added, and the plate was incubated until a color change appeared (typically within 10 min). When the chromogenic reaction was completed, the OD value was read at 450 nm with an ELISA reader. For CLIA, after the final round of washing, luminol substrate was added, and relative light unit (RLU) values were determined using a luminescence microplate reader.

Peptide transfection.
Mitochondrial N-formyl peptides or control unformylated peptides were synthesized and purchased from GenScript (Piscataway, NJ, USA). Each peptide was dissolved in DMSO and stored at − 20 °C until use. For peptide transfection, the peptide was mixed with PULSin reagent (Polyplus Transfection, cat. #501-04, Illkirch-Graffenstaden, France) and added to a dish in which the cell medium had been replaced with fresh serum-free DMEM before transfection. The cell medium was replaced with fresh complete medium after 2 h.
Immunoblotting and immunofluorescence staining. Cell lysates were quantified by Bradford assay.
According to the detected protein size, SDS-PAGE gels at various percentages (8-13.5%) were used to separate proteins.