Rabbit monoclonal anti-phospho-eIF2α (Ser51) antibody, rabbit monoclonal anti-phospho-IRF-3 (Ser396) antibody were acquired from Cell Signaling Technology (Danvers, MA, USA). Rabbit monoclonal anti-TBK1 antibody was purchased from Abcam (Cambridge, UK). Rabbit polyclonal anti-STING antibody, rabbit polyclonal anti-PARP antibody, and rabbit polyclonal anti-tubulin antibody were obtained from Santa Cruz Biotechnology (Santa Cruz, CA, USA). Rabbit polyclonal anti-caspase-3 (p17 specific) antibody, rabbit polyclonal anti-caspase-8 antibody, rabbit polyclonal anti-caspase-9 antibody, rabbit polyclonal anti-caspase-12 antibody, rabbit polyclonal anti-BAX antibody, and rabbit polyclonal anti-cytochrome c antibody were purchased from Proteintech (Wuhan, China). Rabbit polyclonal anti-VDAC1 was acquired from Sangon Biotech (Shanghai, China). Peroxidase-Conjugated Affinipure goat anti-rabbit IgG(H+L) (ZB- 5301) and goat anti-mouse IgG(H+L) (ZB- 5305), Alexa Fluor 488-conjugated Affinipure goat anti-rabbit IgG(H+L) (ZB- 0511) were purchased from ZSGB Biotechnology (Beijing, China). Donkey anti-goat IgG/Alexa Fluor 647 was purchased from Biosynthesis Biotechnology (Beijing, China). Tunicamycin was purchased from Fermentek Ltd (Jerusalem, Israel). BX-795 was purchased from Selleckchem (Houston, TX, USA).

Murine macrophage Raw 264.7 cell line was obtained from China Infrastructure of Cell Line Resources (Beijing, China), and was maintained in Dulbecco's modified Eagle's medium (HyClone, South Logan, Utah, USA) supplemented with 10% fetal bovine serum, 1% L-glutamine (Gibco, Grand Island, NY, USA), penicillin (10,000 units/ml), streptomycin (10,000 μg/ml) (HyClone, South Logan, Utah, USA) at 37°C with 5% CO₂.

M. bovis Beijing strain was obtained from China Institute of Veterinary Drug Control (CVCC, China) and was grown in Middlebrook 7H9 liquid medium (BD Biosciences, NY, USA) with 10% ADC (albumin, dextrose, catalase), and 0.05% Tween-80 and grown to mid-log phase for 1 week at 37°C. Aliquots were frozen at −80°C until used.

Macrophages were infected in vitro with M. bovis at an MOI of 10 and incubated for 3 h at 37°C with 5% CO₂. After allowing 3 h for phagocytosis, the cells were washed three times with warm PBS to remove extracellular bacteria and then incubated with fresh medium without antibiotics for an additional time. They were then lysed in autoclaved distilled water to allow collection of intracellular bacteria. The lysates were plated separately on Middlebrook 7H11 agar plates and incubated at 37°C with 5% CO₂ for 3 weeks. Colony counts were performed in triplicate.

Apoptotic cells were assessed using an Annexin V/propidium iodide (AV/PI) staining kit according to the manufacturer's instructions (KeyGENBioTECH, Nanjing, China). Binding of Annexin V and PI was analyzed by BD FACSCalibur flow cytometer (BD Biosciences, San Jose, CA, USA) with FlowJo software (Tree Star, Ashland, OR, USA).

Cells were infected and then washed three times with PBS, trypsinized, fixed in ice-cold 5% glutaraldehyde in 0.1 M sodium cacodylate buffer (pH 7.4) at 4°C for 15 min, and then centrifuged. Cell pellets were fixed for 4 h. After a complete rinse with sodium cacodylate buffer, the cell pellet was further fixed in 1% OsO4 in 0.1 M sodium cacodylate buffer on ice for 1 h and dehydrated with acetone. The cell pellet was embedded in EM-bed 812 resin and polymerized at 60°C for 48 h. Ultrathin sections (70 nm) were obtained on a Leica Ultracut UCT ultramicrotome (Vienna, Austria) and counterstained with uranyl acetate and lead citrate before observation under a fluorescence microscope (Olympus Fluoview, Japan).

Total RNA was extracted using the EASYspin Plus RNA Extraction Kit (Aidlab, Beijing, China). Then total RNA was reverse-transcribed to cDNA using the Maxima First Strand cDNA synthesis kit (Thermo Scientific, Waltham, MA, USA). For determination of XBP-1 splicing, the cDNA product was subjected to 35 cycles of PCR using the forward primer 5′-AAACAGAGTAGCAGCTCAGACTGC-3′ and the reverse primer 5′-TCCTTCTGGGTAGACCTCaTGGGAG-3′ which are specific for mouse XBP-1. β-actin (forward: CCTTCTGACCCATTCCCACC; reverse: GCTTCTTTGCAGCTCCTTCG) was used as an housekeeping control. Products were separated by electrophoresis through a 3% agarose gel.

For protein localization analysis, Raw 264.7 cells grown on cover slips were washed twice with PBS, fixed by Immunol Staining Fix Solution, blocked 1 h at room temperature by Immunol Staining Blocking Buffer (Beyotime Biotechnology, Shanghai, China) and then incubated overnight at 4°C with the appropriate primary and secondary antibodies. The nuclei were stained with DAPI. Raw 264.7 cells were labeled using anti-phospho-IRF3 antibody, anti-STING antibody, anti-TBK1 antibody, and fluorescein isothiocyanate (FITC)-conjugated donkey anti-goat, Alexa Fluor-conjugated goat anti-rabbit and FITC-conjugated goat anti-rabbit antibodies as the primary and secondary antibodies respectively. Macrophages were finally mounted with glycerin buffer and examined immediately with an OLYMPUS microscope. All experiments were performed on three independent occasions in triplicate at each time.

Cellular proteins were extracted with lysis buffer and total protein concentrations were determined with the Bradford assay and 30 mg of protein was separated with SDS-PAGE, followed by electrotransfer to a Immobilon-P Membrane (Millipore, MA, USA). The blots were probed with primary antibodies and secondary antibodies at optimized concentrations. Enhanced chemiluminescence was used as WB detection system. Image J (NIH, Washington, C, USA) was used for quantification of immunoblots. For immunoprecipitation, cell lysates were pre-cleared using protein A/G agarose (Santa Cruz, CA, USA) beads at 4°C for 30 min on a rocker. The beads were removed by centrifugation at 13,200 rpm at 4°C for 1 min and transferred the supernatant to a fresh centrifuge tube. Then protein A/G agarose/sepharose slurry was incubated with anti-IRF3 antibody overnight at 4°C. Beads were washed 3–4 times with RIPA or NP40 buffer. The agarose/sepharose beats were resuspended in the sample buffer and were boiled for 5 min. The beats were then spinned down and the supernatants were used for SDS-PAGE and western blot analysis.

Isolation of mitochondria and cytoplasm from Raw 264.7 whole cell lysates was performed with the Subcellular Structure Mitochondrial Isolation Kit and the Subcellular Structure Cytoplasmic and Nuclear Isolation Kit respectively according to manufacturer's protocol (Boster, Wuhan, China).

Mitochondrial transmembrane potential was assessed using the cationic fluorescent indicator JC-1 (Molecular Probes, Eugene, OR, USA), which aggregates in intact mitochondria (red fluorescence) with normal ΔΨm, but remains in the monomeric form in the cytoplasm (green fluorescence) of cells with disrupted mitochondrial membrane. Raw 264.7 cells were incubated in DMEM medium containing 10 μM JC-1 at 37°C for 15 min, washed with PBS, and then transferred to a clear 24-well plate. JC-1 aggregate fluorescent emission was measured at 583 nm with an excitation wavelength of 526 nm; and JC-1 monomer fluorescence intensity was measured with excitation and emission wavelengths at 525 and 530 nm. Finally, the cells were mounted with the DakoCytomation fluorescent medium and visualized via fluorescence microscopy (Olympus Fluoview, Japan).

All assays were performed on three separate occasions in triplicate at each time. Results are expressed as means ± S.D. All comparisons of data were made using one-way ANOVA followed by post-hoc Tukey's test or Student's t-test. SPSS software (version 13.0: SPSS Inc., Chicago, IL, USA), GraphPad Prism 5 software (La Jolla, CA, USA) and Image J (National Institutes of Health, USA) were used, and P < 0.05 was considered significant.

Apoptosis plays an important role in host defense against mycobacterial infection (Xu et al., 2014; Liu et al., 2015). We first investigated if intracellular M. bovis could induce apoptosis in macrophages. Initial evaluation of MOI to efficiency of phagocytosis in the RAW 264.7 macrophages showed that 98% of the cells were infected with an MOI of 10 at 24 h. Flow cytometric analysis with Annexin V/propidium iodide (AV/PI) staining was used to quantify cells undergoing apoptosis after 24 h of infection with M. bovis. As shown in Figures 1A,B, the percentage of early apoptotic and late apoptotic cells increased from 0.7% in control cells to 9.5% after M. bovis infection. Caspases were activated in M. bovis infected Raw 264.7 cells. Cleavage of caspase-12 and caspase-9 gradually increased in a time-dependent manner over a 48 h period, while caspase-3 started to decrease by 24 h post infection (pi) (Figures 1C,D). These data indicate that M. bovis induces macrophage apoptosis.

Previous studies have shown that mycobacterium-induced apoptosis is associated with ER stress (Seimon et al., 2010; Sohn et al., 2011). To determine if M. bovis-induced apoptosis was also linked to ER stress, morphology of ER was detected by transmission electron microscopy (TEM). We found expansion of the ER compartment within 6 h after M. bovis infection, a morphological hallmark of ER stress (Figure 2A). Apoptosis of macrophages was significantly decreased when the chemical chaperone 4-phenyl butyric acid (4-PBA), an ER stress inhibitor, was used to treat cells prior to M. bovis infection (Figures 1A,B). ER stress triggers apoptosis mainly through the PERK and IRE1α pathways (Sano and Reed, 2013). To investigate whether these two pathways were activated during M. bovis infection in Raw 264.7 cells, we examined the levels of expression of the markers of ER stress. Expression and the splicing of X-box binding protein 1 (XBP-1) mRNA was markedly increased at 9 h pi and then gradually decreased until 48 h pi (Figures 2B,C). Phosphorylation of eukaryotic translation initiation factor 2A (eIF2α) appeared as early as 6 h pi and reached the peak at 24 h pi (Figures 2D,E). These results suggest that M. bovis interacts with ER and inducing stress.

Based on above studies, we reasoned that M. bovis-induced apoptosis was associated to the ER stress pathways (Lim et al., 2011; Choi et al., 2013). Raw 264.7 cells were treated with 4-PBA before M. bovis infection. Phosphorylation of eIF2α was significantly attenuated by the treatment with 5 mM 4-PBA at 6 h and 24 h pi, compared to respective control cells (Figures 3A,B). In response to the treatment with 2.5 mM and 5 mM 4-PBA, phosphorylation of eIF2α and the cleavage of caspase-12 which resides in the ER, caspase-3 and its downstream cleaved poly (ADP-ribose) polymerase (PARP) were attenuated in a dose-dependent manner at 24 h pi (Figures 3C,D). Then we posited that ER stress could influence survival of intracellular bacteria. ER stress agonist tunicamycin (Tm) was used prior to M. bovis infection. We found that the total numbers of intracellular bacteria were respectively decreased by 44.4% and 33.8% compared to macrophages without Tm treatment at 6 h and 24 h pi (Figure 3E). We further confirmed that intracellular survival of M. bovis was significantly increased by 53.7% in response to 4-PBA at 24 h pi (Figure 3F). These data suggest that M. bovis-induced apoptosis is mediated by ER stress, which can effectively control intracellular mycobacteria.

Previous studies have suggested that IRF3, a transcription factor regulating innate immune responses, was involved in ER stress-mediated apoptosis (Liu et al., 2012; Petrasek et al., 2013; Jin et al., 2015). However, the role of IRF3 in M. bovis infection is unknown. Thus, the phosphorylation of IRF3 was studied during M. bovis infection. Our studies show that phosphorylation of IRF3 ensued M. bovis infection at 24 h pi (Figures 4A–D). Nuclear translocation of IRF3 implies that phosphorylation has occurred (Majumdar et al., 2015). We further confirmed that the activation of IRF3 during M. bovis infection by immunofluorescence. IRF3 was significantly detected in the nucleus, while no fluorescent clusters were found in uninfected control cells (Figure 4E). Raw 264.7 cells were treated with 5 mM 4-PBA before M. bovis infection to verify that phosphorylation of IRF3 was mediated by ER stress. As shown in Figures 4C–E, phosphorylation and nuclear translocation of IRF3 were reduced. These results indicate that ER stress results in the phosphorylation and the nuclear translocation of IRF3 during M. bovis infection.

As an adaptor, STING predominantly resides in the ER. Recent studies have shown that calcium mobilization is involved in ER stress, which causes the translocation of STING and subsequently activates the downstream molecule TBK1 in the cytoplasm. TBK1 is an upstream serine/threonine kinase that plays an essential role in phosphorylating IRF3 (Liu et al., 2012; Tanaka and Chen, 2012; Petrasek et al., 2013). To verify whether the activation of IRF3 was associated with TBK1, Raw 264.7 cells were treated with the potent and specific TBK1 inhibitor BX-795 before M. bovis infection. Levels of phosphorylation of IRF3 decreased in a dose-dependent manner in response to 5 and 10 μM BX-795 (Figures 4A,B).

To investigate whether STING-TBK1 pathway was involved in M. bovis infection, phosphorylation of TBK1 was studied in a time-course macrophage infection model. We found that phosphorylation of TBK1 increased in 24 h pi and decreased by 48 h pi, but no difference was identified at 6 h pi, compared to uninfected control cells (Figures 5A,B). Next, we studied whether the activation of TBK1 was associated with ER stress. Raw 264.7 cells treated with 4-PBA prior to M. bovis infection showed that the phosphorylation of TBK1 was significantly attenuated by the treatment with 4-PBA at 24 h pi (Figures 5C,D).

To determine whether the translocation of STING caused the phosphorylation of TBK1, we interrogated the interaction between STING and TBK1. By immunofluorescence microscopy, a striking co-localization of STING and TBK1 was identified, with association into larger clusters around the nucleus at 24 h pi. But this co-localization was abrogated when cells were pretreated by 4-PBA (Figures 5E,F). Together, these data suggest that the activation of IRF3 requires the mobilization of STING and TBK1, mediated by ER stress during M. bovis infection.

It is well-known that IRF3 plays a critical role in innate immunity by regulating inflammation and type-I interferons (IFNs) (Opitz et al., 2006). Recent studies have identified that activated IRF3 binds cytosolic Bax, which results in Bax translocation to the mitochondria and initiation of the intrinsic apoptotic pathway in response to several types of stimuli such as viruses and alcohol (Chattopadhyay et al., 2010; Petrasek et al., 2013). To verify if this pathway was conserved in M. bovis infection, we first assessed mitochondrial function by JC-1 mitochondrial transmembrane potential (ΔΨm) assay (Figure 6A). After infection with M. bovis for 24 h, green fluorescence (JC-1 monomer form) increased and red fluorescence (JC-1 aggregates form) decreased, indicating low ΔΨm values and thus mitochondrial dysfunction compared against the negative control. Then we further tested the translocation of Bax and cytochrome c. We found that Bax was detected in isolated mitochondria and was reduced in the cytoplasm. In contrast, cytochrome c was released from mitochondria to the cytoplasm after M. bovis infection for 24 h (Figure 6B). To investigate if this process was associated with the phosphorylation of IRF3, Raw 264.7 cells were infected with M. bovis after pre-treatment with BX-795. The translocation of Bax and the release of cytochrome c were significantly inhibited. Furthermore, we detected the interaction between IRF3 and Bax by co-immunoprecipitation. As shown in Figure 6C, the phosphorylated IRF3 bound to Bax during M. bovis infection. Collectively, these data demonstrate that M. bovis stimulated the phosphorylation of IRF3 in the cytoplasm, where IRF3 together with bax translocate into the mitochondria to activate mitochondrial outer membrane permeabilization (MOMP), leading to the release of cytochrome c subsequently damaging the organelle and cause apoptotic cell death.

Previous studies indicate that the extrinsic cell death regulators FADD and caspase-8 are linked to IRF3 activation in hepatocyte apoptosis (Petrasek et al., 2013; Liu et al., 2015). However, this has not been studied in mycobacterial infection. Raw 264.7 cells treated with BX-795 before M. bovis infection were studied on this effect. Caspase-8 and caspase-3 were decreased in a dose-dependent manner in response to BX-795 at 24 h pi (Figures 6D,E). Intracellular survival of M. bovis was sustained when BX-795 was used (Figure 6F). These findings show that the extrinsic cell death pathway may also be involved in IRF3-mediated apoptosis. And IRF3 can partly control intracellular survival of bacteria.

These data highlight that (i) M. bovis-induced apoptosis is mediated by ER stress and (ii) STING-TBK1-IRF3 pathway is activated by ER stress, which finally leads to apoptosis (Figure 7).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.