All the cells including murine macrophage J774.A1 cells, RAW264.7 macrophages, mouse bone marrow derived macrophages (BMDMs), murine embryonic fibroblasts (MEFs), and Drosophila S2 cells were cultured as previously described (Qin et al., 2008, 2011; Pandey et al., 2017).

Brucella melitensis WT strains 16M and 16M-GFP (Qin et al., 2008) were used in this study. Bacterial cultivation and host infection were performed using previously described methods (Qin et al., 2008). Host cells including Drosophila S2 cells, mouse MEFs. RAW264.7 macrophages and BMDMs, were infected with B. melitensis strain 16M (Bm16M) or 16-GFP at an MOI of 100, unless otherwise indicated. After centrifugation for 5 min at 200 × g, infected host cells were then incubated at 29°C (S2 cells) or 37°C (mammalian cells) for 30 min. Culture media were removed, and then the infected cells were rinsed with 1 × phosphate buffered saline (PBS, pH 7.2). Fresh media supplemented with 40 μg/ml gentamicin was then added for 30 min to kill extracellular bacteria. Infected cells were lysed for bacterial internalization assay or continuously incubated in this antibiotic for various lengths of time and performed CFU assay to assess the bacterial intracellular growth as previously described (Qin et al., 2008). Brucella infections, CFU assays, and lysate preparation of Brucella-infected host cells were performed in the BSL3 facilities located in the College of Veterinary Medicine at Texas A & M University.

Murine RAW264.7 macrophages or BMDMs were co-incubated in 24-well plates with assorted chemicals including SP600125 (Bennett et al., 2001), ethyl 2,7-dioxo-2,7-dihydro-3H-naphtho[1,2,3-de]quinoline-1-carboxylate (NQDI1) (Volynets et al., 2011), KIRA6 (Ghosh et al., 2014) at the indicated concentrations. Cells were treated with the drugs 1 h before or after, and during, infection with Bm16M strains. To evaluate Bm16M internalization, after 30 min of infection and three washes with 1 × PBS, fresh media supplemented with the same concentration of drugs and 40 μg/ml gentamicin was added to kill extracellular bacteria. After an additional 30 min of incubation, the cells were lysed and the CFU per well was determined as previously described (Qin et al., 2008). To assess Brucella intracellular replication, CFU analysis was performed at 24, 48, and 72 h.p.i. To investigate whether the tested compounds inhibit Brucella growth, the chemicals were individually added to Brucella TSB cultures at 37°C and incubated for 1, 24, 48, or 72 h. To evaluate the potential inhibitory effects of these tested drugs, Bm16M growth in the presence of these drugs was assessed using a CFU plating approach as previously described (Qin et al., 2008).

dsRNA-mediated knockdown of target genes in Drosophila S2 was performed as previously described (Qin et al., 2008). For RAW264.7 macrophages, lentivirus mediated generation of stable target gene knockdown cell lines was performed as the previously described (Chaki et al., 2013; Pandey et al., 2017). Briefly, the pSuperRetro retrovirial vector system (OligoEngine, Inc.) was used to construct vectors to knockdown expression of target genes in murine RAW264.7 macrophages as per the manufacturer's instructions. The reported shRNAs (Alers et al., 2011; Pandey et al., 2017) were used for knockdown of the expression of ULK1 gene. RAW264.7 macrophages (2.0 × 10⁵) cultured in six-well plates were used for transfection, and puromycin was used for screening clones in which the insert was stably integrated. Western blot was performed to validate the depletion of the targeted proteins.

LysM-Ern1^mut/mut and control mice were generated under the auspices of approval by the Texas A & M University Institutional Animal Care and Use Committee in an Association for Assessment and Accreditation of Laboratory Animal Care International Accredited Animal Facility. BMDMs collected from the femurs of Ern1^mut/mut and control littermates were cultivated in L929-cell conditioned media [DMEM medium containing 20% L929 cell supernatant, 10% (v/v) FBS, penicillin (100 U/ml) and streptomycin (100 U/ml)]. After 3 days of culture, non-adherent precursors were washed away and the retained cells were propagated in fresh L929-cell conditioned media for another 4 days. BMDMs were split in 24-well plates (2.5 × 10⁵ cells/well) in L929-cell conditioned media and cultured at 37°C with 5% CO₂ overnight before use.

Host cells coincubated with or without various drugs and/or infected with Brucella were subjected to 0.2% trypan blue vital stain analysis at various time points post-treatment or -infection as previously described (Qin et al., 2008) to quantify viability or evaluate membrane permeability of drug-treated or Bm16M infected host cells. Host cells in which drug treatment or Brucella infection induced no significant differences in viability and membrane permeability were used in the experiments reported in this work.

To visualize Bm16M intracellular trafficking, the indicated host cells (5.0 × 10⁴) were seeded on 12-mm coverslips placed on the bottom of wells of 24-well plates and infected with Bm16M-GFP or Bm16M for various lengths of time. To analyze infections of less than 0.5 h, the infected cells were rinsed with 1 × PBS and then fixed with 3.7% formaldehyde at 4°C for overnight before immunofluorescence microscopy analysis (IFMA). Otherwise, at 0.5 h.p.i., the infected cells were washed with 1 × PBS and fresh media supplemented with 40 μg/ml gentamicin was added to kill extracellular bacteria. At the indicated time points post-infection, the infected cells were fixed and IFMA was performed as previously described (Qin et al., 2008, 2011; Pandey et al., 2017). The primary antibodies used were as follows: goat-anti Brucella, rabbit anti-LAMP-1; rabbit anti-cathepsin D; rabbit anti-mouse LC3; rabbit anti-Calreticulin; rabbit anti-ULK1, rabbit anti-Beclin 1 (Santa Cruz Biotech., Inc, 1:200-500). Samples were stained with Alexa Fluor 488-conjugated and/or Alexa Fluor 594-conjugated secondary antibody (Invitrogen/Molecular Probes, 1:1,000). Acquisition of confocal images, and image processing and analyses were performed as previously described (Qin et al., 2008, 2011; Pandey et al., 2017).

Protein sample preparation and Western immunoblotting were performed as previously described (Qin et al., 2011; Pandey et al., 2017). Primary antibodies used in the immunoblotting analysis included anti-p-ULK1, anti-ULK1, anti-ASK1, anti-p-ASK1, anti-SAPK/JNK, anti-p-SAPK/JNK (Cell signaling); anti-AMPKα, anti-p-AMPKα, anti-Atg9a (Thermo Scientific); anti-LC3, anti-p-ASK1 and anti-GAPDH (Santa Cruz Biotech., Inc); anti-BAK, anti-BAX (EMD Millipore), anti-p-IRE1α (GeneTex, Inc), anti-IRE1α (Novus Biologicals). Dilution of primary antibodies was 1:1,000. Secondary antibody HRP anti-IgG (Sigma-Aldrich, USA, 1:1,000~5,000) was used in the immunoblotting analysis. Densitometry of blots was performed using the ImageJ software package (http://rsbweb.nih.gov/ij/). The relative expression levels of target proteins and the ratio of blot LC3-II/LC3-I at the indicated time points were calculated as previously described (Qin et al., 2011). All Westerns were performed in triplicate and representative images are shown.

All quantitative data were derived from results obtained in triplicate wells for at least three independent experiments. All the analyzed data were normalized with internal controls before preforming the Student's t-test to assess statistical significance between two experimental data sets or a one-way ANOVA test to evaluate the statistical differences of multiple comparisons of the data sets.

We previously exploited a Drosophila melanogaster S2 cell system and RNAi technology to demonstrate that host IRE1α activity confers susceptibility to B. melitensis strain 16M (Bm16M) infection of the insect cells (Qin et al., 2008). To further dissect the molecular mechanisms by which host IRE1α controls the intracellular lifestyle of Brucella, we employed this convenient and established system to determine whether other UPR-related components regulate Brucella entry or replication in host cells. We found that XBP1 depletion did not affect the uptake or intracellular replication of the pathogen (Figure S1A). In addition, we observed that Bm16M replicated inefficiently (compared to controls) in S2 cells that had been individually depleted of key components in the XBP1-independent kinase cascade, including JAB1, ASK1, and JNK (Figure S1A). We also tested whether several host autophagy-related components, including Atg1 (ULK1), Atg2, Atg18 (WIPI1), Atg9 and Beclin 1, conferred susceptibility to Bm16M infection via CFU assay and image-based host cell infection analysis (Qin et al., 2008). We found that RNAi-mediated depletion of these proteins also reduced Brucella intracellular replication (Figure S1), thereby raising the possibility that an axis linking IRE1α and downstream autophagy proteins controls Bm16M intracellular parasitism.

We previously demonstrated that host factor IRE1α (ERN1) is required for Bm16M intracellular replication in mammalian cells (Qin et al., 2008). To test the hypothesis that Bm16M infection of host cells activates IRE1α activity, we monitored the activation of IRE1α during infection. We observed that the expression level of IRE1α was relatively unchanged during a time course (24 h) of infection in both BMDMs (Figure 1A) and RAW264.7 macrophages (Figure 1C). However, we found that IRE1α phosphorylation was enhanced over the same time period in BMDMs and RAW264.7 cells that had been infected with Bm16M (Figures 1A–D). The finding is consistent with the previous findings of host IRE1α activation by infection with B. abortus (Taguchi et al., 2015; Liu et al., 2016) or B. suis (Wang et al., 2016). As expected, IRE1α phosphorylation was also reduced in RAW264.7 cells treated with KIRA6, a specific inhibitor of IRE1α kinase activity (Ghosh et al., 2014), before and during Bm16M infection (Figures 1C,D). To assess the influence of host IRE1α activity on the intracellular replication of the pathogen, we used gentamicin protection assay (Qin et al., 2008) to examine the pathogen replication in IRE1α^+/+ and IRE1α^−/− MEFs or in RAW264.7 macrophages treated with KIRA6, we found that MEFs derived from IRE1α ^−/− mice or RAW264.7 macrophages treated with the IRE1α kinase inhibitor also displayed lower levels of Bm16M intracellular replication than controls (Figure 1E, and data not shown). Therefore, the data supported the hypothesis that activation of IRE1α contributes to Bm16M intracellular parasitism.

To test the hypothesis that the activity of an IRE1α-activated signaling cascade contributes to the susceptibility of host cells to Bm16M intracellular parasitism, we investigated whether the intracellular replication or trafficking of the pathogen was altered in host cells that harbored deletions in several genes in this cascade, including BAX/BAK, ASK1, and JNK1, that were shown to be important in host cells alleviating UPR (Hetz et al., 2006), and in the S2-Brucella interaction system (Figure S1). Gentamicin protection assays revealed that MEF cells harboring double deletions in BAK/BAX (BAK/BAX DKO) reduced Bm16M intracellular replication compared to controls (Figure 2A). Bm16M also showed reduced trafficking to a calreticulin-positive replicative niche in these cells (Figure 2B). To better evaluate signaling events during Bm16M infection, we determined IRE1α-dependent activation of kinases (including ASK1, JNK1/2, etc.) in the IRE1α-activated signaling cascade in Bm16M-infected mouse Ern1^wt/wt and Ern1^mut/mut BMDMs derived from LysM-Ern1^mut/mut mice in which exons 21–22 of the Ern1 gene encoding IRE1α were deleted from monocytes and macrophages (Iwawaki et al., 2009). Our data demonstrated that during the pathogen infection of mouse Ern1^wt/wt and Ern1^mut/mut BMDMs, ASK1 phosphorylation (Thr845) level increased over the same time period (24 h) in the infected Ern1^wt/wt BMDMs, indicating that IRE1α activity contributes to the infection-dependent phosphorylation of ASK1 (Figures 2C,D). Gentamicin protection assays indicated that host cells (RAW264.7 macrophages) pre-treated with NDQI1, a selective inhibitor of ASK1 kinase activity (Volynets et al., 2011), reduced Brucella intracellular replication (Figure 2E). These data indicated that host BAK/BAX and ASK1 activity support Bm16M intracellular replication.

To determine IRE1α-mediated activity of JNK1/2 in Bm16M infection, we monitored phosphorylation (T183/Y185) of JNK1/2 during Bm16M infection of Ern1^wt/wt and Ern1^mut/mut BMDMs. We also determined the roles of JNK1/2 in Bm16M intracellular parasitism via gentamicin protection assays of the pathogen infection of mouse MEFs harboring JNK deletion or J774.A1 macrophages pre-treated with SP600125, a selective inhibitor of JNK kinase activity (Bennett et al., 2001). We found that Bm16M infection activated JNK1/2 in macrophages (Figures 3A,B) and MEFs (Figure S2), and that JNK1/2 displayed less infection-dependent phosphorylation in Ern1^mut/mut BMDMs than in corresponding controls (Figures 3A,B). Moreover, JNK1^−/− MEFs or J774.A1 macrophages pre-treated with SP600125 showed striking reductions in Bm16M infection compared to controls (Figures 3C,D,F). Consistent with these findings was the observation that when SP600125 was added to host cells after infection was initiated, reductions in Bm16M replication were also observed (Figures 3E,G). These data revealed that JNK1 activity plays a sustained role in enhancing susceptibility to intracellular replication by the pathogen. Taken together, these findings indicate that host IRE1α-to-JNK signaling contributes to the biogenesis of rBCVs as well as the intracellular replication of the pathogen.

Our findings in the insect cells implicated a role for ER-associated autophagy in conferring susceptibility to Bm16M intracellular parasitism (Figure S1). We therefore investigated whether kinase activity of ULK1 (Atg1), one of the main component in the autophagy initiation complex (Mizushima, 2010), played a role in Bm16M infection of mammalian cells. To test this possibility, we infected Ern1^wt/wt and Ern1^mut/mut BMDMs with Bm16M and monitored phosphorylation level of ULK1 during a time course of 48 h of infection. We also performed gentamicin protection assays and immunofluorescence microscopy assays to analyze Bm16M intracellular replication and trafficking in these BMDMs. We found that Ern1^wt/wt and Ern1^mut/mut BMDMs cultivated under low nutrient conditions displayed similar levels of ULK1 phosphorylation on Ser555 (Egan et al., 2011a,b; Kim et al., 2011), an event associated with the activation of this protein (Figure S3A). However, Bm16M infection enhanced the phosphorylation of ULK1 on Ser555 (Figures 4A–D). Moreover, the amount of phosphorylation of this residue was lower in uninfected RAW264.7 macrophages (Figures 4A,B) or in Ern1^mut/mut BMDMs infected with the pathogen (Figures 4C,D). These data indicated that host IRE1α activity controlled Bm16M-induced, but not nutrient deficiency-associated, ULK1 activation. Consistent with these findings was the observation that RAW264.7 macrophages that had been depleted of ULK1 proteins via lentiviral mediated gene silencing (Chaki et al., 2013; Pandey et al., 2017) reduced Bm16M intracellular parasitism (Figure 4E). In these cells, Bm16M trafficking to calreticulin-positive compartments was also reduced (compared to controls) (Figures 4F,G); however, the pathogen displayed enhanced trafficking to LAMP1-positive (Figures S3B,C), and cathepsin D-positive (Figures 4H,I) compartments. In addition, ULK1 co-localized with BCVs following infection (Figures S3D,E). These findings indicate that Brucella infection activation of host ULK1 is IRE1α dependent and promotes the pathogen intracellular trafficking and replication.

To test the hypothesis that Atg9a activity conferred susceptibility to Bm16M infection, we examined Bm16M replication in Atg9a^−/− MEF cells (Saitoh et al., 2009; Figure 5A), as well as in RAW264.7 macrophages depleted of the protein via lentiviral mediated gene silencing (Figure 5B). We found that these Atg9a-deleted or -depleted cells had lower levels of intracellular Bm16M than WT controls (Figure 5). To test the hypothesis that Beclin 1 activity confers susceptibility to Bm16M infection, we examined the replication of the pathogen in Beclin1^−/− BMDMs via gentamicin protection and immunofluorescence microscopy assays. We found that Beclin 1-deficient cells supported lower levels of Bm16M infection than controls (Figure 6A). Moreover, during a time course (48 h) of Bm16M infection, less Beclin 1-positive BCVs were found in the Ern1^mut/mut BMDMs, indicating that Beclin 1 recruitment by BCVs is in an IRE1α-dependent fashion (Figures 6B,C; Figure S4A). These data suggest that IRE1α-mediated Beclin 1 association with BCVs promotes the intracellular trafficking and replication of the pathogen. Finally, to investigate a role of IRE1α-mediated LC3-associataed autophagy in Bm16M intracellular parasitism, we examined the effect of IRE1α activity on LC3 processing. We found that IRE1α-deficiency impaired LC3 processing in Bm16M infected BMDMs (Figures S4B,C). Moreover, endogenous LC3 surrounding or accumulating near BCVs was observed during pathogen infection of host cells (Figure 6D; Figure S4D; Movie S1), thereby suggesting that host LC3 and IRE1α-mediated LC3 processing may contribute to the intracellular lifestyle of Bm16M. Taken together, the data suggest a mechanism whereby the activity of a novel IRE1α-ULK1 signaling axis promotes Bm16M intracellular parasitism.

Liu and colleagues recently showed that IRE1α activity promotes B. abortus growth by activating AMPKα, which suppresses NADPH-derived ROS production and limits infection (Liu et al., 2016). During starvation, activated AMPK regulates autophagy through direct phosphorylation of ULK1 (Kim et al., 2011). To test whether Bm16M-induced activation of ULK1 could also occur through the activation of AMPKα, we examined the phosphorylation AMPKα in the Ern1^wt/wt or Ern1^mut/mut BMDMs during a time course of Bm16M infection. We found that AMPKα displayed enhanced phosphorylated (Thr172) in infected Ern1^wt/wt BMDMs (compared to Ern1^mut/mut controls) (Figures 7A,B), and moreover, that AMPKα^−/− cells (Pandey et al., 2017) reduced Bm16M intracellular replication (Figure 7C). Our findings that AMPKα was phosphorylated in an IRE1α-dependent fashion in Bm16M infected cells, and that mouse BMDMs harboring deletions in AMPKα showed lower levels of infection, are consistent with these findings (Liu et al., 2016). These data therefore suggest that Bm16M activation of host autophagy components may also be partially achieved through the activity a signaling pathway that couples IRE1α and AMPKα.

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.