Unphosphorylated STAT1 represses apoptosis in macrophages during Mycobacterium tuberculosis infection

Mycobacterium tuberculosis (Mtb) is a aggressive and infectious bacterial pathogen. Macrophages are the main host cells for Mtb, and they exert their effects via immune modulation and antigen presentation by ‘programmed’ cell death or autophagy (Cooper and Khader, 2008; Cooper, 2009). However, Mtb is able to survive and proliferate by reducing the level of apoptosis and sensitivity of macrophages to apoptosis-inducing agents (Tiwari et al., 2009; Vergne et al., 2005; Behar et al., 2010). The study of the mechanisms of immune evasion by Mtb and the immunological role of macrophages in tuberculosis infection is important for insights into the anti-tuberculosis immune mechanisms in the host and design of new anti-tuberculosis vaccines.

Cell signaling leading to apoptosis of macrophages can be triggered by Mtb; one of the main signaling pathways involved is the JAK/STAT pathway, which is mediated by the phosphorylation of JAKs and STATs (Rojas et al., 2002; Lim et al., 2016; Rhee et al., 2003). Phosphorylated STAT1 (P-STAT1) has been described as a pathogen suppressor because of its function as an immunosurveillance mediator (Najjar and Fagard, 2010; Lim et al., 2016). The pro-apoptosis activity of P-STAT1 seems to be important at the onset of Mtb infection and is supposed to result in the elimination of infected cells by the innate and adaptive immune system (Rojas et al., 2002). At the cellular level, P-STAT1 functions by upregulating the expression of many pro-apoptosis genes, including those encoding TNF (Gifford and Lohmann-Matthes, 1987; Scuderi et al., 1987; Collart et al., 1986), caspases (Shakhov et al., 1996; Chin et al., 1997), CD95 (also known as FAS) (Xu et al., 1998) and nitric oxide synthase (NOS2) (Gao et al., 1997). The expression of these genes can trigger cell responses to Mtb infection and lead to macrophage activation and apoptosis.

P-STAT1 also induces STAT1 expression, which leads to an accumulation of unphosphorylated STAT1 (Cheon and Stark, 2009). The increase in the level of P-STAT1 only lasts a few hours, whereas the increase in unphosphorylated STAT1 can persist for several days (Cheon and Stark, 2009), raising the possibility that the concentration of unphosphorylated STAT1 might play an important role during the later stages of Mtb infection.

The function of P-STAT1 in regulation of gene expression in Mtb-infected macrophages is well established (Rojas et al., 2002). However, the role of unphosphorylated STAT1 is not well defined. Previous studies have shown that both P-STAT1 and unphosphorylated STAT1 are transcription factors (Chatterjee-Kishore et al., 2000; Cheon and Stark, 2009) so they may both be involved in the cell apoptosis induced by Mtb strain H37Ra. In the present study, we show that there is a correlation between the level of unphosphorylated STAT1 and the intracellular bacterial burden in H37Ra-infected macrophages, which suggests that unphosphorylated STAT1 acts as a negative regulator of apoptosis in the host cell during H37Ra infection. Indeed, H37Ra infection-induced apoptosis was severely compromised in macrophages overexpressing unphosphorylated STAT1 compared to control. Unphosphorylated STAT1 inhibited CD95 signaling, resulting in activation of anti-apoptotic effectors, a decrease in active caspase-3 and cytosolic translocation of cytochrome c. Furthermore, we characterized the unphosphorylated STAT1 protein interaction network with an efficient proteomic approach, and identified 149 proteins that interacted with unphosphorylated STAT1 with high confidence. In addition, we fully verified that the complex between unphosphorylated STAT1 and STAT3 is involved in repressing CD95 expression in macrophages, and we verified that unphosphorylated STAT1 and IFIT1 competed for binding to eEF1A, which was associated with the inhibition of apoptosis. Thus, our data indicate that unphosphorylated STAT1 shows anti-apoptosis activity and promotes immune evasion during Mtb infection, differing from the pathogen suppressor P-STAT1.

Pathogen infection activates the JAK/STAT pathway and its activation is related to apoptosis in macrophages (Najjar and Fagard, 2010; Rojas et al., 2002; Lim et al., 2016). Thus, we investigated whether Mtb strain H37Ra could affect STAT1 signaling in mice bone marrow-derived macrophages (BMDMs). As shown in Fig. 1A, the concentration of P-STAT1 increased within 1 h, and then decreased over the next several hours. As P-STAT1 decreased, there was a reciprocal increase in the concentration of unphosphorylated STAT1 (the monoclonal C-terminal anti-STAT1 reagent used detects only unphosphorylated STAT1 and not P-STAT1; Cheon and Stark, 2009), beginning at about 6 h, which persisted for at least 2 days. At each time point after infection, the viability of H37Ra was determined by a colony-forming unit (CFU) assay and the intracellular bacterial burden showed a similar trend to that of the level of unphosphorylated STAT1 expression, increasing in a time-dependent manner during the later stages of infection (Fig. 1B).

To investigate the role of H37Ra-induced unphosphorylated STAT1, we transduced the mouse macrophage cell line RAW264.7 lentiviruses encoding unphosphorylated STAT1 (Y701F, S727E double mutant) with a FLAG tag to establish a cell line stably overexpressing unphosphorylated STAT1 (denoted RAW-USTAT1); RAW264.7 cells transduced with empty lentiviruses were used as control (denoted RAW-CT). The immunoblot results confirmed that unphosphorylated STAT1 was expressed stably in RAW-USTAT1 cells (Fig. 2A). Next, we assessed P-STAT1 (phosphorylated on Y701) and unphosphorylated STAT1 levels at each time point in both infected RAW-USTAT1 and RAW-CT cells. As shown in Fig. 2B, the result is similar to that detected in BMDMs. Both immunoblotting and immunofluorescence experiments showed that unphosphorylated STAT1 was distributed throughout the whole cell, and the accumulation of unphosphorylated STAT1 within the nuclear compartment increased in response to H37Ra infection (Fig. 2C,D). Next, we analyzed the apoptosis rates in the RAW-CT or RAW-USTAT1 cells infected with H37Ra. At the indicated time points post infection, the Annexin-V staining assay results showed that the cell apoptosis rate declined gradually with prolonged infection time. In addition, the H37Ra-infected RAW-USTAT1 cells had a significantly lower apoptosis rate than that of infected RAW-CT cells (Fig. 2E,F). Next, we analyzed the intracellular bacterial burden of the H37Ra-infected RAW-CT and RAW-USTAT1 cells. As expected, the bacterial burden of the RAW-USTAT1 cells increased significantly in comparison with the RAW-CT cells (Fig. 2G). It has been shown that unphosphorylated STAT1 directly regulates the expression of several constitutive genes (Chatterjee-Kishore et al., 2000; Cheon and Stark, 2009). Real-time quantitative PCR (qPCR) analysis revealed that the mRNA expression levels of Lmp2 (also known as Psmb9), Mcl1 and CIIta were upregulated and those of CD95, Hspa1a and Jak1 were downregulated in RAW-USTAT1 cells. Surprisingly, we also noticed that infection with Mtb strain H37Ra regulated those genes in a similar manner (Fig. 2H).

The viability of intracellular Mtb in macrophages is often associated with CD95 and CD95 ligand (CD95L, also known as FASLG)-mediated apoptosis (Oddo et al., 1998; Kornfeld et al., 1999). As unphosphorylated STAT1 downregulated the mRNA expression of CD95 (Fig. 2F), we suggest that unphosphorylated STAT1 may regulate the CD95/CD95L apoptotic pathway in H37Ra-infected macrophages. To test this hypothesis, we first analyzed the cell surface protein level of CD95 by flow cytometry. The mean fluorescence intensity (MFI) results indicated that overexpression of unphosphorylated STAT1 in RAW264.7 cells led to a lower level of cell surface CD95 protein as compared with that seen in the control cells (Fig. 3A,B). Secondly, we tested the sensitivity of RAW-USTAT1 cells to CD95L-mediated apoptosis. While RAW-CT cells were sensitive to CD95L-induced apoptosis, the proportion of apoptotic cells in the RAW-USTAT1 cell population was diminished compared with that of RAW-CT cells (Fig. 3C,D). Next, we evaluated the mRNA expression level of CD95L in H37Ra-infected RAW-CT and RAW-USTAT1 cells. The qPCR results showed the expression of CD95L was elevated in H37Ra-infected cells but the change did not seem to be caused by unphosphorylated STAT1 (Fig. 3E).

In order to further validate and strengthen our observation, we assessed the expression of cytosolic pro-apoptosis and anti-apoptosis effectors in RAW-USTAT1 cells infected with H37Ra. The loss of integrity of mitochondrial outer membrane is one of the hallmarks of apoptosis because it results in cytochrome c release, which culminates in the activation of caspases (Liu et al., 1996; Bratton et al., 2001). As shown in Fig. 3F, H37Ra infection-induced apoptosis was severely compromised in RAW-USTAT1 cells compared to RAW-CT cells as analyzed by determining the expression levels of pro-apoptosis genes, release of cytochrome c and activation of caspase-3, with a concomitant increase in expression of the anti-apoptosis gene Mcl-1 expression.

Unphosphorylated STAT1 has been shown to function as a novel regulator of several genes and to inhibit cell apoptosis (Cheon and Stark, 2009; Zimmerman et al., 2012; Chatterjee-Kishore et al., 2000), but until now, little has been known about its binding proteins, even for its human homolog. To map the unphosphorylated STAT1 interactome in macrophages, a one-step pull-down procedure, using streptavidin and biotinylated (Bio-tagged) unphosphorylated STAT1, was undertaken in RAW264.7 cells; in this procedure, the Bio tag allows a high affinity binding without affecting protein function (Blumert et al., 2013). We constructed a multi-cistronic lentiviral construct encoding a Bio tag fused to the C-terminus of the unphosphorylated STAT1 protein and Escherichia coli biotin ligase (BirA) separated by a self-cleaving 2A peptide sequence (P2A) derived from porcine teschovirus-1 (Kim et al., 2011) (Fig. 4A). The lentiviral construct encoding BirA alone was used as a control. RAW264.7 cells were transduced with these lentiviral constructs to establish cell lines stably overexpressing BirA-P2A-Bio-USTAT1 (denoted RAW-Bio-USTAT1) or BirA (RAW-Bio-CT) (Fig. 4B). The Bio-USTAT1 protein complexes were isolated from whole-cell extracts by streptavidin-mediated affinity purification. RAW-Bio-CT cells were used as negative controls. After purification, the protein complexes were separated by SDS-PAGE. Staining the gel with Coomassie blue showed that many protein bands were visible in the RAW-Bio-USTAT1 lane, but such bands were seldom visible in the RAW-Bio-CT lanes (Fig. 4C). Individual protein bands were excised, in-gel digested with trypsin and then sequenced by high-performance liquid chromatography-electrospray tandem mass spectrometry. A total of 149 known or predicted proteins were identified, which were repeatedly quantified in at least two of the three independent experiments in RAW-Bio-USTAT1 cells, but not detected or where at relatively low abundance in RAW-Bio-CT cells (Table S1). Among them was the closely related factor STAT3, which forms heterodimers with STAT1, as previous studies suggested (Buro et al., 2010; Blumert et al., 2013). Interaction of unphosphorylated STAT1 with STAT3, eEF1A, Rps3, Anxa2, Ybx1 and Bag2 were confirmed by a further immunoprecipitation (IP) and immunoblotting experiment, indicating that the interaction between these proteins and unphosphorylated STAT1 occurs in vivo (Fig. 4D).

A network map of the identified unphosphorylated STAT1 interactome was constructed using the STRING software (Szklarczyk et al., 2015). The network clusters associated with ribosome, antigen processing, virus infection and presentation, and endoplasmic reticulum were readily identified in the map (Fig. S1). To better clarify the functions and signaling pathways of unphosphorylated STAT1-interacting proteins, we used the DAVID software package (Huang et al., 2009) to classify all identified proteins according to their molecular functions and biological processes (Fig. 4E). The substantial enrichment in terms related to translation, protein transport, macromolecular assembly, macromolecular complex subunit organization and generation of precursor metabolites energy indicated several unexplored functions of unphosphorylated STAT1–KEGG pathway analysis revealed that unphosphorylated STAT1-interacting proteins were enriched in those annotated as belonging to ribosome, oxidative phosphorylation, antigen processing, and glycolysis/gluconeogenesis functions (Fig. 4F). These results confirm that we successfully isolated the unphosphorylated STAT1 protein complex from RAW264.7 cells and identified its interacting proteins. The interactome data suggest that unphosphorylated STAT1 is involved in transcriptional regulation, histone modification and immunological stress, and shuttles between the nucleus and cytoplasm.

We verified that unphosphorylated STAT1 and STAT3 could form a complex. It has been reported STAT3 is an oncogene, regulates many anti-apoptotic proteins (Bromberg et al., 1999; Catlett-Falcone et al., 1999) and elicits suppression of CD95 by cooperating with other transcription factor(s) (Ivanov et al., 2001). Hence, we predicted that the complex between unphosphorylated STAT1 and STAT3 suppresses CD95 transcription. To validate our hypothesis, we silenced unphosphorylated STAT1 and STAT3 expression with siRNA in RAW 264.7 cells (Fig. 5A) and analyzed the CD95 expression both at the mRNA level and cell surface protein level. [Note that the level of phosphorylated STAT1 would also be partly diminished by the siRNA; however, as the phosphorylated STAT1 form is transient and unphosphorylated STAT1 protein is the main form present in later stages of infection (Fig. 2), the siRNA chiefly affects the level of unphosphorylated STAT1.] A clear increase in the expression level of CD95 was observed in both unphosphorylated STAT1 and STAT3 silenced cells. Similarly, silencing both unphosphorylated STAT1 and STAT3 together led to a stronger increase in CD95 expression compared with control (Fig. 5B–D). Furthermore, we tested the apoptosis rate of the cells with silenced unphosphorylated STAT1 or STAT3 with infection of H37Ra. Compared with control, silencing unphosphorylated STAT1 and STAT3 dramatically increased macrophages sensitivity to H37Ra-induced apoptosis in vitro (Fig. 5E,F). This finding points to an inverse correlation between unphosphorylated STAT1–STAT3 complex and CD95 expression, suggesting that these complexes have a role in the inhibition of CD95-mediated apoptosis in H37Ra-infected macrophages.

Previous studies have shown that the –(460–230) sequence [–(360–130) for mouse] in the CD95 promoter (relative to the transcription start site) is the key region containing GAS-, AP1- and NF-κB-binding sites, which represent the activities seen by the full-length 1.7 kb promoter region (Chan et al., 1999). Hence, we considered this region may be the site at which the complex between unphosphorylated STAT1 and STAT3 binds. Further confirmation for binding of the unphosphorylated STAT1 and STAT3 complex to the CD95 promoter was investigated in vivo by performing chromatin immunoprecipitation (ChIP) assays on chromatin samples from H37Ra-infected RAW264.7 cells (Fig. 5G). Both unphosphorylated STAT1 and STAT3 were found in association with CD95 promoter in cells infected for 12 h, consistent with the notion that both transcription factors are bound to the CD95 promoter under conditions in which CD95 expression is suppressed. The binding of unphosphorylated STAT1 to CD95 promoter was lower at 6 h and even lower in uninfected cells (0 h time point) compared to that seen at 12 h after infection (Fig. 5G). The increase in the amount of unphosphorylated STAT1 and STAT3 binding to the CD95 promoter sequences after H37Ra infection, suggests that the inhibition of apoptosis is mediated through a reduction in CD95 expression.

Another unphosphorylated STAT1-interacting partner, eukaryotic elongation factor 1A (eEF1A), is reported to be involved in apoptosis in several cell types (Duttaroy et al., 1998; Chen et al., 2000; Lamberti et al., 2004; Condeelis, 1995; Kato, 1999; Kato et al., 1997). A recent study has pointed out that the interaction of interferon-induced protein tetratricopeptide repeats-1 (IFIT1) and eEF1A complex is involved in the processes of apoptosis induced by TNF in macrophages (Li et al., 2010). In order to assess whether or not IFIT1 and eEF1A protein levels in RAW264.7 cells are affected by H37Ra infection, immunoblotting analysis was performed and the levels of the proteins determined by densitometric analysis of band intensities standardized with internal control. As determined by immunoblot band densitometries, the timecourse of the kinetics of IFIT1 increase is similar to that of eEF1A. The IFIT1 and eEF1A protein expression are both induced after initiation of H37Ra infection. Both IFIT1 and eEF1A protein levels achieved the peak elevation within 6 h, and did not return to normal levels, but persisted at a high levels to 48 h (Fig. 6A,B). Since unphosphorylated STAT1 and eEF1A can also form a complex (Fig. 4), it implies that unphosphorylated STAT1 and IFIT1 might display competitive binding to eEF1A. To validate our hypothesis, we transfected an expression vector encoding 3×FLAG-tagged IFIT1 into RAW-Bio-USTAT1 cell or RAW-Bio-CT cells. Cellular extracts were co-immunoprecipitated with anti-FLAG antibody. eEF1A protein was precipitated with FLAG–IFIT1 in both RAW-Bio-USTAT1 cell and RAW-Bio-CT cells, but the band of eEF1A in RAW-Bio-USTAT1 from the co-immunoprecipitation experiment had a lower intensity than the band of RAW-Bio-CT (Fig. 6C), suggesting that unphosphorylated STAT1 affects the interaction between eEF1A and IFIT1. These observations establish the interaction of IFIT1 and eEF1A is disassociated upon the accumulation of unphosphorylated STAT1 (Fig. 6D).

Following Mtb binding to macrophages through Toll-like receptor (TLRs), there is activation of intracellular cell signaling events leading to macrophage apoptosis or necrosis. Previous reports have documented that the JAK/STAT signaling pathway is dependent on a variety of adapter proteins, such as interferon-γ receptors (IFNγRs) (Dale et al., 1989) and TLRs (Rhee et al., 2003; Kovarik et al., 1999). It is known that the JAK/STAT signaling pathway plays an important role in the Mtb phagocytosis-independent method of macrophage apoptosis induction (Rojas et al., 2002; Lim et al., 2016). Selective gene deletion of STAT1 in mice leads to rapid death from severe infections, including intracellular bacteria and viruses, demonstrating its major role in the response to pathogens (Meraz et al., 1996; Dupuis et al., 2003). However, our study shows that the increase in the phosphorylated STAT1 form in H37Ra-infected mice macrophages is transient, while the expression of STAT1 is substantially increased, and the unphosphorylated STAT1 form accumulates and persists for at least 2 days. It is possible that the repressed expression of Jak1 mediated by unphosphorylated STAT1 might lead to the inhibition of STAT1 phosphorylation, which terminates the P-STAT1 signaling pathway. An analogous negative-feedback mechanism is found in several other immune cells (Zimmerman et al., 2012; Cheon and Stark, 2009). Interestingly, there is a similar trend between the level of unphosphorylated STAT1 and the intracellular bacterial burden in H37Ra-infected macrophages. The bacillary viability is associated with the apoptotic death of infected macrophages (Oddo et al., 1998). Additionally, the inhibition of activation of caspase-3 and cytochrome c translocation, and increase in the level of Mcl-1 in macrophages overexpressing unphosphorylated STAT1, leads to a severe reduction in H37Ra infection-induced apoptosis. Overall, our finding suggest that, in an opposite manner to P-STAT1, unphosphorylated STAT1 inhibits Mtb-induced macrophage apoptosis.

Previous studies have extensively presented the localization of unphosphorylated STAT1 in many cell lines, finding that its nuclear export independently of phosphorylation, is in a cell type-specific manner (Cheon and Stark, 2009; Meyer et al., 2002). Our results indicate that the U-STAT1 induced by Mtb moves into the nuclei, where it can function as a novel transcription factor to increase the expression of immune regulatory genes. It has been shown that unphosphorylated STAT1 directly regulates the expression of several constitutive genes in several types of cells (Zimmerman et al., 2012; Chatterjee-Kishore et al., 2000), and similar results have been validated in mice macrophage cell line RAW264.7. These genes are involved in immune responses, antigen presentation and programmed cell death; for example, Lmp2 and CIIta mediate IFN-induced antiviral, antitumor and immunomodulatory activities (Muhlethaler-Mottet et al., 1998; Chatterjee-Kishore et al., 1998); Mcl1 is an anti-apoptosis member of the B-cell lymphoma 2 (Bcl-2) family (Thomas et al., 2010); Jak1 is the key kinase that is involved in the phosphorylation of STAT1 (Yeh and Pellegrini, 1999); and Hsp70 is associated with the CD95-induced apoptotic cell death (Liossis et al., 1997). Interestingly, expression of these unphosphorylated STAT1-regulated genes is also induced by H37Ra infection (the result is consistent with the previous microarray results on murine macrophages; Zhang, 2013), raising the possibility that the H37Ra-induced gene expression may partly be due to the accumulation of unphosphorylated STAT1 in macrophages. These results provide evidence that unphosphorylated STAT1 may act as a novel transcription factor involved in the expression of a wide variety of genes and regulation of cell destiny in H37Ra-infected macrophages.

It has been observed that the accumulation of unphosphorylated STAT1 participates in resistance to cell apoptosis (Bowman et al., 2000; Kao et al., 2013). Indeed, the aberrant expression of unphosphorylated STAT1 is closely associated with inflammation caused by pulmonary infection (Sampath et al., 1999; Lee et al., 2000). In this study, we showed that increasing unphosphorylated STAT1 concentration exogenously leads to a significantly lower rate of apoptosis in H37Ra-infected or CD95L-initiated cells compared with control. Hence, this indicated that the accumulation of unphosphorylated STAT1 decreases the sensitivity of RAW264.7 cells to CD95/CD95L-mediated apoptosis during H37Ra infection. The CD95/CD95L system is often associated with pathogen infection (Oddo et al., 1998; Zhang et al., 2005). Infected macrophages showed a reduced susceptibility to CD95L-induced apoptosis, correlating with a reduced level of CD95 expression in both our and other studies (Oddo et al., 1998; Zhang et al., 2005). CD95 itself does not initiate apoptosis. CD95L is the physiologic ligand of CD95 and expressed on the surface of cytotoxic T lymphocytes (CTLs) (Kilinc et al., 2009; Fritzsching et al., 2005) and macrophages (Zhang et al., 2005; Guenther et al., 2012). Upregulated CD95L proteins in H37Ra-infected macrophages cells may interact with each other to induce apoptosis (Xu et al., 1998). Overall, our results suggest that the unphosphorylated STAT1 might confer resistance to CD95-mediated apoptosis to macrophages to allow them to escape the CTL- or macrophage-mediated pathogen suppression, which represents a pathogen strategy for survival in the host.

A previous study has shown that unphosphorylated STAT family proteins play a role as active transcription factors: unphosphorylated STAT3 cooperates with NF-κB to induce RANTES (also known as CCL5) gene expression (Yang et al., 2007, 2005) and unphosphorylated STAT6 forms a complex with p300 (also known as EP300) to upregulate the cyclooxygenase-2 gene (Cui et al., 2007). To describe the transcriptional pattern and the complicated cellular biological processes of unphosphorylated STAT1, exploring its binding partners and networks of protein is necessary. In the present study, we used a streptavidin-mediated affinity purification approach to comprehensively analyze the composition of the unphosphorylated STAT1 interactome in RAW264.7 cells, and it appears that the complex between unphosphorylated STAT1 and STAT3 seems to be that primarily involved in the gene regulation mediated by unphosphorylated STAT1. A previous study has suggested that unphosphorylated STAT1 together with other STATs can form homo- or hetero-dimers that bind to overlap GAS elements to regulate gene expression (Chatterjee-Kishore et al., 2000). Our ChIP data demonstrated that the complex between unphosphorylated STAT1 and STAT3 inhibits the expression of CD95 by binding to the key region of the CD95 promoter. Following Mtb infection, there is an increase in unphosphorylated STAT1 and an enhanced binding to the CD95 promoter in vivo, which coincides with decreased CD95 transcriptional expression and concomitant CD95L-dependent apoptosis, which provides important support for a dynamic regulation by recruitment of a complex between unphosphorylated STAT1 and STAT3 to the CD95 promoter during infection.

In addition to the above nuclear roles, the cytoplasm-localized unphosphorylated STAT1 may also play an anti-apoptosis role. Our results show that unphosphorylated STAT1 affects the interaction affinity between eEF1A and IFIT1. eEF1A has a very complicated role not only in the protein elongation step of translation (Browne and Proud, 2002), but also in the process of stress-induced cell apoptosis (Duttaroy et al., 1998; Chen et al., 2000; Lamberti et al., 2004; Condeelis, 1995; Kato, 1999; Kato et al., 1997). IFIT1 acts as an infection resistor by suppressing protein translation (Guo et al., 2000; Hui et al., 2005), and lipopolysaccharide or TNF stimulation rapidly activates IFIT1 transcription (Smith and Herschman, 1996; Wathelet et al., 1987). A previous study has shown that coupling with IFIT1 results in a change of eEF1A activity and subsequent stress-induced cell apoptosis, suggesting that this complex may accumulate during the early stages of infection and be involved in the processes of cell apoptosis (Li et al., 2010). We observed the levels of IFIT1 and eEF1A in RAW264.7 cells infected by H37Ra. Their protein levels elevated (within 6 h) during the early stages of infection, suggesting that the interaction between IFIT1 and eEF1A may augment the apoptosis process in infected cells. Unphosphorylated STAT1 affects the interaction affinity between eEF1A and IFIT1, further inferring that during the later stages of infection increased cytoplasmic unphosphorylated STAT1 is involved in the competitive binding reactions that may inhibit the apoptosis processes mediated by the eEF1A–IFIT1 complex. In order to characterize the function of the complex between unphosphorylated STAT1 and eEF1A, more supporting evidence is required, and further experimental studies are also needed to confirm how unphosphorylated STAT1 and its interacting partners affect H37Ra-induced apoptosis in macrophages.

In conclusion, we propose that STAT1 has a dual role during H37Ra infection in macrophages (Fig. 7). On the one hand, exposure of macrophages to H37Ra induces rapid phosphorylation of STAT1 followed by production of TNF and NOS2, and caspase activation to enhance macrophage sensitivity to induction of apoptosis, which is linked to a good response for the host to in resistance to pathogens. On the other hand, after short-term activation of P-STAT1, unphosphorylated STAT1 starts to accumulate after P-STAT1 start to degrade. The sustainable high levels of unphosphorylated STAT1 suppresses Jak1 to terminate the P-STAT1 signaling in a feedback inhibition manner. The unphosphorylated STAT1 then recruit STAT3, forming a complex that further represses CD95 to inhibit macrophage apoptosis resistance to Mtb. This has a negative impact on the outcome of infection for the host. In addition, the featured binding partners of unphosphorylated STAT1 might play roles as alternative signaling modules responsible for the formation of transcriptional complex and recruitment of this complex to the apoptotic gene promoters. These observations might eventually contribute to a deeper understanding of the immune evasion mechanism and lead to therapeutic modulation of the functions of unphosphorylated STAT1 during Mtb infection in macrophages.

The animal experiments were approved by the Animal Care Commission of the College of Veterinary Medicine, Northwest A&F University. Adult C57BL/6 mice (4–6-weeks-old, weighing 20–25 g) were purchased from the experimental animal center of the Fourth Military Medical University (Xi'an, Shaanxi, China) and maintained on a 14-h-light–10-h-dark cycle with free access to food and water in strict accordance with the Guidelines for the Care and Use of Animals of Northwest A&F University. Every effort was made to minimize animal pain, suffering and distress and to reduce the number of animals used.

The BMDMs of C57BL/6 mice were isolated as described previously (Lim et al., 2016). Briefly, BMDMs were generated by flushing bone marrow cells from femurs and tibias, and culturing for 2 days in RPMI1640 supplemented with 10% fetal bovine serum, penicillin (100 IU/ml), streptomycin (100 μg/ml), 25 ng/ml granulocyte-macrophage colony-stimulating factor (R&D Systems, Minneapolis, MN) or 25 ng/ml macrophage colony-stimulating factor (R&D Systems). The mouse macrophage cell line RAW264.7 was grown in RPMI1640 supplemented with 10% fetal bovine serum. All cells were cultured at 37°C in 5% CO₂ in humidified incubator.

M. tuberculosis strain H37Ra (ATCC 25177) was cultured in Middlebrook 7H9 broth medium supplemented with 10% OADC (Becton, Dickinson and Company, Franklin Lakes, NJ). BMDMs or RAW264.7 cells were infected at a multiplicity of infection of five bacteria per cell (MOI 5:1). After 1 h, the infected cells were washed with RPMI1640 and new medium added. To assay the intracellular bacterial burden of macrophages, the infected cells were incubated for the indicated time (timing started when medium was refreshed) and then lysed in PBS with 0.05% SDS, and intracellular bacteria were plated on Middlebrook 7H10 agar plates supplemented with OADC and incubated for 3 weeks at 37°C to determine the colony number.

The STAT1-coding sequence was amplified by PCR by using cDNA from C57BL/6 mouse lung. The coding sequence of biotin ligase BirA was amplified from genomic DNA of E. coli BL21 (DE3). The STAT1 site-directed mutagenesis (Y701F-S727E) were introduced by using the QuikChange Lightning Site-Directed Mutagenesis kit (Stratagene, Santa Clara, CA). To construct the lentiviral constructs unphosphorylated STAT1, Bio-USTAT1-P2A-BirA or Bio-P2A-BirA, the expression cassettes with FLAG tag, the biotin tag, P2A, USTAT1 and BirA open reading frame (ORF) sequences were assembled according to their relative location by overlap extension PCR, and then cloned into the lentiviral vector pCDH-MCS-T2A-Puro-MSCV. The full-length coding cDNA fragment of mouse IFIT1 was obtained using cDNA from C57BL/6 mouse lung, then cloned into p3×FLAG-CMV-10 vectors. Primer sequences for the plasmid construction are listed in Table S2. All the constructs were confirmed by DNA sequencing. To produce infectious virus, each construct was transfected into 293FT packaging cells by using Lipofectamine^®2000 (Invitrogen, Carlsbad, CA). The supernatant medium, collected once each day, was applied to infect cells. To select stably transfected cells, they were treated with 5 μg/ml puromycin for more than 2 weeks. Proteins or RNAs were extracted and purified from these cell pools for western blot or qPCR analyses.

Nuclear extracts of RAW264.7 cells were prepared as described previously (Rojas et al., 2002). Briefly, macrophages were harvested and resuspended in 1.5 ml cold PBS. The cell pellets were resuspended in 400 μl lysis buffer (10 mM Hepes-KOH, pH 7.9, at 4°C, 1.5 mM MgCl2, 10 mM KCl, 0.5 mM DTT, and 0.2 mM PMSF), incubated on ice for 10 min and then vortexed for 10 s. Samples were centrifuged for 10 s at 15,000 g at room temperature and the supernatants were subjected to western blot analysis to detect cytoplasmic components. The pellet was resuspended in 100 μl of cold nuclear lysis buffer (20 mM Hepes-KOH, pH 7.9, 25% glycerol, 420 nM NaCl, 1.5 mM MgCl2, 0.2 mM EDTA, 0.5 mM DTT, and 0.2 mM PMSF) and incubated on ice for 20 min. Nuclear debris was removed by centrifugation at 15,000 g for 10 s, and the supernatants were used to detect nuclear components.

Cells were fixed with 3.7% formaldehyde in phosphate-buffered saline (PBS) for 30 min, incubated with DAPI (Beyotime, Jiangsu, China) to visualize the nuclei, washed with PBS, and mounted onto microscope slides. Exogenous FLAG-tagged STAT1 stained with anti-FLAG antibody (1:500, F1804, Sigma, St Louis, MO), diluted in blocking solution, was applied. After incubation at 4°C overnight, the cells were incubated with Alexa 555-conjugated goat anti-mouse-IgG antibodies (1:500, Beyotime) diluted in blocking solution for 2 h. After washing twice with PBS, the slides were examined and images acquired by using a Nikon ECLIPSE TE2000 confocal microscope.

Cell apoptosis was determined by staining with Alexa Fluor 488-conjugated Annexin V and propidium iodide (PI) (Molecular Probes, Eugene, OR), and then analyzed by flow cytometry (BD Biosciences, San Jose, CA). CD95L protein was purchased from Sigma. Cells were stained with FITC-conjugated anti-CD95 monoclonal antibody (Biorbyt, Berkeley, CA) and analyzed by flow cytometry. CD95 protein level was quantified through determining the mean fluorescence intensity. Flow cytometry was used to sort viable cells which were further gated on the basis of forward versus side scatter and then analyzed after staining with conjugated dye. In each analysis, appropriate negative or positive controls were used to define gate settings.

Total RNA was extracted from RAW264.7 cells by using Trizol reagent (Invitrogen), and then 1 μg of RNA was reverse transcribed to cDNA by using the SYBR PrimeScript RT reagent kit (Takara, Dalian, China). The qPCR was performed by using SYBR Premix ExTaq II (Takara) on a StepOne Plus PCR system (Applied Biosystems, Foster City, CA). The comparative CT method was employed for quantification of target mRNA expression, and the relative expression of mRNA was normalized to GAPDH expression. Primer sequences for the qPCR are listed in Table S2.

The fractionation of cytosolic extract was performed by following a previously described procedure (Ghorpade et al., 2012). In brief, the cells were harvested and gently resuspended in lysis and extraction buffer (10 mM HEPES pH 7.9, 10 mM KCl, 0.1 mM EDTA, 0.1 mM EGTA, 1 mM dithiothreitol, and 0.5 mM PMSF). Cell membranes were disrupted with 10% NP-40 after incubation on ice for 15 min. Cytosolic extract was finally separated from nuclei and mitochondria by centrifugation at 16,200 g for 15 min at 4°C.

The purification of unphosphorylated STAT1 protein complexes was achieved by following by a described protocol (Kim et al., 2009). The affinity purification protein sample was separated by 4–20% gradient SDS-PAGE (Bio-Rad), and then stained with Acqua stain protein gel dye (Bulldog Bio Inc., Portsmouth, NH). Bands were excised from the gel and digested with trypsin. The digested extraction was analyzed using liquid chromatography coupling electrospray ionization tandem mass spectrometry (LC/MS/MS) as reported previously (Wang et al., 2016). According to the obtained amino acids sequence information, the BLAST was carried out using the NCBI database.

Co-immunoprecipitation was carried out by using a co-immunoprecipitation kit (Pierce, Rockford, IL), following the manufacturer's instructions. The immunoprecipitated protein sample was resolved in a 12% SDS-PAGE gel and transferred onto a PVDF membrane. Membranes were blocked with 10% non-fat dry milk diluted in TBST for 3 h, and probed with primary antibodies (1:1000) overnight at 4°C, and subsequently incubated with HRP-labeled goat anti-mouse-IgG antibodies (1:1000, Beyotime, Jiangsu, China). Finally, blots were developed with ECL chemiluminescence reagent (Beyotime). Primary antibodies used were against pStat1 (Tyr701), pStat1 (Ser727), caspase 3, Mcl-1, cytochrome c, STAT3, eEF1A, Anxa2 and Ybx1 purchased from Cell Signaling Technology (Danvers, MA), monoclonal antibody against STAT1 (C-terminal) purchased from BD Transduction Laboratories (Biocompare, CA), IFIT1 and Bag2 (Abcam, Cambridge, MA), FLAG (Sigma), actin (TransGen Biotech, Beijing, China), PCNA, Cox IV and horseradish peroxidase (HRP)-conjugated streptavidin (Beyotime). The catalog numbers of each primary antibody are listed in Table S3.

siRNA targeting mouse STAT1 and STAT3 and negative control siRNA were purchased from GenePharma (Shanghai, China). siRNA sequences were as follows, si-Stat1: 5′-GCUGAACUAUAACUUGAAA-3′; si-Stat3, 5′-GGGUCUGGCUAGACAAUAUTT-3′; si-control, 5′-UUCUCCGAACGUGUCACGU-3′. RAW264.7 cells were transfected with 50 nM of indicated siRNAs overnight using Lipofectamine 2000 Reagent (Thermo Scientific, Rockford, IL).

The ChIP was carried out by using a ChIP kit (Pierce, Rockford, IL). RAW264.7 cells (2.5×10⁶ cells per 10 cm diameter plate) were infected with Mtb H37Ra strain and cells were fixed with 1% (w/v) formaldehyde at the indicated time points. The reaction was subsequently quenched with 125 mM glycine. Genomic DNA was isolated and sheared to average lengths of 500–1000 bp by ultrasonication. 50 μg of purified chromatin samples were immunoprecipitated with 1 μg of anti-unphosphorylated STAT1, anti-STAT3 or IgG. ChIP enrichment was performed by qPCR (17 cycles of 1 min at 94°C, 53°C, and 72°C) using primers (forward, 5′-TGAGCAGCAGGCAGAAAAAAAAATCTCACTTGAC-3′ and reverse, 5′-TGTCACTTTTTTTCTTTTGAGGGACCAAACACAAT-3′) that amplify the fragment-containing key region of CD95 promoter.

The data are represented as the mean±s.d. and were analyzed using the Student's t-test. A value of P<0.05 was considered significant.