Mycobacterium tuberculosis PE_PGRS20 and PE_PGRS47 Proteins Inhibit Autophagy by Interaction with Rab1A

ABSTRACT Autophagy is a fundamental cellular process that has important roles in innate and adaptive immunity against a broad range of microbes. Many pathogenic microbes have evolved mechanisms to evade or exploit autophagy. It has been previously demonstrated that induction of autophagy can suppress the intracellular survival of mycobacteria, and several PE_PGRS family proteins of Mycobacterium tuberculosis have been proposed to act as inhibitors of autophagy to promote mycobacterial survival. However, the mechanisms by which these effectors inhibit autophagy have not been defined. Here, we report detailed studies of M. tuberculosis deletion mutants of two genes, pe_pgrs20 and pe_pgrs47, that we previously reported as having a role in preventing autophagy of infected host cells. These mutants resulted in increased autophagy and reduced intracellular survival of M. tuberculosis in macrophages. This phenotype was accompanied by increased cytokine production and antigen presentation by infected cells. We further demonstrated that autophagy inhibition by PE_PGRS20 and PE_PGRS47 resulted from canonical autophagy rather than autophagy flux inhibition. Using macrophages transfected to express PE_PGRS20 or PE_PGRS47, we showed that these proteins inhibited autophagy initiation directly by interacting with Ras-related protein Rab1A. Silencing of Rab1A in mammalian cells rescued the survival defects of the pe_pgrs20 and pe_pgrs47 deletion mutant strains and reduced cytokine secretion. To our knowledge, this is the first study to identify mycobacterial effectors that directly interact with host proteins responsible for autophagy initiation. IMPORTANCE Tuberculosis is a significant global infectious disease caused by infection of the lungs with Mycobacterium tuberculosis, which then resides and replicates mainly within host phagocytic cells. Autophagy is a complex host cellular process that helps control intracellular infections and enhance innate and adaptive immune responses. During coevolution with humans, M. tuberculosis has acquired various mechanisms to inhibit host cellular processes, including autophagy. We identified two related M. tuberculosis proteins, PE_PGRS20 and PE_PGRS47, as the first reported examples of specific mycobacterial effectors interfering with the initiation stage of autophagy. Autophagy regulation by these PE_PGRS proteins leads to increased bacterial survival in phagocytic cells and increased autophagic degradation of mycobacterial antigens to stimulate adaptive immune responses. A better understanding of how M. tuberculosis regulates autophagy in host cells could facilitate the design of new and more effective therapeutics or vaccines against tuberculosis.

and mortality. In 2019, 1.4 million deaths were directly attributed to M. tuberculosis infection, and a further 10 million people developed active TB (1,2). M. tuberculosis has evolved many effective survival strategies as a primarily intracellular parasite of host phagocytic cells. One of its best-studied survival mechanisms is phagolysosome maturation and acidification (3). However, substantial evidence shows that other host defense mechanisms, including autophagy and apoptosis, are also inhibited by M. tuberculosis infection (4)(5)(6).
Autophagy is a ubiquitous and fundamental cellular process and is an important control mechanism against intracellular microbes. Autophagy can initiate phagosome maturation and enhance the processing and presentation of antigens to T cells, increasing bacterial clearance through innate and adaptive immunity. Regulation and modulation of autophagy are complex and affected by many feedback mechanisms, most of which are not fully elucidated. Stimulation of autophagy by starvation or by its critical regulator mechanistic target of rapamycin (mTOR) by drugs such as rapamycin results in colocalization of intracellular bacteria to autophagosomes and, consequently, increased bacterial clearance (4)(5)(6)(7)(8)(9)(10).
Bacterial clearance by autophagy is termed xenophagy and will be referred to as autophagy here. M. tuberculosis cytosolic DNA is recognized by the cytosolic DNA sensor, cyclic GMP-AMP synthase (cGAS), resulting in the release of cyclic GMP (cGMP). cGAMP is recognized by the stimulator of interferon genes (STING), leading to the recruitment of ubiquitin receptors p62, NDP52, and optineurin, along with type I interferon (IFN) release (11,12). These receptors are recruited to the ubiquitinated pathogen, thereby allowing for specific targeting by the autophagosome (13)(14)(15). While autophagy is an effective mechanism for clearing bacteria, some bacterial pathogens possess effectors inhibiting its induction from promoting their survival. In mycobacteria, multiple such effectors have been identified, including the secreted protein EIS (enhanced intracellular survival) (16,17) and several virulence-associated M. tuberculosis PE_PGRS and PPE family proteins (6,18,19). The mechanisms by which these mycobacterial effectors may inhibit autophagy are not well understood.
The PE/PPE family of mycobacterial proteins is found most abundantly in slowgrowing pathogenic mycobacteria and constitutes approximately 10% of the M. tuberculosis genome (20). It has been previously demonstrated that the PE/PPE proteins of M. tuberculosis are partially responsible for inhibiting autophagy in phagocytic cells upon infection with mycobacteria (6,18,19,21). This family of proteins has coevolved with the ESX type VII secretion systems present in mycobacteria, which are prominently involved in virulence and intracellular survival of mycobacteria (22)(23)(24). Several PE/PPE proteins have been implicated as autophagy inhibitors in M. tuberculosisinfected cells, thus contributing to the virulence and persistence of the bacteria (6,18,19,21). Our previous screen of a random transposon mutant library of M. tuberculosis identified PE_PGRS20 and PE_PGRS47 as two such autophagy-inhibiting factors.
In the current study, we focused on determining how M. tuberculosis PE_PGRS20 and PE_PGRS47 proteins inhibit autophagy. We found that M. tuberculosis strains with genetic deletions of pe_pgrs20 and pe_pgrs47 induced increased autophagy and reduced bacterial survival in human monocyte-derived macrophages and murine bone marrow-derived macrophages. While confirming that this autophagy induction was reliant on canonical autophagy and the previously described role for PE_PGRS20 and PE_PGRS47 in mTOR modulation (18), the current study identified a functionally relevant direct interaction of these PE_PGRS proteins with Rab1A, a host protein required for autophagy initiation. In addition to blocking autophagy, the interaction of Rab1A with PE_PGRS20 or PE_PGRS47 reduced secretion of proinflammatory cytokines and inhibited major histocompatibility complex (MHC) class II-restricted antigen presentation. These results demonstrate related effects on innate and adaptive immunity that are likely to contribute to bacterial escape from host defenses during M. tuberculosis infection.

RESULTS
PE_PGRS20 and PE_PGRS47 of M. tuberculosis inhibit autophagy in macrophages. Previous studies have implicated several PE_PGRS proteins of M. tuberculosis as host cell autophagy inhibitors (6,19). We identified pe_pgrs20 and pe_pgrs47 as contributors to autophagy inhibition in a loss-of-function screen of a transposon M. tuberculosis library (13). In this study, we demonstrated that both recombinant PE_PGRS proteins are secreted by Mycobacterium smegmatis. M. tuberculosis PE_PGRS20 was expressed in response to oxidative stress, a known autophagy inducer, while PE_PGRS47 was expressed in response to acidic and oxidative stress. PE_PGRS20 and PE_PGRS47 inhibited autophagy induced by M.smegmatis and reduced proinflammatory cytokine secretion during infection (18). To confirm and extend these findings, we generated strains with targeted deletions of pe_pgrs20 (Rv1068c) or pe_pgrs47 (Rv2741) genes in M. tuberculosis H37Rv via homologous recombination using specialized transduction (25). An increase in autophagy following infection with these mutants was observed in macrophages derived from human THP-1 cells, based on immunoblot analysis showing increased levels of LC3B-II, and this effect was reversed by genetic complementation of the deleted genes ( Fig. 1A and B and Fig. S1 in the supplemental material). Inhibition of autophagy during mycobacterial infection has been previously shown to limit bacterial clearance from phagocytic cells (5). Consistent with this, both DPE_PGRS20 and DPE_PGRS47 strains showed reduced intracellular bacterial burdens following infection of THP-1 cells compared to infection with their complemented strains or with wildtype parental M. tuberculosis (Fig. 1C). The DPE_PGRS20 and DPE_PGRS47 M. tuberculosis strains also induced significantly increased autophagy in primary mouse bone marrow-derived macrophages (BMDM) compared to wild-type M. tuberculosis ( Fig. S2A and B). As with THP-1 macrophages, the increased autophagy led to reduced cultivatable intracellular bacteria during DPE_PGRS20 and DPE_PGRS47 infection in BMDMs (Fig. S2C).
To further assess the role of autophagy in reducing the bacterial burden in cells infected with the mutant strains, we used an autophagy reporter RAW 264.7 mouse macrophage cell line that stably expresses the LC3B-GFP fusion protein (26). These were infected with parental M. tuberculosis, DPE_PGRS20, or DPE_PGRS47 bacteria constitutively expressing a dsRed fluorescent marker. At 24 h postinfection, at a multiplicity of infection (MOI) of 10, significantly more green fluorescent protein (GFP)-positive puncta were observed in cells infected with the DPE_PGRS20 and DPE_PGRS47 than cells infected with the wild-type parental M. tuberculosis ( Fig. 1D and E). This correlated with an increased colocalization of GFP-positive puncta with red fluorescent DPE_PGRS20 and DPE_PGRS47 mutant bacteria (Fig. 1F).
To confirm whether these effects of PE_PGRS20 and PE_PGRS47 on autophagy inhibition and decreased bacterial survival were through the regulation of the canonical autophagy pathway, we analyzed the impact of genetic knockdown of autophagyrelated gene 16L1 (Atg16L1), an essential component of this pathway. RAW 264.7 macrophages were stably transfected with inhibitory short hairpin RNA constructs (shAtg16L1). These transfected cells exhibited a significant reduction in Atg16L1 expression and autophagy induction during Torin-1 treatment and M. smegmatis infection (Fig. S3), both known as strong autophagy inducers via the canonical pathway (26,27). Infection of shAtg16L1 macrophages with M. tuberculosis, DPE_PGRS20 and DPE_PGRS47 resulted in limited autophagy induction, suggesting that these PE_PGRS gene products inhibit canonical autophagy ( Fig. 2A and B). As expected, the loss of autophagy induction resulted in the same intracellular bacterial burden in shAtg16L1 macrophages during DPE_PGRS20 and DPE_PGRS47 infection compared to wild-type infection at 48 h postinfection (Fig. 2C). Atg16L1 silencing had no significant effects on autophagy induction and intracellular bacterial burden during wild-type M. tuberculosis infection, indicating that M. tuberculosis actively inhibits autophagy induction.
To further confirm that these PE_PGRS proteins inhibit autophagy induction and not alternative autophagy-related processes such as LC3-associated phagocytosis (LAP) or autophagic flux, immunoblots for additional components of the autophagy machinery were analyzed. Infection of RAW 264.7 macrophages with DPE_PGRS20 and DPE_PGRS47 resulted in increased Atg5 and decreased p62 (SQSTM1) and phospho-Unc-51-like autophagy-activating kinase 1 (Ulk1) compared to wild-type M. tuberculosis infection ( Fig. 2D and E). Atg5 is an essential component of the autophagosome machinery, and increased Atg5 suggests autophagy activation during DPE_PGRS20 and DPE_PGRS47 infection. Reduced induction of the phosphorylation of Ulk1 at the mTORC1 phosphorylation site (Ser-757) also indicated autophagy activation through inhibition of the negative autophagy regulator, mTOR. Reduced p62 (an indicator of autophagic flux) accumulation indicated autophagosomes and their cargo were being degraded by autophagy during DPE_PGRS20 and DPE_PGRS47 infection ( Fig. 2D and E). Bafilomycin A1 inhibits autophagy flux by blocking autophagosome degradation by lysosomes. Accumulation of similar levels of LC3B-II (autophagosomes) was observed in bafilomycin-treated cells ( Fig. 2F and G), once again indicating that autophagy flux was not impaired in infected macrophages. To further support the conclusion that these pe_pgrs gene products did not modulate LAP, rubicon expression was measured ( Fig. 2D and E) since an increase in rubicon levels is associated with LAP but not with canonical autophagy (28,29). No significant differences in rubicon levels were observed following Densitometric summary analysis was calculated by LC3B-II normalized to b-actin, and then the fold change ratio was calculated compared to uninfected control for each assay. The mean 6 SD of 3 independent assays is shown. Significance was calculated by two-way ANOVA corrected by Dunnett's test for multiple comparisons. (C) M. tuberculosis survival was determined in wild-type RAW 264.7 and shAtg16L1 stably transfected macrophages (MOI 10) at 48 h postinfection. The mean 6 SD of 3 independent assays is shown. Significance was calculated by two-way ANOVA corrected by Dunnett's test for multiple comparisons. (D) Immunoblotting of Atg5, phospho-Ulk1 (Ser757), p62, and Rubicon in RAW 264.7 macrophages infected with M. tuberculosis at MOI of 10, 24 h postinfection. (E) Densitometric summary analysis was calculated by the protein of interest normalized to GAPDH, and then the fold change ratio was calculated compared to uninfected control for each assay. The mean 6 SD of 3 independent assays is shown. Significance was calculated by one-way ANOVA corrected by Dunnett's test for multiple comparisons. (F) Immunoblotting of LC3B-II accumulation in RAW 264.7 macrophages infected with M. tuberculosis at MOI of 10, 24 h postinfection with 10mM bafilomycin A1 treatment. (G) Summary densiometric analysis of the ratio between LC3B-II levels from RAW 264.7 macrophages infected at MOI 10 for 24 h with and without bafilomycin A1 treatment (autophagy flux). *, P # 0.05; **, P # 0.01; ***, P # 0.001. infection with either wild-type, mutant, or complemented bacteria, which reinforced the conclusion that pe_pgrs20 and pe_pgrs47 inhibited canonical autophagy but not LAP.
Inhibition of autophagy by direct expression of PE_PGRS proteins in macrophages. To probe the mechanisms by which PE_PGRS20 and PE_PGRS47 inhibited autophagy, we generated RAW 264.7 macrophage lines with an inducible cytosolic expression of these mycobacterial proteins. RAW 264.7 macrophages were stably transduced using a lentiviral transfection system with tetracycline-inducible pe_pgrs20 and pe_pgrs47 genes. Both PE_PGRS20 and PE_PGRS47 were expressed at the expected full-length size following anhydrous tetracycline (aTCN) induction (Fig. 3A). Infection of RAW macrophages induced to express PE_PGRS20 and PE_PGRS47 or empty vector control were infected with M. tuberculosis and respective DPE_PGRS knockouts. Cells were lysed 24 h after infection for immunoblot assay. Immunoblots of LC3B-II accumulation in macrophages expressing PE_PGRS20 or PE_PGRS47 and infected at MOI of 10 followed by harvest at 24 h postinfection are shown. (C) Summary densitometric analysis was calculated by LC3B-II normalized to GAPDH, and then the fold change ratio was calculated compared to uninfected control for each assay. Mean 6 SD of 3 independent assays is shown. Significance was calculated by two-way ANOVA corrected by Dunnett's test for multiple comparisons. (D) Survival of M. tuberculosis and respective DPE_PGRS knockout mutants was determined in RAW 264.7 macrophages expressing PE_PGRS20 or PE_PGR47 (MOI of 10) 24 h postinfection. Mean 6 SD of 3 independent assays shown, and significance was calculated by two-way ANOVA corrected by Dunnett's test for multiple comparisons. (E) Immunoblotting of phospho-S6 (Ser235/236), phospho-4E-BP1 (Thr37/46), phospho-p70S6k (Thr389), phospho-Ulk1 (Ser757), and phospho-mTOR (Ser2448) accumulation in RAW 264.7 macrophages expressing PE_PGRS20 or PE_PGRS47 after 3 h Torin-1 or M. smegmatis (MOI of 1) treatment. (F) Summary densiometric analysis of the ratio between p-S6 and b-actin was calculated. Mean 6 SD of 3 independent assays shown. Significance calculated by two-way ANOVA corrected by Dunnett's test for multiple comparisons. *, P # 0.05; **, P # 0.01; ***, P # 0.001. 264.7 macrophages expressing either of the PE_PGRS proteins with M. tuberculosis or DPE_PGRS strains demonstrated no statistical difference in LC3B-II levels; however, infection of RAW 264.7 cells expressing empty vector control with the DPE_PGRS20 or DPE_PGRS47 strains showed a significantly higher LC3B-II level ( Fig. 3B and C) than wild-type infection. Consistent with our earlier findings, increased autophagy in macrophages infected with DPE_PGRS20 and DPE_PGRS47 mutants resulted in reduced intracellular bacteria 24 h postinfection; the expression of PE_PGRS20 or PE_PGRS47 in macrophages inhibited autophagy activation and subsequent intracellular mutant M. tuberculosis killing (Fig. 3D). Moreover, autophagy inhibition by PE_PGRS expression had no significant effects on autophagy induction and intracellular bacterial burden during wild-type M. tuberculosis infection, further confirming that M. tuberculosis actively inhibits autophagy induction.
We further demonstrated the direct expression of PE_PGRS20 and PE_PGRS47 to inhibit autophagy during potent autophagy induction using Torin-1 and M. smegmatis. With both of these stimuli, PE_PGRS20 or PE_PGRS47 expressing RAW 264.7 macrophages demonstrated reduced autophagy induction compared to the empty vector control RAW 264.7 macrophages. In previous studies, we demonstrated that PE/PPE proteins of M. tuberculosis, when expressed in M. smegmatis, inhibit autophagy by maintaining the activation of mTOR and downstream targets (18). Similar to these results, a significantly increased level of mTOR activation was observed, as measured by S6, 4EBP-1, p70S6 kinase, and Ulk1 phosphorylation during M. smegmatis infection of PE_PGRS expressing RAW 264.7 macrophages compared to the empty vector control macrophages ( Fig. 3E and F).
Interaction of PE_PGRS20 and PE_PGRS47 with Rab1A. Using our stably transduced macrophage cell lines expressing PE_PGRS proteins, we carried out immunoprecipitation assays to identify host cell interacting partners of PE_PGRS20 and PE_PGRS47. We conducted mass spectroscopy (MS) analysis of host proteins immunoprecipitated with hemagglutinin (HA)-tagged PE_PGRS proteins. Proteins identified in cells expressing PE_PGRS proteins, but not in empty vector control by immunoprecipitations, are summarized in Table 1. As our data showed that both PE_PGRS20 and PE_PGRS47 similarly inhibit autophagy, we focused on coprecipitated host proteins identified in both PE_PGRS20 and PE_PGRS47-expressing cells. Among these, we chose to further study Ras-related protein Rab-1A (30), galectin 3 (31), and the Ras GTPaseactivating-like protein (IQGAP1) (32) because of their known roles in autophagy. We excluded the nonspecific contaminant proteins found in previous immunoprecipitation studies (33). Western blot analysis of subsequent immunoprecipitations identified Galectin 3 and IQGAP in induced and uninduced samples, potentially suggesting nonspecific interactions (data not shown). However, Rab1A was confirmed by immunoprecipitation with anti-HA and anti-Rab1A as a specific interacting partner of PE_PGRS20 and PE_PGRS47 proteins ( Fig. 4A; Fig. S4).
To ensure that the PE_PGRS and Rab1A interaction was maintained under infection conditions, we employed the previously characterized recombinant M. smegmatis strains expressing HA-tagged PE_PGRS20 or PE_PGRS47 or PPE51 (18). Rab1A was silenced in RAW 264.7 macrophages by transfection with Rab1A small interfering RNA (siRNA). Rab1A knockdown and control macrophages were infected with recombinant M. smegmatis at an MOI of 10 for 3 h and assayed for colocalization of PE_PGRS and Rab1A by immunofluorescence. Increased colocalization of HA-tagged PE_PGRS proteins and Rab1A was observed during M. smegmatis infection of control scrambled siRNA-transfected cells compared to M. smegmatis empty vector or PPE51-expressing controls. In contrast, Rab1A siRNA-transfected macrophages infected with the recombinant M. smegmatis or uninfected macrophages did not demonstrate increased colocalization of Rab1A and PE_PGRS proteins ( Fig. 4B and C).
Our data support the view that inhibition of autophagy by PE_PGRS20 and PE_PGRS47 provides M. tuberculosis a survival advantage within phagocytic cells. To determine if this survival advantage depended on Rab1A, we examined LC3B-II accumulation and bacterial survival in Rab1A-silenced macrophages (Fig. 5A). Rab1Asilenced macrophages were infected with M. tuberculosis or DPE_PGRS strains at an MOI of 10 for 24 h. No difference in LC3B-II accumulation was observed in Rab1Asilenced cells infected with wild-type M. tuberculosis or the DPE_PGRS mutants ( Fig. 5B and C). At 24 h postinfection, no difference in bacterial survival was evident in Rab1Asilenced cells. In line with previous data, DPE_PGRS strains showed significantly decreased survival compared to the parental M. tuberculosis strain in scrambled siRNA control macrophages (Fig. 5D). As with Atg16L1 silencing, no significant effects on autophagy induction and intracellular bacterial burden during wild-type M. tuberculosis infection were observed in control or Rab1A-silenced cells, further showing that M. tuberculosis actively inhibits autophagy induction via Rab1A. As expected, Torin-1 and M. smegmatis did not increase autophagy when Rab1A was silenced in empty vector control RAW 264.7 macrophages (Fig. S5A). Similarly, PE_PGRS-expressing macrophages transfected with control siRNAs demonstrated less LC3B-II accumulation than empty vector-expressing macrophages when autophagy was induced (Fig. S5B).
Inhibition of antigen presentation and cytokine secretion by PE_PGRS proteins. Mycobacteria are known to induce host proinflammatory cytokines that can stimulate autophagic responses. Autophagy induction modulates the transcription, processing,  and secretion of many proinflammatory cytokines, resulting in a negative feedback regulation (34). The loss of bacterial autophagy inhibition has previously been shown to increase proinflammatory cytokine secretion (16,34,35). RAW 264.7 macrophages infected with DPE_PGRS mutants increased their secretion of interleukin 6 (IL-6), IL-1b, tumor necrosis factor alpha (TNF-a), and IL-12(p70) compared to M. tuberculosis wildtype infected cells, and this increased secretion of proinflammatory cytokines was inhibited by silencing of Rab1A (Fig. 6A). While enhanced proinflammatory cytokine secretion has been shown to be necessary for in vivo control of M. tuberculosis infection, MHC class II antigen presentation also confers host advantage against the mycobacteria (36,37). Autophagy has been identified as a route by which pathogenic antigens are processed and delivered to MHC class II molecules to present them to CD4-positive (CD4 1 ) T cells (38,39). To assess the ability of DPE_PGRS20 and DPE_PGRS47 to modulate MHC class II-mediated antigen presentation during infection, we infected primary murine bone marrowderived dendritic cells (BMDCs) and cocultured infected cells with T cell hybridomas specific for the MHC-II M. tuberculosis Ag85b epitope. These hybridomas showed enhanced IL-2 secretion during infection with DPE_PGRS mutants over M. tuberculosis Significance was calculated by one-way ANOVA and corrected by Dunnett's test for multiple comparisons. *, P # 0.05; **, P # 0.01; ***, P # 0.001. (Fig. 6B). As observed in previous studies with DPE_PGRS47, this established that the inhibition of autophagy by DPE_PGRS20 and DPE_PGRS47 leads to suppression of MHC class II antigen presentation without influencing the total surface level of MHC class II proteins of infected BMDCs (Fig. 6C).
In total, these data support an essential role for PE_PGRS20 and PE_PGRS47 interactions with the host protein Rab1A. Based on our findings, we propose a model for autophagy inhibition by these PE_PGRS proteins by interaction with Rab1A (illustrated schematically in Fig. 7). In this model, PE_PGRS20 and PE_PGRS47 are expressed and secreted by the wild-type M. tuberculosis. These proteins directly interact with Rab1A, found on the preautophagosome (also known as the phagophore), at the endoplasmic reticulum. This interaction between mycobacterial proteins and host Rab1A prevents translocation of the Ulk1 complex to the preautophagosome, which is the first step in autophagy initiation. Inhibition of autophagy by M. tuberculosis PE_PGRS proteins resulted in secondary effects, including reduced proinflammatory cytokine secretion and reduced MHC class II-restricted antigen presentation, which combine to provide the bacteria with immune evasion and survival advantages.

DISCUSSION
Research has revealed that autophagy is an essential host pathway for controlling intracellular bacteria. However, many pathogenic intracellular bacteria, including M. tuberculosis, have evolved various molecular mechanisms to evade and subvert this autophagy-dependent control (6,16,19,26,29,40,41). In the case of M. tuberculosis, a previous study identified the M. tuberculosis protein SapM as an inhibitor of autophagy flux in macrophages by interaction with Rab7 (10,42). Another mycobacterial effector, Eis, inhibits autophagy dependent on NADPH oxidase and mitochondrial reactive oxygen species. Induction of IL-10 by M. tuberculosis Eis protein also increases autophagy potential in macrophages during infection (16,17). Similar to inhibition of autophagy and autophagy flux, the M. tuberculosis protein CpsA inhibits LAP. Like Eis, CpsA inhibition of LAP is dependent on NADPH oxidase and reactive oxygen species (29). Thus, M. tuberculosis has multiple interactions to block or subvert the autophagy pathway, underscoring the importance of this pathogen to control host cell autophagy at multiple levels. From a high-throughput loss-of-function screening of an M. tuberculosis transposon mutant library, we previously identified 16 additional genes that contribute to autophagy inhibition. Six autophagy-inhibiting M. tuberculosis DNA regions encoded PE/PPE family proteins (18). The expression of these M. tuberculosis PE/PPE proteins in M. smegmatis confirmed that they indeed inhibit autophagy and increase intracellular bacterial persistence or replication. These effects were correlated with increased mTOR activity and also with decreased production of TNF-a and IL-1b. One of the identified PE/PPE genes from the screen was PE_PGRS47. We have previously demonstrated that mice infected with the DPE_PGRS47 deletion mutant showed enhanced autophagy induction, attenuated growth, and a significant increase in CD4 1 T cell responses against MHC class II presented antigens. These findings suggested a nonredundant role for PE_PGRS47 in modulating innate intracellular host responses and the generation of adaptive immune responses that are critical to the control of M. tuberculosis (6). However, the mechanism by which PE_PGRS47 or other PE_PGRS proteins mediate autophagy-regulating function(s) remains to be determined.
This study revealed that PE_PGRS20 and PE_PGRS47 regulate host autophagy by the same molecular mechanism. Immunoprecipitation and colocalization by microscopy of macrophages stably expressing PE_PGRS20 and PE_PGRS47 demonstrated that the host protein Rab1A interacts with these PE_PGRS proteins. Rab proteins are a superfamily of small GTPases that can recruit effector proteins, forming complexes. Many studies have established that these GTPase proteins play an essential role in autophagy (30,43,44). Several bacterial effectors target Rab proteins to inhibit autophagy during infection. One of the host proteins frequently targeted by these bacterial effectors is Rab1A, which is essential for recruiting the Ulk1 complex to the preautophagosome to initiate autophagy (45). Intracellular pathogens such as Legionella pneumophila, Listeria monocytogenes, Salmonella enterica serovar Typhimurium, Streptococcus pyogenes, and various species of Chlamydia and Brucella all secrete bacterial effectors that interact with or inhibit Rab1A activity and thus enhance bacterial survival (41,44,(46)(47)(48)(49)(50). For example, Salmonella effectors SseG and SseF bind to Rab1A, subsequently inhibiting Ulk1 activation, as evidenced by increased phosphorylation of Ulk1 and its downstream targets. Unlike SseG and SseF, SseK3 of S. Typhimurium modifies Rab1A by arginine GlcNAcylation (51). Interestingly, PE_PGRS20 and PE_PGRS47 share approximately 40% similarity to SseG and SseF regions of homology (52), suggesting that these proteins may have related functions.
We have previously observed that the expression of PE_PGRS proteins during mycobacterial infection of macrophages results in increased phosphorylation of mTOR (18). While the inhibition of Ulk1 recruitment by the interaction of Rab1A is established, how this mediates mTOR activation is unknown. mTORC1 impacts autophagosome biogenesis through phosphorylation of the autophagy regulatory complex formed by Ulk1 and its interacting proteins, the autophagy-related protein 13 (Atg13) and the focal adhesion kinase family interacting protein of 200 kDa (FIP200). Previous studies have demonstrated that translocation of the Ulk1 complex to the preautophagosome requires the loss of mTOR-dependent phosphorylation of Ulk1 (41,53,54). A possible explanation for the role of PE_PGRS proteins on mTOR regulation is provided by cancer studies. In those studies, Rab1A was shown to be a mTOR activator and an oncogene, frequently overexpressed in human cancer (55)(56)(57). Rab1A overexpression promotes mTORC1 signaling and oncogenic growth in an amino acid-and mTORC1-dependent manner. Conversely, Rab1A knockdown selectively attenuates the oncogenic growth of Rab1-overexpressing cancer cells (55,58). Thus, we hypothesize that the interaction of PE_PGRS proteins with Rab1A activates Rab1A GTP activity, resulting in mTOR activation and autophagy inhibition. In our study, Rab1A knockdown cells (siRab1A) infected with the wild-type M. tuberculosis or two complementing strains demonstrated a reduced mTOR activation (detected by p-S6) compared to the RablA-positive cells infected with the identical mycobacteria (Fig. S6), which suggests mTOR regulation by Rab1A. However, this result needs to be further confirmed to elucidate a positive correlation between Rab1A expression levels and that of different upstream or downstream mTOR targets during M. tuberculosis infection.
Silencing of Rab1A or Atg16L1 did not affect mycobacterial survival or autophagy induction during wild-type M. tuberculosis infection (Fig. 2 and 5), providing further evidence that M. tuberculosis actively inhibits autophagy, highlighting the importance of autophagy modulation for mycobacterial survival. However, the importance of autophagy as a cellular defense mechanism in response to M. tuberculosis infection in vivo (59). Despite numerous in vitro studies emphasizing a role for autophagy in macrophages during M. tuberculosis infection (4, 5), loss of genes essential for canonical autophagy does not correlate with susceptibility to M. tuberculosis infection in vivo (59). The apparent insignificance of autophagy for controlling M. tuberculosis in vivo is likely to reflect that M. tuberculosis encodes very efficient inhibitors of canonical autophagy. Further characterization of M. tuberculosis DPE_PGRS20 and DPE_PGRS47 mutants using conditional Atg gene or Rab1A knockout mice could help elucidate a role of autophagy during M. tuberculosis infection in vivo.
Autophagy is essential for innate defense and serves as a method to capture pathogens inside autophagosomes, with subsequent fusion with lysosomes (4-6). In this way, epitopes of pathogens and intracellular antigens can be delivered to MHC class II molecules (38,39). Therefore, it is not surprising that autophagic pathways promote MHC class II antigen presentation and may serve to alarm the adaptive immune system against pathogens. We observed increased antigen presentation in macrophages and dendritic cells (DCs) infected with DPE_PGRS20 and DPE_PGRS47, suggesting that both PE_PPE proteins restrict replication of intracellular bacteria and autophagic degradation of mycobacterial antigens for MHC class II presentation to the adaptive immune system. The DPE_PGRS47 mutant has previously been demonstrated to display attenuated growth and enhanced CD4 1 T cell responses in mouse infection models (18,60). Inhibiting antigen presentation enables M. tuberculosis to persist in early infection relatively undetected by the host immune system. Developing treatment options based upon mycobacterial autophagy inhibition would likely lead to increased bacterial clearance as well as improved long-term adaptive immune responses. These findings improve our understanding of autophagy's role during M. tuberculosis infection and provide further information about early events in innate immunity that are important for bacterial replication and disease progression.
Gene-specific DPE_PGRS knockouts were made by allelic exchange via specialized transduction as previously described (61). Knockouts were selected for hygromycin resistance and checked by PCR (primers cataloged in Table S1 in the supplemental material). For complementation, full-length PE/PPE genes from M. tuberculosis H37Rv were amplified by PCR and cloned in-frame into pMV261, an episomal M. tuberculosis vector with a 39 His 6 -hemagglutinin (HA) tag (6,62). The expression constructs were electroporated into M. smegmatis or M. tuberculosis DPE_PGRS strains and selected for kanamycin resistance. The expression of PE_PGRS proteins was confirmed by immunoblotting using an anti-HA antibody.
Antibodies and other reagents. Antibodies were purchased from Cell Signaling Technology unless indicated otherwise and are catalogued in Table S2. All reagents and media purchased from Sigma-Aldrich unless otherwise stated. Torin-1 and bafilomycin A1 were purchased from Cell Signaling Technology and dissolved in dimethyl sulfoxide (DMSO).
Cell culture. All murine macrophage assays were conducted in 12-well plates seeded at 2.5 Â 10 5 cells/well 16 h before infection or treatment. Mycobacteria were grown to an optical density at 600 nm (OD 600 ) of 0.6 to 0.8. Mycobacteria suspended in RPMIc or DMEMc were used to infect the cells at an MOI of 10 unless otherwise stated. Infection was carried out for 3 h, after which cells were washed 3 times with phosphate-buffered saline (PBS) and then treated with 50 mg/ml gentamicin in complete media for 1 h to kill extracellular bacteria. Assays were conducted in RPMIc or DMEMc with 20 mg/ml gentamicin. Cells were harvested at indicated time points in radioimmunoprecipitation assay (RIPA) buffer (150 mM NaCl, 1% NP-40 or Triton X-100, 0.5% sodium deoxycholate, 0.1% SDS, 50 mM Tris-HCl, pH 8.0, and 20 mM Tris-HCl, pH 7.5) for plating for intracellular survival or analysis of autophagy induction by immunoblotting. For CFU enumeration, lysates were serially diluted and plated on 7H10 with appropriate antibiotics. Macrophages treated with Torin-1 were washed three times with PBS and then incubated in DMEMc with 10 mM Torin-1 for 3 h. Macrophages treated with M. smegmatis were washed three times with PBS and incubated with media containing M. smegmatis at indicated MOI for 3 to 6 h.
Murine primary cell isolation. Bone marrow-derived primary cells were derived according to previously published methods (63). Briefly, marrow was flushed from tibias and femurs of 6-to 8-week-old C57BL/6J mice aseptically and cultured in non-tissue culture-treated serological plates in RPMIc supplemented with 100 U/ml penicillin and 100 mg/ml streptomycin (RPMIcAbx). For macrophage differentiation, cells were seeded in 100-mm plates at 2 Â 10 5 cells/ml and differentiated by the addition of 15% L929 fibroblast-conditioned media for 6 days, followed by feeding with fresh media every 2 days. On day 6, adherent cells were washed with ice-cold PBS and detached by incubation on ice for 20 to 30 min in ice-cold PBS. For dendritic cell differentiation, cells were seeded in 100-mm plates at 2 Â 10 5 cells/ml and differentiated by the addition of 20 ng/ml recombinant granulocyte-macrophage colony-stimulating factor (rGM-CSF) (BioLegend), with feeding every 3 days. On day 10, adherent cells were washed with ice-cold PBS and detached by incubation with ice-cold PBS containing 5 mM EDTA for 5 min.
RAW 264.7 macrophages were transduced using recombinant lentivirus. Lentivirus was generated using ViromerRed (OriGene Technologies, MD) according to the manufacturer's instructions in the human embryonic kidney (HEK) 293T cells. Briefly, expression constructs (pInducer or shRNA, 1.64 pmol), packaging (psPAX2, 1.3 pmol), and envelope (pMD2.G, 0.72 pmol) plasmids were added to ViromerRed and then added to cells. HEK293T cell culture supernatant was collected at 48 h posttransfection and polybrene added at 10 mg/ml. Recombinant lentivirus was added to RAW 264.7 macrophages and incubated for 24 h, after which media containing lentivirus was removed and fresh DMEMc added. At 48 h postaddition of lentivirus, macrophages were treated with 10 mg/ml puromycin or 400 mg/ml G418. Single puromycin-or G418-resistant cells were expanded and assayed for Atg16L1 expression or PE_PGRS expression after induction with 500 ng/ml anhydrous tetracycline (aTcn) for 16 h.
Rab1A knockdown. Rab1A was silenced in RAW 264.7 macrophages by transfection with Rab1A siRNA (Thermo Fisher Scientific). Rab1A siRNA or scrambled siRNA (Cell Signaling Technology) at a concentration of 30 nM was added to Lipofectamine 2000 (Invitrogen) as per the manufacturer's instructions and incubated for 10 min at room temperature. Lipidated siRNA was diluted 1:10 with Opti-MEM (Gibco) and added to RAW 264.7 macrophages. Before adding siRNA, RAW 264.7 macrophages grown in DMEMc were transferred to Opti-MEM media supplemented with 5% fetal bovine serum (FBS) and incubated for 3 h. Macrophages were incubated for 48 h and then seeded for infection as described above. The efficiency of Rab1A knockdown was assayed by immunoblot for each transfection.
Immunoblotting. Cellular protein was prepared in 1Â radioimmunoprecipitation assay (RIPA) buffer, and the protein concentration was determined by bicinchoninic acid (BCA) assay (Pierce). Aliquots of lysates containing 1 to 10 mg of protein were resolved on 12% SDS-PAGE gels at 180 V for 40 min. Proteins were transferred to 0.2 mm polyvinylidene difluoride (PVDF) using a Bio-Rad TransBlot Turbo at 2.5 amps and 25 V for 5 to 10 min depending on molecular weight. PVDF membranes were blocked in 5% nonfat dry milk in 1Â Tris-buffered saline (TBS) plus 0.1% Tween 20 (TBST) or OneBlock Western-CL blocking buffer (Genesee Scientific) for LC3B blots at room temperature for at least 1 h. Primary antibodies at 1:5,000 dilution were incubated overnight at 4°C in TBST. Anti-rabbit IgG-horseradish peroxidase (HRP) antibody (1:10,000) was added to membranes for 45 min in TBST. Proteins of interest were revealed using Clarity ECL (Bio-Rad) according to the manufacturer's instructions. Films were scanned, and densitometric analysis was conducted by ImageJ software (https://imagej.nih.gov). The protein of interest was normalized to b-actin or GAPDH (glyceraldehyde-3-phosphate dehydrogenase) loading control to calculate autophagy levels (6).
Confocal microscopy imaging. Macrophage assays were carried out as described above, with macrophages allowed to adhere to glass coverslips. At 24 h postinfection, macrophages were washed with PBS and fixed in 4% paraformaldehyde for 48 h before mounting coverslips. For antibody staining assays, macrophage assays were carried out as described above, with macrophages allowed to adhere to glass coverslips. Coverslips were fixed in 4% paraformaldehyde for 15 min and then permeabilized using 0.02% saponin in PBS with 5% bovine serum albumin (BSA) for 20 min. Coverslips were then washed and blocked in PBS with 5% BSA, 0.01% saponin, and 0.2% Tween 20 for 20 min. Macrophage-coated coverslips were incubated with anti-Rab1A (1:100) in blocking solution overnight at 4°C. Coverslips were washed three times in blocking solution and incubated for 1 h in Alexa Fluor 555-conjugated anti-rabbit secondary antibody (1:1,000) and Alexa Fluor 488-conjugated anti-HA (1:500). Coverslips were then mounted in ProLong Gold with DAPI (49,6diamidino-2-phenylindole; Cell Signaling Technology) at 4°C overnight. Images were taken on a Zeiss laser scanning microscope 880 and analyzed with ImageJ software.
Immunoprecipitation. RAW 264.7 macrophages induced to express PE_PGRS proteins were lysed in lysis buffer (20 mM Tris-HCl [pH 7.5], 150 mM NaCl, 0.2% Nonidet P-40, and cOmplete protease inhibitor [Roche]) and centrifuged at 12,000 Â g for 20 min. The supernatant was cleared using rabbit serum conjugated to magnetic beads for 16 h at 4°C. Cleared lysates were incubated with magnetic beads conjugated to anti-HA or anti-Rab1A antibodies for 5 h at 4°C. Beads were washed 3 times using lysis buffer. Bound proteins were analyzed by Western blotting.
ELISA. At indicated time points, cell culture supernatants were collected for cytokine analysis. BioLegend enzyme-linked immunosorbent assay (ELISA) Max Deluxe kits were used as per the manufacturer's instructions. Briefly, plates were coated in capture antibody for 16 h and blocked in blocking buffer for 1 h at room temperature. A total of 100 ml of standards and culture supernatant was added and incubated for 2 h at room temperature. The detection antibody was incubated for 1 h, followed by avidin-HRP for 30 min. Next, TMB (3,39,5,59-tetramethylbenzidine) substrate was incubated for 20 min, and sulfuric acid was added to stop the reaction. Absorbance was read at 450 nm, and background absorbance (570 nm) was subtracted.
Antigen presentation assay. Dendritic cells were seeded in 96-well plates at 5 Â 10 5 cells/ml and allowed to adhere overnight. Cells were infected as per macrophage infection as described above. After washing infected cells 3 times with PBS, T cell hybridomas specific for the I-Ab Ag85B epitope (33) were seeded at 5 Â 10 5 cells/ml on top of the infected dendritic cells and incubated for a further 21 h. Culture supernatants were collected, and assays for IL-2 secretion were performed by ELISA.
Statistics. GraphPad Prism 8 was used for all analyses. Analysis of variance (ANOVA) was used to determine significance with Dunnett correction for multiple comparisons. A P value of ,0.05 was considered to be significant.

SUPPLEMENTAL MATERIAL
Supplemental material is available online only.