Phagocytosis enhances lysosomal and bactericidal properties by activating the transcription factor TFEB

Summary

Macrophages internalize pathogens through phagocytosis, entrapping them into organelles called phagosomes. Phagosomes then fuse with lysosomes to mature into phagolysosomes, acquiring an acidic and hydrolytic lumen that kills the pathogens. During an ongoing infection, macrophages can internalize dozens of bacteria. Thus, we hypothesized that an initial round of phagocytosis might boost lysosome function and bactericidal ability to cope with subsequent rounds of phagocytosis. To test this hypothesis, we employed Fcγ receptor-mediated phagocytosis and endocytosis, which respectively internalize immunoglobulin G (IgG)-opsonized particles and polyvalent IgG immune complexes. We report that Fcγ receptor activation in macrophages enhanced lysosome-based proteolysis and killing of subsequently phagocytosed E. coli compared to naïve macrophages. Importantly, we show that Fcγ receptor activation caused nuclear translocation of TFEB, a transcription factor that boosts expression of lysosome genes. Indeed, Fc receptor activation was accompanied by increased expression of specific lysosomal proteins. Remarkably, TFEB silencing repressed the Fcγ receptor-mediated enhancements in degradation and bacterial killing. In addition, nuclear translocation of TFEB required phagosome completion and failed to occur in cells silenced for MCOLN1, a lysosomal Ca²⁺ channel, suggesting that lysosomal Ca²⁺ released during phagosome maturation activates TFEB. Finally, we demonstrated that non-opsonic phagocytosis of E. coli also enhanced lysosomal degradation in a TFEB-dependent manner suggesting that this phenomenon is not limited to Fcγ receptors. Overall, we show that macrophages become better killers after one round of phagocytosis and suggest that phagosomes and lysosomes are capable of bi-directional signaling.

Introduction

During phagocytosis, phagocytic cells like macrophages can engulf dozens of particles such as bacteria, fungi, polluting particles and apoptotic bodies to help clear an infection and maintain tissue homeostasis [1,2]. Phagocytosis is a receptor and ligand-mediated process that can broadly be grouped as non-opsonic and opsonic phagocytosis. The former occurs when phagocytic receptors engage ligands endogenous to the particle, while the latter ensues through host-derived ligands that adhere to the particle [1,2]. The best-studied example of opsonic phagocytosis employs Immunoglobulin G (IgG) antibodies that bind to specific epitopes displayed on the particle [3]. Fcγ receptors then recognize IgG-opsonized pathogens [4,5]. Regardless of route, ligand-receptor engagement elicits a complex signaling pathway that restructures the plasma membrane of the phagocyte to ultimately internalize and sequester the particle within a phagosome [1,2]. Phagosomes then fuse with lysosomes to mature into phagolysosomes, acquiring an acidic and hydrolytic lumen by acquiring the V-ATPase H⁺ pump and a multitude of hydrolytic enzymes – the engulfed particle is consequently destroyed [1,6,7]. Unfortunately, many human pathogens, including Mycobacteria, Listeria, and Salmonella, have evolved mechanisms to alter phagosome maturation and ensure their survival within host cells [8–11].

Phagosome maturation depends on a variety of lysosomal regulators including the Rab7 and Arl8b GTPases, and the PIKfyve lipid kinase, which synthesizes phosphatidylinositol-3,5-bisphosphate [PtdIns(3,5)P₂] [12–14]. In part, PIKfyve is required for phagosome-lysosome fusion by stimulating MCOLN1/TRPML1 (herein MCOLN1), a lysosomal Ca²⁺ channel that binds to PtdIns(3,5)P₂ to release lysosomal Ca²⁺ and trigger membrane fusion [15,16]. Indeed, silencing of MCOLN1 trapped lysosomes and phagosomes in a futile, docked step [16]. Interestingly, phagocytosis caused a prolonged increase in cytosolic Ca²⁺ that depended on MCOLN [16]. Given that Ca²⁺ is a versatile second messenger that controls many cellular functions [17], this suggests that lysosomal Ca²⁺ released during phagosome maturation may have additional functions other than triggering phagosome-lysosome fusion.

MCOLN1 is also required to activate the transcription factor EB (TFEB) in autophagy, a process during which cytosolic components are sequestered into autophagosomes that then fuse with lysosomes to degrade and release energy sources during starvation [18]. TFEB governs expression of the “Coordinated Lysosomal Expression And Regulatory” (CLEAR) gene network, which includes numerous lysosomal hydrolases, membrane proteins and acidification proteins [19–22]. Thus, activation of TFEB enhances lysosome gene expression to serve the increased catabolic demand during autophagy [19–22]. Additionally, TFEB and the bHLH-30 ortholog in C. elegans, was shown to enhance expression of host defense genes including cytokines, chemokines and anti-microbial peptides in response to bacterial infection, suggesting that the CLEAR network is not limited to lysosomal genes [23]. However, this study did not investigate changes to lysosomal activity in response to infection.

Given that both autophagosomes and phagosomes fuse with lysosomes, and that a single macrophage can internalize dozens of bacteria, we hypothesized that phagocytosis activates TFEB. Indeed, here we show that phagocytosis stimulates lysosome-based degradation and bacterial killing by activating TFEB. This represents a feedback mechanism between phagosomes and lysosomes that likely helps to resolve infections.

Results

Fcγ receptor-mediated signaling enhances proteolytic and bactericidal activities in macrophages

We first examined if stimulation of Fcγ receptors enhanced lysosome-based degradation in macrophages by treating primary bone marrow-derived and RAW macrophages with aggregated IgG (AIgG) immune complexes and measuring the fluorescence intensity of DQ-BSA. DQ-BSA is internalized by pinocytosis and accumulates in lysosomes, where its degradation is tracked by increased fluorescence [24]. To account for possible effects on pinocytic uptake that might confound our interpretation, we normalized DQ-BSA fluorescence against the fluorescence signal of co-endocytosed fluorescent dextran for each experiment. Strikingly, we observed a significant increase in the normalized DQ-BSA signal 4 h post-treatment with AIgG relative to control cells (Fig. 1A, B). We then predicted that this enhanced lysosome-based degradation would strengthen bacterial killing. To test this, we allowed macrophages to endocytose AIgG immune complexes or engulf IgG-opsonized beads, followed by phagocytosis of live E. coli 4 h post-Fcγ receptor stimulation. While there was no difference in the uptake of E. coli, we found that relative to non-stimulated macrophages, Fcγ receptor-activated macrophages killed E. coli more effectively (Fig. 1C, D). Similarly, macrophages that internalized IgG-opsonized E. coli were better at killing subsequently internalized E. coli expressing an antibiotic-resistance gene (Fig. 1D). Overall, our data show that engagement of Fcγ receptors, either by endocytosis or phagocytosis, increases lysosome-based degradation and bacterial killing.

Figure 1.

Fcγ receptor enhances the degradative and bactericidal capacity in macrophages. (A, B): Quantitation of lysosome-based proteolysis in bone marrow-derived macrophages (A) and RAW macrophages (B). Macrophages remained resting or were stimulated with AIgG for the indicated times, followed by co-endocytosis and chase of DQ-BSA and fluorescent-dextran into lysosomes. Using microscopy or flow cytometry, the fluorescence intensities for DQ-BSA and dextran per cell were quantified and the former normalized against the latter for each cell to account for possible differences in pinocytosis caused by AIgG. (C, D): Quantitation of bactericidal activity in primary mouse macrophages (C) and RAW macrophages (D). Macrophages were either untreated cells (control), stimulated with AIgG, or allowed to ingest IgG-opsonized beads (OB) or IgG-opsonized E. coli, followed by phagocytosis of live E. coli as described in Methods. Macrophages were then immediately lysed or allowed to mature their phagosomes before lysis to respectively estimate the rate of phagocytosis (Internalized) and killing (Survived). Data are shown as a normalized mean ± SEM for three to six independent experiments, where * indicates statistical significance (p<0.05) against control conditions using Student’s t-test (A,B) or ANOVA followed by Tukey’s post-hoc test (C,D).

Fcγ receptor engagement causes TFEB translocation into the nucleus

To explain the enhanced lysosome-based degradation and bacterial killing, we postulated that Fcγ receptor signaling stimulated TFEB [20,21]. Indeed, AIgG or IgG-beads increased the percent of cells with nuclear TFEB-GFP after 40 min (Fig. 2A,B), followed by a decrease at 12 h (Fig. 2B). These kinetics were similar for cells treated with torin1, an inhibitor of mTOR [22,25]. To measure the robustness of TFEB activation, we quantified the nuclear-to-cytosolic ratio of TFEB-GFP. AIgG and torin1 both increased this ratio significantly, though torin1 was more robust at 4 and 12 h (Fig. S1A).

Figure 2.

Fcγ receptor engagement causes TFEB translocation into the nucleus. (A) RAW cells were transfected with TFEB-GFP and were then exposed to vehicle (control), the mTOR inhibitor torin1 (TOR), aggregated IgG (AIgG), or IgG-opsonized beads (OB). Cells were then fixed and counter-stained with DAPI to identify the nucleus. Arrows denote beads. Scale bar, 10 μm. (B) Percent of cells with nuclear localization of TFEB-GFP after the indicated post-stimulation time. (C,D) The nuclear to cytoplasmic fluorescence intensity ratio of endogenous TFEB in RAW (C) and primary macrophages (D). Data are shown as mean ± SEM from three independent experiments, and statistically assessed as above (* indicates p<0.05).

To ensure that our observations were not an artifact of TFEB-GFP overexpression, we assessed the localization of endogenous TFEB. We obtained a weak diffused signal when cells were stained with anti-TFEB antibodies. To ascertain that this corresponded to endogenous TFEB and not solely non-specific background fluorescence, we silenced TFEB expression. Both Western blotting and immunofluorescence analyses showed that we could strongly silence TFEB relative to non-targeting oligonucleotides (Fig. S2). This confirms that the immunofluorescence signal, albeit weak, was partially derived from endogenous TFEB (Fig. S2). We then showed that AIgG immune complexes increased the nuclear-to-cytosolic signal of RAW cells stained for TFEB in non-targeting siRNA-treated cells, while little response was observed in cells silenced for TFEB (Fig. 2C, Fig. S2B). Similarly, we showed that primary macrophages exhibited an increase in nuclear-to-cytosolic endogenous TFEB after AIgG stimulation (Fig. 2D). Overall, these data suggest that FcγR engagement stimulates endogenous TFEB in macrophages. Interestingly, TFEB activation does not appear to be a general response to macrophage activation since treatment with interferon-γ (IFNγ), interleukin 12 (IL-12), interleukin 10 (IL-10), or interleukin 4 (IL-4) did not induce TFEB-GFP nuclear accumulation (Figure S3A). Moreover, Fcγ receptor engagement specifically acts on TFEB since NFAT-GFP remained cytosolic after AIgG stimulation (Fig. S3B, C).

Fcγ receptor-mediated signaling increases expression of specific TFEB target genes

To determine if TFEB nuclear accumulation after Fcγ receptor activation enhances lysosome gene expression, we quantified the relative expression of a select group of TFEB target genes [20]. We found that phagocytosis of IgG-opsonized beads caused a significant increase in expression of Ctsd and Atp6v1h mRNA, which respectively encode for the lysosomal protease cathepsin D (CTSD) and the H subunit (ATP6V1H) of the vacuolar-type proton ATPase (V-ATPase) (Fig. 3A). However, we did not observe a significant change in expression of Lamp1 and Mcoln1, which encode for the lysosomal-associated membrane protein 1 (LAMP1) and MCOLN1, respectively (Fig. 3A). Moreover, we also examined the relative expression of specific autophagy-related genes and found that the expression of Map1lc3, encoding LC3, was boosted by phagocytosis of IgG-opsonized beads, but not that of Atg5 and Becn1 [19,21]. This contrasts with mTOR inhibition using torin-1, which significantly stimulated expression of most genes tested (Fig. 3A). The enhanced transcription of Atp6v1h and Ctsd was accompanied by a significant increase in ATP6V1H and CTSD protein levels, while LAMP1 remained constant, for cells stimulated with IgG-beads (Fig. 3B). Notably, Fcγ receptor-stimulation failed to increase the protein levels of CTSD and ATP6V1H in TFEB-silenced cells (Fig. 3C). Moreover, TFEB-silencing did not alter the basal levels of CTSD, ATP6V1H or LAMP1 (Fig. 3C), which suggests that i) TFEB may not control basal lysosome biogenesis or ii) that residual TFEB may suffice to maintain basal lysosome function. Regardless, our results indicate that Fcγ receptor signaling does not appear to globally control the CLEAR gene network, but targets a subset of lysosomal enzymes. Consistent with this, we did not observe a difference in lysosome number between resting, torin1-treated or AIgG-treated cells (Fig. S3D).

Figure 3.

Fcγ receptor stimulation enhances gene and protein expression of specific lysosomal genes. (A) Relative mRNA expression of select lysosomal (left) and autophagy (right) genes measured by qRT-PCR in RAW macrophages that were untreated (control), exposed to torin1 (TOR) or IgG-opsonized beads (OB). (B) Protein expression of ATP6V1H, CTSD and LAMP1 in RAW macrophages treated as above and detected by Western blotting (left) and quantified relative to HSP60 (right panel). The 34 and 52 kDa bands in the CTSD blot respectively correspond to the mature and pro-CTSD isoforms and were combined for quantification. Data are shown as mean ± SEM from seven independent experiments for CTSD and ATP6V1H and three experiments for LAMP1, and statistically tested as above (* indicates p<0.05). (C) Quantification of protein expression of ATP6V1H, CTSD, and LAMP1 in RAW macrophages electroporated with non-targeting siRNA (NT siRNA) or TFEB siRNA and treated as indicated. Data are shown as mean ± SEM from four independent experiments and statistically tested as above (* indicates p<0.05).

TFEB is required for the Fcγ receptor-dependent augmentation in bacterial killing

We next postulated that TFEB was necessary for the enhanced lysosome-based degradation and bacterial killing after Fcγ receptor engagement. To test this hypothesis, we silenced TFEB expression in RAW macrophages and re-deployed the DQ-BSA degradation and bactericidal assays before and after Fcγ receptor stimulation. Cells electroporated with non-targeting siRNA oligonucleotides displayed superior DQ-BSA proteolysis in response to AIgG exposure relative to unstimulated cells (Fig. 4A, Sup. Fig. S4). Remarkably, non-stimulated and AIgG-stimulated cells had similar DQ-BSA degradation when silenced for TFEB expression using either pooled siRNA oligonucleotides or either one of two independent oligonucleotides against TFEB (Fig. 4A, , Sup. Fig. S4). Furthermore, cells silenced for TFEB expression were also impaired for the enhanced bacterial killing in response to AIgG exposure, while cells electroporated with non-targeting oligonucleotides preserved this phenomenon (Fig. 4B). As before, none of the treatments employed affected E. coli engulfment (Fig. 4B). Notably, basal proteolytic and bactericidal activity remained unchanged after TFEB silencing, possibly because i) residual TFEB is sufficient for this or ii) that TFEB does not control basal lysosome biogenesis. Overall, these data support the hypothesis that TFEB directly mediates the enhancement of lysosome proteolysis and bacterial killing in macrophages whose Fcγ receptors were activated.

Figure 4.

TFEB is necessary for Fcγ receptor-mediated enhancements in degradation and bactericidal activities in macrophages. (A) Quantitation of DQ-BSA proteolysis in control and TFEB-silenced cells. (B) Quantification of bacterial uptake and survival in control and TFEB-silenced cells. Shown is the normalized mean ± SEM from three independent experiments. All experimental and control conditions were statistically assessed as above (* indicates p<0.05).

Fcγ-receptor signaling is insufficient for TFEB nuclear translocation

We next examined whether TFEB nuclear translocation occurs upon Fcγ receptor activation but in the absence of uptake by employing frustrated phagocytosis [26]. RAW macrophages transfected with TFEB-GFP were parachuted onto BSA-coatd or IgG-coated coverslips and allowed to attach. Since the coverslip surface is too large, phagosomes cannot form leading to frustrated phagocytosis. Macrophages that attached to IgG-opsonized coverslips acquired phosphorylated Syk, a proximal signaling event downstream of Fcγ receptors (Fig. 5A; [27,28]). In comparison, there was no detectable phospho-Syk in cells parachuted onto BSA-coated coverslips (Fig. 5A). Notwithstanding the remarkable accumulation of phospho-Syk, frustrated phagocytosis did not elicit nuclear translocation of TFEB-GFP (Fig. 5B, C, Sup. Fig. S1B). These data suggest that activation of Fcγ receptors is not sufficient to trigger nuclear translocation of TFEB. We propose that TFEB activation may require internalization of activated Fcγ receptors and/or particle enclosure.

Figure 5.

Frustrated phagocytosis does not elicit nuclear translocation of TFEB. (A) Western blotting (left) and blot densitometry (right) showing that frustrated phagocytosis on IgG-coated surfaces elicits Syk phosphorylation, while cell attachment to surfaces coated with BSA do not (control). The quantification of total Syk and phospho-Syk was done against HSP60 and is illustrated as the normalized mean ± SEM from seven independent experiments. (B) Frustrated phagocytosis by RAW cells failed to induce nuclear localization of TFEB-GFP. TFEB-GFP transfected cells were parachuted onto BSA-only (control) or anti-BSA antibody-coated coverslips (FP). After attachment, cells were fixed and stained with DAPI and Alexa647-conjugated secondary antibodies to anti-BSA antibodies to visualize the nuclei and demonstrate the presence of opsonizing antibodies during frustrated phagocytosis. Torin1 was used as positive control to show that TFEB-GFP can enter the nucleus. Scale bar = 10 μm. (C) Quantitation of nuclear localization of TFEB-GFP. Shown is the mean percent of cells with nuclear TFEB-GFP ± SEM from three independent experiments. Conditions were statistically assessed as above (* indicates p<0.05).

TFEB nuclear accumulation is dependent on lysosomal calcium release

The above data suggest that some aspect of phagocytosis or phagosome maturation helps control TFEB activation. It is unlikely that this happens by repressing mTOR since phagocytosis increases the levels of phospho-S6K, a substrate of mTOR [29], (Sup. Fig. S5A). Instead, given that i) MCOLN1 releases lysosomal Ca²⁺ during autophagy to activate the phosphatase calcineurin that then de-phosphorylates and activates TFEB [18] and ii) that phagosome maturation triggers a prolonged increase in cytosolic Ca²⁺ in a MCOLN1-dependent manner [16], we postulated that MCOLN1 and Ca²⁺ may be required for phagocytosis-dependent control of TFEB. To assess this, we first treated cells with BAPTA-AM to chelate Ca²⁺ and showed that this significantly reduced TFEB-GFP nuclear accumulation after phagocytosis (Fig. 6A, B, Sup. Fig. S1). To show that the release of lysosomal Ca²⁺ was at least in part responsible for TFEB-GFP nuclear accumulation, we employed siRNA against MCOLN1. Indeed, while Fcγ receptor-mediated phagocytosis retained the ability of to induce TFEB nuclear accumulation after non-targeting siRNA oligonucleotide treatment, MCOLN1 silencing arrested TFEB-GFP nuclear accumulation in response to phagocytosis (Fig. 6C, D, Sup. Fig. S1). TFEB-GFP translocation was rescued in cells expressing human MCOLN, which is resistant to the murine silencing oligonucleotides (Sup. Fig. S5B,C). Overall, our data suggest that TFEB nuclear translocation depends on lysosomal Ca²⁺ released by MCOLN1, likely during phagolysosome biogenesis.

Figure 6.

MCOLN1 and Ca²⁺ are required for TFEB nuclear localization in response to phagocytosis. (A) RAW cells were transfected with TFEB-GFP and then exposed to vehicle (control) or BAPTA-AM, followed by phagocytosis of IgG-opsonized beads (OB). (B) RAW cells were electroporated with MCOLN1 siRNA or non-targeting (NT) siRNA oligonucleotides, followed by transfection with TFEB-GFP. They were then exposed to vehicle (control) or allowed to phagocytose IgG-opsonized beads (OB). For A and B, cells were fixed and counter-stained with DAPI to identify the nucleus. Arrows denote internalized beads. Scale bar =10 μm. (C,D). Percent of cells with nuclear localization of TFEB-GFP after the indicated stimulus. For B and D, shown are the percent mean ± SEM from three independent experiments. Conditions were statistically tested as before (* indicates p<0.05).

Non-opsonic phagocytosis of E. coli activates TFEB and enhances lysosomal degradation

We next pondered whether other phagocytic signals might activate TFEB to enhance lysosomal activity. To assay this, we exposed macrophages to non-opsonized E. coli. Since E. coli exhibit multiple endogenous ligands, then phagocytosis of E. coli likely proceeds through cooperative engagement of several receptors including lectins, scavenger receptors and TLR4 [2,30–32]. Indeed, while TFEB-GFP was predominantly cytosolic in resting macrophages, E. coli engulfment caused rapid translocation of TFEB into the nucleus (Fig. 7A, B). Importantly, macrophages that internalized unopsonized E. coli displayed a significant increase in lysosome-based proteolysis using the DQ-BSA assay relative to resting macrophages (Fig. 7C). Strikingly, cells electroporated with the non-targeting oligonucleotides, but not those silenced for TFEB, showed a robust enhancement in lysosome-based degradation after phagocytosis of E. coli (Fig. 7D). Overall, this suggests that phagocytosis of non-opsonized bacteria can also stimulate TFEB to enhance lysosome-based degradation.

Figure 7.

Phagocytosis of non-opsonized bacteria augments lysosomal-based degradation in a TFEB-dependent mechanism. (A) TFEB-GFP remains cytosolic in control cells (top) but is nuclear in macrophages that internalized unopsonized E. coli. Nuclei and E. coli were respectively visualized with DAPI and anti-E. coli antibodies as described in Methods. Scale bar = 10 μm. (B) Quantitation of TFEB-GFP nuclear localization in cells untreated or exposed to unopsonized E. coli. Data is shown as the percent mean ± SEM from three independent experiments. (C, D) Quantitation of DQ-BSA proteolysis in macrophages. Phagocytosis of E. coli enhanced DQ-BSA degradation relative to control cells over time (C). Similarly, cells electroporated with non-targeting siRNA (NT) oligonucleotides, but not TFEB-silenced cells, show enhanced DQ-BSA degradation in response to unopsonized E. coli. For C and D, data are shown as mean normalized fluorescence ± SEM from three independent experiments and was statistically analysed as above (* indicates p<0.05).

Discussion

Immune cells have evolved intricate molecular networks to sense and heighten their ability to resolve an infection. This is understood to occur through paracrine and autocrine signals governed by the release of cytokines that orchestrate inflammation and its resolution [33]. Cytokine production is tightly modulated at the transcriptional level by various transcription factors including the canonical NF-κB and NF-AT proteins [reviewed in [34,35]. However, much less is known about how transcriptional switches help cells adapt their organellar function in response to infection and other immune cues. Here, we show that Fcγ receptor engagement activates the transcription factor TFEB to boost lysosome-based degradation and bacterial killing. Thus, we have uncovered a novel phenomenon during which the uptake of IgG immune complexes trains macrophages to become better killers, in part by super-activating their lysosomes through a transcriptional circuit. Importantly, a similar process occurs in macrophages undertaking phagocytosis of non-opsonized E. coli, suggesting that this is not limited to Fcγ receptor signaling. We speculate that this “training” is important to help macrophages dispose of the dozens of pathogens that can be engulfed by a single macrophage, particularly during a persistent infection [36,37].

Our results further bolster recent reports that link TFEB to immune processes. First, TFEB expression toggles the ability of dendritic cells to mediate antigen cross-presentation. In particular, TFEB over-expression favours MHC-II-dependent antigen presentation by enhancing lysosome hydrolytic activity and acidification, while its suppression prioritizes MHC-I-antigen presentation by reducing lysosome activity and preserving exogenous antigens [38]. Secondly, Visvikies et al. elegantly showed that the TFEB ortholog in C. elegans, bHLH-30, is activated by infection with S. aureus and is responsible for up-regulating the vast majority of immune and cytoprotective genes in that model organism [23]. They also showed that S. aureus activated TFEB in RAW macrophages and was necessary for increased expression of a variety of immune-related genes [23]. It is thus evident that opsonins like IgG and endogenous ligands on S. aureus and E. coli can stimulate TFEB in macrophages. It will now be important to identify the range of immune signals that modulate TFEB, including other Fc receptors and Toll-like receptors.

Interestingly, we showed that Fcγ receptor engagement did not lead to a global up-regulation of lysosome and autophagy-related genes known to respond to TFEB stimulation [20,21]. Indeed, while mTOR inhibition increased five out of the seven genes tested, only Ctsd and Atp6v1h exhibited significant increases in expression and protein levels after Fcγ receptor engagement relative to control cells. This suggests that IgG immune complexes do not induce full-scale lysosome biogenesis to increase lysosome number, but rather may bolster the ability of existing lysosomes or serve to replace enzymes that might be lost during particle breakdown. Additionally, MITF and TFE3 also translocate to the nucleus upon Fcγ receptor activation (data not shown). Together, this suggests that TFEB and its related transcription factors may differentially modulate different sets of genes. This is consistent with the evidence that TFEB and related factors respond to diverse physiological cues including autophagy, lysosome stress, WNT signaling, S. aureus, E. coli and IgG immune complexes [19,21,23,39]. Given this diversity of inputs, one might expect different sets of signaling pathways to combine to allow TFEB to express subsets of genes in response to a specific input. For example, this could be achieved by differential phosphorylation of 12 potential sites on TFEB [40]. Future research, including transcriptomics, will be required in order to understand how TFEB can selectively target gene subsets under various conditions.

Finally, TFEB activation may require phagosome formation, Ca²⁺ and MCOLN1. With respect to the former, we expect that signals that arrest engulfment will interfere with TFEB activation. With respect to the latter, we speculate that Ca²⁺-released by MCOLN1 during phagosome maturation activates the protein phosphatase calcineurin, causing dephosphorylation at inhibitory sites of TFEB and ultimately its nuclear translocation. This model is consistent with two keys observations. First, we previously showed that MCOLN1 is necessary for phagosome-lysosome fusion, causing an apparent decrease in lysosomal Ca²⁺ stores and an increase in cytosolic Ca²⁺ levels [16]. Second, TFEB nuclear translocation during autophagic stress also required MCOLN1 and the consequent calcineurin stimulation [18]. It will be interesting to see if other forms of phagocytosis also induces TFEB activation through a similar mechanism and whether pathogens that interfere with phagosome-lysosome fusion such as Mycobacterium tuberculosis prevent TFEB stimulation [41]. Indeed, given that TFEB over-expression seems to help resolve some defects in lysosomal storage diseases [42,43], it will be critical to evaluate the therapeutic power of TFEB during infection. While our studies employ siRNA-based technology, future studies will benefit from CRISPR/Cas9-based gene manipulation and/or macrophage-specific gene deletion of TFEB.

Experimental Procedures

Antibodies

Rabbit polyclonal antibodies used were against ATP6V1H (GeneTex, CA), TFEB (Bethyl, TX), BSA (Sigma-Aldrich, MO), phospho-T³⁸⁹-S6K (Cell Signaling) and E. coli (AbD Serotec, NC). Rabbit monoclonal antibodies used were against CTSD (GeneTex, CA), SYK, phospho-Tyr^525/526-SYK, S6 kinase (49D7) and HSP60 (Cell Signaling, MA). A rat monoclonal antibody was used against LAMP1 (Developmental Studies Hybridoma Bank, University of Iowa, IA). HRP-conjugated, Alexa 647-conjugated, Alexa 546-conjugated and Alexa 488-conjugsted donkey polyclonal antibodies were also used (Cedarlane, Burlington, ON).

Cell culture, transfection and electroporation

RAW 264.7 mouse macrophages were cultured in DMEM supplemented with 5% (v/v) fetal bovine serum (FBS; Gibco, ON). Transfection of plasmids encoding GFP-fusion of TFEB [22] were carried out using PolyJet DNA transfection reagent (FroggaBio, ON) or FuGene HD transfection reagent (Promega, WI) following manufacturers’ instructions for 12-well plates. siRNA-mediated gene silencing was carried out using the Neon Electroporation System (Life Technologies, ON) following the manufacturer’s instructions. To silence TFEB, we used the ON-TARGETplus SMARTpool or individual oligonucleotides from the SMART pool against murine Tfeb (Oligo 1 is J-050607-17 and Oligo 2 is J-050607-20; GE Dharmacon, ON), while the non-targeting control cells were electroporated with ON-TARGETplus Non-targeting Pool oligonucleotides (GE Dharmacon, ON). Cells underwent siRNA electroporation twice, separated by 24 h, with a 24-h period following the second round of electroporation prior to use in assays. MCOLN1 silencing and rescue was carried out as previously described ([16]; see supplemental text).

Primary mouse macrophages were differentiated from bone marrow-derived monocytes from femurs and tibias dissected from euthanized mice (see supplemental text for details). Mice were used according to ethical protocols for research animals approved by Ryerson University and St. Michael’s Hospital Animal Care Committee.

Treatments

Aggregated IgG (AIgG) was prepared by heating human serum IgG (Sigma Aldrich, MO) at 62°C for 20 min and then centrifuged at 16,000 xg for 10 min. IgG-opsonized beads (OB) were generated by rotating 3.87 μm or 2.08 μm poly(styrene/divinylbenzene) beads (Bangs Laboratories, IN) with IgG from human serum for 20 min at room temperature, prior to 3× wash with PBS. Torin1 was applied at 100 nM for the period of time specified for each individual experiment. 1,2-bis(o-aminophenoxy)ethane-N,N,N′,N′-tetraacetic acid acetoxymethyl (BAPTA-AM) was applied at 10 μM for 40 min immediately after stimulation with AIgG or IgG-opsonized beads. E. coli phagocytosis was carried out by growing DH5α E. coli to log phase (OD₆₀₀ 0.6), then applying to macrophages at 1000 bacteria/macrophage for TFEB-GFP nuclear localization or at 50 bacteria/macrophage for the DQ-BSA assay. E. coli phagocytosis was synchronized through centrifugation of the plates at 400 xg for 5 min at room temperature.

Fluorescence staining and microscopy

TFEB-GFP transfected cells were left untreated or treated with torin1, AIgG, IgG-opsonized beads or E. coli. AIgG and IgG-opsonized beads were applied to cells for 20 min, followed by a chase (30 min to 24 h), while torin1 was applied continuously. After the indicated time, cells were fixed with 4% (v/v) paraformaldehyde, stained with 0.4 μg/ml 4′,6-diamidino-2-phenylindole (DAPI). For phagocytosis of live E. coli, TFEB-GFP expressing macrophages were allowed to internalize for 5 min after spinning bacteria onto the macrophage surface. Macrophages were then treated for 30 min with media containing antibiotics to kill external bacteria. Antibiotic containing media was then removed, and cells were incubated for the indicated time prior to fixation. E. coli were stained using an anti-E. coli primary antibody and an Alexa-647-conjugated secondary antibody. For endogenous TFEB localization, cells remained resting or were treated as described above with AIgG for 20 min followed by a 1.5 h chase or torin1 for 2 h continuously. Cells were then fixed, permeabilized with 0.1% (v/v) Triton-X100 in PBS, and TFEB was labeled with a rabbit anti-TFEB primary antibody (1:1000 dilution in 0.5% BSA in PBS) followed by an Alexa-546 conjugated donkey anti-rabbit secondary antibody (1:1000 dilution in 0.5% BSA in PBS). Cells were then DAPI stained prior to mounting. Imaging was done with either a Leica DM5000 B epifluorescence microscope (Leica, ON) or an Olympus IX83 Inverted Microscope linked to a Hamamatsu ORCA-Flash4.0 digital camera.

Images acquired with the Olympus system were then deconvolved using CellSens Dimension module (Olympus, ON) with the advanced maximum likelihood estimation algorithm (ADVMLE). Image analysis was carried out using ImageJ (NIH, MD) prior to contrast enhancement or application of pseudo-colour filters for presentation.

DQ-BSA Proteolysis

Cells were treated with AIgG or live E. coli continuously for indicated times prior to lysosome loading. Lysosome labelling was done by co-endocytosis of 2 μM fixable Alexa-647-conjugated 10 kDa dextran (Life Technologies, ON) and 10 μg/ml DQ Green BSA (Life Technologies, ON) for 15 min, followed by a 1 h chase in label-free media. Cells were then scraped from wells in PBS and analyzed for whole cell fluorescence using the BD FACSCalibur flow cytometer (BD Bioscience, CA), where 10,000 events were counted per sample per condition using the FITC (FL1) and Cy5 (FL4) channels. Background signal was determined with non-labelled cells. The DQ-BSA signal (FITC/FL1 channel) was then normalized against the Alexa-647-dextran signal (Cy5/FL4 channel). Alternatively, DQ-BSA fluorescence was analysed by microscopy using the Olympus system by stimulating cells with a 30-min pulse of AIgG followed by a 3.5-h chase. Cells were then loaded with DQ-BSA as above but using fixable Alexa-546-conjugated 10 kDa dextran (Life Technologies, ON) in place of Alexa-647 dextran. Following a chase period of 1-h, cells were fixed and imaged. Fluorescent intensities were analyzed using ImageJ by normalizing the DQ-BSA fluorescence signal against the Alexa-546 signal after background correction.

Quantitative real-time-polymerase chain reaction (qRT-PCR)

Total RNA was collected, using the GeneJet RNA Extraction Kit (ThermoScientific, ON), following stimulation with AIgG or IgG-opsonized particles for 20-min and a 3.5 h chase, or 4 h torin1-treatment. cDNA was generated using SuperScript VILO cDNA Synthesis Kit and Master Mix (Life Technologies, ON). qRT-PCR was carried out using the StepOnePlus Real-Time PCR System (Applied Biosystems, ON) utilizing TaqMan Fast Advanced Mastermix and the following TaqMan Gene Expression Assays: Abt1 (Mm00803824_m1), Lamp1 (Mm00495262_m1), Atp6v1h (Mm00505548_m1), Ctsd (Mm00515586_m1), Becn1 (Mm01265461_m1), Atg5 (Mm00504340_m1), Map1lc3 (Mm00458725_g1), Mcoln1 (Mm00522550_m1).

Western Blotting

After the respective treatments, cells were lysed using 2× Laemmli sample buffer, passed through a 27-gauge needle and boiled for 5 min at 95 °C. After a 2-min centrifugation at 13,000 xg, the samples were loaded onto a 10% SDS-polyacrylamide gels, and electrophoresed. To detect phospho-Syk and phospho-S6K, lysates contained phosphatase inhibitors (phosSTOP, Roche, Mississauga, ON). Samples were then transferred onto a PVDF membrane, followed by blocking and incubation with primary and secondary antibodies at recommended dilutions. Protein levels were visualized by enhanced chemiluminescence using a ChemiDoc XRS+ Imaging System (BioRad, ON). Band intensities were determined using Image Lab software, background-corrected and normalized to loading control proteins (BioRad, ON).

Bactericidal Colony Assay

Bacteria killing assays were carried out as previously described [16], with some alterations. Prior to phagocytosis of unopsonized E. coli, RAW or primary macrophages remained untreated, or were activated with AIgG or IgG-opsonized beads for 20 min, followed by a 3.5 h chase. For RAW cells treated with opsonized E. coli, DH5α E. coli was opsonized with an anti-E. coli rabbit polyclonal antibody at a dilution of 1:100 in PBS. Bacteria were then applied to macrophages at 50 bacteria/macrophage and synchronized through centrifugation. Non-adherent bacteria were removed through rinsing. Following 4 hours, macrophages were then allowed to phagocytose DH5α E. coli carrying an ampicillin-resistant plasmid. Each treatment was carried out in duplicate, with one replicate being lysed and plated immediately after gentamicin treatment, which was used to determine E. coli uptake. The remaining replicates were incubated for 5 h, to allow for bacterial digestion, prior to lysis and plating. Colonies were counted on each set of plates after an overnight incubation. Lysates from cells treated with opsonized E. coli were plated on LB-ampicilin plates to select for bacteria carrying the ampicillin-resistance plasmid.

Frustrated Phagocytosis

Glass coverslips were treated with 0.1 mg/ml poly-D-lysine (Sigma-Aldrich, MO) for 1 h, followed by treatment with 2.5% glutaraldehyde for 15 min. Coverslips were then washed with double distilled water, followed by washes with PBS, then treatment with 1 mg/ml BSA in PBS for 30 min. Coverslips were then washed with PBS, incubated in 0.1 M glycine in PBS for 2 h, then incubated with 5.8 μg/ml of anti-BSA antibody in PBS. RAW 264.7 cells previously transfected for TFEB-GFP were then scraped, applied to the coverslips and then the plates were spun at 400 xg for 5 min to synchronize adherence, followed by a chase of 1 h. Cells were then fixed and coverslips were stained with Alexa-647 secondary antibody to visualize anti-BSA antibodies. For Syk and phospho-Syk analysis, non-transfected cells were applied to coated coverslips as described above, followed by lysis and Western blotting as described above.

Statistical Analysis

All data was subject to statistical analysis using Student’s t-test or one-way ANOVA coupled to Tukey’s post-hoc test. Number of independent experiments performed is indicated in figure legends. For experiments using single-cell analysis, at least 100 cells were scored per condition per experiment.