The SapM phosphatase can arrest phagosome maturation in an ESX-1 independent manner in Mycobacterium tuberculosis and BCG

ABSTRACT

Mycobacterium tuberculosis (Mtb) is an intracellular pathogen that survives and grows in macrophages. A mechanism used by Mtb to achieve intracellular survival is to secrete effector molecules that arrest the normal process of phagosome maturation. Through phagosome maturation arrest (PMA), Mtb remains in an early phagosome and avoids delivery to degradative phagolysosomes. One PMA effector of Mtb is the secreted SapM phosphatase. Because the host target of SapM, phosphatidylinositol-3-phosphate (PI₃P), is located on the cytosolic face of the phagosome, SapM needs to not only be released by the mycobacteria but also travel out of the phagosome to carry out its function. To date, the only mechanism known for Mtb molecules to leave the phagosome is phagosome permeabilization by the ESX-1 secretion system. To understand this step of SapM function in PMA, we generated identical in-frame sapM mutants in both the attenuated Mycobacterium bovis bacille Calmette-Guérin (BCG) vaccine strain, which lacks the ESX-1 system, and Mtb. Characterization of these mutants demonstrated that SapM is required for PMA in BCG and Mtb. Further, by establishing a role for SapM in PMA in BCG, and subsequently in a Mtb mutant lacking the ESX-1 system, we demonstrated that the role of SapM does not require ESX-1. We further determined that ESX-2 or ESX-4 is also not required for SapM to function in PMA. These results indicate that SapM is a secreted effector of PMA in both BCG and Mtb, and that it can function independent of the known mechanism for Mtb molecules to leave the phagosome.

INTRODUCTION

Tuberculosis (TB) is a forgotten pandemic that continues to be a significant world health problem, with 1.6 million deaths from TB in 2021 (1). The pathogenesis of Mycobacterium tuberculosis (Mtb), the bacteria responsible for TB, depends on the ability of Mtb to survive and replicate in macrophages (2, 3). To survive in macrophages, Mtb must subvert the antimicrobial immune responses of macrophages. One such way that Mtb promotes its survival in macrophages is by arresting the normal process of phagosome maturation [hereafter called phagosome maturation arrest (PMA)]. As a result of PMA, Mtb avoids being delivered to acidic degradative phagolysosomes. Instead, Mtb remains in an early phagosome that is permissive for Mtb replication. Although it is central to Mtb pathogenesis, there remain significant gaps in the knowledge of how Mtb achieves PMA. This is partly due to PMA being a complex process that relies on multiple Mtb effector molecules, many with redundant effects (2, 3). The redundancy highlights the importance of PMA to Mtb intracellular survival. It also means that elimination of a single effector may not necessarily eliminate PMA. All of the Mtb PMA effector proteins that are currently known are secreted from the bacteria as a necessary step for reaching their host targets. Consistent with this finding, the Mtb-specialized SecA2 secretion system is required for PMA (4). Two SecA2-secreted proteins, SapM and PknG, contribute to PMA, but the data suggest the existence of at least one additional SecA2-secreted PMA effector (5).

SapM was the first secreted phosphatase discovered in Mtb (6). In vitro experiments with purified SapM demonstrate SapM is capable of dephosphorylating phosphotidylinositols, with the highest specificity for phosphatidylinositol-3-phosphate (PI₃P) and phosphatidylinositol-4,5-bisphosphate (PI (4, 5)P₂) (7, 8). PI₃P is a regulatory lipid with a key role in phagosome maturation. During phagosome maturation, PI₃P is generated on phagosomal membranes where it recruits PI₃P-binding proteins, such as early endosome antigen 1 (EEA1) and the hepatocyte growth factor-regulated tyrosine kinase substrate (Hrs), which promote downstream fusion events between early phagosomes and late endosomes/lysosomes (9–13). Consistent with the ability of SapM to dephosphorylate PI₃P and the role of PI₃P in phagosome maturation, in in vitro assays purified SapM inhibits fusion between phagosomes and late endosomes (7). PI₃P is located on the cytosolic-facing leaflet of the phagosome membrane (12, 14, 15). Therefore, for SapM to function in PMA, the SapM protein must not only be secreted from Mtb but also transit out of the Mtb-containing phagosome in order to reach its PI₃P target (16). In Mtb, the ESX-1 Type VII secretion system promotes phagosome permeabilization, which enables interactions between Mtb molecules and host cytosolic pathways (17–20). ESX-1 is also necessary for PMA (21). These properties have led to the idea that phagosome permeabilization by the ESX-1 system is the likely mechanism for effectors to reach the macrophage cytosol (2, 3). The mycobacterial lipid phthiocerol dimycocerosate (PDIM), which is also necessary for PMA, potentiates the membrane permeabilization effects of ESX-1 (22–26).

The early studies of SapM did not involve sapM mutants (6, 7); thus, a requirement for SapM during mycobacterial infection was not addressed until recently. In Mtb, ΔsapM mutants are reported to be defective in PMA (27, 28). However, a sapM::tn mutant in the live-attenuated TB vaccine Mycobacterium bovis BCG strain, which is being developed as a TB vaccine candidate, is reported to not exhibit a PMA defect (29, 30). One of a number of differences between BCG and Mtb is that BCG is missing the region of difference 1 (RD1) that is present in the Mtb genome and encompasses genes encoding the ESX-1 system (20, 31). Thus, the lack of a PMA defect for a BCG sapM mutant was interpreted to reflect BCG SapM being unable to reach its cytosolic target, due to the absence of ESX-1-mediated phagosome permeabilization (2, 29). However, there are differences in the nature of the Mtb and BCG sapM mutants studied and the methods used to monitor PMA (27, 28, 30). Therefore, to definitively address the issue of the required role of SapM in PMA as well as the role of the ESX-1 system in SapM function, we constructed identical in-frame ΔsapM mutants in Mtb and BCG and carried out side-by-side comparisons. Our results indicate that SapM is necessary for PMA in both Mtb and BCG and that SapM does not require the ESX-1 system to reach the macrophage cytosol to carry out its role in PMA. Our results also demonstrate that neither ESX-2 nor ESX-4 is required for PMA in the absence of ESX-1 (i.e., in BCG) arguing against the possibility that these other ESX systems provide a route for SapM to access the host cytoplasm.

MATERIALS AND METHODS

Bacterial strains and media conditions

In this study, we used the Mtb H37Rv strain and ΔsapM mutant generated in the H37Rv background (Table S1). Additionally, we used the BCG Pasteur strain and ΔsapM mutant generated in the BCG Pasteur background. Mtb and BCG strains were cultured using Middlebrook 7H9 (liquid) or 7H10 (agar) supplemented with 1× albumin dextrose saline (ADS), 0.5% glycerol, 0.05% Tween 80, and kanamycin (20 µg/mL) or hygromycin (50 µg/mL) when necessary (32).

ΔsapM mutant construction in Mtb and BCG by allelic exchange using a counterselectable suicide plasmid

Unmarked ΔsapM mutants in Mtb H37Rv and BCG Pasteur were generated by two-step allelic exchange using a suicide counterselectable vector (32). The ΔsapM allelic exchange plasmid, pSEB1, was created using the hygromycin-resistant suicide sacB counterselectable plasmid pMP62 (33). pMP62 was cut with SpeI and NheI, and the resulting 8.8 kb vector backbone fragment was used in a 3-part Gibson Assembly reaction with PCR products amplified from Mtb DNA upstream or downstream of sapM. The sapM upstream PCR product of 892 bp was amplified with PCR primer 5b [5′-gacgtatctagacacgtctgaaggctgaagtgctacttggagattc-3′] and primer 2 [5′-ttattggcggaccgcggagcatgccgggagtat-3′]. The sapM downstream PCR product of 751 bp was amplified with PCR primer 3 [5′-atgctccgcggtccgccaataaccgatatttggg-3′] and primer 4 [5′-ccatcccagctcggcaaggatcactagttgcccacctgcaaccagaagtc-3′]. The upstream and downstream PCR products were designed to have an overlapping region of sequence to promote 3-part Gibson assembly (NEB Gibson assembly kit) and primers 5b and 4 had extensions to anneal to vector pMP62. The resulting ΔsapM deletion is unmarked and encodes for a severely truncated SapM protein with the first four amino acids of SapM fused to the final nine amino acids of SapM (MLRG-PPITDIWGD).

pSEB1 was electroporated into Mtb H37Rv or BCG Pasteur, and the first recombination event was selected for on 7H10 plates supplemented with 50 µg/mL hygromycin. Following growth in 7H9 media without antibiotics, bacteria that underwent a second recombination event between the deletion and wild-type alleles of sapM were selected by their growth on 7H10 plates supplemented with 3% sucrose to select for bacteria that lost the integrating sacB and hygromycin resistance containing plasmid. Sucrose-resistant and hygromycin-sensitive colonies were screened by PCR to identify ΔsapM deletion mutant versus wild-type recombinants. ΔsapM mutants were confirmed by immunoblot analysis of whole cell lysates with anti-SapM antibodies. Complementation strains were generated by electroporation of corresponding plasmids into mutant strains (Table S2) as previously described (32).

ΔsapM/ΔeccD1 Mtb double mutant construction by phage transduction

An ESX-1 (ΔeccD1) mutant was generated in the ΔsapM Mtb background to yield a ΔsapM/ΔeccD1 Mtb double mutant. The double mutant was constructed by specialized phage transduction using a temperature sensitive mycobacteriophage phAE159 to deliver a hygromycin-marked ΔeccD1 allelic exchange substrate (Table S2), as described previously (4, 32). Mtb cultures were resuspended in MP phage buffer (50 mM Tris/HCl pH7.5, 150 mM NaCl, 10 mM MgSO₄, 2 mM CaCl₂) and high titer phage stock, prepared in Mycobacterium smegmatis, added to a desired multiplicity of infection (MOI) of 10 phages to bacteria. The cell-phage mixture was incubated at 37–39°C for 4–6 h and then plated onto 7H10 plates with hygromycin (50 µg/mL) to select for allelic exchange mutants. Hygromycin-resistant colonies were screened by PCR to confirm mutants.

Δesx2 and Δesx4 BCG mutant construction by phage transduction

esx-2 and esx-4 operons, from rv3884c to rv3895c and rv3342c to rv3452, respectively, were deleted in BCG Pasteur by specialized phage transduction (Table S2), as described above. To make the Δesx-2 and Δesx-4 operon allelic exchange substrates, PCR reactions were performed using Mtb genomic DNA as a template. Forward and reverse flanking regions of Mtb esx-2 and esx-4 operons (1 kb each) were PCR amplified using primers ESX-2–1_LL [5′-ttttttttgcataaattgttccagccgcgtgggtcgc-3′], ESX-2–1_LR [5′-ttttttttgcatttcttgcaatgccatgaaatagctcgggatgtctcactgaggtctctagccgcatatcggctagtgcgg-3′], ESX-2–1_RL [5′-ttttttttgcatagattgcctcctaatgcgttaagctcacgagtgtctggtctcgtagtctcaataccccctggggctgc-3′], ESX-2–1_RR [5′-ttttttttgcatcttttgcagtatggtcgtggcgcagggg-3′], ESX-4_LL [5′-ttttttttccataaattggatcaatgcccgggcgataccca-3′], ESX-4_LR [5′-ttttttttccatttcttgggaaagtcttactgccgtgttgatgtctcactgaggtctctttccgggccccggagcgc-3′], ESX-4_RL [5′-ttttttttccatagattggaactcgtataacatccgcagcgagtgtctggtctcgtagctcccacggtggcgcgctg-3′], and ESX-4_RR [5′-ttttttttccatcttttggcggcaccggctcgccagc-3′]. Amplicons were either digested with Van91I or BstAPI based on the restriction site included in the primers and cloned into pYUB1471 backbone followed by ligation into phAE159 as previously described (34). The phages were propagated to high titers in M. smegmatis and verified by amplifying the flanking regions by PCR followed by Sanger sequencing. The Δesx2 and Δesx4 BCG mutant strains were confirmed by PCR.

Culture supernatant preparation

For preparing short-term culture supernatants from Mtb and BCG, cultures were first grown to log-phase in supplemented 7H9 media. Cultures were then pelleted at 3,000 rpm and washed once in phosphate buffered saline (PBS) with 0.05% Tween 80 and then sub-cultured to a starting OD₆₀₀ of 0.25 in fresh supplemented 7H9 media with 0.05% Tween 80 for 2 days. Cultures were pelleted at 3,000 rpm and supernatants were double filtered with 0.2 µm filters to remove residual cells.

Secreted SapM phosphatase activity assay

Secreted SapM phosphatase activity was measured as previously described (5, 35). Briefly, Mtb and BCG cultures were grown to mid-log phase and supernatants were normalized by culture OD₆₀₀. Normalized filtered supernatants (160 µL) were added in triplicate to a 96-well plate with 20 µL 0.1 mM Tris base pH 6.8 and 50 mM p-nitrophenyl phosphate (pNPP) and 20 µL 2 mM sodium tartrate for a final volume of 200 µL. SapM enzymatic activity is not inhibited by tartrate (6). Therefore, sodium tartrate was added to each sample to inhibit tartrate-sensitive phosphatases. Samples were incubated at 37°C, and the absorbance of the reactions was measured at OD₄₀₅ every 3 min for 7 h to calculate the rate of pNPP conversion, which was normalized to the amount of SapM secreted by wild-type Mtb or BCG.

Macrophage infections

Bone-marrow derived macrophages (BMDMs) were isolated from femurs of C57/Bl6 mice (approved by the University of North Carolina IACUC) by flushing with complete Dulbecco’s modified Eagle medium (DMEM) supplemented with 10% fetal bovine serum, 5 mM non-essential amino acids, and 5 mM L-glutamine. Cells were then washed, resuspended, and plated in complete DMEM supplemented with 20% L-929 conditioned media (LCM) without antibiotics and incubated at 37°C with 5% CO₂ for 6 days. Macrophages were then harvested using cold 5 mM EDTA in PBS and seeded in complete DMEM supplemented with 10% LCM without antibiotics. For monitoring intracellular growth of Mtb or BCG, macrophages were seeded at 2 × 10⁵ macrophages/well in 8-well chambered slides. For LysoTracker microscopy experiments, macrophages were seeded at 1 × 10⁵ macrophages/well in 8-well chambered cover glass. After 24 h, the macrophages were infected. For Mtb, macrophages were infected at MOI 0.1 for intracellular growth kinetics experiments or MOI 1 for LysoTracker staining experiments. For BCG, macrophages were infected at an MOI 1 for both growth kinetics experiments and LysoTracker staining. Before infection, the bacterial strains were grown to mid-log phase, washed twice using PBS containing 0.05% Tween 80, and then diluted in complete DMEM supplemented with 10% LCM. Macrophages were infected for 4 h and then washed three times with PBS to remove extracellular bacteria. For growth kinetics experiments, media were removed, cells were lysed using 0.1% Triton X-100 at various time points, and macrophage lysates were diluted and plated on supplemented 7H10 agar for intracellular CFU enumeration. All data shown are representative of a minimum of two independent experiments.

LysoTracker staining and co-localization

For LysoTracker staining, after 4 h of infection followed by washing the infected macrophage monolayer to remove extracellular bacteria, DMEM complete media containing 100 nM LysoTracker Red DND99 (Invitrogen) was added to each well and incubated at 37°C with 5% CO₂ for 1 h. The media were then removed, and the chambered cover glasses were submerged and fixed in 4% paraformaldehyde (PFA) for 1 h at room temperature. The fixed slides were then washed with PBS, and Fluoromount-G (Southern Biotech) was added to each well to protect the fluorescent signal. Widefield microscopy was performed using an Olympus IX-81 and Metamorph software. Images were taken using a 40× oil objective and the same exposure time was used across all images from a single experiment. To visualize the Mtb or BCG, we used one of two methods. Either we took advantage of the endogenous autofluorescence of mycobacteria, using a cyan fluorescent protein (CFP) filter cube (4) or we stained the bacteria with DMN (4 N,N-dimethylamino-1,8-naphthalimide)-trehalose (DMN-Tre) obtained from OliLux Biosciences Inc. (36). For DMN-Tre staining, DMN-Tre was added to mycobacteria-infected macrophages after the initial 4 h infection and washes, stained for 24 h, and subsequently stained with LysoTracker for 1 h as previously described (37). A minimum of 10 images per well containing 250 mycobacteria-containing phagosomes were scored for LysoTracker co-localization. Two replicate wells (minimum of 500 phagosomes) were analyzed for each experiment. All data shown are representative of a minimum of two independent experiments.

SapM antiserum production

To generate polyclonal antibodies against SapM, a peptide corresponding to amino acids 286–299 of SapM was synthesized. This peptide was injected into a specific pathogen-free New Zealand white rabbits using TiterMax Gold adjuvant, at Pierce Custom Antibody Services (Rockton, IL, USA). Antibodies were tested against wild-type Mtb and the ΔsapM mutant for specificity and purified by affinity chromatography.

Immunoblots

10–25 mL of bacterial cultures were pelleted and resuspended in extraction buffer (50 mM Tris-HCl pH 7.5, 5 mM EDTA, and 6% SDS) with protease inhibitor cocktail (Roche; 1 mg/mL Aprotinin, 1 mg/mL E-64, 1 mg/mL Leupeptin, 50 mg/mL Pefabloc SC, and 1 mg/mL Pepstatin A). Bacterial whole cell lysates were made by bead beating twice for 90 s with <106 µm glass beads using a MiniBeadbeater-16. Bacterial whole cell lysates, normalized by OD₆₀₀ to load equal protein, were separated by SDS-PAGE and transferred to nitrocellulose membranes. Membranes were blocked and subsequently proteins were detected using anti-rabbit primary antibodies specific for SapM (1:10,000) or SatS (1:10,000) (35). An α-Rabbit IgG-conjugated horseradish peroxidase secondary antibody (Bio-Rad) was used, and the signal was detected using Western Lightning Plus-ECL chemiluminescent detection reagent (Perkin-Elmer). Alternatively, membranes were incubated with the α-Rabbit IRDye 800CW (LI-COR) and imaged using a LI-COR imager.

Whole genome sequencing (WGS) of Mtb and BCG strains

WGS was used to verify that all strains used in this project are free of spontaneous mutations in PDIM lipid synthesis genes. 10 mL of Mtb or BCG cultures was grown to saturation and pelleted for isolation of genomic DNA as previously described (38). Genomic DNAs were sent for WGS at the Microbial Genome Sequencing Center (MiGS) (seqcenter.com), and sequences were analyzed using Breseq. The reference genome used for Mtb was NCBI RefSeq assembly GCF_000195955.2 and the reference genome used for BCG was GenBank: AM408590.1.

Statistical analyses

For comparisons between groups for secreted SapM phosphatase activity and LysoTracker co-localization experiments, one-way analysis of variance (ANOVA) with the Tukey’s multiple comparisons test was used. For intracellular growth experiments over time in macrophages, two-way ANOVA with the Tukey’s multiple comparisons test was used. For the statistical analysis and graph generation, Prism 9 software (GraphPad Software Inc., CA, USA) was used. All data shown are representative of a minimum of two independent experiments.

RESULTS

sapM mutants of Mtb and BCG lack secreted SapM phosphatase activity

To study the role of SapM in Mtb PMA and intracellular growth in macrophages, we generated an in-frame, unmarked ΔsapM deletion mutant in the virulent H37Rv Mtb strain by two-step allelic exchange. In parallel, we generated an identical ΔsapM mutant in the M. bovis BCG Pasteur strain to enable comparison of the role of SapM in virulent Mtb and in the attenuated BCG vaccine. Successful deletion of sapM in both backgrounds was confirmed by PCR. Further validation of the ΔsapM mutants was carried out by immunoblot analysis on whole cell lysates using a rabbit polyclonal antibody raised against the last 14 amino acids of SapM (Fig. 1A), which demonstrated absence of SapM in the ΔsapM mutants compared to wild type. In an operon and immediately downstream of sapM is the gene satS, which encodes a chaperone protein that stabilizes and promotes secretion of a subset of proteins by the SecA2 pathway, including SapM. SatS is also required for Mtb growth in macrophages (35). Thus, it was important to additionally confirm that the ΔsapM mutation did not disrupt satS. For this reason, whole cell lysates of ΔsapM mutants were additionally evaluated for SatS protein, using anti-SatS antibodies. Immunoblot analysis revealed SatS was present at wild-type levels in both ΔsapM Mtb and BCG mutants (Fig. 1A), confirming the sapM deletion did not disrupt downstream satS expression.

FIG 1.

ΔsapM mutants in Mtb and BCG lack secreted phosphatase activity. (A) Equal protein from whole-cell lysates prepared from wild-type Mtb or BCG and corresponding sapM mutants were examined for SapM or SatS protein by immunoblot with anti-SapM or anti-SatS antibodies. ΔsapM mutants in Mtb (B) or BCG (C) and their corresponding parental strains were grown in supplemented 7H9 media with 0.05% Tween 80 over 10–12 days. Optical density was used as a measure of in vitro growth. (D and E) Secreted phosphatase activity in culture supernatants from wild-type Mtb or BCG, the corresponding ΔsapM mutant, and complemented strains containing an empty vector (EV) or plasmid-expressing wild-type SapM were assessed using pNPP as a substrate, in triplicate. Activity was normalized to wild-type Mtb or BCG levels. Data shown are representative of at least two independent experiments. *P < 0.05, **P < 0.001 by ANOVA with Tukey’s post hoc test when compared to wild-type Mtb or BCG. WT, wild type.

The ΔsapM Mtb and BCG mutants grew equally well as the wild-type parental strains in vitro in Middlebrook 7H9 media. Thus, at least under these conditions, SapM is not required for in vitro growth (Fig. 1B and C). We next examined filtered short-term culture supernatants prepared from wild type, ΔsapM mutant, and complemented strains for secreted SapM-specific phosphatase activity using pNPP as a substrate. Because SapM phosphatase activity is resistant to sodium tartrate (6), sodium tartrate was included in the phosphatase assay to inhibit other secreted phosphatases and focus on SapM-specific activity. Both wild-type Mtb and BCG culture supernatants exhibited secreted phosphatase activity (Fig. 1D and E), while ΔsapM Mtb and BCG mutants exhibited a significant decrease in secreted phosphatase activity, as expected. The ΔsapM phosphatase defect could be complemented back by overexpressing wild-type sapM (ΔsapM+psapM) in both Mtb and BCG ΔsapM mutants (Fig. 1D and E). These data demonstrated that the Mtb and BCG ΔsapM mutants constructed in this study lack secreted SapM tartrate-resistant phosphatase activity.

SapM of Mtb is required for PMA and growth in a macrophage

Biochemical analysis of SapM demonstrates phosphatase activity on PI₃P (7, 8 ). Because PI₃P drives phagosome maturation events, the PI₃P dephosphorylation activity of SapM suggests a role of SapM in arresting phagosome maturation. Using our ΔsapM mutant of Mtb, we directly addressed the requirement of SapM for PMA by Mtb. Murine BMDMs were infected with wild type, ΔsapM mutant, or complemented strains of Mtb. To assess PMA in infected macrophages, we used fluorescent LysoTracker, a dye that accumulates in acidified compartments, and fluorescence microscopy to determine Mtb localization to mature acidified phagosomes. As expected, the majority of wild-type Mtb did not localize to acidified (LysoTracker positive) phagosomes demonstrating the ability of Mtb to avoid/arrest the normal process of phagosome maturation. Approximately 20% of intracellular wild-type Mtb was observed in acidic phagosomes. In comparison to wild-type Mtb, the ΔsapM Mtb mutant exhibited a significantly higher association with acidified phagosomes indicating a requirement of SapM for Mtb PMA (Fig. 2A and B). The increase in localization to acidified phagosomes could be partially complemented by expressing wild-type sapM in the ΔsapM Mtb mutant (ΔsapM+psapM).

FIG 2.

Mtb SapM arrests phagosome maturation and promotes growth in macrophages. BMDMs were infected with wiltype Mtb, ΔsapM mutant-containing EV plasmid or a complemented strain containing a plasmid expressing wild-type SapM. (A) The percentage of acidified mycobacteria-containing phagosomes at 1 h postinfection was determined by immunofluorescence microscopy using endogenous autofluorescence of mycobacteria and LysoTracker staining of quadruplicate wells. A minimum of 500 phagosomes were counted per condition. (B) Representative images used to quantify co-localization are shown. Mtb autofluorescence was re-colored green to highlight the co-localization in merged images. (C) BMDMs were infected at an MOI of 0.1 and intracellular growth was monitored over the course of 5 days. Triplicate wells were lysed and plated for CFU. Not significant (ns), *P < 0.05, **P < 0.001 by ANOVA with Tukey’s post hoc test when compared to wild-type Mtb or between two strains connected by a line when otherwise noted. Data shown are representative of at least two independent experiments. WT, wild type.

As PMA is important for Mtb intracellular growth in a macrophage, we next determined whether the deletion of sapM in Mtb attenuates growth in a macrophage. BMDM were infected with the same strains as above and macrophage lysates were plated for CFU over a 5-day time course. The ΔsapM Mtb mutant was significantly attenuated for growth at 5 days postinfection when compared to wild-type and complemented strains (Fig. 2C) indicating a requirement for SapM in intracellular growth. In addition to an effect on PMA, an effect on host cell death can impact the intracellular burden of Mtb. To address whether the growth defect of the ΔsapM mutant in macrophages might be attributed to host cell death, we performed a lactate dehydrogenase (LDH) release assay with macrophages infected with wild type, ΔsapM mutant, and the complemented strain. Under our experimental conditions, no cytotoxicity was observed during macrophage infection with any of these strains. Thus, the intracellular growth defect of the ΔsapM mutant is not due to a cell death effect and it is most likely the consequence of the PMA defect (Fig. S1).

SapM of BCG is required for PMA and growth in a macrophage

We next sought to determine whether SapM is required for BCG to arrest phagosome maturation. As with the Mtb experiments, BMDM were infected with wild type, ΔsapM, and complemented strains and PMA was determined by quantitating LysoTracker co-localization with BCG (Fig. 3A; Fig. S2). In comparison to Mtb, BCG was not as efficient at PMA. Wild-type BCG localized to more acidic compartments (approximately 40%) compared to wild-type Mtb (approximately 20%). This finding is consistent with previous reports that show BCG arrests phagosome maturation even though BCG is an attenuated strain of M. bovis (7, 16, 39). Nonetheless, even in the attenuated BCG strain background, a ΔsapM mutant exhibited a significantly higher association with acidic phagosomes compared to wild-type BCG (Fig. 3A; Fig. S2). This defect in PMA arrest was partially complemented in the ΔsapM + psapM BCG complementation strain. This result indicates that SapM is also required for PMA in BCG.

FIG 3.

BCG SapM arrests phagosome maturation and promotes growth in macrophages. BMDMs were infected with wiltype BCG, ΔsapM containing an EV plasmid or complemented strains containing a plasmid expressing wild-type SapM. (A) The percentage of acidified mycobacteria-containing phagosomes at 1 h postinfection was determined by immunofluorescence microscopy using endogenous autofluorescence of mycobacteria and LysoTracker staining of quadruplicate wells. A minimum of 500 phagosomes were counted per condition. (B) BMDMs were infected at an MOI of 1 and intracellular growth was monitored over the course of 5 days. Triplicate wells were lysed and plated for CFU. Not significant (ns), *P < 0.05, **P < 0.001 by ANOVA with Tukey’s post hoc test when compared to wild-type BCG or between two strains connected by a line when otherwise noted. Data shown are representative of at least two independent experiments. WT, wild type.

Next, we determined whether SapM was required for BCG growth in BMDMs. Wild-type BCG grew in macrophages over 5 days albeit to a lesser degree than observed with Mtb. Compared to wild type and the complemented ΔsapM BCG strain, the ΔsapM BCG mutant was significantly attenuated for intracellular growth (Fig. 3B).

Taken together, the results from studying ΔsapM Mtb and BCG mutants are significant in demonstrating a conserved role for SapM in both Mtb and BCG in PMA and intracellular growth in a macrophage.

The role of SapM in PMA can occur independent of the ESX-1 pathway

Because PI₃P, the target of SapM and a factor promoting downstream phagosome maturation events, is located on the cytosolic leaflet of the phagosome (12, 14, 15), SapM must move beyond the phagosome to reach PI₃P to play its role in PMA (16). The ESX-1 secretion system, which permeabilizes the phagosome membrane, is commonly assumed to be responsible for enabling secreted Mtb effectors to translocate from the Mtb-containing phagosome and carry out their functions on cytoplasmic targets of the macrophage (2, 3, 17, 20). However, because BCG lacks RD1, which encompasses genes encoding the ESX-1 system, a role for SapM in BCG suggests that SapM does not require ESX-1 to reach the macrophage cytoplasm (Fig. 3A). To determine whether SapM in Mtb can also act independent of ESX-1, we deleted eccD1, which encodes a critical component for a functioning ESX-1 system. We constructed an ΔsapM/ΔeccD1 Mtb double mutant and compared its ability to carry out PMA to that of wild-type Mtb, single ΔsapM Mtb, or single ΔeccD1 Mtb mutants in BMDM. As observed earlier, compared to wild-type Mtb, the single ΔsapM Mtb mutant exhibited a PMA defect (i.e., localized to significantly more acidic phagosomes) (Fig. 2A and 4). The single ΔeccD1 Mtb mutant also exhibited a defect in PMA. However, the PMA defect of the single ΔeccD1 Mtb mutant was significantly greater than that of the single ΔsapM Mtb mutant, which is indicative of role(s) of ESX1 in PMA that are distinct from delivering SapM to the macrophage cytoplasm. Finally, the ΔsapM/ΔeccD1 Mtb double mutant exhibited the most severe PMA defect compared to either the single ΔsapM or ΔeccD1 Mtb mutants (Fig. 4; Fig. S3). The finding that the double mutant defect is greater than either single mutant demonstrates that SapM and ESX-1 have unrelated roles in PMA (i.e., the role of ESX-1 in PMA must include roles other than delivering SapM to its cytoplasmic target). This result confirms that, as seen in BCG, SapM does not require ESX-1 to carry out its role in PMA. Introduction of wild-type sapM on psapM to the ΔsapM/ΔeccD1 mutant restored the ΔsapM/ΔeccD1 Mtb mutant phenotype back to the level of the single ΔeccD1 Mtb mutant, further demonstrating that SapM can perform its PMA function in the absence of ESX-1 (Fig. 4; Fig. S2). Together, our data in BCG and the ΔsapM/ΔeccD1 Mtb double mutant indicate that SapM can function independent of ESX-1.

FIG 4.

SapM can arrest phagosome maturation independent of ESX-1. BMDMs were infected with wild-type Mtb, ΔsapM, ΔeccD1, or ΔsapM/ΔeccD1 mutants containing an EV plasmid or complemented strains containing a plasmid expressing wild-type SapM. The percentage of acidified mycobacteria-containing phagosomes at 24 h postinfection was determined by immunofluorescence microscopy using DMN-Tre and LysoTracker staining of quadruplicate wells. A minimum of 500 phagosomes were counted per condition. Not significant (ns), *P < 0.05, **P < 0.001 by ANOVA with Tukey’s post hoc test when compared to wild-type Mtb or between two strains connected by a line when otherwise noted. Data shown are representative of at least two independent experiments. WT, wild type.

ESX-2 and ESX-4 are not required for the role of SapM in PMA

A recent study suggests that ESX-2 and ESX-4 work in concert with ESX-1 to promote permeabilization of the Mtb-containing phagosome membrane (40). As the SapM function in PMA can occur in an ESX-1-independent manner, we sought to determine whether ESX-2 or ESX-4 is responsible for SapM cytosolic access, as a step in PMA, in the absence of ESX-1. To do this, we generated esx2 and esx4 operon deletion mutants in BCG using a specialized phage transduction system. Successful mutants were confirmed by PCR. Because BCG lacks ESX-1, the resulting mutants were missing ESX-1 and ESX-2 or ESX-1 and ESX-4. We examined the ability of these mutants to block phagosome maturation compared to wild-type BCG and the ΔsapM BCG mutant with LysoTracker. As before, the ΔsapM BCG mutant localized to a significantly higher percentage of acidic phagosomes compared to wild-type BCG (Fig. 5A and B; Fig. S4). However, the Δesx2 BCG mutant showed no significant difference in PMA compared to wild-type BCG (Fig. 5A; Fig. S4). Similarly, macrophages infected with the Δesx4 BCG mutant did not reveal any significant difference in PMA compared to wild-type BCG (Fig. 5B). Together, these data suggest that neither ESX-2 nor ESX-4 alone is responsible for the ESX-1 independent translocation of SapM to the cytosol for its role in PMA.

FIG 5.

ESX-2 and ESX-4 are not required for SapM cytosolic access. BMDMs were infected with wild-type BCG, ΔsapM, Δesx2, or Δesx4 mutants containing EV plasmid or complemented strains containing a plasmid expressing wild-type SapM. For ESX-2 (A) or ESX-4 (B), the percentage of acidified mycobacteria-containing phagosomes at 24 h postinfection was determined by immunofluorescence microscopy using DMN-Tre and LysoTracker staining of quadruplicate wells. A minimum of 500 phagosomes were counted per condition. Not significant (ns), *P < 0.05, **P < 0.001 by ANOVA with Tukey’s post hoc test when compared to wild-type Mtb or between two strains connected by a line when otherwise noted. Data shown are representative of at least two independent experiments. WT, wild type.

DISCUSSION

Purified SapM is able to dephosphorylate phosphatidylinositol lipids in vitro, most notably PI₃P (6–8). Normally, PI₃P accumulates on phagosomes during phagosome maturation and phagosome-lysosome fusion (12). However, on Mtb and BCG-containing phagosomes, PI₃P levels are kept low (7). Initial studies by Vergne et al. (7) were the first to suggest a role for SapM as an effector molecule of PMA that dephoposphorylates PI₃P on the phagosome. It is noteworthy that wild-type BCG was used in these early studies. Later, published studies of ΔsapM mutants of Mtb supported a role for SapM in PMA. However, in BCG, which retains the ability to carry out PMA, although at a reduced efficiency (4, 41, 42), evaluation of a sapM::tn mutant did not lead to the same conclusion (i.e., a BCG sapM mutant was not defective in PMA) (29). BCG lacks the ESX-1 secretion system, which is required for PMA in Mtb and is often assumed to be the means by which Mtb effectors traffic from the phagosome. Therefore, the lack of an obvious role for SapM in BCG PMA was interpreted as reflecting an inability of BCG to transport SapM beyond the phagosome to act on PI₃P (2, 27–29) However, in the past studies of Mtb and BCG sapM mutants, the nature of the mutants and methods for measuring PMA differed, which could have impacted the outcome of assessing the role of SapM. Because sapM is immediately upstream and in an operon with satS, there was the potential for sapM mutations to have downstream effects on satS, which encodes a SecA2 chaperone that promotes secretion of SapM in addition to other SecA2-dependent proteins (35). For these reasons, here, we sought to definitively determine the role of SapM in PMA in Mtb and BCG in a side-by-side comparison using identical in-frame, unmarked sapM mutations that do not impact satS.

Our results with an in-frame Mtb ΔsapM deletion mutant affirmed a role for SapM in Mtb in PMA and growth in macrophages (Fig. 2A and C). However, in contrast to published studies of a sapM::tn transposon mutant in BCG (29), our evaluation of an in-frame ΔsapM mutation in BCG demonstrated a role for SapM in BCG PMA and growth in macrophages (Fig. 3A and B). Importantly, the PMA and intracellular growth defects of the ΔsapM mutants could be partially complemented by introduction of a plasmid expressing wild-type sapM (Fig. 2A and 3A). These results demonstrate that SapM has a conserved role in PMA and intracellular growth in Mtb and BCG.

The kinase that converts phosphatidylinositol to PI₃P and PI₃P itself is located on the cytosolic-facing leaflet of the phagosome membrane (12, 14). Additionally, SapM is recently reported to dephosphorylate Raptor, a member of the cytoplasmically localized mTORC1 complex (43–45). For SapM to act on these host cell targets, it will need to reach the cytosolic side of the phagosomal membrane. It is worth noting that our attempts to use immunofluorescence microscopy to study SapM localization in host cells, using different anti-SapM antibodies or tagged SapM proteins, were unsuccessful. Without such a type of direct visualization, we cannot completely rule out the possibility that SapM does not leave the phagosome to carry out its function in PMA. However, the established cytoplasmic nature of SapM substrates makes that possibility unlikely, which raises the question of how secreted SapM reaches the macrophage cytosol. The ESX-1 secretion system of Mtb is required for PMA (4, 21, 46) and, working with PDIM, it is known to permeabilize the phagosome membrane to allow interactions between Mtb products and host cytosolic molecules, including innate immune sensors (18, 19). Thus, it is generally assumed that secreted Mtb effector proteins will exit the phagosome with the assistance of the ESX-1 secretion system (2, 3, 17, 20). However, currently, there are only two Mtb proteins, Tuberculosis Necrotizing Toxin (TNT) and Mpt64, demonstrated to require ESX-1 for their delivery to the macrophage cytosol and neither protein has a known role in PMA (40, 47). In the study presented here, we provide evidence to indicate that SapM does not require ESX-1 to reach its host cell target and function in PMA. First, we demonstrated a role for SapM of BCG in PMA even though BCG lacks a functional ESX-1 system (Fig. 3A). Second, with a sapM complementation plasmid and an ΔsapM/ΔeccD1 double mutant in Mtb we demonstrated SapM can function in PMA in the absence of ESX-1 in Mtb (Fig. 4). Further, the ΔsapM/ΔeccD1 double mutant exhibited a more pronounced PMA defect than either an ΔeccD1 or ΔsapM single mutant (Fig. 4). The more severe phenotype of the ΔsapM/ΔeccD1 double mutant indicates that ESX-1 and SapM have independent functions in PMA, and it argues for a mechanism other than ESX-1 being able to deliver SapM to its host cell target for its role in PMA. While these results do not rule out the possibility that ESX-1 contributes to SapM release to the cytosol, the results provide strong evidence for SapM being able to reach and act on its host target in PMA in an ESX-1 independent manner. This conclusion is also supported by a study performed with Mycobacterium marinum, in which phosphatase overexpression (SapM, PtpA, and PtpB) is associated with reduced phagosome PI₃P levels in an esx-1 mutant background (48). The 19 kDa lipoprotein is another example of a mycobacterial protein reaching the cytoplasm in an ESX-1 independent manner as shown by immuno-gold labeling and electron microscopy of BCG infected cells (49). Thus, the idea that all mycobacterial effectors rely exclusively on ESX-1 to confer cytosolic access and activity on host targets needs to be reconsidered. Further, the role of ESX-1 in PMA remains to be understood as PMA effectors that do depend on ESX-1 for their activity remain to be identified (3). Interestingly, partial inhibition of phagosome acidification (i.e., maturation) is proposed to be a prerequisite for ESX-1-mediated phagosome permeabilization (50). This raises the interesting possibility that ESX-1 independent transport of SapM could promote initial reduction of phagosome acidification as a step toward ESX-1 permeabilization and robust PMA.

Recently, ΔeccC2 or ΔeccC4 mutants of Mtb, which encode ATPases essential for the ESX-2 and ESX-4 secretion systems, were shown to have phagosome permeabilization defects (40). Thus, suggesting that ESX-2 and ESX-4 systems work with ESX-1 to permeabilize the phagosome. Therefore, we explored the possibility that ESX-2 or ESX-4 secretion systems account for SapM gaining cytosolic access. To address this possibility, we constructed deletion mutants, in which the entire esx2 or esx4 operon sequences were deleted in BCG. Because of the absence of ESX-1 in BCG the resulting mutants can be considered Δesx1/Δesx2 or Δesx1/Δesx4 double mutants. The level of PMA observed in macrophages infected with these strains was no different than that was associated with the parental BCG strain. Thus, BCG does not require ESX-2 or ESX-4 for PMA thereby ruling out the possibility that these other ESX systems are responsible for the ESX-1 independent SapM delivery to its host cell target in PMA.

The question of how SapM gains cytosolic access to dephosphorylate PI₃P remains to be answered. One possibility is that the PDIM lipid, which enhances ESX-1-dependent phagosome rupture (26, 51), can act on its own (i.e., in the absence of ESX-1) to enable SapM transit from the phagosome. An alternate possibility is that SapM may enable its own transit from the phagosome. SapM belongs to the phospholipase C/phosphatase superfamily, which includes the Francisella tularensis acid phosphatase AcpA (8, 52). AcpA is involved in disrupting the phagosome membrane to allow F. tularensis cytosolic access (53). Thus, it is possible that the SapM protein may enable its own transit from the phagosome to allow it to act on phagosomal PI₃P. Alternatively, SapM may gain cytosolic access via an unknown mechanism that we have yet to discover.

In conclusion, here, we demonstrated that SapM is required for PMA and growth in macrophages in both Mtb and BCG. It is noteworthy that sapM mutants of BCG and Mtb are being studied for their potential to be used as live-attenuated TB vaccines (27, 29, 30, 54). Thus, the knowledge that a role for SapM in PMA is conserved in both species may help explain how sapM mutations improve antigen presentation and vaccine efficacy. We also demonstrated that the function of SapM in PMA does not require ESX-1, nor does it require ESX-2 or ESX-4. Thus, our data report on the existence of a mechanism other than ESX-1-mediated phagosome permeabilization as a way for Mtb effectors to gain cytosolic access and reach host targets.

SUPPLEMENTAL MATERIAL

The following material is available online at https://doi.org/10.1128/iai.00217-24.

Fig. S1. iai.00217-24-s0001.pdf.

LDH assay.

Fig. S2. iai.00217-24-s0002.pdf.

Representative images for sapM BCG mutants.

Fig. S3. iai.00217-24-s0003.pdf.

Representative images for eccD1 Mtb mutants.

Fig. S4. iai.00217-24-s0004.pdf.

Representative images for esx2 and esx4 Mtb mutants.

Supplemental methods. iai.00217-24-s0005.docx.

Method for Fig. S1.

Table S1. iai.00217-24-s0006.xlsx.

Strains.

Table S2. iai.00217-24-s0007.xlsx.

Plasmids and phages.

ASM does not own the copyrights to Supplemental Material that may be linked to, or accessed through, an article. The authors have granted ASM a non-exclusive, world-wide license to publish the Supplemental Material files. Please contact the corresponding author directly for reuse.