Introduction

Tuberculosis (TB), caused by the pathogen Mycobacterium tuberculosis (Mtb), continues to pose challenges to global communicable disease management and is one of the leading causes of mortality due to infectious diseases. The ability of the bacteria to alter and/or hijack various host cellular responses in its favour is one of the major factors contributing to the success of Mtb as a human pathogen. Arguably, understanding the host physiological processes hijacked or altered by Mtb also presents an opportunity to develop unconventional strategies to target and control the infection. In the past decade, several host-dependency factors have been identified, through large genome-wide screens, high-throughput transcriptome and systematic analysis of bacterial trajectory inside the host macrophages, the primary cells infected by Mtb (1).

The role of lysosomes in killing bacterial pathogens, including Mtb is well-known (2). Macrophages phagocytose bacterial pathogens, forming phagosomes, which mature from early to late phagosomes, eventually fusing with the lysosomes, where the pathogen is degraded (3). The phagosome maturation involves RAB5 (early phagosome) to RAB7 (late phagosome) exchange as a critical rate-limiting step (4). Mtb is known to block this RAB-conversion by excluding RAB7 recruitment, which helps them evade the lysosomal killing (5). Several mechanisms, including the role of lipid mediators and other signalling events, have been shown to contribute towards the phagosome maturation arrest (PMA) caused by Mtb (6).

Autophagy is another important anti-bacterial process that kills bacteria through a lysosomal targeting mechanism (7, 8). Autophagy performs crucial housekeeping functions like degradation of misfolded proteins or damaged organelles like mitochondria, peroxisomes, Endoplasmic Reticulum etc., in addition to its role in response to stresses like oxidative, nutritional or infection-induced stress (9). Understanding how autophagy machinery distinguishes and segregates its homeostatic cargos, as mentioned above, from those like pathogens is of considerable fundamental significance (10).

The cargos targeted for autophagic degradation are usually ubiquitinated, followed by their recognition by the autophagy adaptors, which interact with the ubiquitinated cargos and ATG8 or LC3B to capture the cargos into the growing autophagosomes (11). Autophagy adaptors are a family of proteins with ubiquitin-binding domains and LC3/ATG8 interaction domains. These adaptors primarily include p62/SQSTM1, NDP52, NBR1, OPTN and TAX1BP1 (12). Each of these adaptors is reported to have its roles defined in cargo selectivity. However, some reports suggest functional redundancy among them (13, 14).

While autophagic targeting of bacteria via xenophagy is part of the defence arm of autophagy, degradation of cellular cargos like mitochondria via mitophagy constitutes the homeostatic arm (10, 15). Mitophagy is a key pathway for mitochondrial quality control (MQC) through which damaged mitochondria are removed from the cell (16). Since mitochondrial damage is a normal consequence of respiration and other oxidative stresses, their quality control through mitophagy and other pathways is fundamental to cellular survival (17, 18).

In this study, while studying the effect of depletion of autophagy adaptor proteins like p62/SQSTM1 on intracellular Mtb survival, we unravel an unexpected MQC that facilitated Mtb targeting to the lysosomes by reversing the PMA. Our data provide mechanistic insights on the crosstalk between MQC and conventional host anti-bacterial pathways that are of both fundamental and biomedical relevance.

Results

Selective recruitment of autophagy adaptors to Mtb phagosomes and mitochondria

Since the autophagy adaptors target the autophagic cargo during selective autophagy, we investigated how the expression and recruitment of these adaptors to the phagosomes and mitochondria are affected upon Mtb infection. We first analysed for any perturbation in the expression of the major autophagy adaptors, such as p62/SQSTM1, OPTN, NBR1, NDP52 and TAX1BP1, in THP-1 macrophages upon infection with the Mtb strain, H37Rv (MOI 1:10, see methods). For all the infection experiments in this study, the time 0 hpi (hours post-infection) includes incubation with H37Rv for 4 hours followed by 2 hours of amikacin treatment to kill extracellular bacteria (detailed in the methods section). At the transcript level, OPTN and p62/SQSTM1 showed increased expression at 0-, 24- and 48-hpi (Fig. S1A). However, immunoblotting indicated higher expression of only p62/SQSTM1 proteins upon infection (Fig. S1B-C). To target the bacteria for autophagic degradation, autophagy adaptors need to get recruited onto the bacteria or bacterial phagosomes. Therefore, we evaluated the recruitment of autophagy adaptors on mCherry-H37Rv phagosomes by immunostaining and confocal microscopy at 0-, 6-, 12-, 24- and 48-hpi (see methods). While all of the five autophagy adaptors being studied were recruited onto the bacterial phagosomes (Fig.1A-E), we did observe some selectivity in the extent of colocalisation. For instance, OPTN showed the least recruitment to bacterial phagosomes, ranging from 2-4% colocalisation across all the time points (Fig. 1B), whereas p62/SQSTM1 showed maximum recruitment starting at ∼12% colocalisation at 0 hpi and increasing to ∼40% colocalisation at 48-hpi (Fig. 1A). The other three adaptors showed intermediate levels of recruitment: NBR1 5-20%, NDP52 10-15% and TAX1BP1 5-15% (Fig. 1C-E). Notably, only p62/SQSTM1 and NBR1 showed increased recruitment to the bacterial phagosomes over time as the infection progressed (Fig. 1A, 1C).

Fig. 1.

Since there are common autophagy adaptors involved in the autophagic targeting of bacteria (for xenophagy) as well as other cellular cargos like mitochondria (for mitophagy) (10, 19), we studied the distribution of these autophagy adaptors to a major cellular autophagic cargo such as mitochondria, during infection. For this, we analysed TOM20, an integral outer membrane mitochondrial protein that can be reliably used to track mitochondrial dynamics inside the cell (20). In uninfected as well as mCherry-H37Rv infected THP-1 cells, the recruitment of p62/SQSTM1 to TOM20⁺ mitochondria was significantly higher than any other autophagy adaptors (Fig. 1F). Upon H37Rv-infection, the pattern of autophagy adaptor recruitment on TOM20⁺ mitochondria mainly remained unchanged, except that there was a marginal increase in OPTN recruitment. Thus, no significant re-distribution of autophagy adaptors on mitochondria occurred during Mtb infection.

Role of autophagy adaptors in intracellular bacterial growth and survival

All five autophagy adaptors being studied were recruited to the bacteria and mitochondria in the macrophages, albeit differentially. To investigate their involvement in the bacterial targeting for autophagic degradation, we individually depleted the autophagy adaptors in the H37Rv-infected macrophages. Each one of the five autophagy adaptors (p62/SQSTM1, OPTN, NBR1, NDP52 and TAX1BP1) was knocked down in separate experiments using siRNAs (see methods) with 60-80% knockdown efficiency (Fig. S2A). While different autophagy adaptors showed differential recruitment to the phagosomes (Fig. 1), surprisingly, all except NDP52 knockdown resulted in a lower intracellular bacterial CFU (Fig. 2A).

Fig. 2.

Paradoxically, while the autophagy adaptors were expected to facilitate xenophagy, knocking them down reduced the bacterial CFU, implicating that these adaptors could be possibly helpful in Mtb growth in the host macrophages. Likewise, knocking down p62/SQSTM1 or OPTN in another human monocytic cell line U937 also reduced intracellular bacterial load (Fig. S2B).

To gain further mechanistic insights into the role of autophagy adaptors in governing bacterial growth and survival inside macrophages, we focused on p62/SQSTM1 since it showed the highest recruitment to both mCherry-H37Rv phagosomes and TOM20⁺ mitochondria. We targeted p62/SQSTM1 by CRISPR/Cas9-mediated gene knockout in THP-1 macrophages (see methods) using gRNA designed against exon 3 of the primary protein-coding transcript (see methods). We achieved ∼80% knockdown using p62/SQSTM1 gRNA and hCAS9 co-transfection followed by selection as assessed by immunoblotting (Fig. S2C) and immunofluorescence (Fig. S2D). We continued using these stable cells as p62 knockdown (p62^KD) cells and repeatedly analysed p62 expression to ensure efficient downregulation even after infection with Mtb (Fig. S2E). Knockdown of p62/SQSTM1 did not overtly alter the cellular autophagy response since no discernible alterations in LC3-II levels were observed in the absence or presence of bafilomycin A₁ (BafA₁) (Fig. S2F).

Likewise, p62/SQSTM1 knockdown did not affect the xenophagy response against H37Rv, as measured by colocalisation studies with LC3B (Fig. S2G). Similar to siRNA knockdown experiments, in p62^KDTHP-1 cells, bacterial CFU was consistently lower than in the WT cells (Fig. 2B). To confirm whether the reduction in bacterial CFU was due to slowing down of replication or enhanced killing in p62^KD cells, we used the replication clock plasmid (21). The replication clock plasmid (pBP10) is unstable and contains the Kanamycin resistance gene that gets lost from Mtb during active replication. Loss of Kan^R during assay is indicative of active replication. A time-course experiment with pBP10 containing H37Rv clearly showed that the gradual loss of Kan^r over time in both WT and p62^KD cells was indistinguishable, suggesting that the lower Mtb CFU observed in p62^KD cells compared to the WT cells is due to a loss of survival and not due to a slowing down of replication (Fig. 2C).

Furthermore, we knocked down p62 using siRNAs in human Monocyte-derived Macrophages (hMDMs) from three independent donors. In each case, we observed significant but varying degrees of decline in bacterial CFU in the knockdown cells (Fig. 2D). Thus, the absence of autophagy adaptor p62 appears to hamper Mtb growth and survival in THP-1 cells, U937 cells and also in the human primary macrophages, indicating it to be a general phenomenon in human macrophages.

In other animal models like zebrafish and mice, there are reports on p62/SQSTM1 and OPTN being host resistance factors against Mtb infection (22). We knocked down p62/SQSTM1 and OPTN in the mouse BMDMs and did not observe any significant difference in Mtb CFU (Fig. S2H-I). Similarly, in a HOXB8-derived p62^KO mouse monocytic cell line, there was no effect on Mtb CFU between WT and p62^KO cells (Fig S2J). These results were consistent with another report, where bacterial CFU load in the lungs of p62 knockout mice was not significantly different from the wild-type animals (23). Therefore, autophagy adaptors like p62/SQSTM1 and OPTN may have some distinct species-specific functions in mouse and human macrophages.

Perturbations in mitochondria-associated gene expression in p62^KD cells upon Mtb infection

To understand the role of autophagy adaptors, specifically p62/SQSTM1, in regulating bacterial survival, we sought to elucidate the underlying mechanism behind the bacterial killing in the p62^KD cells. In an unbiased approach, we performed gene expression analysis of WT and p62^KD cells under uninfected and infected conditions at 0-, 24- and 48- hpi (methods). Overall, both WT and p62^KDcells showed similar gene expression patterns across the three-time points and broadly similar expression patterns upon H37Rv infection.

However, the magnitude of regulation showed some differences between the WT and p62^KD cells at later time points (Fig. S3A, Table S1). This was reflected in the gene ontology analysis. In both WT and p62^KD cells, functional classes like a response to cytokine stimulus, inflammatory responses and response to LPS stimulation were equally upregulated (Fig. S3B). There were also some unique upregulated events in the p62^KD cells, like NFκB pathway upregulation at 0 hpi, chemokine response pathways at 24 hpi and inflammatory responses at 48 hpi (Fig. S3B). In the cellular component GO class, most regulated genes belonged to trafficking vesicles, Golgi apparatus and Mitochondria (Table S2). Since p62/QSTM1 was the central autophagy adaptor getting recruited at the mitochondria, we specifically examined mitochondrial genes (Table S3). While at 0 hpi, WT and p62^KD cells appeared almost identical in terms of gene expression, as the infection progressed, some differences emerged at later time points (24 and 48 hpi) (Fig. 2E-G). While genes like RP2, GSR and OXR1 were upregulated only in the WT cells at 24 hpi, several other genes like ETFBMKT, SLC25A19, SOD2, FTH1 and BNIP3 were upregulated only in the p62^KD cells (Fig. 2F). Increased BNIP3 expression in the p62^KD cells was exciting since it along with proteins like FUNDC1 and NIX constitutes ubiquitin independent autophagy adaptors and can directly tag damaged mitochondria for mitophagy-mediated degradation (24). Another interesting observation was the significantly enhanced expression of MGME1, a mitochondrial genome maintenance protein, in the p62^KD cells at 48 hpi that gets downregulated upon Mtb infection (Fig. 2G). Overall, the mitochondria-targeted genes showing differential regulation upon Mtb infection at later time points included functional classes like lipid metabolism, mitochondrial nucleotide metabolism, immune response, redox regulation, mitophagy, autophagy and trafficking (Fig. 2H-I, Table S4). Thus, p62^KDcells could exhibit a few differences in mitochondrial functioning or dynamics upon infection compared to the infected WT cells. In addition, p62^KDcells showed dysregulated cytokine expression upon H37Rv infection compared to the WT cells (Fig. S3B, Table S5). Some of these cytokines, which consistently showed increased expression in p62^KD cells compared to the WT cells, were IL1B, TNF, CXCL6 and CXCL2, suggesting a more inflammatory phenotype in the p62^KD cells upon H37Rv infection. The results above indicate that p62/SQSTM1 is involved in mitochondrial dynamics and controlling inflammatory pathways during infection.

Mitochondrial quality is maintained during H37Rv infection in p62/SQSTM1-depleted cells

The gene expression analysis suggested that the WT and p62^KD cells responded differently to Mtb infection, particularly for genes involved in mitochondrial dynamics and function. However, how that could undermine mycobacterial growth and survival in those cells is unclear. Given the role of autophagy adaptors, specifically p62/SQSTM1 in the process of mitophagy (25–28) and also the fact that it was the maximally recruited autophagy adaptor on the mitochondria (Fig. 1F), we measured the parameters for assessing the mitochondrial function and dynamics in the cell. MitoTracker green (MTG) staining showed that there was a marginal increase in the total mitochondrial content in the p62^KD cells upon H37Rv infection at early time points but not at the later time points (Fig. 3A). Moreover, there was no significant change in the overall status of mitochondrial depolarisation, as observed by TMRE staining at 0-, 24- and 48- hpi in p62^KD and WT cells (Fig. 3B). Next, we measured mitochondrial superoxide that indicates the presence of active mitochondria, using MitoSox dye. While there was not much difference between the WT and p62^KDcells, both cells, however, showed a tendency for an increase in the mitochondrial superoxide upon Mtb infection (Fig. 3C). The infection-dependent increase in MitoSox signal increased as the infection progressed from 0- to 24- and 48 hpi (Fig. 3C). However, only at the later time-point (48 hpi), this increase was significant in p62^KD cells (Fig. 3C). The seahorse extracellular flux analyzer further examined mitochondrial respiration by performing the mitostress test, where we measured the cellular oxygen consumption rate (OCR) in response to the sequential addition of inhibitors of the electron transport chain (ETC). Overall, the OCR was found to be similar between WT and p62^KD cells, although the latter showed a non-significant increase in basal and maximal respiration (Fig. 3D). Additionally, there was no significant difference in the mitochondrial DNA content and the expression of genes involved in mitochondrial fission/fusion in WT versus p62^KD cells even after infection (Fig. S4A-B).

Fig. 3.

Our data suggest that p62/SQSTM1 depletion does not overtly alter the mitochondrial function and dynamics, like content and mitochondrial outer membrane permeabilization (MOMP). However, we did observe a delayed increase in the mitochondrial respiration in Mtb-infected p62^KD cells, indicating involvement and/or activation of specific quality control pathways in these cells.

Mtb witnesses increased redox stress in p62^KD cells

While the usual mitochondrial quality parameters were unperturbed in the p62^KD cells upon H37Rv infection, an infection-induced increase in the production of mitochondrial superoxide indicated the possibility of a redox imbalance in these cells. Indeed, there was a significant increase in the reactive oxygen species (ROS), as measured by DCFDA staining in the p62^KDmacrophages in the uninfected state or upon H37Rv infection compared to the corresponding WT conditions (Fig. 3E). Increased ROS in the infected p62^KD cells was also corroborated with significantly higher CellRox fluorescence observed in the Mtb infected p62^KDcells compared to the WT cells (Fig. 3F). Next, we examined, whether Mtb too experienced a more oxidising environment in p62^KD cells and if that contributed to their diminished growth in these cells. To do this, we utilized a genetically encoded biosensor of mycothiol redox potential in Mtb (mrx1-roGFP2) (29). The dynamic changes in mycothiol redox potential serve as a proxy for changes in the cytoplasmic redox state of Mtb. The redox state of Mtb was quantified by measuring the Mrx1-roGFP2 fluorescence ratio (excitation: 405 nm/488 nm; emission: 510 nm). Upon oxidation, the biosensor ratio increases, whereas the ratio decreases in response to the reductive stress (30). Flow cytometry and imaging data revealed that Mtb was exposed to a highly oxidising environment in the p62^KD cells compared to the WT cells (Fig. 3G-I). To test whether increased oxidative stress impacted bacterial growth in these cells, we used a redox-sensitive strain of Mtb- H37RvΔaosR. The H37RvΔaosR lacks a critical cysteine biosynthesis pathway, thereby rendering the strain sensitive to oxidative stress (31). Similar to the WT H37Rv strain, the H37RvΔaosR strain showed a significant decline in CFU in the p62^KD cells compared to the WT THP-1 cells (Fig. 3J). Interestingly, the CFU count of redox-sensitive strain H37RvΔaosR was significantly lower than the WT H37Rv strain in p62^KD cells (Fig. 3J). Moreover, complementation of the deleted strains with WT aosR rescued the bacterial CFU in the p62^KOcells (Fig. 3J). While the ΔaosR::WT strain showed higher CFU in the p62^KO cells than the H37Rv, it was significantly lower than the aosR::WT CFU in the WT cells (Fig. (3J). These results suggest that the elevated oxidative stress witnessed by Mtb may contribute to a decline in its CFU count in p62^KD cells. To test this hypothesis, we used the ROS scavenger N-acetyl cysteine (NAC), which significantly rescued the bacterial CFU in the NAC-treated p62^KD cells compared to the untreated p62^KD cells (Fig. 3K). Using mrx1-roGFP2 reporter, via confocal microscopy we verified that NAC treatment indeed led to a decrease in the level of oxidative stress witnessed by Mtb in the p62^KD cells. (Fig. 3L) These data suggest that the depletion of p62/SQSTM1 in THP-1 macrophages enhances redox stress, contributing to reduced mycobacterial growth and survival in these cells.

Understanding mitochondrial quality control in p62^KD cells

We next studied if mitochondria contributed to the enhanced cellular ROS in p62^KD cells, despite showing normal mitochondrial parameters. Mitochondria may act as the direct source of ROS for bacterial killing upon their recruitment to the bacterial phagosome (32). In p62^KD cells, there was significantly enhanced recruitment of the mitochondria to the bacterial phagosomes, as evident from mitochondrial marker protein TOM20 (translocase of outer mitochondrial membrane 20) immunostaining and 3D-reconstructions to analyse TOM20 recruitment on H37Rv structures (Fig. 4A). Since TOM20 recruitment to Mtb correlated with reduced bacterial growth, we suspected that the phagosome-recruited mitochondria could contribute to the bacterial killing. We also hypothesised that the fraction of mitochondria recruited to the phagosomes could be mostly depolarised, considering they are the major source of mitochondria-generated ROS (33, 34). The ubiquitin ligase Parkin is known to get stabilised on the depolarised mitochondria and ubiquitinates mitochondrial outer membrane proteins for mitophagy targeting (35, 36). Contrary to our expectation, there was a decline in Parkin-TOM20 co-localization in p62^KD cells (Fig. 4B). However, unlike Parkin, there was an increase in the ubiquitin-TOM20 interaction in the p62^KD cells (Fig. 4C). Curiously, the pattern of Ubiquitin and Parkin recruitment to TOM20 mirrored with their recruitment to H37Rv, where there was an increase in the Ubiquitination of Mtb but a decline in Parkin recruitment (Fig. 4D-E). The contrasting observation of enhanced ubiquitination and reduced Parkin recruitment at TOM20⁺-mitochondria and Mtb-containing phagosomes was intriguing. Since ubiquitinated mitochondria could undergo autophagy-mediated degradation or mitophagy, we next checked the status of mitochondrial localization of LC3B in p62^KDcells upon infection. We analysed the co-localization of TOM20 with LC3B in the WT and p62^KD cells. There was a significant increase in TOM20-LC3B colocalisation in p62^KD cells compared to the WT cells (Fig. 4F). We also included a BafA₁ treatment group for both WT and p62^KDcells in this assay to examine whether LC3B⁺-mitochondria positive vesicles matured and fused with the lysosomes distinctly in the p62^KD cells compared to the WT cells. We expected an increase in TOM20-LC3B colocalization in the BafA₁-treated group compared to the corresponding untreated controls, as BafA₁ would block both LC3B and TOM20 degradation. Surprisingly, in either the WT or p62^KD cells, BafA₁ treatment did not increase the TOM20-LC3B colocalisation (Fig. 4F). However, BafA₁-treated p62^KD cells had significantly higher TOM20-LC3B colocalisation compared to the corresponding WT cells. Thus, whereas more mitochondrial cargos were targeted towards LC3B⁺-compartments in the p62^KD cells, the mitochondrial degradation was less efficient in both WT and p62^KD cells.

Fig. 4.

Since similar BafA1 treatment shows increased LC3II upon immunoblotting (Fig. S2F), we infer that autophagic degradation of cargos other than mitochondria continues in p62^KD cells. Could an increase in ubiquitination alone explain the enhanced LC3B recruitment to TOM20 in p62-depleted cells? LC3B can also be recruited to mitochondria through ubiquitin-independent autophagy receptors like BNIP3, FUNDC1, NIX, etc. Incidentally, BNIP3, but not FUNDC1 or NIX, was also among the highly upregulated genes in Mtb-infected p62^KD cells (Fig. 2F), a finding that we confirmed at the protein level by immunoblotting (Fig. S4C). The increased BNIP3 levels in the p62^KD cells were also associated with a significantly increased BNIP3 recruitment to both TOM20 and Mtb as compared to the WT cells (Fig. 4G-H).

Greater recruitment of BNIP3 and also TOM20, both of which are mitochondrial outer membrane proteins, to Mtb-phagosomes indicated that in p62^KD cells, damaged mitochondria or possibly mitochondria-derived vesicles (MDVs) could be recruited to Mtb-phagosomes on their way to being targeted for mitophagy. While significantly higher BNIP3 recruitment, as well as ubiquitination, could explain enhanced LC3B recruitment to TOM20, overall, the pattern of recruitment of several key molecules to mitochondria and Mtb phagosomes in p62^KDcells was incongruent with the anticipated mitophagy pathways and indicated that there were additional mechanisms operational to ensure MQC in p62^KD cells. Results so far, while strongly suggesting crosstalk between mitochondria and Mtb phagosomes, were inconclusive regarding the nature of mitochondrial recruitment to the Mtb phagosomes and whether and how that contributed to the enhanced redox stress to Mtb in the p62^KD cells.

H37Rv-infected p62^KD cells show enhanced biogenesis of TOM20⁺-Mitochondria-derived vesicles (MDVs)

While mitophagy is the primary mechanism for MQC in diverse cell types, increasingly, MDVs are being reported as an important mechanism for MQC, specifically in those cells where mitophagy is compromised (37, 38). MDVs are small vesicles (50-200 nm in size) derived exclusively from the outer or inner mitochondrial membrane that buds off from the damaged parts of mitochondria and fuses with the lysosome for autophagy-independent degradation (39, 40). Analysis of TOM20 immunostaining in p62^KD THP-1 cells under high-resolution microscopy revealed a large number of smaller TOM20⁺ vesicles-like structures in a size range of 200 nm or smaller compared to the WT cells (Fig. 5A). Dual staining with TOM20 and MCU, which respectively are mitochondrial outer membrane (OM) and inner membrane (IM) proteins, strongly indicated that indeed the smaller TOM20 structures seen in p62^KD cells represented MDVs. While we could notice both IM-derived (TOM20^-MCU⁺) and OM-derived (TOM20⁺MCU^-) MDVs (Fig. 5B) in WT and p62^KD cells, we counted the number of TOM20⁺ vesicles in the size range of 100-200nm and observed more than two-fold increase in the number of these vesicles in p62^KD cells compared to the WT cells (Fig. 5C). Interestingly, Mtb infection did not alter the MDV numbers in the WT or p62^KD cells further (Fig. 5C), suggesting that MDV biogenesis reflected an infection-independent mechanism, which most likely was activated as an MQC pathway in the absence of autophagy adaptor p62/SQSTM1. To further confirm that in Mtb-infected p62^KD cells, there was an increase in the formation of TOM20⁺-MDVs, we analysed TOM20 co-localization with additional markers from the mitochondrial IM (MCU and ECSIT) or the mitochondrial matrix (PDHA1, HSP60, SOD2) (Fig. 5D). We anticipated that TOM20⁺ MDVs being exclusively derived from the OM of the mitochondria should show selectivity in terms of its association with IM proteins and matrix proteins. Among the three mitochondrial matrix proteins studied, only SOD2 showed a marginal increase (from ∼66% in the WT cells to ∼73% in p62^KD cells) in its association with TOM20 in the p62^KD cells compared to the WT cells (Fig. 5E). In contrast, PDHA1 and HSP60 showed no significant difference (Fig. 5E, S5A-C). Unlike the matrix proteins, there was a substantial decline in TOM20 colocalisation with the mitochondrial IM proteins ECSIT and MCU (Fig. 5F and S5D-E). The clear distinction in the association between TOM20 and IM proteins (ECSIT and MCU) and OM proteins (BNIP3, Fig. 4), coupled with the increased TOM20⁺-vesicles observed in p62^KD cells, further suggested an increase in OM-derived MDV-biogenesis in p62^KDcells. Taken together from the results in the above sections, it was evident that p62-depleted macrophages trigger TOM20⁺-MDV biogenesis.

Fig. 5.

TOM20 recruited at Mtb phagosomes in p62^KD cells represent TOM20+-MDVs

Given the results above, it was imperative to assess whether the enhanced TOM20 recruitment on Mtb phagosomes in p62^KD cells was actually due to the recruitment of TOM20⁺-MDVs in these cells. We examined the recruitment of mitochondrial IM and matrix markers shown in Fig. 5E-F for their recruitment to Mtb phagosomes. There was a significant decline in the recruitment of mitochondrial IM marker ECSIT to Mtb-phagosomes in the p62^KD cells, a pattern similar to ECSIT-TOM20 colocalisation (Fig. 5G, S5D). Unlike ECSIT, there was no decline in the recruitment of the other IM protein MCU on Mtb-phagosomes (Fig. 5H, S5E). Since MCU⁺TOM20^- MDVs are also formed in both WT and p62^KD cells (Fig. 5B), it is possible that some of the MCU⁺MDVs continue getting recruited to the phagosomes in both WT and p62^KD cells. Additionally, it is possible that ECSIT proteins were not included in the MCU⁺ IM-derived MDVs. The matrix proteins PDHA1, SOD2, and HSP60 did not entirely follow the pattern for Mtb colocalisation vis-à-vis TOM20 colocalisation (Fig. 5I-K, SA-C). Overall, these colocalisation studies reveal that TOM20 and Mtb showed a remarkable similarity in increased recruitment of OM (BNIP3) and decreased recruitment of IM (ECSIT) proteins. However, the matrix proteins did not show a strong correlation. These results suggest that in p62^KD cells, there is a significant increase in OM-derived TOM20⁺-MDVs formation and that a large fraction of these MDVs gets recruited to the Mtb⁺-phagosomes upon infection.

TOM20⁺-MDVs overcome phagosome maturation arrest imposed by Mycobacterium tuberculosis

We next investigated whether MDV recruitment to the Mtb-phagosomes contributed to bacterial growth and survival phenotype observed in these cells. MDVs, once formed, can either be secreted out of the cells as extracellular vesicles (EVs) or fuse with intracellular compartments, such as lysosomes, where they are degraded (41–43). Since TOM20 levels on Mtb-phagosomes were increased in the p62^KD cells, it is possible that TOM20⁺-MDVs were recruited to the Mtb-phagosomes while en route to being targeted to the lysosomes. The lysosomal targeting of MDVs requires proteins from the Sortin-Nexin family, like SNX9 and the small GTPase RAB7 (37, 44). Incidentally, the exclusion of RAB7 from mycobacterial phagosomes and autophagosomes is a widely studied mechanism that helps the bacteria evade the lysosomal targeting (45). In p62^KD THP-1 cells, there was a significant increase in RAB7 recruitment to Mtb-phagosomes compared to the WT cells (Fig. 6A). Increased RAB7 recruitment on the bacterial phagosomes should also result in increased lysosomal targeting of Mtb in these cells. In agreement with that, there was a significant increase in mycobacterial localisation to LysoTracker, Cathepsin D, and LAMP1 positive compartments in p62^KDcells (Fig. 6B-D). It is, therefore, possible that TOM20⁺-MDVs, while getting recruited to the Mtb phagosomes for lysosomal targeting in p62^KD cells, also help recruit RAB7, which unblocks the maturation arrest imposed by Mtb on these phagosomes. This was further corroborated by an increase in LAMP1-TOM20 colocalisation in p62^KDcells, confirming the targeting of MDVs to the lysosomes (Fig. 6E). To verify if indeed RAB7 recruitment was involved in enhanced lysosomal targeting and impacted Mtb survival in p62^KD cells, we knocked down RAB7 using siRNA (Fig. S5F). In the WT cell, the knockdown of RAB7 did not have much impact on H37Rv survival (Fig. 6F), which is expected since very little RAB7 is known to get recruited to Mtb under WT conditions (45). However, the knockdown of RAB7 in p62^KD cells substantially improved Mtb growth and survival, suggesting that RAB7-recruitment indeed played a role in combating Mtb growth in these cells (Fig 6F). In agreement with this notion, the knockdown of RAB7 in p62^KDcells decreased lysosomal targeting of Mtb, as evident from lower colocalisation of Mtb with LAMP1 (Fig. 6G). Curiously, RAB7 knockdown in the p62^KD cells also led to a decline in the TOM20 recruitment on Mtb phagosomes, suggesting that TOM20⁺-MDV recruitment to Mtb phagosomes was somewhat dependent on RAB7 (Fig. 6H).

Fig. 6.

While it was now evident that Mtb was driven towards the lysosomes by TOM20+MDVs, we also witnessed ROS-mediated killing of Mtb in the p62^KD cells (Fig. 3). We hypothesised that the increased oxidative stress seen by Mtb in p62^KD cells could also be attributed to the recruitment of TOM20⁺-MDVs. At least one recent study shows similar MDV-mediated oxidative stress and killing of Staphylococcus sp. Bacteria (46). Since RAB7 knockdown in p62^KD cells decreased TOM20⁺MDV recruitment on Mtb phagosomes, using H37Rv::mrx1-roGFP2 bacteria, we found that the loss of TOM20⁺-MDV recruitment reversed the oxidative stress witnessed by Mtb in the p62^KD cells (Fig. 6I). Our data implicate that TOM20⁺MDV recruitment on Mtb phagosomes could have anti-bacterial functions due to a combination of both redox stress and reversal of PMA.

MQC pathways intricately impact anti-bacterial responses in the macrophages

As evident from our data, TOM20⁺-MDV-dependent facilitation of lysosomal targeting of Mtb in p62^KD cells showed a unique perspective in the crosstalk between mitochondrial homeostasis and cellular anti-bacterial responses. It is also evident that increased TOM20⁺MDVs observed in p62^KD cells could be helping MQC in these cells since we did not observe any significant loss of MQC in the p62^KDcells upon Mtb infection (Fig. 3). We therefore next studied whether the loss of MQC under these circumstances could further aid to the bacterial killing. Parkin is an important mediator of MQC, specifically since it triggers ubiquitin-mediated mitophagy and is also implicated in MDV biogenesis under diverse conditions (47, 48).

Expectedly, the knockdown of PRKN in p62^KD THP-1 cells led to a loss of MQC, as was evidenced by a significant increase in mitochondrial depolarisation in p62^KD cells (Fig. 7A). We also observed round, spherical TOM20 structures, typical of depolarised mitochondria upon PRKN knockdown in p62^KD cells (Fig. 7B). Since the presence of depolarised mitochondria could cause more oxidative damage to the cell, we investigated the oxidative stress witnessed by Mtb in p62^KDcells upon PRKN knockdown. We observed a slight but significant increase in the redox stress witnessed by Mtb in p62^KD cells upon PRKN knockdown compared to p62^KD cells alone (Fig. 7C). To test whether the loss of MQC in p62^KD cells upon PRKN knockdown had any impact on intracellular bacterial growth and survival, we investigated Mtb CFUs in these cells. We observed significantly reduced Mtb survival in PRKN knockdown p62^KD cells compared to the p62^KDcells (Fig. 7D). PRKN knockdown did not impact MDV biogenesis as inferred from the absence of any difference in the recruitment of TOM20⁺-MDVs to the Mtb phagosomes (Fig. 7E-F). In agreement to that, there was no change in the lysosomal targeting of Mtb upon PRKN knockdown in p62^KD cells (Fig. 7G). Thus, enhanced redox stress due to mitochondrial depolarisation, along with continued lysosomal targeting, caused excessive Mtb-killing in these cells. This was further corroborated by using Nigericin, a K⁺/H⁺ ion exchanger, which causes mitochondrial depolarisation (49). Treatment with nigericin led to a dramatic loss of mitochondrial outer membrane permeability (Fig. S5H) and showed significantly increased oxidation of Mrx1-RoGFP2 reporter (Fig. S5I). Moreover, similar to the results in PRKN knockdown in p62^KD cells, nigericin treatment also led to a significantly reduced Mtb survival in the p62^KD cells (Fig. S5J). At the doses used, nigericin did not impact host cell survival (Fig. S5K).

Fig. 7:

MIRO1, a Rho-family type GTPase, is required for enhanced MDV biogenesis in p62^KD cells

Next, we wanted to understand what regulates MDV biogenesis in p62^KDcells. Along with PRKN1 and DRP1, another family of proteins that have recently been implicated in MDV biogenesis is MIRO1/2 (50). At the same time, SNX9 is reported to help in the trafficking of MDVs (38). We knocked down DRP1 (DNML1), SNX9, MIRO1 (RHOT1), and MIRO2 (RHOT2) in separate experiments using WT and p62^KD cells, and screened for conditions that could reverse the decline in Mtb CFU in the p62^KD cells (Fig. S6A). Out of these, siMIRO1 and siMIRO2 were able to reverse the Mtb growth in the p62^KD cells back to the levels seen in the WT cells (Fig. S6B-C, Fig. 7H). We next investigated whether the increase in the bacterial CFU in MIRO1/2 knockdown cells was due to a direct role of MIRO1 and 2 in MDV biogenesis. We observed a significant decline in the cellular MDV count in p62^KDcells where MIRO1 or MIRO2 were knocked down (Fig. 7I). This also led to a significant decline in the recruitment of TOM20 to Mtb phagosomes (Fig. 7J), which in turn led to reduced recruitment of RAB7 on Mtb phagosomes in the MIRO1 or MIRO2 knocked down p62^KD cells (Fig. 7K). Thus, MIRO1/2-mediated MDV biogenesis in p62^KD cells contributes to enhanced TOM20 and RAB7 recruitment to Mtb-phagosomes.

Discussion

Autophagy adaptors recognise and target the cargos for autophagy-mediated degradation (12). The damaged mitochondria are one of the significant cellular cargos that must get degraded via mitophagy to ensure mitochondrial quality and cellular health (51). Since the autophagy adaptors also play a crucial role in targeting the bacterial pathogens for lysosomal degradation, in this study, we wanted to study the distribution of the autophagy adaptors between mitochondria, the cellular cargo and Mtb, the pathogenic cargo, in the infected macrophages. Understanding this functional segregation of the autophagy adaptors is crucial since several host-directed therapies against tuberculosis envisage targeting the autophagy (52). However, most autophagy regulators do not distinguish between the homeostatic and defence arm of autophagy, potentially limiting the evolution of these regulators to clinically acceptable uses. Our data suggest p62/SQSTM1 as the major autophagy adaptor recruited to both Mtb and mitochondria in the infected macrophages. Contrary to our hypothesis, infection with Mtb did not significantly alter, except for a minor change in OPTN, recruitment and the distribution of autophagy adaptors on mitochondria. However, irrespective of their high or low association with the Mtb or mitochondria, individual depletion of each autophagy adaptor, except for NDP52, resulted in reduced bacterial survival in THP-1 macrophages. These early results showed that autophagy adaptors in the human macrophages regulate some housekeeping functions that crisscross with the anti-bacterial pathways. Moreover, it was also apparent that p62/SQSTM1 regulates some distinct downstream pathways in humans versus other vertebrates like mice and zebrafish. That is because, in the zebrafish model, p62/SQSTM1 and OPTN were previously shown to play a critical role in host resistance to mycobacterial infections (22). In mice, there are reports of an inverse correlation between Mtb-p62/SQSTM1 colocalisation and the bacterial CFU in the BMDMs (53). In yet another study, the absence of p62/SQSTM1 in mouse BMDMs did not affect Mtb CFU (54). The same study reported that starvation, which reduced Mtb CFU in the WT cells failed to do the same in the absence of p62/SQSTM1(54). In our experiments, mouse macrophages, knocking down autophagy adaptors in BMDM (p62 and Optn) or knocking out in HOXB8-derived macrophages (p62), did not show any impact on intracellular Mtb growth, which is consistent with earlier reports (54). In addition, even in vivo studies in p62^-/-animals did not show any significant impact on Mtb survival or killing (23). Curiously emerging evidences suggest that p62/SQSTM1could have tissue-specific functions (55). How p62/SQSTM1 regulates the anti-bacterial pathways in a host-specific manner has significant implications for developing new host-directed therapy against tuberculosis.

Autophagy, through mitophagy, ensures the maintenance of mitochondrial quality -a serious challenge for the cells with respiring mitochondria since the process of oxidative phosphorylation itself generates free radicals that can damage the mitochondria (17). The damaged mitochondria must get degraded via mitophagy to keep cellular apoptotic pathways under check (56). The damaged mitochondria can also cause bacterial killing, primarily through enhanced redox stresses (32, 57). While p62/SQSTM1 was highly recruited on mitochondria in the macrophages, we did not observe any loss to mitochondrial quality in p62^KD cells. However, loss of p62/SQSTM1 in human cortical neurons is associated with an aberrant mitochondrial functionality (58). In HeLa cells, only OPTN, NDP52, and, to some extent, TAX1BP1 were essential for the mitophagy (59). Therefore, it is possible to have tissue/species-specific functions for each autophagy adaptor. Despite maintaining MQC, Mtb witnessed enhanced oxidative stress in the p62^KDcells; which also contributed to the reduced bacterial survival in these cells. The increased redox stress could also be due to the proven role of p62/SQSTM1 in regulating the KEAP1-NRF2 antioxidant pathway (60, 61).

Interestingly, in addition to its role in mitophagy, p62/SQSTM1 also regulates peroxisome quality control by recognising ubiquitinated peroxisome protein PEX5 and targeting the peroxisomes for pexophagy (62). We cannot rule out peroxisomal dysfunction in p62^KD cells, which could also contribute towards enhanced oxidative stress witnessed by the Mtb. The involvement of pexophagy in regulating redox stress during mycobacterial infections is only beginning to emerge and needs further explorations (63).

The more exciting discovery in this study was the determination that in the absence of p62/SQSTM1, these cells maintained mitochondrial quality through MDV biogenesis. As described by others, MDV biogenesis takes over in the absence of mitophagy to ensure MQC (37). The MDV biogenesis in p62^KD cells was independent of infection, suggesting the homeostatic role of this process. Targeting MDVs to the lysosome for degradation ensures MQC and depends on RAB7 (64). Since MDV biogenesis in p62^KD cells is independent of infection, we believe that RAB7 recruitment to Mtb phagosome occurs inadvertently during the transit of TOM20+-MDVs towards the lysosomes. Phagosome maturation arrest, or inhibition of xenophagy flux, caused by Mtb, specifically acts via the exclusion of RAB7 from the Mtb-containing phagosomes and autophagosomes (5) (45). Since RAB7 recruitment was the effector that impacted Mtb survival in p62^KD cells, depletion of RAB7 rescued Mtb survival only in p62^KD cells. Mechanistically, RAB7 depletion reduced the lysosomal targeting of Mtb and rescued Mtb from the redox stress, coinciding with a reduction in TOM20 recruitment to the Mtb phagosomes. Thus, MDV recruitment to Mtb phagosomes facilitated bacterial targeting to the lysosomes and also brought redox mediators to the bacterial compartments. At least one report in the literature suggests redox mediators brought in by MDVs to the Staphylococcus-containing phagosomes, which helps in the bacterial killing (46).

Could MDV biogenesis be blocked in p62^KD cells to rescue Mtb survival? Among many regulators, Parkin, DRP1 and MIRO1/MIRO2 are implicated in MDV biogenesis (39, 50). Depleting Parkin in p62^KD cells provided yet another intriguing insight. In the absence of Parkin, MQC was affected in both WT and p62^KD cells, as expected from what is known about Parkin’s function in MQC (64). However, this also led to a further decline in bacterial survival in WT cells, contrary to earlier results using PRKN^-/-BMDMs (65). Crucially, in the p62^KDcells, the loss of Parkin synergised and further enhanced bacterial killing. Contrary to the perceived role of Parkin in MDV biogenesis, loss of Parkin in p62^KD cells had no impact on TOM20 recruitment to Mtb phagosomes and lysosomes. In fact, increased Mtb killing in Parkin depleted condition was more consistent with other studies that show possible negative influence of Parkin on MDV formation (50, 66). In contrast, knocking down MIRO1 and MIRO2 in p62^KDcells inhibited MDV biogenesis, reduced TOM20 and RAB7 recruitment on Mtb phagosomes and rescued intracellular Mtb CFU. It appears that cells can activate yet another MQC mechanism in p62^KD cells independent of MIRO1/2. Further studies are needed to understand the complex interplay between diverse MQC pathways. Collectively, we conclude that the cumulative effect of lysosomal targeting and mitochondrial damage compromises bacterial survival in p62^KD cells (Fig. 8).

Fig. 8.

This study provides an as-yet unrecognised crosstalk between mitochondrial quality control and the phagosome maturation pathway. The reversal of phagosome maturation reported here is unprecedented and underscores the need to dissect further the MQC pathways vis-à-vis anti-bacterial pathways in the host. We propose that these results provide novel ways to target MQC to achieve bacterial killing as a potential host-directed therapy strategy.

Materials and Methods

Ethics Statement

The Institutional Ethics Committee, ILBS (IEC/2020/81/MA03) and ICGEB (ICGEB/IEC/2020/25, version 3) approved studies on healthy human samples (healthy adults following informed consent). For BMDM experiments, animals were approved by IAEC of ICGEB (ICGEB/IAEC/18092021/CI-15).

Cell Culture

Human monocytic cell line THP-1 was procured from American Type Culture Collection (ATCC) and cultured in RPMI 1640 medium (Cell Clone, CC3014) along with 10% Fetal Bovine Serum (Gibco). Cells were regularly checked for mycoplasma contamination. The cells were incubated in a humidified 5% CO₂ chamber at 37 °C. THP-1-derived macrophages were obtained by treating THP-1 cells with 20 ng/ml phorbol myristate acetate (Sigma, P8139) for 24 hours, followed by washing and maintenance in complete media. Human MDMs (hMDMs) were isolated from human PBMCs and separated from healthy human donors’ blood using PolymorphPrep (PROGEN,1895) layering and centrifugation. The heparinised blood was layered on PolymorphPrep in a 1:1 ratio and centrifuged at 500g for 45 min at 25°C. The interface containing PBMC was isolated carefully and treated with RBC lysis buffer. Subsequently, cells were washed twice with DPBS. PBMCs were diluted in 10% RPMI 1640 media to a concentration of 1 × 10⁶ cells/ml. Cells were put in a six-well cell culture plate and incubated for 4 hours at 37°C in a humidified 5% CO₂ chamber. Non-adherent cells were removed, followed by two washes with RPMI media. Complete RPMI media containing 50 ng/ml recombinant human M-CSF (R & D systems, 216-MC-025) was added and cells were allowed to differentiate for a week into macrophages in a humidified 5% CO₂ incubator at 37 °C. For BMDMs, bone marrow was isolated from the femur and tibia (hind limb) of mice (6-8 wk old, female, C57BL/6 mice) and passed through a 23G needle. The RBCs were lyzed by RBC lysis buffer and then washed with 1x PBS. The isolated Bone marrow cells were cultured in 10% DMEM containing 20% L929 supernatant for 5-7 days. All mice were housed in the Bio experimentation animal facility in a sterile environment.

Generation of Hoxb8 progenitors cell lines and differentiation in macrophages

The method was previously described (67). Briefly, BM cells were collected from the femurs of 8-week-old mice and separated on a Ficoll gradient. Cells were resuspended in RPMI containing 15% FBS and supplemented with 15% of Flt3L and 30% GM-CSF containing supernatant. After 2 days of cell culture, cells were collected and resuspended in RPMI supplemented with 10% FBS, 1 μM β-estradiol (MilliporeSigma, E-2758), 1% gentamicin and infected with MSCV herad-3HAHOXB8 retroviral particles. For 1-month, cells were dispensed in fresh medium every 3–4 days and transferred into new wells. After this period, cells were expanded in Hoxb8 Medium supplemented with 15% Flt3L supernatant. For subsequent differentiation experiments, cells were washed twice with warm PBS containing 1% FBS and resuspended in a concentration of 0.5 × 10⁵ cells/mL in IMDM containing 20% M-CSF. On day 4 medium was replaced with a fresh one, and 1/3 of starting media was added on day 6. Differentiated cells were used between days 7 and 8. Flow cytometry confirmed homogeneous differentiation between different genotypes using F4/80 and CD11b markers.

Mycobacterium tuberculosis culture and infection

Mycobacterial strain H37Rv was obtained from Colorado State University. The mCherry-H37Rv was made using pMSP12: mCherry plasmid (Addgene No. 30169). H37Rv culture was grown in Middlebrook 7H9 (BD, 271310) media supplemented with 10% ADC (Albumin Dextrose Catalase, BD, 211887), 0.4% glycerol, and 0.05% Tween 80 until the mid-log phase.

The bacteria were then pelleted down by spinning at 4000 rpm for 5 min, washed with RPMI 1640 and their single-cell suspension was obtained by passing the culture through needles of 3 different gauge (G) sizes in the following way- 23 G (7 times), 26 G (5 times) and 30 G (3 times). For Mtb quantification, OD600 was considered equivalent to 100 million bacteria. The required number of bacteria was added onto the seeded cells in complete media at MOI of 1:10, followed by spinning the plates at 800 rpm for 5 min. After 4 hrs, 200ug/ml amikacin sulphate (Sigma, A2324) treatment was done for 2 hrs. Subsequently, cells were washed, fresh media was added, and cells were kept at 37°C for the desired time point of processing. Thus, the duration to complete the infection protocol was 6 hours, and the time point immediately after the amikacin step was considered 0 hours post-infection (hpi). All Mtb work was carried out in the Tuberculosis Aerosol Challenge Facility (TACF), a national BSL3 facility for tuberculosis research.

CFU Experiment

For CFU experiments, the cells were seeded in 96-well cell culture plates. At desired time point, culture media was removed, and cells were washed with 1X PBS. Cells were lysed with 100 μl of 0.06% SDS lysis buffer for 10 minutes at room temperature. The lysate was then diluted in 7H9 media, and 10μl of this sample was plated on 7H11 agar (BD, 8801671) square plates supplemented with 10 % OADC (BD, 211886). Plates were dried and incubated at 37°C humidified incubator for a minimum of two weeks or till the appearance of Mtb colonies. CFUs were calculated by multiplying the actual colony count with the dilution factor and the volume of the diluted sample used for plating. For the replication clock experiment, pBP10 containing H37Rv was used for infection, and at the designated time points, CFU plating was performed parallelly on normal or kanamycin-containing 7H11-OADC plates. The fraction of CFU obtained in Kan⁺ plates compared to the Kan^- plates was calculated to determine the %pBP10⁺ Mtb.

Flow Cytometry Assays

Cells were seeded in a 12-well culture plate for all the flow cytometry experiments. The H37Rv used was labelled with CellVue Claret Far red dye (Sigma, MINCLARET-1KT) according to the manufacturer’s protocol. After processing the samples at the desired time, 500 μl of 1X PBS was added to the wells, and cells were scrapped off using a scrapper and added to fresh round bottom polystyrene FACS tubes. The samples were immediately acquired in BD FACSAria Flow cytometer. For most experiments, 30000 events were captured and analysed by FlowJo v10 software. The assay for DCFDA (10 μM, 30 minutes) (Sigma, D6883), TMRE (100 μM, 30 minutes) (Invitrogen, T669), Mitotracker Green (200 nM, 40 minutes) (Invitrogen, M7514), and MitoSOX Red (5 μM, 20 minutes) (Invitrogen, M36008) were performed according to manufacturer’s protocol.

Immunoblotting

For the lysate preparation, cells were seeded in a 6-well cell culture plate, and at the desired time, cells were washed with ice-cold 1X PBS. The lysis of the cells was performed by incubating them with the lysis solution containing RIPA buffer (Sigma-Aldrich, R0278) and 1X Protease Arrest (G⁻Biosciences) for 20 minutes on ice. The lysate was collected in microcentrifuge tubes and centrifuged at 6000 g for 10 minutes at 4 °C. The supernatant (cytoplasmic extract) was collected, and protein quantification was done by Bradford Assay. The protein sample was mixed with 6× loading dye, and denaturation was done at 95 °C for 10 minutes. The protein sample was subjected to SDS-PAGE for separation and blotted onto the nitrocellulose membrane (BIO-RAD) using BIO-RAD Trans-Blot Turbo Transfer System.

Blocking was performed for an hour with Odyssey blocking buffer (LI-COR Biosciences) and 1X PBS in the ratio of 1:1 at room temperature. Blots were stained overnight with primary antibody at 4 °C and then with secondary antibody for 1.5 hours at RT, with good washes with 1X PBST in between. The IR-conjugated secondary antibodies anti-mouse IR780 (926–32210) and anti-rabbit IR780 (925–68071) were obtained from LI-COR Biosciences. The dilutions of the antibodies were made according to the supplier’s protocol in the Odyssey blocking buffer. The blots were scanned with the Odyssey Infra-Red Imaging system (LI-COR Biosciences).

siRNA Transfection

Cells were either transfected with control or specific siRNA using Dharmafect-2 reagent (Dharmacon Inc, T-2002-02) according to the manufacturer’s protocol. 50nM siRNA was added to cells at 0-hour post infection for 24 hours. All the siRNA used in our study were siGenome siRNA SmartPool (Dharmacon Inc). The following were used: p62/SQSTM1 (M-010230-00-0005), NDP52 (M-010637-01-0005), OPTN (M-016269-02-0005), TAX1BP1 (M-016892-01-0005), NBR1 (M-010522-01-0005), RAB7A (M-010388-00-0005), PARKIN (PRKN) (M-003603-00-0005), MIRO1 (RHOT1) (M-010365-01-0005), MIRO2 (RHOT2) (M-008340-01-0005), DRP1 (DNM1L) (M-012092-01-0005), and SNX9 (M-017335-02-0005). In BMDM, mouse-specific p62/sqstm1 (M-047628-01-0005) and optn (M-041705-01-0005) siRNA was used.

Confocal Microscopy

For microscopy experiments, cells were seeded on the sterile coverslips in a 24-well tissue culture plate. All the staining steps were performed at room temperature. At the requisite time, cells were fixed with 4% PFA (Sigma, 158127) for 20 minutes, followed by 2 washes with 1X PBS. The cells were permeabilised with 0.2% Triton-X100 (Sigma, T8787) (in 1X PBS) for 15 minutes with two subsequent washes with 1X PBS. Blocking was performed with the buffer of 3% BSA (Sigma, A7906) and 0.5% Tween 20 (Sigma, P1379) in 1X PBS for 1 hour. The cells were stained with respective primary antibodies for 2 hours and then washed twice with 1X PBST. This was followed by staining with individual secondary antibodies tagged with the fluorophore of choice for 1 hour. Both primary and secondary antibodies were made in the blocking buffer at the required dilution. The cells were washed twice with 1X PBS before mounting the coverslips on glass slides with ProLong Gold anti-fade reagent (Life Technologies, P36930). Confocal images were acquired with a Nikon A1R microscope equipped with 100X/1.4 NA plan-apochromat VC, DIC N2 objective lens. The analysis was performed on Imaris 9.6 imaging software using several tools like deconvolution, 3D reconstruction and shortest distance calculation. The research for the roGFP2 redox sensor was conducted with Nikon NIS Element. For MDVs spot analysis, Deep learning-based TruAI tool of cellSens Imaging software was used.

The primary antibodies used were: α-p62/SQSTM1 (Abcam, ab155686), α-OPTN (Abcam, ab23666), α-NDP52 (Abcam, ab68588), α-TAX1BP1 (Abcam, ab121812), α-NBR1 (Novus Biologicals, NBP1-71703), α-MAP1LC3B (NB100-2200), α-BNIP3 (Abcam, ab109362), α-FUNDC1 (ab74834), α-NIX (Abcam, ab8399), α-TOM20 (Abcam, ab56783, ab78547), α-Parkin (Novus Biologicals, NBP2-67017), α-HSP60 (Abcam, ab13532), α-SOD2 (Novus Biologicals, NB100-1992), α-PDHA1 (Abcam, ab110330), α-DRP1 (Abcam, ab184247), α-SNX9 (Abcam, ab181856), α-ECSIT (Novus Biologicals, NBP1-918858), α-LAMP1 (Santa Cruz Biotechnology, Sc-20011), MIRO1 (Novus, NBP1-89011), MIRO-2 (Novus, NBP1-88982), α-RAB7 (Abcam, ab50533), α-CATHEPSIN-D (Santa Cruz Biotechnology, Sc-10725), α-Ubiquitin (Abcam, ab134953), α-MCU (Cell Signaling Technology, D2Z3B), α-TUBULIN (Abcam, ab6046), α-ACTIN (Abcam, ab8226), α-GAPDH (Abcam, ab9485)

All secondary antibodies used were from Invitrogen, conjugated with Alexa Flour dye: anti-mouse Alexa Fluor 405 (A31553), anti-mouse Alexa Fluor 488 (A28175), anti-mouse Alexa Fluor 568 (A11031), anti-mouse Alexa Fluor 647 (A21235), anti-rabbit Alexa Fluor 405 (A31556), anti-rabbit Alexa Fluor 488 (A11034), anti-rabbit Alexa Fluor 568 (A11011), anti-rabbit Alexa Fluor 647 (A21245). CellROX deep red reagent (C10422) and LysoTracker green (L7526) were from Invitrogen. For microscopy experiments, mCherry-H37Rv was used.

p62^KD THP-1 monocytic cell line generation

The p62^KD cells were generated using the CRISPR/Cas9 system. The guide RNA was designed against exon 3 of p62 and cloned into a gRNA cloning vector plasmid (Addgene No. 41824). The Cas9 enzyme encoding hCas9 plasmid was from Addgene (No. 41815). Both the plasmids were co-transfected into THP-1 monocytes using the Invitrogen Neon Transfection system according to the manufacturer’s protocol. At 24 hours post-transfection, the cells were selected on Neomycin (Sigma, G418 A1720) at 750ug/ml for 21 days, followed by checking the absence of p62 in knockdown cells by microscopy and immunoblotting.

RNA Sequencing

Total RNA was extracted from THP-1 macrophages, following the double extraction protocol: RNA isolation by acid guanidinium thiocyanate-phenol-chloroform extraction (TRIzol) followed by a Qiagen RNeasy Micro clean-up procedure. RNA was analysed on Agilent Bioanalyser for quality assessment with RNA Integrity Number (RIN) ranging from 7.5 to 10 and a median of RIN 9.8. cDNA libraries were prepared using 2 ng of total RNA using the SMARTSeq v2 protocol with the following modifications: 1. Addition of 20 µM TSO; 2. Use of 200 pg cDNA with 1/5 reaction of Illumina Nextera XT kit. The length distribution of the cDNA libraries was monitored using a DNA High Sensitivity Reagent Kit on the Perkin Elmer Labchip. All samples were subjected to an indexed paired-end sequencing run of 2x151 cycles on an Illumina HiSeq 4000 system (28 samples/lane).

RNA-Seq analysis

Reads were quality filtered using Trimmomatic (v.0.39, LEADING:10 TRAILING:10 SLIDING WINDOW:4:15 MINLEN:75). They were assessed for quality using FastQC (v.10.1))] and visualised with MultiQC (v.1.9). Salmon (v.1.8, –numGibbsSamples 30) was used to quantify quality filtered reads against the human genome (GRCh38, primary assembly) with Gencode annotation (v.40). Profiles were imported using tximeta (v.1.10) and SummarizedExperiment (v.1.22) which were then summarised to gene-level expression. Non-coding genes were filtered out, and the following gene biotypes were retained: protein-coding, pseudogenes (translated and transcribed), Immunoglobulin (IG) and T cell receptor (TCR) genes). Dataset was further filtered to retain only those genes whose expression was detected in at least 3 samples. Statistical analysis was done using swish methodology from fishpond (v.1.8) (68) and significantly differentially expressed genes (DEGs) were defined based on qvalue ≤ 0.1 and absolute log₂ fold change ≥ 1. Heatmap of the DEGs was generated with heatmap (v.1.0.12) using mean z-scaled normalised scaled TPM values (summarised to gene level using summarizeToGene() function) and clustered based on Euclidean distance and Ward.D2 methodology. Enrichment analysis was done separately on overexpressed and repressed DEGs from each pairwise comparison using enrichR (v.3.0) (69) with KEGG Human 2021 and Gene Ontology (GO) (70) (Biological Process (BP), Cellular Component (CC) and Molecular Function (MF)) 2021 database. Mitochondrial DEGs and their associated pathways were identified using Mitocarta 3.0 (71). Cytokines were identified from the comprehensive list as described by Pro et al.(72). Figures were generated using ggplot2 (v.3.3.5). Analyses were done in R statistical software (v.4.1).

Mrx1-roGFP2 Redox Sensor Assay

H37Rv cultures were electroporated with Mrx1-roGFP2, and colonies were selected on a hygromycin 7H11 plate.THP-1 macrophages were infected with Mrx1-roGFP2-H37Rv at M.O.I of 10. At the requisite time, cells were washed with 1X PBS, with subsequent treatment of 10 mM NEM (N-ethylmaleimide, Sigma, E3876) for 10 minutes at 37 °C. To the positive and negative control groups, 5 mM CHP (Cumene hydroperoxide, Sigma) and 10 mM DTT (Dithiothreitol, Sigma, D0632) were added respectively for 5 minutes at 37 °C. The cells were washed twice with 1X PBS, followed by NEM-PFA fixing. For FACS, 2% PFA was used, whereas for confocal microscope acquisition, 4% PFA was used.

Extracellular flux analysis

At 24 hours post-infection, THP-1 macrophages were scrapped off from a 6-well culture plate, and 4 x 10⁴ cells were seeded in Cell-Tak (Corning, 354240) coated Agilent Seahorse XFp cell culture mini plate (Agilent, Santa Clara, CA). The plates were centrifuged at 200g for 1 min with minimum deceleration. The plates were parafilm-wrapped and incubated in a non-CO₂ incubator at 37°C for 25-30 minutes. After adding 130μl warm assay media and incubation for for 15-25 minutes at 37°C, the plates were ready for the assay. The Agilent Seahorse XFp analyser was used to assess the oxygen consumption rate. The modulators included in this assay were oligomycin (1.5 μM, Sigma, #75351), Carbonyl cyanide-4 (trifluoromethoxy) phenylhydrazone (FCCP) (1 μM; Sigma, C2920), Rotenone (0.5 μM; Sigma, R8875) and antimycin A (0.5 μM; Sigma, A8674) (73). Initially, the baseline OCR of the cells was measured, from which the basal respiration was determined by eliminating the non-mitochondrial respiration. Maximal respiration was calculated by subtracting the non-mitochondrial respiration from the maximal OCR after FCCP addition. Non-mitochondrial respiration represents the minimum OCR after rotenone and antimycin A treatment.

mtDNA/nDNA

The cellular DNA was extracted from the DNeasy kit (Qiagen) according to the manufacturer’s protocol. To quantify the mtDNA/nDNA ratio, qPCR was performed to amplify one gene from the mitochondrial genome (mtCO1) and one gene from the nuclear genome (RPL13A). Primer sequences were as: mtCO1_forward: 5′-CAGGAGTAGGAGAGAGGGAGGTAAG-3′, mtCO1_reverse: 5′-TACCCATCATAATCGGAGGCTTTGG-3′; RPL13A_forward:5′-CGCCCTACGACAAGAAAAAG-3′, and RPL13A_reverse: 5′ CCGTAGCCTCATGAGCTGTT-3′

Inhibitor or Activators treatment

The reagents were used according to the following conditions before processing the samples: Bafilomycin A1 (B1793) 100 nM for 3 hrs, Carbonyl Cyanide m-Chlorophenylhydrazine (CCCP, Sigma, C2759): 5 μM for 12 hours for CFU and 20 μM for 20 min for mitochondrial depolarisation assay (TMRE), Nigericin (Sigma, N7143-5MG): 2 μM for 24 hours, N-acetyl cysteine (NAC, Sigma, A7250), 2mM for 24 hours.

RT-PCR

Total RNA was extracted at the required time point from THP-1 macrophages using MDI total RNA Miniprep kit (MTRK250) according to the manufacturer’s protocol. cDNA was synthesised from 500 ng of total RNA by reverse transcriptase PCR using a BIO-RAD iScript cDNA synthesis kit (170–8891). The qPCR was set using the mixture of cDNA sample, SYBR green dye (QuantiNova) and desired primers, where 18s rRNA was used as an internal control. The acquisition was made on the Bio-Rad CFX 96 Real-time PCR system. The primers used are Fw_p62/SQSTM1: 5’-GCACCCCAATGTGATCTGC-3’,

Rv_p62/SQSTM1:5’-CGCTACACAAGTCGTAGTCTGG-3’,

Fw_NBR1:5’-AGGAGCAAAACGACTAGCTGC-3’,

Rv_NBR1:5’-TCTGGGGTCTTCATGTCTGAT-3’,

Fw_NDP52:5’-ATTTCATCCCTCGTCGAAAGGA-3’,

Rv_NDP52:5’-TGAAGGTGTAATACTCACGGGTT-3’,

Fw_OPTN:5’-CCAAACCTGGACACGTTTACC-3’,

Rv_OPTN:5’-CCTCAAATCTCCCTTTCATGGC-3’,

Fw_TAX1BP1:5’-TAAAGGAGCAACTTCGTAAAGCA-3’,

Rv_TAX1BP1:5’-CTGCCATCGTTCTGTCTCGT-3’,

Fw_MFN1:5’-TGGCTAAGAAGGCGATTACTGC-3’,

Rv_MFN1:5’-TCTCCGAGATAGCACCTCACC-3’,

Fw_MFN2:5’-CTCTCGATGCAACTCTATCGTC-3’,

Rv_MFN2:5’-TCCTGTACGTGTCTTCAAGGAA-3’,

Fw_DRP1:5’-CTGCCTCAAATCGTCGTAGTG-3’,

Rv_DRP1:5’-GAGGTCTCCGGGTGACAATTC-3’,

Fw_OPA1:5’-TGTGAGGTCTGCCAGTCTTTA-3’,

Rv_OPA1:5’-TGTCCTTAATTGGGGTCGTTG-3’,

Fw_FIS1:5’-GTCCAAGAGCACGCAGTTTG-3’,

Rv_FIS1:5’-ATGCCTTTACGGATGTCATCATT-3’,

Fw_18s:5’-GCTTAATTTGACTCAACACGGGA-3’,

Rv_18s: 5’-AGCTATCAATCTGTCAATCCTGTC-3’.

Statistical Analysis

Statistical significance for comparisons between two sets of the experiment was performed using an unpaired two-tailed Student’s t-test. For multiple treatment experiments, one-way ANOVA followed by multiple comparison analyses was conducted in GraphPad PRISM 8.

Data availability

All data are available in the main text or the supplementary materials. The RNAseq raw data is available at GEO database, accession number GSE216985.

Acknowledgements

We acknowledge Dr David Sherman for the replication clock plasmid (pBP10). We acknowledge the Tuberculosis Aerosol Challenge Facility (TACF), supported by the Department of Biotechnology, Govt of India (BT/INF/22/SP43067/2021), for all BSL3 experiments. This work was supported by the Department of Science and Technology through a DST-UKIERI collaborative grant (DST/INT/UKP-136/2016, DK and SS) and partly by the Department of Biotechnology, Govt of India (BT/IC-06/003/91 and RAD-40/104/2023-MED-DBT; Flagship program, DK), DBT-Wellcome Trust India Alliance Senior Fellowship (IA/S/17/1/503071, DK) and Science and Engineering Research Board (SERB), Govt of India (EMR/2016/005296, and CRG/2022/008256; DK). University of Birmingham India Institute Visiting Fellowship to S.V. and S.S.; LifeArc, Wellcome Trust and Birmingham Fellowship to S.S. and DST-INSPIRE fellowship to S.V. A.S is supported by A*STAR ID Labs, R01HL152078 and NMRC/OFIRG/may-0096.

Additional information

Author contributions

Conceptualization: DK, SS, SV

Methodology: SV, ST, RS, VY, GMP, MZK, MM, SF, SH, PS, ST

Investigation: SV, ST, RS, MD, VY

Visualization: SV, ST, MD, DK

Resources: VN, BM, MB, JSM, FB

Supervision: DK

Writing—original draft: DK, SV

Writing—review & editing: DK, SV, ST, MD, VY, FB, AS, SS, AS

Funding

Department of Science and Technology, Ministry of Science and Technology, India (DST) (DST/INT/UKP-136/2016)

• Dhiraj Kumar

Department of Biotechnology, Ministry of Science and Technology, India (DBT) (BT/IC-06/003/91 and RAD-40/104/2023-MED-DBT; Flagship program)

• Dhiraj Kumar

Wellcome Trust DBT India Alliance (India Alliance) (IA/S/17/1/503071)

• Dhiraj Kumar

DST | Science and Engineering Research Board (SERB) (EMR/2016/005296)

• Dhiraj Kumar

DST | Science and Engineering Research Board (SERB) (CRG/2022/008256)

• Dhiraj Kumar

Additional files

Supplementary figure. Fig S1. Expression profile of autophagy adaptors in H37Rv infected and uninfected THP-1 Macrophages. (A) Gene expression of NBR1, NDP52, TAX1BP1, OPTN, and p62/SQSTM1 (p62) by RT-PCR at 0-, 24-, and 48-hpi. Data show mean ± SD from three independent experiments. (B) Immunoblots represent protein expression of TAX1BP1, NDP52, OPTN, NBR1, and p62 at 0-,24- and 48-hpi; GAPDH is used separately as the loading control for each condition; (C) bar graphs on the right show the quantification of the protein levels of these adaptors after normalising with GAPDH. Data show mean ± SD from three independent experiments. Fig S2. Depletion of autophagy adaptors in human and mouse macrophages and their effect on H37Rv survival. (A) Immunoblot represent the siRNA-targeted depletion of adaptors TAX1BP1, p62, NDP52, OPTN and NBR1 in THP-1 monocytes in the separate experiment for each adaptor at 24 hours post-transfection. The numbers depict the ratio of intensities of specific proteins to that of GAPDH. (B) CFU analysis of H37Rv in control and siRNA-targeted-OPTN and -p62 U937 macrophages at 24 hpi. Right panel: Immunoblot represents the siRNA-targeted depletion of p62 and OPTN at 24 hours post-transfection. The numbers depict the ratio of intensities of specific proteins to that of GAPDH. (C-D) Respective immunoblot and IFA images for p62 in p62^KD THP-1 cells generated using CRISPR/Cas9 and their quantification graphs (Right Panel). Plots represent data from three experiments (C) and 5 fields from two different experiments (D). (E) Immunoblot depicts the levels of p62 in uninfected and infected cells in WT and p62^KD cells at 24- and 48 hpi and separate quantification plots from two independent experiments (Right). (F) Immunoblot represents LC3B levels in the lysates of WT and p62^KD macrophages at 0-,24-, and 48-hpi in the presence and absence of BafA1 (100nM for 3 hours before processing). (G) Confocal images show the interaction of LC3B spots (blue spots, at <0.2 um from bacteria) and H37Rv. The bar graph represents the quantification of blue spots on H37Rv in WT and p62^KDcells in the presence and absence of BafA1 at 0- and 24-hpi. Data show mean ± SD, n> 9 Fields, 20-35 per field, from three experiments. (H-I) Respective CFU analysis of H37Rv in control and siRNA-targeted p62/sqstm1 depleted and Optn depleted mouse bone marrow monocyte-derived macrophages at 24 hpi. Data show mean ± SD, from three independent experiments, right panel shows immunoblot confirming the p62 and OPTN depletion upon siRNA transfection. The numbers depict the ratio of intensities of specific proteins to that of GAPDH. (J) CFU analysis of H37Rv in WT and p62^KOHOXB8 derived macrophages at 24 hpi. Data show mean ± SD, from three independent experiments. The immunoblot shows the levels of p62 in these cells. Fig S3. Gene expression patterns and functional class enrichment. (A) DEGs (differentially expressed genes, absolute log2FC ≥1 & q-value ≤ 0.1) at each time point. Row annotation represents DEGs identified when compared between infected versus uninfected in WT and p62^KD macrophages at 0-, 24- and 48-.hpi. Row clustering was done using ward. D2 methodology on Euclidean distance calculated from the z-scaled values. Heatmap key reflects expression intensity based on z-scaled values (B) shows Gene Ontology Biological Process (GOBP) enrichment of the upregulated or downregulated genes identified based on infected versus uninfected in WT and p62^KD macrophages at 0-, 24- and 48-.hpi. Significance is determined by nominal p.value ≤ 0.05. Only the top 15 enriched GOBPs from each comparison are visualised. Coverage represents the proportion of genes identified in a particular GOBP out of the total genes associated with that process. Colour intensity represents the nominal p-value gradient. (C) is a subset of (A) which shows an expression of the identified DE cytokine genes between infected versus uninfected in WT and p62^KD macrophages at 0-, 24- and 48-.hpi. Fig S4. Impact of p62/SQSTM1 depletion on mitochondrial DNA content and fission-fusion pathways. (A) Quantification of mtDNA by normalising the gene expression of the mitochondria encoded protein (mtCO1) with nuclear genome encoded RNA polymerase (RPL13A) in WT and p62^KDTHP-1 macrophages in infected and uninfected THP-1 macrophages at 0-, 24-, and 48-hpi. (B) Fold change in gene expression of mitochondrial fusion proteins MFN1, MFN2 and OPA1, and fission proteins DRP1 and FIS1 by RT-PCR in H37Rv infected WT and p62^KD THP-1 macrophages at 24 hpi. (C) Immunoblots represent protein expression of BNIP3 in infected WT and p62^KD cells at 24 hpi. Fig S5. Effect of mitochondrial quality control in H37Rv infected THP-1 macrophages. (A-E) Confocal images depict the interaction of TOM20 (green) with proteins from three different mitochondrial compartments (red). Representative TOM20 3D structures and mitochondrial protein spots: (A) PDHA1, (B) HSP60, (C) SO2, (D) ECSIT, and (E) MCU are shown at the right. The yellow spots are at <0.2 μm, and the magenta spots are at > 0.2 μm from TOM20 structures. (F) Immunoblot show the levels of RAB7 in siControl and siRAB7 in THP-1 macrophages at 24 hours post-siRNA transfection. (G) Immunoblot show the levels of PARKIN in siControl and siPRKN in THP-1 macrophages at 24 hours post-siRNA transfection. (H) Per cent MFI change in TMRE of infected THP-1 macrophages in the presence and absence of nigericin. (I) The relative oxidation state of Mrx1-roGFP2 reporter as the ratios of emission MFI at 510nm when excited at 405nm to 488nm in nigericin-treated WT and p62KD cells at 24 hpi. Data show mean ± SD, n=3. (J) H37Rv CFU in Nigericin-treated and untreated WT and p62KD cells. Data show mean ± SD, n=4, from two independent experiments. (K) The graph represents the MTT cell viability assay of nigericin-treated and untreated WT and p62KD THP-1 macrophages. Scale bar: 10 μm. Fig S6. Role of MIRO1 and MIRO2 in MDV biogenesis. (A) CFU analysis of H37Rv in control and siRNA-targeted-MIRO1, MIRO2, DNM1L (DRP1) and SNX9 WT and p62^KD macrophages at 24 hours post-transfection, Data show mean ± SD, n=5, from two independent experiments. (B-C) Immunoblot represents the siRNA targeted depletion of MIRO1 and MIRO2, respectively at 24 hours post-transfection.

Supplementary tables. Table S1: Infection-associated changes at different hours. Table S2: Enrichment analysis of the DEGs. Table S3: DEGs identified to be associated with mitochondria based on Mitocarta 3.0 (a subset of total DEGs). Table S4: Mitochondrial DEGs associated with mitochondria pathways based on Mitocarta 3.0. Table S5: DEGs identified to cytokine genes (a subset of total DEGs).