INTRODUCTION
Tuberculosis is one of the top ten causes of death worldwide (WHO, 2019). While a vaccine against the etiological agent, the intracellular bacterium Mycobacterium tuberculosis (Mtb), has been available for almost one hundred years, it is ineffective for the population most at risk for pulmonary infection (Pai et al. , 2016, Philips & Ernst, 2012). Furthermore, microbial resistance against frontline antibiotics continues to rise (Gygli et al. , 2017, WHO, 2019). Therefore, increasing knowledge about how Mtb interacts with – and survives within – host cells during infection may lead to new vaccine targets and treatment options, which are desperately needed to control the global tuberculosis incidence.
As an intracellular pathogen, M. tuberculosis has been shown to manipulate or hijack cellular processes within host cells to promote its own survival upon phagocytic uptake. One of the best studied strategies employed by the bacteria to evade destruction is to prevent fusion of the bacteria-containing compartment with lysosomes (Armstrong & Hart, 1971, Via et al. , 1997). This process is dependent on bacterial factors secreted through the SecA2 pathway of Mtb (Zulauf et al. , 2018) as well as perturbation of phagosomal pH (reviewed in (Westman & Grinstein, 2020)). Due to their essential roles in host cellular processes such as energy production, calcium storage, and induction of cell death pathways, mitochondria are also key organelles targeted by pathogens. M. tuberculosis disrupts mitochondrial membrane potential post-infection resulting in the aberrant release of cytochrome c (Abarca-Rojano et al. , 2003, Chen et al. , 2006, Duanet al. , 2002) and disruption of cellular bioenergetics parameters, including ATP and ROS concentrations (reviewed in (Mohareeret al. , 2020)). Strikingly, these Mtb-dependent biochemical changes of mitochondrial metabolism are often accompanied by gross alterations in mitochondrial morphology. In healthy cells, mitochondria are dynamic organelles concurrently exhibiting both tubular and spherical morphologies as a result of continuous homotypic fusion and fissioning to maintain proper metabolic homeostasis. During Mtb infection of the human monocytic cell line, THP-1 (Asalla et al. , 2017) and A549 airway epithelial cells (AECs) (Fine-Coulson et al. , 2015), mitochondria appeared fragmented with small spherical structures clustered around the nucleus. In contrast, TEM micrographs of THP-1 macrophages infected with virulent Mtb caused mitochondria to become elongated and increased in electron density (Jamwal et al. , 2013).
Although less studied, lipid droplet (LD) homeostasis is also disrupted during Mtb infection, but the role that LDs play in the infection has not been well characterized. Once believed to be solely involved in lipid storage, it is now recognized that lipid droplets are dynamic organelles that respond to the conditions of the cellular environment. LDs are composed of a phospholipid monolayer and a neutral lipid core, which contains triacyglycerol (TAGs) and cholesterol esters (CE) (Walther et al. , 2017) and can interact with other organelles through membrane contact sites, expanding or contracting depending on cellular metabolism (Pol et al. , 2014). Several studies have shown that M. tuberculosis modifies the host environment to induce the formation of lipid droplets and increase bacterial intracellular survival, most notably within foamy macrophages (D’Avilaet al. , 2006, Peyron et al. , 2008, Singh et al. , 2012, Singh et al. , 2015). In conditions that mimic the hypoxic environment within granulomas, PBMC-derived macrophages infected with Mtb accumulate lipid bodies and utilize host lipids for accumulating triacylglycerol inside the bacteria (Daniel et al. , 2011). The mechanism underlying the manipulation of host cell lipids and induction of lipid body synthesis by Mtb requires further investigation.
Mtb proteins that are secreted into the host cell cytoplasm are perfectly localized to interact with host organelles and induce the changes seen in the endo-lysosomal system, mitochondria and lipid droplets (reviewed in (Augenstreich & Briken, 2020)). Early efforts to identify these proteins relied mainly on biochemical isolation and characterization (Wiker & Harboe, 1992). One of the earliest characterized and most intensively studied of the cohort of secreted Mtb proteins is the 6 kDa early secretory antigenic target (ESAT-6) (Sorensen et al. , 1995). Several studies have shown that ESAT-6 plays an integral role in M. tuberculosis escape into the host cytosol by rupturing phagosomal membranes (Houben et al. , 2012, Smith et al. , 2008). In addition, the actions of ESAT-6 may also be important in allowing cytosolic access to other secreted proteins whose targets lie outside of the bacterium-containing compartment (Stammet al. , 2019).
Alternative methodologies designed to identify secreted Mtb proteins and their interactors have been developed using genetic, proteomic and transposon screens (Barczak et al. , 2017, Cao et al. , 2019, McCann et al. , 2007). A unique approach to characterizing some of these secreted proteins was developed by Stamm et al (Stammet al. , 2019) and specifically identified mycobacterial secreted proteins that interact with yeast membranes and disrupt eucaryotic secretory pathways. Using GFP fusions of predicted Mtb secreted effectors found to bind membranes in yeast, they then localized each of these proteins to organelles within human HeLa cells. Although the largest proportion of the fusion proteins colocalized with the ER resident protein, calreticulin, several others were observed to localize to the Golgi apparatus, mitochondria, and peroxisomes. When considering secreted Mtb proteins that could both bind membranes and disrupt host protein secretion pathways, only five such proteins remained. Four localized to the ER, but one was found to localize to mitochondria:Rv1646 (PE17).
PE17 belongs to the family of PE (Pro-Glu) proteins, which along with the PPE (Pro-Pro-Glu) family, account for approximately 7% of theM. tuberculosis genome (Cole et al. , 1998, Fishbeinet al. , 2015). PE17 is evolutionarily grouped with the PE_PGRS proteins, although it lacks the characteristic PGRS domain (Talaricoet al. , 2008). PE_PGRS proteins have been found to be secreted through ESX-5, and contain a YXXXE/D ESX general signal sequence C-terminal to the PE domain (Daleke et al. , 2012). While many of the PE/PPE proteins remain uncharacterized, there are some indications that they may be involved in virulence. The best studied member of the PE family, PE_PGRS33, has been shown to mediate M. tuberculosisentry into the macrophage, induce TNF-α expression, and cause cell death (Basu et al. , 2007, Cadieux et al. , 2011, Palucci et al. , 2016). Several studies have reported the colocalization of PE_PGRS33 with mitochondria and effects on mitochondrial morphology (Aguilar-Lopez et al. , 2019, Cadieux et al. , 2011, Moreno-Altamirano et al. , 2012). Additionally, there is evidence that PE/PPE family proteins co-evolved with the ESX gene regions (Gey van Pittius et al. , 2006) and that PE_PGRS and PPE_MTPR subfamilies are secreted in an ESX-5 dependent manner, including PE17 (Abdallah et al. , 2009). Despite these intriguing data, many members of the PE/PPE families remain uncharacterized.
Recently, Li et al. published the first article directly characterizing PE17 (Li et al. , 2019). The authors expressed PE17 in non-pathogenic Mycobacterium smegmatis , which does not natively encode the protein and does not persist intracellularly. It was found that expression of PE17 significantly increased bacterial survival and caused higher rates of necrosis in mouse peritoneal macrophages. Expression of PE17 also significantly decreased the degree to which the proinflammatory cytokines IL-6, IL-2, and TNF-α were induced during infection. This finding, coupled with the fact that PE17 was identified as being expressed during M. tuberculosis infection in macrophages (Vipond et al. , 2006), strengthens the hypothesis that PE17 plays an important role in Mtb pathogenesis.
As we have previously observed that an Mtb ESAT-6 deletion strain does not disrupt mitochondrial dynamics or mass upon infection of A549 AECs when compared to the wild type H37Rv strain (Fine-Coulson et al. , 2015), we addressed the following two hypotheses in the present study: 1) ESAT-6 directly disrupts mitochondrial membranes, particularly in light of its pore-forming ability; 2) ESAT-6 pore formation in the bacteria-laden compartment is required for the release of another protein that directly impacts mitochondrial dynamic, such as PE17. Accordingly, we sought to learn more about mitochondrial changes duringM. tuberculosis infection by heterologous expression of both ESAT-6 and PE17 in A549 AECs. Surprisingly, we report that expression of ESAT-6 alone does not result in mitochondrial changes. However, PE17 expression leads to alterations in mass and distribution not only of mitochondria, but of a variety of different organelles. Furthermore, PE17 specifically localizes to intracytoplasmic lipid droplets, suggesting that PE17 has the capacity to directly alter host membrane dynamics.