Autophagic clearance of bacterial pathogens: molecular recognition of intracellular microorganisms

Pathogens that lyse and escape from the containing vacuole

Pathogens that modify the phagosomal compartment

Pathogens that survive in acidic/degradative compartments

Molecular components

In eukaryotic cells three major types of autophagy have been described: macroautophagy, microautophagy, and chaperone-mediated autophagy (CMA). In microautophagy, portions of the cytoplasm are engulfed by invagination, protrusion and fission of the lysosomal membrane for a review see (Mijaljica and Devenish, 2011; Li et al., 2013). In contrast, in CMA unfolded and soluble proteins are translocated directly across the limiting membrane of the lysosome. On the other hand, macroautophagy is a conserved cellular degradation process in which portions of cytoplasm and organelles are sequestered into a double-membrane vesicle (i.e., autophagosome), and delivered by a vesicular transport event to a degradative organelle (i.e., the vacuole/lysosome), for breaking down and eventual recycling of the resulting macromolecules. In this review we will focus on macroautophagy, hereafter referred to as autophagy. The basic function of autophagy is in the turnover of long-lived proteins and the removal of protein aggregates or damaged organelles but it also has diverse roles in innate and adaptive immunity, such as resistance to pathogen invasion. Autophagy can be induced as a cellular response to various stress conditions, such as nutrient limitation and oxidative stress (Levine and Klionsky, 2004).

In mammalian cells, the first structural step in autophagy is the formation of the phagophores which are initiated as membrane curved prolongations called isolation membranes. These structures later become autophagosomes with a double membrane. The endoplasmic reticulum (ER) is believed to be involved in the origin of the phagophore although it has been demonstrated that other compartments also take part (Axe et al., 2008; Zoppino et al., 2010; Militello and Colombo, 2011). Subsequently, the autophagosomes undergo maturation into autolysosomes by fusion with endo/lysosomal organelles, the inner single membrane is released into the lumen and subsequently the captured cytoplasmic targets are degraded (Ohsumi, 2001; Levine and Klionsky, 2004; Klionsky, 2007; Mizushima et al., 2008).

There are a number of signaling complexes and pathways involved in the regulation of the initiation and maturation of autophagy. Also several autophagy-related (ATG) genes essential to drive this cellular process have been identified, both for selective and nonselective types of autophagy. Identification of the ATG genes in yeast, and the finding of orthologs in other organisms, has revealed the conservation of the autophagic machinery in all eukaryotes.

The core proteins and complexes that constitute the autophagic machinery and their main roles in autophagy are described below. (Figure 1).

Autophagy regulation

Autophagy is a highly regulated process that plays an important role in cellular homeostasis under basal levels through the elimination of damaged organelles as well as the turnover of long-lived proteins. On the other hand, under stress conditions, such as nutrient starvation, hypoxia, oxidative stress, pathogen infection, the level of autophagy elevates as a response, resulting in adaptation and survival; however, deregulated or excessive autophagy may lead to cell death (Yoshimori, 2004). Thus, defective autophagy has been implicated in the pathogenesis of diverse diseases, such as certain types of neuronal degeneration and cancer, and also in aging. More recently, in epithelial cells from cystic fibrosis (CF) patients, characterized by a dysfunctional CF trans-membrane conductance regulator (CFTR), a marked inhibition of autophagy has been observed (Luciani et al., 2010). Therefore, CF is considered a disorder with defective autophagy which is believed to contribute to the pathogenesis of the CF.

The regulation of the autophagic activity to prevent an unbalanced situation can be controlled by several complexes. As mentioned above, mTOR is a kinase that acts as a central sensor of growth factors, nutrient signals, and energy status and serves as a master regulator of autophagy. mTORC1 downregulates autophagy through the class I PI3K-protein kinase B (PKB, also known as Akt) pathway. Thus, the PI3K-I/PKB pathway is involved in the negative modulation of autophagy and it functions at the plasma membrane.Class I PtdIns3K is activated by activated receptor tyrosine kinases that autophosphorylate upon association with growth factors leading to an activation of PKB that phosphorylates a downstream protein complex, TSC2, activating mTORC1 (Vanhaesebroeck and Alessi, 2000; Brazil and Hemmings, 2001; Axe et al., 2008).

Apart from being a nutrient sensor, mTOR can also sense changes in the cellular energy levels via AMPK (AMP-activated protein kinase) (Noda and Ohsumi, 1998).AMPK is a sensor of hypoxia bioenergetics, and is activated by a decreased ATP/AMP ratio produced during nutrient and energy depletion. Active AMPK leads to phosphorylation and activation of TSC1/TSC2 thus, inhibiting mTORC1 activity (Hoyer-Hansen and Jaattela, 2007). On the other hand, AMPK also inactivates mTORC1 by direct phosphorylation of the mTOR binding partner raptor which is important for the inhibition of mTOR (Gwinn et al., 2008).In addition, AMPK directly phosphorylates and activates the complex ULK1/2, resulting in the activation of autophagy (Kim et al., 2011).

Beclin 1/hVps34 complex contributes to autophagosome formation by allowing other Atg proteins to relocate to the pre-autophagosomal structure. Beclin 1 acts as a platform, recruiting activators or repressors of Beclin 1/hVps34-dependent autophagy and the whole complex can be modulated by negative or positive regulators. Beclin1 was initially identified as a binding partner of the anti-apoptotic protein Bcl-2, an interaction that inhibits autophagy by preventing the association of Beclin1 with Vps34. Bcl-2 proteins do not directly compete with Vps34 for binding to Beclin 1, since they bind to the Bcl-2 binding domain of Beclin 1, whereas Vps34 is thought to bind to its EC domain. Hence, the binding of Bcl-2 to Beclin1 is tightly regulated and widely exploited by pathogens and by other regulatory processes. On the other hand, several proteins have been discovered to be active components of the pro-autophagic multimolecular complex: (i) UVRAG and Beclin 1 interact directly through their coiled-coil domain (Itakura et al., 2008; Liang et al., 2008) and is negatively regulated by Rubicon. (ii) AMBRA1 (the Activating Molecule in Beclin 1-Regulated Autophagy) stabilizes the association of Vps34 with Beclin 1(Fimia et al., 2007). (iii) The small GTPase Rab5, a regulator of early endocytosis, also interacts with and activates Vps34 in its complex with Beclin 1 and increases autophagosome formation. The PI3P generated by the Beclin 1/hVps34 complex allows the recruitment of Wipi (Atg18) via a PH domain and DFCP1 via a FYVE domain (Itakura and Mizushima, 2010; Matsunaga et al., 2010), which are important players in autophagosome formation.

It is important to take into account that a significant crosstalk between apoptosis and autophagy exits. As explained above the interaction between Bcl-2and Beclin1is key for the switch between autophagy and apoptosis. However, proteins that were originally thought to belong to the autophagic pathway have been found to modulate apoptosis either by inhibiting or inducing this type of cell death. For example ATG5, that plays an essential role in inducing autophagy, also functions as a pro-apoptotic protein when is cleaved by calpain (Yousefi et al., 2006).

As indicated above, autophagy is defective in epithelial cells from CF patients (Luciani et al., 2010). Interestingly, as a consequence of mutations on the CFTR the conformational defective protein leads to autophagy inhibition through the sequestration of Beclin 1 into aggresomes mediated by a ROS-transglutaminase complex [for a review see Villella et al. (2013)]. In addition, these autophagy deficient cells also accumulate the ubiquitin binding protein p62 (please see below), causing proteosome overload and aggresome formation contributing to hamper the clearance of the misfolded proteins. This situation generates a vicious loop with impeded autophagy and proteosomal degradation.

Selective recognition of bacterial pathogens by autophagy: molecular machinery involved

Autophagic adaptors are proteins that participate in the recognition of pathogens by the autophagic machinery. These adaptors interact with both, components of the autophagic machinery and specific cargos targeted for autophagy degradation (Bjorkoy et al., 2005; Kirkin et al., 2009).The adaptor proteins binds both ubiquitin by means of an ubiquitin-associated (UBA) domain and ATG8/LC3 via an identified WXXL-like sequences known as AIM (Atg8-family interacting motif) and LIR (LC3 interacting domain region), respectively. LIRs consist of a beta-strand containing the WxxL motif that forms an intermolecular beta-sheet with Atg8/LC3 (Noda et al., 2010). These motifs are present in the protein p62 and neighbor of BRCA1 gene1 (NBR1) involved in autophagic degradation of protein aggregates, as well as in Atg32 and Nix required for mitophagy and in Atg19 for the cytoplasm-to-vacuole targeting pathway, linking the cargo to the autophagosome generating machineries [for a revision see Noda et al. (2010)].

As mentioned above, p62 (also known as SQSTM1) is an adaptor protein with multiple protein-protein interaction domains, including an UBA domain for binding to ubiquitinated cargo and a LIR domain for binding to LC3.The protein p62 was initially known as a molecular adaptor involved in the recognition of toxic aggregates to be deliver to the autophagic pathway (Bjorkoy et al., 2005; van der vaart et al., 2008; Komatsu and Ichimura, 2010). However, soon a wider role was uncovered demonstrating its participation in the detection of intracellular pathogens, as a molecular component responsible for the targeting of certain microorganism toward the autophagy pathway, revealing a novel function for p62 in innate immunity. Thus, the cell uses a conserved pathway for both the surveillance of misfolded proteins and intracellular bacteria.

Salmonella enterica serovar Typhimurium (S. Typhimurium) is known to manipulate the fate of its SCV by injecting effector proteins into the host cytoplasm via its Type III secretion system (Hueck, 1998; Knodler and Steele-Mortimer, 2003), and (Steele-Mortimer et al., 2002; Knodler et al., 2003). However, a small population of Salmonella that is unable to establish a stable SCV, is released into the cytoplasm after damaging its vacuolar niche. Upon entry into the mammalian cytosol the bacteria become decorated by a layer of polyubiquitinated proteins (Perrin et al., 2004; Birmingham and Brumell, 2006). It was shown that p62 is recruited to S. typhimurium and associates with ubiquitinated proteins localized to the bacteria (Zheng et al., 2009).Expression of p62 is required for efficient autophagy of the pathogen and for the control of bacteria intracellular replication. Interestingly, p62 is not the only adaptor protein to recognize intracellular Salmonella. The ubiquitin-coated cytosolic bacteria are also recognized by NDP52 (nuclear dot protein 52 kDa), an adaptor protein that binds both ubiquitin and LC3, which in turn, targets them for autophagy contributing to innate immunity (Thurston et al., 2009).NDP52 by binding the adaptor proteins Nap1 and Sintbad, recruits TBK1 (Tank-binding kinase 1)an IKK family kinase. Knockdown of NDP52 and TBK1 facilitates bacterial proliferation and impairs autophagy of Salmonella (Thurston et al., 2009). Interestingly, it has been shown that the two adaptor proteins, p62 and NDP52, are recruited at the same time (i.e. 60–90 min post infection) to bacteria-associated microdomains independently of each other and that both are required for antibacterial autophagy of Salmonella enterica (Cemma et al., 2011).Of note, double knockdown of both effectors do not have an additive effect on autophagy impairment indicating that they function in the same pathway and are not redundant (Cemma et al., 2011).Other pathogens that occasionally invade the cytoplasm such as Streptococcus pyogenes are also restricted by NDP52 and TBK1. Recently, the mechanism by which TBK1 restricts bacterial proliferation has been uncovered. Bacterial wall components such as lipopolysaccharide (LPS) activates TBK1 via the Toll-like receptor 4 (TLR4), a pattern recognition receptor (Wild et al., 2011).The activated TBK1 phosphorylates another autophagic receptor, optineurin (OPTN) which in turn binds to LC3. Indeed, TBK1-mediated phosphorylation of OPTN at Ser177 increases its affinity for LC3, targeting ubiquitin-coated cytosolic Salmonella to autophagosomes and favoring its elimination by autophagy. As expected, silencing of OPTN leads to Salmonella proliferation confirming the requirement for OPTN for bacterial restriction by autophagy (Wild et al., 2011). A role for TBK-1 in the autophagic elimination of Mycobacterium tuberculosis var. bovis BCG has been also recently reported (Pilli et al., 2012). It was shown that TBK-1 coordinated the assembly and function of the autophagic machinery and phosphorylated the autophagic adaptor p62.

Interestingly, OPTN colocalized with TBK1 and NDP52 but not with p62 on the surface of Salmonella that has escaped to the cytoplasm, supporting the idea that several molecular mechanisms participate in bacteria clearance (Weidberg and Elazar, 2011). In a recent publication it was shown that the cytosolic lectin galectin 8 detects damaged pathogen-containing vacuoles. Galectin 8 mediates the recruitment of NDP52 at early infection times followed by an ubiquitin-dependent NDP52 recruitment (Thurston et al., 2012).Thus, this study highlights the recruitment of NDP52 mediated by two different molecules in a sequential fashion to ensure the proper clearance of cytoplasmic Salmonella. (Figure 2).

It was expected that those intruders that are well-adapted to thrive in the cytoplasm such as Shigella flexneri were not restricted by NDP52. Indeed, in a previous publication S. flexneri was indicated not to colocalize with ubiquitin or NDP52 (Thurston et al., 2009) and consistently, NDP52 depletion did not affect S. flexneri proliferation. However, more recently Pascal Cossart and collaborators have shown that p62 and NDP52 target Shigella to an autophagy pathway dependent upon actin and septin, a newly characterized component of the cytoskeleton. Septin filaments form a cage in a myosin-dependent manner which traps cytosolic Shigella leading to bacteria restriction by autophagy (Mostowy et al., 2011). Indeed, the formation of septin cages wrapping bacteria that colonize the cytosol occurred concurrently with the acquisition of autophagy markers (Mostowy et al., 2010). In contrast, the Listeria ActA mutant is targeted to an autophagy pathway by p62 or NDP52 but independently of septin or actin (Mostowy et al., 2011). Thus, it seems that the proteins p62 and NDP52 drive intracytoplasmic Shigella and Listeria to different autophagic pathways.

In a study by Ogawa and collaborators (Ogawa et al., 2011), another autophagy cargo receptor has been identified but in this case the binding partner is not LC3 but upstream molecular components of the pathway. Tecpr1 (Tectonin domain-containing protein) binds Atg5 and WIPI-2, a PI(3)P-interacting protein involved in phagophore formation. Tecpr1, colocalized with Atg5 at Shigella-containing phagophores and its activity is necessary for efficient autophagic targeting of bacteria. Interestingly, Tecpr1 is also involved in autophagy clearance of protein aggregates and damaged mitochondria, but it is not required for canonical autophagy induced by starvation or rapamycin (Ogawa and Sasakawa, 2011).

In addition to ubiquitin, a role for diacylglycerol (DAG)-dependent signaling cascade in antibacterial autophagy has been described (Shahnazari et al., 2010). In this report the authors have shown that DAG production was necessary for efficient autophagy of Salmonella. DAG was localized to bacteria-containing phagosomes and preceded the recruitment of autophagic markers. Of note, the Salmonella-containing autophagosomes colocalized independently with either ubiquitin or DAG, indicating that the DAG-signaling pathway is an alternative mechanism to control autophagic clearance of bacteria.

Another signaling molecule that has an important role in both phagosome and autophagosome formation and maturation is ROS (Reactive oxygen species). Sources of ROS are, among others, mitochondria, ER, and peroxisomes. In phagocytic cells an important source of ROS is the NADPH oxidase associated with the phagosome which has a key function as a microbicidal molecule (Babior, 2000; Bylund et al., 2010; Kotsias et al., 2013) DAG is an upstream activator of the NADPH oxidase and ROS production is also required for LC3 recruitment to a self-tailored niche containing L. monocytogenes (Lam et al., 2013); please also see below next section. Interestingly, ROS actively regulates autophagy by targeting autophagy gene products such as the protein Atg4, the hypoxia inducible transcription facto Hif-1a and also the kinases mTOR and MAPKs (mitogen-activated protein kinases). In contrast, ROS production can be controlled via the removal of mitochondria by autophagy (i.e., mitophagy) (for a comprehensive review see Vernon and Tang, 2013).

One important question is how this autophagic process is initiated, since the xenophagic event commonly takes places in cells incubated in full nutrient condition. A critical role for amino acid and mTOR signaling modulation was recently uncovered. In a very recent study, it was demonstrated that infection with Shigella and Salmonella triggered an early state of intracellular amino acid starvation causing the dissociation of MTOR from endomembranes and the downregulation of its activity (Tattoli et al., 2012). Of note, this amino acid starvation signal was caused by host membrane damage, which was differentially elicited by Salmonella (a transient response) in contrast to the sustained starvation signal in Shigella-infected cells.

Evasion from autophagy recognition

If autophagy is such a powerful defense mechanism as described above, how is it possible that certain pathogens have been able to adapt to survive and defend against this weapon machinery? What molecular mechanism are they using to be protected? In this review we will present same examples of how intracellular pathogens of either a vesicular or cytoplasmic lifestyle evade autophagy recognition.

A very interesting autophagy escape mechanism has been unveiled in the case of S. flexneri. Soon after internalization Shigella disrupt and break out from the containing membrane vacuole toward the cytoplasm where the bacteria active multiply and move infecting neighboring cells. This movement is carried out by directing local actin polymerization at one pole of the bacterium which is dependent on the exposure and accumulation of the bacterial protein VirG (Goldberg et al., 1993).VirG interacts with N-WASP which in turn via the Arp2/3 complex induces actin polymerization. A study revealed that VirG, via interaction with the autophagy protein Atg5, is a key molecule in the targeting of cytoplasmic Shigella by autophagy (Ogawa et al., 2005). IcsB an effector protein secreted by Shigella via the T3SS (type three secretion system) is capable of binding VirG inhibiting competitively the binding of Atg5 to VirG, preventing autophagy targeting of the bacterium. In contrast, in the icsB mutant, Atg5 binds to the VirG and cytoplasmic Shigella are recognized by autophagy being trapped in acidic LC3-positive compartments. As a consequence the icsB mutant presents a phenotype characterized by deficient ability to multiply intracellularly. Thus, the binding of IcsB to VirG which prevents the interaction with Atg5 at one pole of the bacterium would serve as a shield protecting Shigella against autophagic recognition. Interestingly, it has been recently shown that Shigella uses another device to escape from autophagy (Dong et al., 2012). The VirA effector exhibits potent RabGAP activity and specifically inactivates the Rab protein, Rab1 disrupting ER-to-Golgi trafficking. Rab1 has been shown to be required for autophagosome formation (Zoppino et al., 2010) thus, via VirA Shigella negatively modulates the autophagic pathway avoiding this pathogen clearance mechanism.

Listeria monocytogenes also uses an actin-based motility to escape from autophagy. ActA is the bacterial protein responsible for actin polymerization once the bacterium has reach the cytoplasm. It has been shown that a non-motile actA mutant of L. monocytogenes treated with the bacteriostatic antibiotic chloramphenicol is trapped by autophagy in the cytoplasm of infected macrophages (Rich et al., 2003). Although, ActA expression was sufficient to prevent bacteria targeting by autophagy in the cytoplasm of macrophages, ActA expression was not strictly necessary since an actA mutant was competent to evade autophagy in the absence of chloramphenicol treatment, indicating that other virulence factors were involved (Birmingham et al., 2007). It was found that the bacterial phospholipases, PI-PLC, and PC-PLC play an important role in autophagy evasion. More recently Yoshikawa and collaborators have demonstrated that recruitment to the bacterial surface of the Arp2/3 complex and Ena/VASP, mediated by ActA, camouflages the bacteria from autophagic recognition, regardless of the bacterial capacity for movement. Thus, an ActA mutant bacteria unable to recruit these actin nucleating proteins were ubiquitylated and subsequently, recruited p62 and LC3, undergoing autophagy (Yoshikawa et al., 2009a,b).

Interestingly, Listeria employs another strategy to avoid autophagy. Internalins are a family of proteins mostly expressed at the surface of Listeria (Glaser et al., 2001). The role of one member of this family, InlK was recently demonstrated as a virulence factor important for bacteria infection in vivo (Dortet et al., 2011). The Major Vault Protein (MVP) was identified as a host partner of InlK. The interaction between InlK/MVP takes place at the bacterial surface in the cytoplasm of infected cells independently of actin polymerization. Of note, it was found that the recruitment of MVP prevents L. monocytogenes from autophagic recognition, leading to enhanced bacterial survival in infected cells. Thus, L. monocytogenes utilizes more than one mechanism to avoid targeting by the autophagic pathway during colonization of host cells (Figure 3). Of note, a population of L. monocytogenes that seems to remain in the phagosome forms Spacious Listeria-containing Phagosomes (SLAPs), which are large endosome- or autophagosome-like compartments containing multiple bacteria. These self-tailored compartments do not mature but require bacterial LLO and host autophagy to form. L. monocytogenes replicate in these phagosomes, though at a much slower rate compared to cytoplasmic bacteria and these compartments have been associated with persistent infection (Lam et al., 2013)

The Gram-negative tularemia-causing bacterium Francisella tularensis is a facultative intracellular pathogen capable of surviving and growing in several mammalian host cells. The bacteria are phagocytosed but escape the phagosome a few hours after infection (Clemens et al., 2004; Chong et al., 2008) actively multiplying in the cytoplasm although the pathogen seems to occupy different compartments during its intracellular cell cycle [for a revision see Chong and Celli (2010)]. Prior to escaping from the phagosome, Francisella initially resides within a vacuole that interacts with early and late endosomes but it does not recruit lysosomal markers (Clemens et al., 2004; Checroun et al., 2006; Chong et al., 2008; Santic et al., 2008; Wehrly et al., 2009). Regarding the relationship with the autophagic pathway apparently there is no evidence for an autophagic response, suggesting that the pathogen avoids its recognition by the autophagic machinery or it may inhibits the autophagic response. Indeed, it has been shown that several autophagy-related genes (ATG) are downregulated during infection of human monocytes suggesting that Francisella suppresses autophagy at the gene expression level (Butchar et al., 2008; Cremer et al., 2009). Interestingly, it has been shown in murine BMMs that a subpopulation of cytosolic F. tularensis is capable of reentering the endocytic pathway residing in a compartment with autophagosomal features by 24 h after infection (Checroun et al., 2006). In a recent publication it has been shown that a replication-deficient mutant is captured from the cytoplasm of murine and human macrophages into double-membrane vacuoles colocalizing with the late endosomal marker, LAMP1, and the autophagic protein LC3, limiting intracellular bacteria viability (Chong et al., 2008). These vacuoles were also labeled with ubiquitin and the autophagy adaptors p62 and NBR1. Thus, these findings suggest that Francisella avoids autophagic recognition to ensure bacterial growth in the host cells. Nevertheless, additional work is required to fully elucidate the complex interplay between Francisella and the autophagic pathway.

Similar to Francisella that suppresses autophagy by downregulating autophagy at the gene expression level, another bacterial pathogen Burkholderia cenocepacia, a pathogen that causes severe and persistent infections in CF patients, also downregulates the expression of critical autophagy genes such as Atg12, Atg5, and Atg8 (Abdulrahman et al., 2011). It is interesting to mention that autophagic dysfunction in infected cells is even more pronounced in CF macrophages compared to WT macrophages. This is consistent with the observation that epithelial cells from CF patients have a defective autophagic response even in non-infected cells as we have described above (Luciani et al., 2010). Thus, limited formation of autophagosomes in CF macrophages promotes B. cenocepacia survival. In contrast, autophagy stimulation by rapamycin treatment enhances the targeting of B. cepacia to the lysosomal compartment. (Abdulrahman et al., 2011; Li et al., 2013) In addition, it has been shown that B. cepacia inactivates the small GTPase Rab7 (Huynh et al., 2010), a key molecule required for the maturation and completion of the autophagic pathway (Jager et al., 2004; Gutierrez et al., 2005). Thus, inactivation of this Rab protein is part of the strategy used by B. cepacia to avoid autophagic clearance.

In a previous work we have shown that activation of autophagy by starvation or by other means (i.e., rapamycin-treatment) results in autophagic clearance of Mycobacterium bovis BCG and M. tuberculosis (Gutierrez et al., 2004). In addition, we have also shown that M. marinum, a fish pathogen that causes similar granulomas in the hands of infected humans to those caused by M. tuberculosis in the lungs, was able to induce the recruitment of LC3 to the M. marinum-containing phagosomes (Lerena and Colombo, 2011). However, these compartments were devoid of lysosomal enzymes indicating that fusion with lysosomes was prevented. Interestingly, this LC3 recruitment was dependent on a functional Esx-1 secretion system (Lerena and Colombo, 2011). Consistent with our results, in a recent publication Romagnoli and collaborators have found that the pathogenic M. tuberculosis strain Mtb H37Rv hampers the fusion of autophagosomes with lysosomes (Romagnoli et al., 2012). In contrast, the attenuated strains Mtb H37Ra or BCG, which are deficient in components of the ESX-1 secretion system (incompetent to secrete the protein ESAT-6) were unable to prevent autophagosome maturation. The ability to inhibit the autophagic flux was recovered in recombinant BCG and Mtb H37Ra strains in which the ESX-1 region was restored by genetic complementation. In another report it has been shown that ectopic expression of the ESAT-6/CFP-10 fusion in macrophages hampered autophagosome formation increasing M. tuberculosis viability. Interestingly, expression levels of ATG also diminished, suggesting that the fusion protein may modulate autophagy through the regulation of ATGs (Zhang et al., 2011). Of note, in a recent publication it was shown that phagosomal piercing mediated by the ESX-1 secretion system allows cytosolic components of the ubiquitin-mediated autophagy pathway access to phagosomal M. tuberculosis and also the recognition of extracellular bacterial DNA by the STING-dependent cytosolic pathway (Watson et al., 2012)

It has also been reported that the “enhanced intracellular survival” (eis) gene of M. tuberculosis plays essential roles in modulating autophagic and inflammatory responses in macrophages (Shin et al., 2010). Cells infected with an Mtb eis-deletion mutant H37Rv (Mtb- Δeis) presented a marked accumulation of autophagic vacuoles. In addition, the presence of this mutant bacterium in typical double-membrane autophagic structures was visualized by electron microscopy analysis. Thus, the protein eis seems to act as a negative regulator of autophagy contributing to avoid autophagic clearance of M. tuberculosis. In a recent publication Roy and collaborators have found that Legionella secretes via the Type IV secretion system an effector protein called RavZ (Choy et al., 2012).This factor disrupts the autophagy pathway by cleaving the C-terminal region of lipid-conjugated LC3/Atg8. Thus, Legionella inhibits autophagy in the host cell by direct manipulation of a key protein required for autophagosome formation.

Conflict of interest statement

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.