SEARCH

SEARCH BY CITATION

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. The general secretion pathway
  5. The twin-arginine transporter
  6. The ESAT-6 secretion system
  7. Remaining questions
  8. Acknowledgements
  9. References

Mycobacteria have a unique cell-envelope structure which protects the bacteria from the extracellular environment by limiting access to noxious molecules from the outside. This extremely hydrophobic and thick barrier also poses a unique problem for the export of bacterial products. Here we review the multiple protein secretion pathways in Mycobacteria, including the general secretion pathway and the Twin-Arginine Transporter, with an emphasis on the ESX-1 alternate secretion system. This newly identified protein secretion system is required for growth during infection and has provided insight into how M. tuberculosis manipulates the host immune response during infection.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. The general secretion pathway
  5. The twin-arginine transporter
  6. The ESAT-6 secretion system
  7. Remaining questions
  8. Acknowledgements
  9. References

Mycobacteria have a unique cell-envelope structure which insulates the bacteria from the extracellular environment. This envelope includes the inner plasma membrane and a unique cell wall composed of dual polymer layers (peptidoglycan and arabinogalactan) surrounded by a lipid-rich mycolate layer (Brennan and Nikaido, 1995). In the case of mycobacterial pathogens the cell wall plays a critical role in protecting the bacteria from the physical assaults mediated by the mammalian immune system. However, while limiting access to noxious molecules from the outside, this extremely hydrophobic and thick barrier also poses a unique problem for the export of bacterial products. Here we review the multiple protein secretion pathways in Mycobacteria with an emphasis on systems important for virulence. We highlight a newly identified protein secretion system required for growth during infection that has provided insight into how M. tuberculosis manipulates the host immune response during infection.

The general secretion pathway

  1. Top of page
  2. Summary
  3. Introduction
  4. The general secretion pathway
  5. The twin-arginine transporter
  6. The ESAT-6 secretion system
  7. Remaining questions
  8. Acknowledgements
  9. References

Like all other bacteria, Mycobacteria have an essential general secretion pathway (GSP, or the Sec secretion system), which functions to secrete unfolded proteins with N-terminal signal sequences across the cytosolic membrane. The Sec system consists of the SecY, SecE, SecG, SecD and SecF membrane components and the SecA ATPase, which recognizes the signal sequence. As expected, all indications are that this highly conserved system functions in Mycobacteria analogously to that of other bacteria. Mycobacteria, like Gram-positive bacteria, appear to lack the SecB chaperone and thus may use other cytosolic chaperones to escort proteins in their unfolded state from the ribosome to SecA (see Scott and Barnett, 2006). In Gram-negative bacteria, additional pathways are used to direct proteins across the outer membrane, including the Omp85/YaeT complex (see Gentle et al., 2005). However, in Mycobacteria, it is unknown how GSP-dependent substrates get across the mycolylarabinogalactan (mAGP) layer. It seems likely that comparable systems exist in Mycobacteria but this area requires further study.

As in other prokaryotes, the Mycobacterial Sec signal sequence contains three domains including a positively charged N-terminus, followed by a hydrophobic region, and an uncharged polar region. In Gram-negative bacteria, signal sequences are approximately 20 amino acids long, while in Gram-positive bacteria they can be up to 60 amino acids long. Mycobacterial signal sequences resemble those of Gram-positive bacteria, although the functional consequence of this is unclear (Wiker et al., 2000). Interestingly, most proteins secreted from the M. tuberculosis cell contain an aspartic acid-proline sequence at the N-terminus (the DP motif) that is revealed after signal sequence cleavage. Wiker et al. suggest that this motif may be recognized by a second system for translocation across the cell wall. This raises a tantalizing model in which bipartite signal sequences dictate a protein's final localization, analogous to mitochondrial import signals. However, this notion remains untested.

In stark contrast to most Gram-negative bacteria, all sequenced Mycobacterial genomes encode for a second, non-essential homologue of SecA, termed SecA2 (Braunstein et al., 2001; McDonough et al., 2005). An emerging number of Gram-positive bacteria have been found to contain both SecA1 (the SecA homologue with the highest sequence similarity to that of E. coli) and SecA2, including the human pathogen Listeria monocytogenes (Braunstein et al., 2001; Lenz and Portnoy, 2002). In M. tuberculosis, SecA2 is only ∼50% similar and ∼38% identical to SecA1, suggesting that it is not simply a duplicate allele (Braunstein et al., 2001). In Mycobacterium smegmatis, the secA2 mutant has a growth defect on rich media, suggesting that its removal may have pleiotropic effects on cell fitness (Braunstein et al., 2001). The M. tuberculosis mutant grows normally in minimal liquid media but poorly during infections of macrophages and mice (Braunstein et al., 2003; Kurtz et al., 2006). It is tempting to speculate that the SecA2 requirement during infection reflects a specific interaction of SecA2-dependent substrates with host cells. However, the specific role of SecA2 in pathogenesis is unclear.

What are the possible ways that the SecA2 pathway could promote virulence? Clues from two other pathogens with SecA2 raise interesting possibilities. First, in L. monocytogenes, SecA2 functions as an accessory SecA to secrete enzymes that remodel the peptidoglycan layer (Lenz et al., 2003). Breakdown products of digested peptidoglycan catalysed by these enzymes are recognized by host cells and modify the innate immune response to create an environment favouring bacterial colonization (Lenz et al., 2003). Interestingly, the secA2 mutant of M. tuberculosis elicits a more pronounced inflammatory immune response in infected macrophages, consistent with this hypothesis (Kurtz et al., 2006). Second, the SecA2 of Streptococcus gordonii functions to accomodate the secretion of a highly glycosylated adhesion molecule, GspB (Bensing and Sullam, 2002; Bensing et al., 2004; 2005). Thus, the evolution of SecA2 may reflect structural constraints of certain substrates, and perhaps is suited to preferentially translocate proteins that cannot be accommodated by SecA1, such as those with extensive post-translational modifications. Ultimately, understanding the specific effects of this secretion pathway in M. tuberculosis pathogenesis will require identification of the substrates secreted by SecA2.

The twin-arginine transporter

  1. Top of page
  2. Summary
  3. Introduction
  4. The general secretion pathway
  5. The twin-arginine transporter
  6. The ESAT-6 secretion system
  7. Remaining questions
  8. Acknowledgements
  9. References

Mycobacteria also use a twin-arginine transporter (Tat) pathway, a Sec-independent secretion machine that has the unique ability to translocate folded protein substrates across the plasma membrane. Substrates are targeted to the Tat machine by a signal sequence with the same overall structure of the Sec signal sequence, but include a double arginine motif followed by two uncharged residues near the N-terminus (Berks, 1996). Tat is best understood in E. coli, but is also present and functional in Gram-positive bacteria, and is required for the virulence of many pathogens including Pseudomonas aeruginosa, Agrobacterium tumefaciens, enteropathogenic E. coli and Legionella pneumophila (Voulhoux et al., 2001; Ding and Christie, 2003; Pradel et al., 2003; Rossier and Cianciotto, 2005) (for a recent review see Lee et al., 2006). The Mycobacterial Tat system is homologous to Tat systems in other bacteria, and has been shown to be functional in both M. tuberculosis and M. smegmatis (McDonough et al., 2005; Posey et al., 2006). In M. smegmatis, deletion of any of the three main tat genes yields colonies that grow slowly on solid media. However, in liquid media, the mutant cells have no apparent growth defect (McDonough et al., 2005). In M. tuberculosis, the orthologous genes appear to be essential for growth in culture, precluding phenotypic analysis of null alleles in the pathway (Saint-Joanis et al., 2006).

Predictions using various algorithms have estimated from 11 to 31 potential Tat substrates encoded by the M. tuberculosis genome (Cole, 2002; Dilks et al., 2003). Two classes of substrates have received attention and suggest that this pathway may be used to interact directly with the host. First, four phospholipase C enzymes are secreted by Tat and, together, are required for full virulence in the mouse model of M. tuberculosis (Raynaud et al., 2002; McDonough et al., 2005). Although their role in pathogenesis is not entirely clear, two compelling hypotheses are that they act on host membranes either to alter host signalling pathways or to release lipids making them available for the bacterium to catabolize (Munoz-Elias and McKinney, 2006). Second, the Rv2525c protein was recently shown to be secreted by Tat and a Rv2525c mutant displayed a hyper-virulent phenotype during infection of immuno-deficient mice (Saint-Joanis et al., 2006). The mechanism of Rv2525c is not understood and homology searches of the primary amino acid sequence have not revealed any putative functional domains. However, using the Phyre secondary structure matching program (Kelley et al., 2000), we find that this protein likely has structural homology to the transglycosidase protein superfamily, enzymes involved in peptidoglycan synthesis. This is seemingly consistent with the observation that the mutant is slightly more sensitive to certain beta-lactam antibiotics and suggests yet another possible link between peptidoglycan metabolism and virulence.

The ESAT-6 secretion system

  1. Top of page
  2. Summary
  3. Introduction
  4. The general secretion pathway
  5. The twin-arginine transporter
  6. The ESAT-6 secretion system
  7. Remaining questions
  8. Acknowledgements
  9. References

Over the past decade of bacterial pathogenesis research, the alternative secretion systems of Gram-negativebacteria have generated a remarkable amount of attention. In addition to Sec and Tat, Gram-negative pathogens use a variety of secretion systems (Types I-V), which function to deliver bacterial proteins into the host cell and manipulate the host response to infection (for recent reviews on alternate secretion see Backert and Meyer, 2006; Galan and Wolf-Watz, 2006). Substrates of the individual systems differ between pathogens but the secretory systems themselves are mostly conserved. Although homologues of these systems are absent from the M. tuberculosis genome, the likely functional equivalent of these systems, termed ESX-1, has recently been identified and has been a burgeoning focus in the Mycobacterial field.

Initial clues that a specialized secretion system exists in M. tuberculosis came from studies that identified secreted proteins that lack obvious Sec signal sequences (Sorensen et al., 1995). Most notably, ESAT-6 (Early secreted antigen target 6 kDaA) and CFP-10 (Culture filtrate protein, 10 kDa) are two secreted proteins of unknown function originally identified as immunodominant antigens of M. tuberculosis. Since their identification, several studies suggested that these proteins are important for virulence. Deletion of the genes encoding ESAT-6 and CFP-10 from the virulent Mycobacterium bovis strain results in a diminution of virulence (Wards et al., 2000). Furthermore, all strains of the attenuated vaccine strains of bacillus Calmette–Guerin (BCG) have deletions encompassing the ESX-1 locus, also known as the RD1 region (Mahairas et al., 1996). Importantly, deletion of RD1 from M. tuberculosis attenuates the organism and, conversely, incorporation of the RD1 region from M. tuberculosis into BCG restores ESAT-6 and CFP-10 expression and increases virulence and immunogenicity (Pym et al., 2002; Lewis et al., 2003; Pym et al., 2003).

Components of the ESX-1 secretion system

In silico analysis led to the suggestion that the genes surrounding esxBA, the operon which encodes for CFP-10 and ESAT-6, may be important for secretion of these proteins (Gey van Pittius et al., 2001; 2002; Pallen, 2002). Importantly, this notion was proven to be true when individual genes at the RD1 locus were identified in genetic screens to define virulence genes of M. tuberculosis (Hsu et al., 2003; Sassetti et al., 2003; Stanley et al., 2003; Guinn et al., 2004). Disruption of individual genes (Rv3870, Rv3871 and Rv3877) within this locus prevented secretion of ESAT-6 and CFP-10, providing the first genetic evidence that this region encodes for a secretion system (Stanley et al., 2003; Guinn et al., 2004). Rv3870, Rv3871 or Rv3877 mutant strains are phenotypically indistinguishable from RD1 deletion strains; they are attenuated for growth in macrophages and elicit an altered immune response during macrophage infection (Stanley et al., 2003).

The ESX-1 locus is conserved in several pathogenic and non-pathogenic Mycobacterial species, allowing for more rapid progress to be made than in the slow-growing and highly infectious M. tuberculosis. Using Mycobacterium marinum, Rv3868, Rv3878 and Rv3879 were shown to be required for ESAT-6/CFP-10 secretion (Gao et al., 2004). In M. smegmatis, the homologues of Rv3866, Rv3869, Rv3882c and MycP1 are also required for ESAT-6/CFP-10 export (Converse and Cox, 2005) (see Fig. 1).

image

Figure 1. A model for ESX-1 mediated secretion in Mycobacteria. All known components of the ESX-1 system from a number of Mycobacteria are shown here. Those discovered in M. tuberculosis are annotated with the Rv number. Those from M. marinum are designated ‘Mh’ for Marinum homologue. Components identified in M. smegmatis are noted with ‘Sm’. Proteins of unknown function are in light blue. Putative AAA ATPases are in orange. The three known substrates are in Red. SmMycP1 is a putative mycosin-like protease, Rv3866 has potential DNA binding domains, and Rv3872 is a PE related protein. Although no components outside of the cytoplasm are known to date, we have indicated potential components with a question mark.

Download figure to PowerPoint

Finally, using M. bovis BCG and Mycobacterium microti strains complemented with an ‘extended RD1’ region from M. tuberculosis, each gene was systemically disrupted and monitored for changes in ESAT-6/CFP-10 secretion (Brodin et al., 2006). Deleting Rv3872 (PE35) resulted in no expression of ESAT-6 and CFP-10. As found previously, Rv3868 through Rv3871, and Rv3877 were required for secretion of ESAT-6/CFP-10, although not for their expression. Strains missing Rv3864, Rv3867, Rv3873 (ppe68) and Rv3876 had no virulence or secretion defects whereas disruptions in Rv3865 and Rv3866 resulted in attenuation but did not affect ESAT-6/CFP-10 secretion. In contrast to M. smegmatis and M. marinum, Rv3866, Rv3878 and Rv3879 were not required for ESX-1 secretion in these studies. Therefore, although there is some disagreement between these studies, it is clear that most of the genes at the RD1 locus are required for ESX-1 secretion.

In addition to RD1, another genetic other locus has recently been identified that is required for ESAT-6/CFP-10 secretion in M. tuberculosis (MacGurn et al., 2005; Sassetti and Rubin, 2003). Strains bearing disruptions in Rv3616c-Rv1614c are attenuated for growth in the mouse lung and spleen. Rv3616c (EspA) and Rv3615c were found to be secreted into the culture filtrate in an ESX-1-dependent manner (Fortune et al., 2005; J.A. MacGurn and J.S. Cox, unpubl. obs.). Rv3614c is also required for secretion of ESAT-6 and CFP-10 (MacGurn et al., 2005). This operon likely arose from a gene duplication event originating from the RD1 locus as Rv3616c-Rv1614c are homologous to Rv3864, Rv3865 and Rv3866.

Clearly, there are many genes, identified in a number of distinct mycobacterial species, which are required for the export of ESAT-6, CFP-10 and other small proteins from Mycobacteria. Some of these genes encode for proteins with readily predictable function. These include the ATPases Rv3868, Rv3870 and Rv3871, which likely provide the energy required for translocation. MycP1 is a putative transmembrane serine protease whose active site is predicted to be extracytoplasmic. Finally, there are a number of transmembrane proteins without obvious functional domains. Although the role of these proteins is unknown, these findings suggest that the ESX-1 system is very complex and requires multiple protein complexes for its proper functioning. However, a complete catalogue of the required ESX-1 components is likely still not complete.

Molecular mechanisms of ESX-1 secretion

How do the large number of components required for ESX-1 secretion function together to target and translocate substrates across the cytosolic membrane and the cell wall? Studies on how the components and substrates interact have provided insight into the molecular mechanisms of the ESX-1 system. First, ESAT-6 and CFP-10 interact to form a tight dimer (Renshaw et al., 2002; Stanley et al., 2003), and in the mycobacterial cell, these two proteins are interdependent on each other for stability. Second, yeast two-hybrid experiments revealed that Rv3870 interacts with Rv3871, a cytosolic protein, and together these two proteins are thought to function as an AAA ATPase of the SpoIIIE/FtsK family. Rv3871 also interacts with CFP-10, and was hypothesized to escort CFP-10 and ESAT-6 to Rv3870 and Rv3877, a multitransmembrane protein, which may make up the pore that spans the cytosolic membrane (Stanley et al., 2003; Champion et al., 2006). Therefore, it is likely that Rv3871 functions to recognize the CFP-10/ESAT-6 substrate pair, and deliver it in an ATP-dependent manner to Rv3870 which is at the membrane (Fig. 1).

The interaction of Rv3871 with CFP-10 was exploited to identify a C-terminal signal sequence on CFP-10, which is necessary for targeting both CFP-10 and ESAT-6 for secretion (Champion et al., 2006). In fact, single amino acid changes at the extreme C-terminus of CFP-10 blocks the secretion of both proteins. This seven amino acid signal sequence is also sufficient for targeting an unrelated protein, ubiquitin, for secretion (Champion et al., 2006). Interestingly, the solution structure of the ESAT-6/CFP-10 pair revealed that the C-terminal 15 amino acids of CFP-10 is unstructured and does not participate in interactions with ESAT-6 (Renshaw et al., 2005). Thus, the C-terminal signal sequence is unstructured, and resembles a handle by which Rv3871 can grab onto the ESAT-6/CFP-10 dimer and target it for secretion.

A curious observation is that all the known ESX-1 substrates, Rv3616c, Rv3615c, ESAT-6 and CFP-10 are mutually dependent on each other for secretion. Although the basis for this phenomenon is unknown, the answer to this puzzle will likely shed light on the mechanism of substrate recognition and secretion. For example, it may indicate that these four substrates interact prior to secretion, perhaps via Rv3871. Alternatively, these proteins may be components of the secretion machine itself and the true substrates have yet to be identified (Ize and Palmer, 2006).

Role of the ESX-1 secretion system

What role does the ESX-1 system play in pathogenesis? Clearly, this system is a major determinant of Mycobacterial pathogenesis but the way in which this system affects the biology of the host cell is unknown. Many groups have suggested that this system functions to modulate early events during M. tuberculosis infection. ESX-1 mutants are attenuated for growth during the first few days of infection of mice and cultured macrophages, after which the bacteria eventually begin to grow. Despite the late growth, however, these mutants are severely attenuated (Brodin et al., 2006; Guinn et al., 2004; Stanley et al., 2003).

We believe that the ESX-1 secretion system mediates early contact with the host cell, and functions to modulate the host cell immune response. Indeed, ESX-1 mediated secretion is important for controlling the macrophage cytokine response during infection by M. tuberculosis. Whether this is a functional intention of the bacterium or an inadvertent readout of other perturbations on the host cell remains to be elucidated. Additionally, ESX-1 is responsible for the elicitation of the cytokine interferon-beta and the resulting induction of a set of interferon responsive genes by wild-type M. tuberculosis (Stanley et al., 2007). We hypothesize that these differences in macrophage signalling reflects active perturbations of the host cell by molecules secreted by ESX-1. In contrast, Hsu et al. suggest that ESAT-6 functions as a toxin to directly lyse cellular membranes (Hsu et al., 2003). This is consistent with studies demonstrating that the RD1 region is required for tissue necrosis in lungs of infected mice (Junqueira-Kipnis et al., 2006). Further work is required to reconcile these two seemingly very different models of ESX-1 function.

Mycobacterium marinum causes a systemic tuberculosis-like infection in ectotherms and is closely related to M. tuberculosis phylogenetically. Real time studies of M. marinum infections in a zebrafish model revealed that wild-type M. marinum recruits macroph-ages to form granulomas in an RD1-dependent manner (Volkman et al., 2004). This is consistent with ESX-1 functioning to manipulate macrophage responses, perhaps by activating cytokines or chemokines early during infection that lead to macrophage aggregation later during infection. However, M. marinum differs from M. tuberculosis in that following replication in the host phagosome, M. marinum escapes and propels itself through the host cytosol via actin-based motility, much like L. monocytogenes (Stamm et al., 2003). It is thought that the ESX-1 system plays a role in phagosome escape and cell to cell spread in M. marinum, which may be distinct from how this system functions in M. tuberculosis (Gao et al., 2004). These studies also provided no evidence that ESAT-6 functions directly as a pore-forming cytotoxin.

It has also been suggested by many groups that, like alternate secretion systems in Gram-negative bacteria, the ESX-1 system functions to secrete effector proteins, including ESAT-6/CFP-10, directly into the host cell phagosome or cytosol (Lewinsohn et al., 2006; Stanley et al., 2007). The only evidence for this, however, comes from the ESX-5-dependent secretion of PPE41 into the host macrophage by M. marinum (Abdallah et al., 2006). Proving that ESX-1 substrates are secreted directly into the host cell has proven difficult as the known substrates are likely secreted into the macrophage at very low levels and are extremely difficult to epitope tag without disrupting secretion.

The role of the ESX-1 secretion machines in Gram-positive pathogenic bacteria is less clear. ESX-1 secretion is absolutely required for Staphylococcus aureus pathogenesis (Burts et al., 2005) but, in contrast, is not required for L. monocytogenes virulence in mice (Way and Wilson, 2005). This disparity may be due to differences in the pathogenic strategies employed by these two different pathogens or due to limitations in the virulence models (Way and Wilson, 2005).

Any model of the role of ESX-1 in pathogenesis must take into account the fact that this pathway is also present in non-pathogenic Gram-positive and mycobacterial species. One possibility is that the pathway plays a fundamental role in both pathogenic and non-pathogenic organisms, for example, in cell-to-cell communication. In support of this, Flint et al. reported that M. smegmatis ESX-1 mutants display increased conjugation efficiency compared with wild-type cells, and they present indirect evidence that ESAT-6/CFP-10 secretion in trans suppresses this hyperconjugation phenotype (Flint et al., 2004). Alternatively, the ESX-1 pathway may be modular, allowing substrates to evolve for the particular needs of each organism (Converse and Cox, 2005). It is therefore necessary to study the ESX-1 system in both pathogenic and non-pathogenic bacteria to further elucidate the roles this system may play in addition to promoting bacterial virulence.

Parallels between ESX-1 and Gram-negative secretion systems

Although it appears that ESX-1 evolved independently of the Gram-negative systems, we believe there are significant parallels between the mechanisms of the ESX-1 system and Type IV secretion at the molecular level. First, like CFP-10, Type IV substrates, including those of the Dot/Icm system in L. pneumophila and the VirB/D4 system in A. tumefaciens, are targeted for secretion using unstructured C-terminal transport signals (Amor et al., 2005; Christie et al., 2005; Nagai et al., 2005; Vergunst et al., 2005).

Second, Type IV translocators use the coupling protein (CP) to facilitate the interaction between the substrate and the secretion machine, and to link cytosolic and membrane components. CPs are integral membrane proteins of the SpoIIIE/FtsK family of ATPases. Rv3870 and Rv3871 are members of the same family of ATPases, and it has been previously suggested that Rv3870 and Rv3871 function together as a single membrane bound AAA ATPase (Stanley et al., 2003; Guinn et al., 2004).

Finally, ESAT-6 and CFP-10 are targeted for secretion through the ESX-1 system as a substrate pair. Type IV secretion systems have examples of chaperone-substrate pairs that are targeted and secreted from the bacterial cell (Dumenil and Isberg, 2001; Sundberg and Ream, 1999). Specifically, like ESAT-6 and CFP-10, the chaperones of these systems are small proteins (10–15 kDa) with an acidic pI, are encoded by a gene adjacent to the gene that encodes the secreted substrate, and bind the substrate with high affinity. These chaperones typically function to keep the substrate in a secretion competent conformation and to prevent interaction with other proteins or aggregation.

It is therefore likely that the ESX-1 machine represents a novel class of secretion machines in Mycobacteria and Gram-positive bacteria but which share some characteristics of Type IV machines of Gram-negative bacteria.

Additional ESX systems in Mycobacterial genomes

In addition to the ESX-1 locus, there are 10 additional operons paralogous to the esxBA operon as a result of numerous duplications in the M. tuberculosis genome (Cole et al., 1998). Many of these CFP-10/ESAT-6 paralogues (Esx proteins) have been identified in the secreted proteome of M. tuberculosis but are not ESX-1 substrates (Champion et al., 2006). Interestingly, five of the paralogous CFP-10/ESAT-6 pairs are embedded within loci with synteny to the ESX-1 locus (ESX-2-ESX-5) (Cole et al., 1998; Gey Van Pittius et al., 2001), raising the possibility that other ESX-1-like secretion systems function to secrete these Esx paralogues.

The ESX-5 locus has recently been shown to operate in M. marinum and is the first report of a functional ESX locus besides ESX-1 (Abdallah et al., 2006). They suggest that elements of this system are essential for viability, and that this system functions to secrete proteins of the PPE family during infection. However, because many ESAT-6/CFP-10 paralogues are secreted from the Mycobacterial cell, it is likely that at least some of the other ESX loci encode for functional secretion systems, some of which may be essential for virulence or even viability.

Remaining questions

  1. Top of page
  2. Summary
  3. Introduction
  4. The general secretion pathway
  5. The twin-arginine transporter
  6. The ESAT-6 secretion system
  7. Remaining questions
  8. Acknowledgements
  9. References

Despite many recent advances in the field of protein secretion in Mycobacteria, many questions remain. First, how are proteins that are destined for export out of the bacterial cell secreted across the cell wall and lipid layers? Clearly, Mycobacteria use many different secretion systems to export proteins from the cytosol. In Gram-positive bacteria, chaperones and foldases are localized to the Sec machinery and assist in folding and directing proteins to the final destination (e.g. Rosch and Caparon, 2004; 2005). In Gram-negative bacteria, following translocation to the periplasm by the Sec system, proteins destined for the outer membrane interact with a series of chaperones that fold the proteins and escort them to sites in the outer membrane which functions to target and assemble the OMPs (e.g. Werner and Misra, 2005). There are no known components for any secretion machine in Mycobacterium outside of the cytoplasmic membrane, and how these systems interact with proteins outside of the cytosol has yet to be discovered.

Type III and Type IV systems in Gram-negative bacteria have protein components that span the periplasm and the outer membrane to deliver bacterial effectors into the host cell. It is possible that the secreted proteins including ESAT-6, CFP-10, Rv3616c and Rv3615c, are actually extracytoplasmic components of the ESX-1 secretion system (Ize and Palmer, 2006). This could explain the mutually dependent secretion of these substrates. It could also be that the extracytoplasmic components of the secretion system are as of yet unknown, and further work is required to identify these. Finally, it is possible that ESX-1 substrates, all of which are small, can passively work their way through the cell wall.

Second, how does the ESX-1 secretion machine assemble, and how is it regulated? The Type III and IV machines require the ordered assembly of multiprotein complexes and are activated upon host cell contact. Similarly, it is likely that the many ESX-1 components assemble and are regulated by environmental signals during infection.

Third, in addition to ESAT-6, CFP-10 and EspA, what are the other substrates secreted by the ESX-1 system? Proteomic analysis of the ESX-1 system has generally failed to identify substrates of the ESX-1 system outside of ESAT-6, CFP-10 and Rv3616c. This may be because the other substrates, if they exist, are present at very low levels. It is also possible that these substrates are not present under the in vitro growth conditions used in the laboratory, and are only secreted during host-cell infection. This would suggest that substrates of this system are regulated by some as yet undiscovered mechanism.

Certainly, many more questions will arise as the exciting field of Mycobacterial protein secretion unfolds. There is clearly much fertile ground for discovering both basic mechanisms of protein secretion and new insights into how these systems mediate interactions between host and pathogen.

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. The general secretion pathway
  5. The twin-arginine transporter
  6. The ESAT-6 secretion system
  7. Remaining questions
  8. Acknowledgements
  9. References

P.A.C. is supported by NIH under Ruth L. Kirschstein National Research Service Award A105155. J.S.C gratefully acknowledges the support of the Sandler Family Supporting Foundation, the W. M. Keck Foundation, and NIH Grants AI63302 and AI51667.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. The general secretion pathway
  5. The twin-arginine transporter
  6. The ESAT-6 secretion system
  7. Remaining questions
  8. Acknowledgements
  9. References