Autophagy subversion of burkholderia pseudomallei infection

Burkholderia pseudomallei, a gram-negative bacterium, is the causative agent of melioidosis, which is endemic in tropical areas such as South East Asia and Northern Australia and results in significant mortality. B. pseudomallei is able to invade both non-phagocytic cells and phagocytic cells, escape from the endosome/phagosome and replicate within the cytosol (Wiersinga et al., 2012). Once in the cytosol, B. pseudomallei expresses the protein, BimA, that affords the bacteria actin-based motility allowing them to spread to neighbouring cells, resulting in cell fusion and then ultimately the formation of multinucleated giant cells (MNGCs) (Stevens et al., 2005a, Kespichayawattana et al., 2000). The signalling pathways and mechanisms active during bacterial invasion and the host response have yet to be fully determined. Autophagy is a multi-functional, intracellular process that eukaryotic cells use to maintain intracellular homeostasis. The core autophagy machinery executes several steps that eventually lead to the sequestration of the targeted cellular components within double-membrane autophagosomes which subsequently fuse with lysosomes to degrade the cargo (Shibutani and Yoshimori, 2014). Autophagy is often involved in the removal of old or damaged organelles, mis-folded proteins or toxin-conjugating elements (Rogov et al., 2014). Recent research has found that autophagy is associated with the host immune system to act against invading bacteria. Induction of autophagy might directly target bacteria in phagosomes or ‘free’ in the cytosol, or indirectly trigger the innate immune responses, including inflammation, in order to obtain the maximum bactericidal activity (Huang and Brumell, 2014). However, some bacterial pathogens, for example Shigella flexneri and Listeria monocytogenes, have evolved strategies to avoid autophagy or manipulate the autophagic process to facilitate survival. The fate of B. pseudomallei after escaping from phagosomes and its interactions with the host autophagic system are yet to be fully clarified. Cullinane et al. (2008) provided evidence that lack of bacterial BopA, a type III secretion system cluster 3 (T3SS-3) effector, increased the proportion of bacteria co-localised with LC3. Electron microscopy analysis of infected macrophage cell sections demonstrated that intracellular BopA mutant bacteria are located within single-membrane phagosomes rather than double-membrane autophagosomes, indicating that LC3 is recruited directly to the phagosomes, and the bacteria are subject to LC3-associated phagocytosis (LAP) (Gong et al., 2011). Furthermore, bacteria that have escaped from phagosomes are not then targeted by canonical macroautophagy. BopA shows 23% homology with IcsB of S. flexneri, which is employed in a particular system that involves competitive binding between bacterial IcsB and the host Atg5 to the bacterial surface protein IcsA, to avoid autophagy of targeted bacteria (Ogawa and Sasakawa, 2006). Nevertheless, BopA does not act in the same way as IcsB in S. flexneri (Cullinane et al., 2008). Recent research in the Monash University laboratories has focussed on the identification of potential mechanisms by which host autophagy is avoided by B. pseudomallei following infection and then escape from phagosomes. The research reported in this thesis seeks to address this topic by investigating bacterial proteins encoded by four open reading frames, bpsl1631, bpss0180, bpss1512 (tssM) and bpss1513 (tssN), each a potential virulence factor associated with the manipulation of the host autophagy system. The basis of the experimental strategy applied here involved the modification of the B. pseudomallei genome by double-crossover allelic exchange, to generate the particular gene knockout mutant strains by which to examine whether the loss of individual encoded proteins altered bacterial invasion and intracellular growth. Further analysis included an assessment of pathogenesis using an in vivo mouse model of acute bacterial infection. The application of confocal fluorescence microscopy revealed the bacterial association with cellular proteins (such as ubiquitin or p62), and particularly LC3 as an indicator of bacterial phagosome escape and/or autophagy. Transmission electron microscopy (TEM) provided direct ultrastructure observation of bacteria-infected mammalian cells in order to identify the membrane structure surrounding intracellular bacteria. Thus five mutant strains have been generated from the wild-type strain B. pseudomallei K96243 namely: ?bpsl1631 (Chapter 3), ?bpss0180 (Chapter 4), ?bpss1512 (?tssM), ?bpss1513 (?tssN) and ?bpss1512-1513(?tssMN) (Chapter 5). Chapter 3 addresses BPSL1631, a putative outer membrane protein, which was chosen as an experimental target on the basis of bioinformatics analysis indicating the homology between BPSL1631 and S. flexneri IcsA is 19%. However, when the Atg5-binding region within IcsA was set as the reference sequence, three regions within BPSL1631 were found to have 27% amino acid identity with IcsA. The presence of this sequence homology suggested BPSL1631 may have the capacity to bind Atg5. It was therefore of interest to know if B. pseudomallei used BPSL1631 as part of a mechanism to avoid autophagy and/or LAP. The results obtained from this study indicated that bacteria lacking BPSL1631 can still escape from phagosomes and replicate in RAW264.7 macrophage-like cells. Furthermore, loss of BPSL1631 did not affect virulence in vivo. Therefore, it is unlikely that this protein plays a role in the avoidance of anti-bacterial autophagy. Chapter 4 focuses on BPSS0180, a B. pseudomallei protein that was identified by the research group of Dr Patrick Tan (Singapore Genome Institute), as having the potential to induce autophagy in mammalian cells. They observed that overexpression of a green fluorescent protein (GFP) fusion with BPSS0180 increased the number of green puncta which co-localised with LC3 in both non-phagocytic and phagocytic cells (Singh et al., 2010a). In order to determine whether the induction of autophagy by BPSS0180 might affect B. pseudomallei pathogenesis, collaboration between the Singapore Genome Institute and Monash University research groups was established. A bpss0180-knockout mutant was generated and then used to determine if the potential induction of autophagy by BPSS0180 is associated with bacterial virulence. Although bacteria lacking BPSS0180 showed a significant decrease in co-localisation between bacteria and LC3 puncta compared to wild-type infected cells, bacteria were still able to escape from phagosomes. Intracellular survival was significantly impaired by knockout of bpss0180, however bacteria were still capable of replicating. Overall the results suggested that B. pseudomallei BPSS0180 may serve as a virulence factor to induce bulk non-selective autophagy of host components to facilitate the supply of nutrients available for bacterial growth in the host cells, or otherwise aid bacteria in adapting to the host cytosolic environment. Chapter 5 involved investigation of the genes bpss1512 (tssM) and bpss1513 (tssN), with three knockout mutants being generated, ?tssM, ?tssN and a double-knockout ?tssMN. Current reports suggest that for several different bacterial pathogens cytosolic ‘free’ bacteria can be ubiquitinated and targeted to autophagy by different autophagy receptors, such as p62 and NDP52 (Randow and Youle, 2014). However some bacteria may have evolved a mechanism to avoid ubiquitination, for example by secretion of a deubiquitinase. There is evidence to suggest that B. pseudomallei TssM which contains an ubiquitin C-terminal hydrolyase (UCH) domain, functions as a bacterial deubiquitinase. Secretion of TssM has been demonstrated to interfere with stimulation of the host innate immune response by deubiquitination of essential factors contributing to the downstream inflammatory response, such as TRAF6, I?Ba and TRAF3. (Tan et al., 2010). Therefore it could be hypothesised that B. pseudomallei may also utilise this deubiquitinase to escape from host autophagy by removing ubiquitin from intracellular bacteria that have escaped from phagosomes. The adjacent gene locus bpss1513 displays 99% homology to tssN (bmaa0728) of B. mallei that has been reported to interact with the components of the host ubiquitination pathway, such as ubiquitin precursor (UBB) and ubiquitin ligases (E3s). Also, bioinformatics analysis indicated that B. mallei TssN might interact with mammalian proteins having a UCH domain, which then suggests TssN is a possible TssM-binding partner (Memisevic et al., 2013). Loss of TssM and/or TssN did not show a clear effect in altering bacterial survival in host cells, indicating that they are probably not major virulence factors. Bacteria lacking TssM or TssN showed higher co-localisation with either ubiquitin or p62 at certain time points after infection, hinting that these two proteins might partly function to interfere with ubiquitin recognition of invading bacteria. Triple labelling of bacteria, ubiquitin and p62 also showed a similar pattern, indicating that p62 was recruited to ubiquitinated bacteria. However, only a small portion of mutant bacteria were targeted by ubiquitin and p62 and subsequent LC3 targeting was rarely seen. Interestingly, TEM observation revealed that mutant bacteria were occasionally found within multi-membrane structures, implying that a low level of xenophagy may be taking place, although the vast majority of intracellular bacteria were not associated with any membrane structure following phagosome escape. It remains unknown how B. pseudomallei avoids host autophagy following escape from phagosomes. The results presented in this thesis suggest that B. pseudomallei does not employ a Shigella-like IcsB autophagy avoidance mechanism. (...)