Infection with Shigella flexneri bacteria is a major cause of infant death. It emerges that S. flexneri evades intracellular defences by releasing a protein that triggers the destruction of members of a key family of host enzymes. See Letter p.378
Pathogenic microbes that invade a human host must overcome many obstacles to establish infection. One of the first is an immune response known as cell-autonomous defence, which occurs in a cell under pathogenic attack1. This form of innate defence operates in most human cells, and signalling from immune-system proteins, including those of the interferon family, helps to mobilize it2. Interferons rally hundreds of other proteins into action, including a group of enzymes called guanylate-binding proteins (GBPs), which have potent antimicrobial activity against a range of intracellular pathogens2,3. Whether pathogens promote successful infection by disarming GBPs has been a key unanswered question. Li et al.4 report on page 378 that the answer is yes.
The bacterium Shigella flexneri causes a gut infection called shigellosis, also known as bacillary dysentery, which is associated with symptoms including diarrhoea. More than 160 million cases of shigellosis are estimated to occur globally each year, resulting in some 600,000 deaths, mostly in children less than 5 years old in developing countries5. Infection occurs when ingestion of faecally contaminated food or water enables S. flexneri to invade epithelial cells of the intestine, causing severe inflammation and disrupting the gut's cellular integrity.
When S. flexneri enters a host cell (Fig. 1), it is initially trapped inside a membrane-bound compartment called a vacuole, which it disrupts to escape into the cytoplasm6. The bacterium then encounters defence responses that can destroy or disable it, for example through an intracellular degradation pathway called autophagy1,2. However, invading bacteria can transfer proteins directly into their host's cytoplasm through a needle-like structure called the type III secretion system6, helping it to combat the immune response and promote bacterial replication and spread. Shigella flexneri transfers more than a dozen such bacterial proteins, including one called IcsB that helps it to evade autophagic destruction6. GBPs can interact with autophagic proteins that might target bacteria for destruction2,3,7,8,9,10. This raises the possibility that GBPs also target S. flexneri, and, if this is so, poses the question of whether the bacterium has evolved a counterpunch.
To investigate whether S. flexneri combats GBPs, Li and colleagues used a fluorescent tag to monitor the recruitment of human GBP1 (hGBP1) protein to bacteria in the host-cell cytoplasm. They also tracked the galectin-3 protein, which enabled them to identify damaged vacuoles from which the bacterium had escaped.
What they found was unexpected. Epithelial cells infected with S. flexneri were completely devoid of hGBP1, whereas galectin-3 was still present, suggesting that hGBP1 had been selectively degraded. Other bacteria that can also enter the epithelial-cell cytoplasm, such as Listeria monocytogenes and Salmonella enterica Typhimurium, did not degrade hGBP1; instead, they were quickly surrounded by it. Such action could kill the bacteria by breaking down the microbial cell wall9,10 or by recruiting other antibacterial proteins2,7,11.
The authors next investigated why hGBP1 disappeared from cells upon infection with S. flexneri. One mechanism for targeting intracellular protein destruction is a pathway in which a ubiquitin-protein tag is added to a protein in a process called ubiquitination. The addition of many ubiquitin tags marks the protein for destruction in a multi-protein complex called the proteasome. When the authors treated cells with a proteasomal inhibitor or with an inhibitor that blocks a type of enzyme, known as an E1 ubiquitin ligase, that acts in the first step of ubiquitination, it prevented hGBP1 loss on S. flexneri infection.
To determine how S. flexneri achieves proteasomal-mediated GBP destruction, Li and colleagues tested about 13,000 strains of S. flexneri that had mutant versions of different genes. One strain that was unable to destroy hGBP1 had a disrupted version of the ipaH9.8 gene. This gene encodes an enzyme known as an E3 ubiquitin ligase (IpaH9.8), which aids the process leading to the final step of ubiquitin tagging for proteasomal destruction of a protein. The authors demonstrated that IpaH9.8 could add ubiquitin tags to certain lysine amino-acid residues of hGBP1 in cells grown in vitro. The ipaH9.8 gene lies in a region of the genome harbouring genes that form components released by the type III secretion system, suggesting that the protein encoded by ipaH9.8 is directly delivered into the host cytoplasm on infection.
Shigella flexneri contains a small family of IpaH E3 ubiquitin ligases6, but the other IpaH family members tested by the authors did not degrade hGBP1. Moreover, Li and colleagues demonstrated that if they engineered S. Typhimurium to express ipaH9.8, this sufficed for the bacterium to gain the capacity to destroy hGBP1 in human cells.
Notably, when the authors conducted in vitro tests (in which they also provided the E1 and E2 needed for ubiquitination), they found that IpaH9.8 enabled the ubiquitination of more than half of all the GBP family members found in humans and mice (which have 7 and 11, respectively)3,7,11. Hence, IpaH9.8 could enable S. flexneri to degrade many GBPs that often act in concert to restrict bacterial replication3,7,8,9,10,11. Yet microbial interference with this enzyme family has not been observed previously for any bacterial pathogen.
GBP ubiquitination by IpaH9.8 occurred independently of GBP activation, which is needed for the enzymes to target bacteria after infection3,7,11. This suggests that IpaH9.8 can engage GBPs before they act to restrict microbial growth. Such early intervention would block not only GBP-dependent bacterial destruction through enzyme-mediated changes to the bacterial cell wall, but also activation of a host multi-protein complex called the inflammasome, which can sense microbial products released by bacterial destruction3,8,9,10 and enlists GBPs in the assembly of specific inflammasome complexes3,11,12 that trigger a systemic immune response beyond the infected cell. Shigella flexneri therefore targets a chink in its host's armour that potentially allows it to disable several defences in one fell swoop.
The authors investigated the in vivo function of IpaH9.8. Mice exposed to S. flexneri that express IpaH9.8 rapidly succumbed to infection, whereas animals given S. flexneri that lacked IpaH9.8 survived. However, mice that lacked five members of the GBP family died when infected with the S. flexneri missing IpaH9.8, because the removal of the host defence provided by these GBPs enabled the S. flexneri mutant to now cause infection. Together, these results demonstrate that IpaH9.8 is needed for bacteria to specifically evade a GBP-mediated immune response.
This result highlights the co-evolutionary battle between host defence proteins and bacterial proteins to gain ascendancy during infection. Indeed, a rapidly expanding repertoire of host proteins targeted by bacterial E3 ubiquitin ligases has been identified. These include: NF-kB, which is part of a key immune signalling pathway; autophagic proteins; and proteins that contribute to an inflammation-triggered cell-death process13. GBPs now join that list.
Li and colleagues' discovery of a microbial strategy that disarms GBPs provides a major conceptual advance that is reinforced by a similar study just published14. Both reports highlight the increasing importance of this group of enzymes in intracellular immune defences. How GBPs initially recognize bacteria in the cytoplasm, and how they drive bacterial destruction, are key unanswered questions. Determining why IpaH9.8 targets some GBPs and not others might require structural studies that could also provide results with implications for antibiotic drug design. With the identification of the complete GBP family and other closely related immune-system enzymes several years ago3,7,11, an exciting chapter in host–pathogen biology has begun. Expect more plot twists and intriguing characters as this fascinating story of intracellular skirmishes during infection continues to unfold.
Randow, F., MacMicking, J. D. & James, L. C. Science 340, 701–706 (2013).
MacMicking, J. D. Nature Rev. Immunol. 12, 367–382 (2012).
Kim, B.-H. et al. Nature Immunol. 17, 481–489 (2016).
Li, P. et al. Nature 551, 378–383 (2017).
Kotloff, K. L. et al. Bull. World Health Organ. 77, 651–666 (1999).
Ogawa, M., Handa, Y., Ashida, H., Suzuki, M. & Sasakawa, C. Nature Rev. Microbiol. 6, 11–16 (2008).
Kim, B.-H. et al. Science 332, 717–721 (2011).
Meunier, E. et al. Nature 509, 366–370 (2014).
Man, S. M. et al. Nature Immunol. 16, 467–475 (2015).
Meunier, E. et al. Nature Immunol. 16, 476–484 (2015).
Kim, B. H., Shenoy, A. R., Kumar, P., Bradfield, C. J. & MacMicking, J. D. Cell Host Microbe 12, 432–444 (2012).
Shenoy, A. R. et al. Science 336, 481–485 (2012).
Ashida, H. & Sasakawa, C. Curr. Opin. Microbiol. 35, 16–22 (2017).
Wandel, M. P. et al. Cell Host Microbe 22, 507–518 (2017).
Related links in Nature Research
About this article
Current Opinion in Immunology (2019)
Phagocytosis of Leptospira by leukocytes from mice with different susceptibility to leptospirosis and possible role of chemokines
BMC Microbiology (2019)
Journal of Experimental Medicine (2019)
Science Signaling (2017)