Divide and conquer: Salmonella move into both daughter cells during mitosis

Keywords:

• cell division
• mitosis
• Salmonella
• Salmonella induced tubules
• type III secretion system

This article refers to:

Effector proteins support the asymmetric apportioning of Salmonella during cytokinesis

The recent paper published in Virulence by Zhao et al. characterizes a novel and interesting finding: Salmonella forms tubules that ensure bacteria are divided between mitotic daughter cells, a process that is dependent on bacterial effectors.¹

Over the course of evolution, pathogens have adopted different strategies to cause disease. One key feature that contributes to the virulence of the global pathogen Salmonella enterica is the ability to invade and replicate within host cells.² Salmonella mediates uptake into non-phagocytic cells, such as intestinal epithelial cells; it can also survive and replicate within phagocytic immune cells, such as macrophages.² Salmonella infection causes enteric fever and/or gastroenteritis.³ During human infection, certain Salmonella serovars, for example Typhimurium, normally remain within the intestine causing severe gastroenteritis, and occasionally cause invasive disease.³

Salmonella have 2 Type III Secretion Systems (T3SS), encoded by Salmonella Pathogenicity Islands 1 and 2 (SPI1 and SPI2). These T3SSs inject bacterial effectors into host cells, inducing multiple cellular changes. In epithelial cells, T3SS1 induces the formation of ruffles on the host cell surface, resulting in bacterial uptake.² Once inside the cell, Salmonella reside within a special replicative niche, called the Salmonella containing vacuole (SCV).² Each bacterium resides within a single SCV.⁴ Some Salmonella escape from the vacuole, replicate to high levels within the cytoplasm, and are primed for further invasion.^5,6

From within the SCV, Salmonella secrete effectors that promote further remodelling of the host cell, including SCV juxtanuclear positioning, evasion of immune functions, modulation of vesicle transport and formation of Salmonella induced tubules (SIT, which includes Salmonella induced filaments).² The existence of SITs has been known for some time,⁷ and characteristic markers for these structures include LAMP1, SCAMP3, and Rab7.⁸ SITs are highly dynamic, and different types of SITs have been identified.⁸ Formation of SITs is dependent on T3SS effectors including SifA, SseF, SseG, SopD2 and PipB2.^2,9,10 An extensive study using electron microscopy to examine the ultrastructure of the SCV and SITs showed the presence of single and double membrane SITs in infected cells.¹¹ The authors proposed that SITs begin as single membrane compartments and develop into double membrane structures, dependent on SPI2 effectors SseF and SseG.^8,11 Interestingly, they also found Salmonella in the lumen between the 2 membranes, moving along the tubule.¹¹

In addition to mediating the formation of SITs, effectors can also modulate the cell cycle. For example, the effector SpvB is an ADP-ribosylating protein which depolymerizes actin and is important for virulence.^12,13 Expression of spvB in HeLa epithelial cells results in G¹/S or G²/M cell cycle arrest and apoptosis.¹⁴ Another effector, SspH2 interacts with SGT1, a host protein which is required for the G¹/S and G²/M progression in yeast.^15,16 The Salmonella effector PheA localizes to the nucleus of RAW264.7 macrophages, where it perturbs the cell cycle by increasing the proportion of cells in G⁰/G¹, and reduces the proportion of cells in G²/M phase.¹⁷

The cytolethal distending toxin (cdt) is known to cause cell cycle arrest due to DNA damage.¹⁸ In Salmonella serovar Typhi, CdtB is present and functional.^19,20 CdtB delivery is mediated by PltA and PltB, the genes for which are encoded on the same pathogenicity islet.²¹ However, this gene cluster is not present in many clinically important serovars, including Typhimurium, Enteritidis, Paratyphi B and Paratyphi C.²²

Salmonella preferentially invade mitotic cells.²³ A few studies have recently focused on understanding how Salmonella selects target cells. SipB is a component of the T3SS tip, which binds to cholesterol.²⁴ During mitosis, the amount of cholesterol on the cell surface increases, which may explain Salmonella's preference for mitotic cells.²³ Another study examined Salmonella target cell selection and found that rounded cells (that occur during mitosis), or cells with physical features (such as ruffles) increase stopping and docking of Salmonella.²⁵ Along these lines, invasion of polarized epithelial cells by Salmonella has been shown to occur in a cooperative manner, with epithelial cell ruffles engulfing additional bacteria.²⁶

Together, these data support several key findings: (1) Salmonella form SITs in a T3SS-dependent manner within infected cells. (2) Salmonella actively modulate the host cell cycle using T3SS effectors, which are important for virulence. (3) Salmonella preferentially invades mitotic cells. However, there is still a great deal that we do not yet understand about the function of SITs and Salmonella interactions with cell cycle.

In this issue of Virulence, Zhao and colleagues address some of the key gaps in our knowledge regarding what happens to Salmonella during mitosis.¹ They describe a particularly interesting finding: Salmonella Typhimurium form membranous tubules between mitotic non-polarized epithelial daughter cells. Furthermore, they demonstrate that this process is dependent upon T3SS2 effectors and results in the asymmetric distribution of bacteria between daughter epithelial cells.

The authors observed membranous tubules extending between adjacent HeLa cells. Further examination revealed striking similarities with SITs; they extended from the SCV, they contained LAMP1 and SCAMP3, and were enriched for Salmonella effectors. Interestingly, they were not found in cells infected with a strain lacking T3SS2 (ΔssaV). Zhao and colleagues have termed these structures intercellular tubules (ICTs).

To determine if ICTs were formed between adjacent cells, Zhao et al. performed an elegant experiment where 2 separate HeLa cell populations were infected with either gfp or dsRed expressing bacteria. The epithelial cells were then trypsinized and mixed. This allowed the authors to determine if ICTs were forming between independently infected cells (i.e. red and green bacteria), or if the ICTs were in fact forming between daughter cells. They found strong evidence in support of the latter hypothesis. Furthermore, they found that 67% of infected and cytokinetic epithelial cells had ICTs.

Among the Salmonella-infected mitotic cells, Zhao et al. identified several key differences. In only a minority of cases, they observed after mitosis bacteria were found in only one daughter cell. They termed this negative cytokinesis. The majority of the time, after cytokinesis bacteria were in both daughter cells, this was referred to as positive cytokinesis. They noticed that bacteria were unevenly distributed between the 2 daughter cells.

In order to explore the role of Salmonella secreted effectors in ICT formation, the authors compared wild-type Salmonella with a SPI1 mutant (ΔprgH), and a SPI2 mutant (ΔssaV). They demonstrate that both the wild-type and SPI1 mutant bacteria were preferentially found in G²/M phase cells. However, the SPI2 mutant was evenly distributed among the various cell phases, showing no preference for cell cycle phase.

Specific effector deletion mutants were tested to determine their impact on the formation of ICTs and distribution of bacteria among daughter cells. Deletion of sopD, sifB, sseJ, and pipB had no effect. Conversely, strains with deletions of sifA, sopD2, sseF, sspH2, and pipB2 behaved similarly to the T3SS2 deletion strain, with few ICTs and less positive cytokinetic events. With these data, the authors show a correlation between the formation of ICTs and the distribution of bacteria between daughter cells, and demonstrate that this process is dependent on the activity of specific Salmonella effectors. It should be noted that Zhao and colleagues found that Salmonella did not significantly impact HeLa cell transition through the cell cycle during their experimental time frame.

In addition to HeLa cells, the authors examined Salmonella infected macrophages and found no ICTs, and no difference in the number of mitotic cells between infected and uninfected macrophages. Interestingly, they did see a higher proportion of infected macrophages in G²/M phase, suggesting that Salmonella may be preferentially taken up by mitotic and pre-mitotic macrophages.

This study has made a significant contribution to the Salmonella virulence field because it provides unique insights into how Salmonella is distributed when the cell it is in divides. The findings by Zhao et al. that describe how Salmonella uses effectors to generate ICTs that ensure unequal distribution of bacteria among daughter cells is striking. Future research in this area should address questions pertaining to the role played by ICTs during infection of the intestine or in polarized epithelial cells. In the intestine, epithelial cells replicate in crypts and then migrate toward the villus tip.²⁷ In a rabbit ileal loop model, Salmonella first infected cells on the villus tip, and then reached the villus base 18 hours post-infection.²⁸ Can ICTs be seen in a model similar to this? Do they play a physiological role in intestinal infection?

Some of the effectors involved in formation of ICTs and positive cytokinesis are known to contribute to virulence.¹² This paper by Zhao et al. highlights new roles for these effectors. These will need to be examined more closely to answer questions about how these effectors modulate ICT formation and bacterial distribution, and which host proteins or processes are involved. It is intriguing to hypothesize that in addition to manipulating the host cell cycle, Salmonella is also able to intervene during host cell division, ensuring that bacteria are present in daughter cells of subsequent generations. This may have important implications for infection, with highly infected daughter cells being sloughed off before their less-infected counterparts. Because of this work by Zhao and colleagues, we now know which questions to ask next.

Disclosure of potential conflicts of interest

No potential conflicts of interest were disclosed.

Funding

MMCB is supported by a fellowship from the AXA Research Fund.