Molecular syringes scratch the surface

New data suggest that the most recently discovered class of bacterial 'molecular syringes' inject proteins only across the outer membrane of target cells during interbacterial competition. See Article p.343

Many bacteria use specialized secretion systems to inject proteins or DNA into cells of eukaryotic organisms (such as animals and plants) or into other bacteria. Little is known about the mechanism of secretion in the most recently discovered class of these molecular syringes, the type VI secretion system (T6SS)1. On page 343 of this issue, Mougous and colleagues (Russell et al.2) show that the T6SS of the bacterium Pseudomonas aeruginosa delivers two proteins into target bacteria. These proteins degrade peptidoglycan, a highly cross-linked lattice that lies just below the outer membrane of Gram-negative bacteria in a region called the periplasm, causing lysis of the target cell. These findings strongly suggest that the T6SS 'needles' puncture only one membrane (the bacterial outer membrane in this case), providing substantial insight into this system's mechanism of action.

T6SSs were discovered on the basis of their contribution to symbiosis and virulence in bacterial interactions with eukaryotes. Until recently, the only proteins that had been shown to enter host cells through T6SSs were the VgrG proteins of the bacterium Vibrio cholerae, which seem to form the membrane-puncturing tip of the T6SS needle3.

Last year, Mougous and co-workers4 identified three candidate T6SS-dependent 'effector' proteins in P. aeruginosa. One of these proteins, called Tse2, was toxic to both mammalian and bacterial cells if the cells were engineered to produce it intracellularly. The co-production of an immunity protein called Tsi2 prevented this toxicity. Surprisingly, however, P. aeruginosa itself could not intoxicate mammalian cells if co-cultured with them, but it could outcompete other Gram-negative bacteria in a manner dependent on both cell–cell contact and Tse2.

Subsequently, Pukatzki and colleagues5 showed that V. cholerae could also out-compete other Gram-negative bacteria, and that it did so using the same T6SS that it uses to inject proteins into amoeba and mammalian macrophages. These results suggested that at least some T6SSs can deliver proteins into both mammalian cells and Gram-negative bacteria.

That T6SSs can inject proteins into both eukaryotic cells and Gram-negative bacteria is remarkable, because the membrane structures of these cells are very different. Eukaryotic cells are bounded by a single phospholipid bilayer (the plasma membrane). By contrast, Gram-negative bacteria are bounded by two different phospholipid bilayers — the cytoplasmic membrane and the outer membrane, which contains lipopolysaccharides in its outer leaflet. The two membranes are separated by the periplasm, which contains the peptidoglycan lattice. So how could a single protein-delivery system function on such disparate substrates?

Mougous and colleagues2 now characterize the other two candidate effectors that they discovered in their original screen4 — Tse1 and Tse3. They demonstrate that Tse1 and Tse3 lyse target bacterial cells in a T6SS-dependent manner, that both are enzymes that degrade peptidoglycan, and that neither protein can access the periplasm when produced intracellularly or when added to the outside of intact bacterial cells. Moreover, they show that P. aeruginosa produces specific immunity proteins that protect it against Tse1- and Tse3-mediated interbacterial lysis, but only when these immune proteins are directed into the periplasm. The P. aeruginosa T6SS, therefore, seems to penetrate only the outer membrane, delivering effector proteins into the periplasm, and not directly into the cytoplasm, of target bacteria. T6SS-dependent delivery of proteins into eukaryotic cells and Gram-negative bacteria may therefore not be so different, in that it may require injection across only a single membrane in both cases.

These results provide functional support for the hypothesis — based on structural similarities — that T6SSs operate similarly to bacterial viruses (bacteriophages)6,7. Bacteriophages of the Myoviridae family initiate infection by injecting their DNA into host bacteria using their long contractile tails8. These tails have a 'tube within a tube' structure in which the outer tube is 'spring-loaded' to contract when the proper host cell is recognized, forcing the rigid inner tube through the outer membrane and into the periplasm (Fig. 1a). The peptidoglycan-degrading activity of one of the membrane-puncturing tip proteins allows the tube to penetrate the peptidoglycan lattice. Contact between the inner tube and the cytoplasmic membrane then initiates DNA transfer into the cytoplasm, but the tube does not perforate the cytoplasmic membrane.

Figure 1: Similar injection systems.

a, The bacteriophage T4 binds to the surface of a target bacterium with its tail fibres and baseplate, causing a conformational change in the outer-tube proteins and so outer-tube contraction. Consequently, the rigid inner tube is forced through the outer membrane (OM) of the bacterium. The peptidoglycan-degrading domains of the tip proteins then allow penetration through the periplasmic peptidoglycan lattice, and finally interaction with the cytoplasmic membrane (CM) initiates translocation of viral DNA into the cytoplasm. b, The bacterial type VI secretion system (T6SS) is proposed to function similarly, because several proteins in the two systems are structurally similar (shown in the same colours)1,6,7. Mougous and colleagues2 show that two T6SS effectors of Pseudomonas aeruginosa (Tse1 and Tse3) degrade peptidoglycan; their associated immunity proteins (Tsi1 and Tsi3) are protective only when located in the periplasm. These data support the model (shown) that T6SS delivers proteins across only one membrane, and that secreted proteins are delivered from the cytoplasm of the injecting cell into the periplasm of the target cell without entering the periplasm of the injecting cell. How Tse2, which seems to have a cytoplasmic target4, crosses the cytoplasmic membrane of a target bacterium is not known.

Mougous and colleagues' results2 indicate that, like the tail tube of bacteriophage T4, the needle of the P. aeruginosa T6SS is thrust only through the bacterial outer membrane (or the eukaryotic cell plasma membrane) when the outer tube contracts (Fig. 1b). The predicted membrane-puncturing tip proteins of T6SSs (VgrG proteins) lack obvious peptidoglycan-degrading activity, so the T6SS needle presumably cannot penetrate the peptidoglycan lattice. Although P. aeruginosa injects peptidoglycan-degrading effector proteins (Tse1 and Tse3), their function seems to be to kill the target cell rather than just to advance the needle. The authors also show that although a functional T6SS is required for Tse2-mediated interbacterial competition, Tse1 and Tse3 are not. This suggests that Tse2 is somehow translocated independently into the cytoplasm, and that an intact peptidoglycan lattice does not prevent this movement.

Although Mougous and co-workers' results shed light on how T6SSs function, many questions remain. For example, how is specificity for the target determined? Is it mediated by T6SS components or by separate proteins such as pili or other adhesins on the bacterial surface? In cases where a T6SS (such as the T6SS of V. cholerae) can inject different cell types, are different subsets of effector proteins involved? And if so, how is this controlled?

The biophysics of injection is also intriguing. In bacteriophage T4 infection, the phage's baseplate binds to lipopolysaccharides of the host bacterium's outer membrane, and this interaction is apparently stronger than the force required for the inner tail tube to puncture the outer membrane. For T6SSs, the baseplate seems to be located in the bacterial cytoplasm or possibly in the periplasm. What then holds the opposing membranes together at the injection site so that the T6SS needle punctures the membrane of the target cell rather than deforming either cell's membrane or, perhaps, just pushing the cells apart?

Finally, the most difficult question may be why bacteria use contact-dependent delivery of proteins for interbacterial competition. Wouldn't secreting antibacterial agents (such as antibiotics or bacteriocins) into the extracellular environment be a more efficient way to eliminate one's neighbours? Does contact-dependence allow discrimination — to determine which bacteria to kill and which to spare? Is competition even the goal? Perhaps T6SS-dependent interbacterial interactions are involved in developmental processes such as formation of multicellular structures or of organized communities? Answering these questions will require a way to study T6SS-dependent behaviour of bacteria in their natural habitats.


  1. 1

    Records, A. R. Mol. Plant Microbe Interact. 24, 751–757 (2011).

  2. 2

    Russell, A. B. et al. Nature 475, 343–347 (2011).

  3. 3

    Ma, A. T. et al. Cell Host Microbe 5, 234–243 (2009).

  4. 4

    Hood, R. D. et al. Cell Host Microbe 7, 25–37 (2010).

  5. 5

    MacIntyre, D. L. et al. Proc. Natl Acad. Sci. USA 107, 19520–19524 (2010).

  6. 6

    Kanamaru, S. Proc. Natl Acad. Sci. USA 106, 4067–4068 (2009).

  7. 7

    Bonemann, G. et al. EMBO J. 28, 315–325 (2009).

  8. 8

    Rossmann, M. G. et al. Curr. Opin. Struct. Biol. 14, 171–180 (2004).

Download references

Author information

Correspondence to Peggy Cotter.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Cotter, P. Molecular syringes scratch the surface. Nature 475, 301–303 (2011). https://doi.org/10.1038/475301a

Download citation

Further reading


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.