<<

NEWS & VIEWS NATURE|Vol 437|15 September 2005

MICROBIOLOGY vesicle production virtually ceased. Addition of extra PQS restored vesicle formation, even if synthesis was blocked by an . Bacterial speech bubbles So, whatever PQS is doing, it is probably hap- Stephen C. Winans pening at the post-translational level, rather than causing alterations in . Many socialize using diffusible signals. But some of these Most of the added PQS was found in vesicles, messages are poorly soluble, so how do they move between bacteria? which also contained other quinolones pos- sessing antimicrobial activity. It seems that It seems they can be wrapped up in membrane packages instead. when PQS is inserted into the outer membrane — whether it was made by that bac- On page 422 of this issue, Mashburn and terium or added from another source — Whiteley1 describe the unexpected con- it can destabilize the membrane and vergence of two seemingly unrelated areas spontaneously form vesicles containing of microbiological research: how bacteria itself, other quinolones and presumably

talk to their friends, and how they attack endotoxin and periplasmic . EYE OF SCIENCE/SPL their enemies. The authors studied the These vesicles can then signal to other bacterial aerugi- pseudomonads, kill unrelated bacteria, nosa, which releases a hydrophobic mole- and deliver toxic lipopolysaccharides cule called the ‘pseudomonas quinolone and proteins to host cells. signal’ (PQS) to send messages to other Pseudomonas aeruginosa is an oppor- bacteria of the same . The surprise tunistic human pathogen that is a par- is that, rather than being secreted as single ticular problem in the of cystic molecules, PQS is released in bubble-like fibrosis patients. Greater understanding ‘vesicles’ that also contain antibacterial about how these bacteria communicate agents and probably toxins aimed at host with each other and attack host cells tissue cells as well. could lead to novel treatment strategies Various groups of bacteria use diffusible Figure 1 | Friend or foe? Gram-negative such for P. aeruginosa . The way in chemicals to signal to their own kind, and as Pseudomonas aeruginosa communicate with each other by which other bacteria generate these this method of communication seems to secreting chemical messengers. Mashburn and Whiteley1 ubiquitous membrane vesicles, and what have evolved independently several times2. show that they also release such signals in vesicles that contain they might contain, is unknown, but Pseudomonas aeruginosa is a Gram- toxins against other bacteria and the cells of their host. most proteobacteria do not synthesize negative proteobacterium and, like many PQS, so other mechanisms must be proteobacteria, it secretes molecules called by fusing with the lipid bilayer of target cells. involved. Moreover, hydrophobic long-chain acylhomoserine lactones (AHLs). The concen- Mashburn and Whiteley show that these vesi- AHLs were not found in membrane vesicles in tration of these signal molecules increases with cles can also release the PQS signal to other this study, so how these signalling molecules cell population density, enabling the bacteria to bacteria, thereby sidestepping the problems of diffuse or are transported between bacteria sense how many of their kind surround them poor water-solubility and degradation. has still to be discovered. ■ — a phenomenon known as . Remarkably, P. aeruginosa has three quorum- Stephen C. Winans is in the Department of Whole sets of genes are switched on only at sensing systems that use AHL signals. Apart , Cornell University, Ithaca, high cell densities, including genes for bio- from the PQS system, there is the LasR/LasI New York 14853, USA. luminescence, DNA transfer, pigment produc- system, which signals using 3-oxo-dode- e-mail: [email protected] tion, and genes required for infecting plants, canoylhomoserine lactone (OdDHL), and the animals and humans2. But AHL molecules are RhlR/RhlI system, which uses butyryl- 1. Mashburn, L. M. & Whiteley, M. Nature 437, 422–425 (2005). readily broken down by other bacteria, and homoserine lactone (BHL). Of these three 2. Waters, C. M. & Bassler, B. L. Annu. Rev. Cell Dev. Biol. (in the some AHL signals are poorly soluble in water, signals, OdDHL and PQS are poorly water- press) so they cannot travel far in an aqueous envi- soluble, and all three can be degraded by other 3. Kadurugamuwa, J. L. & Beveridge, T. J. J. Antimicrob. Chemother. 40, 615–621 (1997). ronment — both factors that would seem to bacteria. Mashburn and Whiteley found that 4. Wai, S. N. et al. Cell 115, 25–35 (2003). limit their potential as communication signals. PQS (but not OdDHL or BHL) is not dis- PQS, although not an AHL-type signal, is also persed between bacteria as single, water-solu- poorly water-soluble and readily broken down ble molecules, but rather is concentrated Correction by other bacteria. So how do these molecules within membrane vesicles, presumably buried In the News & Views article “Cell biology: Powerful curves” by L. Mahadevan and T. J. manage to spread the word from one within the lipid bilayer. 1 bacterium to another? PQS is one of at least 55 quinolones and Mitchison , which discussed a paper published in the same issue2, the authors put forward an Mashburn and Whiteley found clues from quinolines made by P. aeruginosa, some of elastic-sheet model of microtubule structure. another bacterial communication system — which have potent antibacterial activity. The Such a model was previously proposed3 and one aimed at enemies rather than friends. membrane vesicles, in addition to signalling to applied 3–5 by Flyvbjerg and colleagues in earlier Many Gram-negative bacteria release minute kindred P. aeruginosa cells, showed potent papers. The authors apologize for the oversight vesicles (about 50 nm in diameter)3 from their antibacterial activity against the Gram-posi- in not citing these earlier quantitative studies. outer membranes as a means of delivering tox- tive bacterium . Most of 1. Mahadevan, L. & Mitchison, T. J. Nature 435, 895–897 ins to host cells and other bacteria (sometimes the antimicrobial activity could be extracted (2005). referred to type VI protein export4). The vesi- by organic solvents, indicating that toxicity 2.Wang, H.-W. & Nogales, E. Nature 435, 911–915 (2005). 3. Jánosi, I. M., Chrétien, D. & Flyvbjerg, H. Eur. Biophys. J. cles consist of a lipid bilayer surrounding an was due to quinolines rather than to the pro- 27, 501–513 (1998). aqueous core and they can therefore transport tein component of the vesicles. 4.Chrétien, D., Jánosi, I. M., Taveau, I. & Flyvbjerg, H. Cell lipid-soluble toxins (lipopolysaccharide endo- The production of the vesicles seems to be Struct. Funct. 24, 299–303 (1999). toxin) on their surface and protein toxins in regulated by PQS, because when PQS synthesis 5.Jánosi, I. M., Chrétien, D. & Flyvbjerg, H. Biophys. J. 83, 1317–1330 (2002). their core. They release their poisonous cargo was abolished by a in the pqsH gene,

330 © 2005 Nature Publishing Group