Nature Structural Biology
9, 83 - 84 (2002)
doi:10.1038/nsb0202-83
Bacterial EsperantoStephen C. WinansStephen C. Winans is in the Department of Microbiology, Cornell University, Ithaca, New York, USA 14853 and is supported by grants from NIH and NSF.
scw2@cornell.edu The structure of a bacterial pheromone bound to its receptor sheds light on cell−cell signaling in bacteria.Many groups of bacteria, which until recently were thought to live rather reclusive lives, are now known to communicate with each other by means of diffusible chemical signals. Bacteria are thought to use chemical pheromones to take a census of their population size and to express particular target genes only at high cell densities, a phenomenon sometimes referred to as quorum sensing1. These bacterial pheromones are required for diverse behaviors, including bioluminescence, the production of pathogenesis factors, antibiotics and other secondary metabolites, the horizontal transfer of DNA, and the formation of biofilms2. Although most known examples of bacterial signaling occur within a single species, Bonnie Bassler and colleagues3 at Princeton University several years ago discovered a chemical signal, called autoinducer-2 (AI-2), that is released by many groups of eubacteria, raising the possibility of a universal chemical lexicon employed by multitudes of diverse bacteria. Although the gene encoding the AI-2 synthase (luxS) was subsequently identified and found to be widely conserved4, the chemical structure of this signal has until now remained elusive. In a recent issue of Nature, Hughson (also at Princeton), Bassler and colleagues5 solved the structure of AI-2, remarkably through X-ray crystallography of its protein receptor.
A rich chemical lexicon for bacterial signaling
Chemical signaling has evolved many times among the prokaryotes. Many groups of Gram-positive bacteria use either oligopeptides or  -butyrolactones as signals2. At least one cyanobacterium also signals via oligopeptides6. In contrast, most signaling in proteobacteria is accomplished using N-acylhomoserine lactones7. However, earlier studies of AI-2, first discovered in the bioluminescent marine bacterium Vibrio harveyi, suggested that it was unlikely to resemble any of these molecules. Bassler and colleagues8 previously showed that AI-2 is formed during the catabolism of S-adenosylmethionine (SAM), a universal donor of methyl groups. Demethylation of SAM yields a compound, S-adenosylhomocysteine (SAH) that is further processed to yield adenine, homocysteine, and a compound (derived from the ribosyl moiety of SAM) that spontaneously rearranges to yield AI-2 (Fig. 1). AI-2 should therefore structurally resemble ribose.
 | | Figure 1. Production and detection of a universal bacterial pheromone. | | |  | In most bacteria, utilization of S-adenosylmethionine (SAM) as a donor of methyl groups results in production of S-adenosylhomocysteine (SAH), which is metabolized in two steps by Pfs and LuxS to create homocysteine, adenine, and a molecule that spontaneously rearranges to form Pro-AI-2. Pro-AI-2 is released from the bacteria and accumulates in the cell supernatant. Vibrio harveyi and probably many other bacterial genera can detect pro-AI-2 as a pheromone. In the case of V. harveyi, pro-AI-2 forms a complex with borate, and this complex binds to LuxP, an extracytoplasmic protein that resembles the ribose binding protein of enteric bacteria. LuxP−AI-2 complexes transduce a signal to LuxQ, a transmembrane kinase, which phosphorylates LuxO (indirectly via LuxU, not shown). LuxO-P indirectly causes induction of a biolumenscence operon, resulting in light production. LuxO can also be phosphorylated by a second kinase, LuxN, in response to a separate pheromone (not shown).
 Full Figure and legend (87K) |
| |
AI-2 is detected by V. harveyi through binding to an extracytoplasmic receptor protein (LuxP) that resembles bacterial ribose binding proteins9. LuxP then transduces this signal to a membrane-spanning two-component histidine protein kinase (LuxQ). LuxP-type proteins are generally involved both in nutrient uptake through ATP-driven permeases and in chemotaxis via interactions with membrane-spanning methyl-accepting chemoreceptor proteins. AI-2 can be released from purified LuxP by thermal denaturation, indicating that the two molecules can be purified as a complex. The recent structural studies showed that, as expected, LuxP has close similarity to the ribose binding protein in that it possesses two domains linked by a three-stranded hinge and a deep cleft containing bound ligand10. Also as predicted, LuxP contains a cyclized carbohydrate that resembles ribose.
An unexpected role for boron
Unexpectedly, the LuxP structure contained additional unassigned electron density, showing an atom that formed a diester with the carbohydrate moiety. This atom was proposed to be boron, indicating that AI-2 is a furanosyl borate diester. Although boron is difficult to distinguish from carbon by electron density alone, a carbon atom at this position would be extremely unstable chemically, while boron would be more stable. This surprising finding was confirmed by 11B NMR, electrospray ionization mass spectrometry, and by showing that addition of borate to whole cells stimulates this signaling pathway. Boron is an abundant component of seawater, so its availability should not pose a challenge for the bacteria. It remains to be determined whether pro-AI-2 (that is, the carbohydrate portion of the AI-2−borate complex) binds borate as it diffuses between cells, or whether borate binds to LuxP independently of AI-2. Either way, this report proves that AI-2 bears no resemblance to any previously characterized bacterial pheromone. Furthermore, the structure suggests a potential novel biological role for boron, an element required by a number of organisms but for unknown reasons.
A metabolite co-opted as an intercellular signal
Although we now know the structure of AI-2, it is far from clear how widely this molecule is used in bacterial signaling. Many groups of bacteria are known to synthesize AI-2, and in several cases luxS orthologs have been shown to be essential for AI-2 synthesis11. However, it is not clear that all the bacteria that produce AI-2 actually use it as a signaling molecule. The available evidence suggests that AI-2 may be partly an intercellular signal and partly a metabolic waste product. Evidence for the latter comes from the fact that, in LuxS-containing bacteria, AI-2 synthesis is stoichiometrically linked to the utilization of SAM. Since most SAM is utilized in the synthesis of the choline moiety of phospholipids, AI-2 is indirectly tied to membrane synthesis and therefore to bacterial growth. Additionally, at least in Salmonella typhimurium, release of AI-2 into culture broth (as opposed to its production within the bacterium) occurs preferentially in carbohydrate-rich media12,
13. Carbon-limited bacteria release very little AI-2 and can scavenge it from the growth medium, possibly for use as a nutrient14. It seems that AI-2, like acetate, can be produced and excreted in the presence of preferred carbon sources, and can be utilized when preferred nutrients are exhausted.
However, even if AI-2 is partly a waste product, this does not preclude a role in signaling. Metabolites could well act as signals, since they accumulate at high cell densities and provide information about the size and metabolic status of a population of cells. Evidence suggesting that AI-2 is released with the goal of signaling comes from its route of biosynthesis. A few groups of eubacteria, as well as archaeal and eukaryotic cells, metabolize SAH via a one-step reaction that yields homocysteine and adenosine (rather than adenine) and does not create AI-2 (ref. 15). Conceivably, the two-step metabolism of SAH that produces AI-2 evolved at least partly for the purpose of generating a diffusible signal.
Several examples of AI-2 mediated signaling are beginning to emerge. It has been well established that V. harveyi uses AI-2 as one of two cell density signals to regulate bioluminescence9. Pathogenic E. coli increase the expression of a Type III protein translocation system in response to AI-2 (ref. 16). In transcriptional profiling experiments, large numbers of E. coli genes are either up-regulated or down-regulated by AI-2 (refs 17,18). In apparent contrast, extensive screens for AI-2 inducible genes in S. typhimurium yielded just one operon, whose products direct AI-2 uptake and metabolism14. AI-2 also controls hemin acquisition genes in Porphyromonas gingivalis19, the expression of the VirB virulence factor in Shigella flexneri20, and the secretion of the SpeB cysteine protease virulence determinant of Streptococcus pyogenes21.
Perhaps AI-2 was originally a metabolic end product with no role in signaling, and some groups of bacteria subsequently evolved to detect it as a signal. If so, this evolutionary process may still be underway. Clearly, V. harveyi, and possibly other species of Vibrio, use AI-2 as a quorum signal. E. coli, S. typhimurium, P. gingivalis, S. flexneri and S. pyogenes also appear to do so. The AI-2 detection systems have evolved independently in Vibrios and in enteric bacteria9,
14, and have not been characterized in other bacteria. Since such a large number of bacterial species release AI-2, it seems highly likely that some use it for quorum sensing, while it remains equally likely that other groups of bacteria may not yet have evolved mechanisms to decode this signal.
REFERENCES
- Fuqua, W.C., Winans, S.C. & Greenberg, E.P. J. Bacteriol. 176, 269−275 (1994). | PubMed | ISI | ChemPort |
- Cell-Cell Signaling in Bacteria. (Dunny, G. & Winans, S.C. (eds); ASM Press, Washington, D.C.; 1999).
- Bassler, B.L., Greenberg, E.P. & Stevens, A.M. J. Bacteriol. 179, 4043−4045 (1997). | PubMed | ISI | ChemPort |
- Surette, M.G., Miller, M.B., & Bassler, B.L. Proc. Natl. Acad. Sci. USA 96, 1639−1644 (1999). | Article | PubMed | ChemPort |
- Chen, X. et al. Nature in the press (2002)
- Yoon, H.S. & Golden, J.W. Science 282, 935−938 (1998). | Article | PubMed | ISI | ChemPort |
- Whitehead, N.A., Barnard, A.M., Slater, H., Simpson, N.J. & Salmond, G.P. FEMS Microbiol. Rev. 25, 365−404 (2001). | Article | PubMed | ISI | ChemPort |
- Schauder, S., Shokat, K., Surette, M.G. & Bassler, B.L. Mol. Microbiol. 41, 463−476 (2001). | Article | PubMed | ISI | ChemPort |
- Bassler, B.L., Wright, M. & Silverman, M.R. Mol. Microbiol. 13, 273−286 (1994). | PubMed | ISI | ChemPort |
- Quiocho, F.A. & Ledvina, P.S. Mol. Microbiol. 20, 17−25 (1996). | PubMed | ISI | ChemPort |
- Winans, S.C. & Bassler, B.L. J. Bacteriol. in the press (2002). | Article | PubMed |
- Surette, M.G. & Bassler, B.L. Mol. Microbiol. 31, 585−595 (1999). | Article | PubMed | ISI | ChemPort |
- Surette, M.G. & Bassler, B.L. Proc. Natl. Acad. Sci. USA 95, 7046−7050 (1998). | Article | PubMed | ChemPort |
- Taga, M.E., Semmelhack, J.L. & Bassler, B.L. Mol. Microbiol. 42, 777−793 (2001). | Article | PubMed | ISI | ChemPort |
- Shimizu, S., Shiozaki, S., Ohshiro, T. & Yamada, H. Eur. J. Biochem. 141, 385−392 (1984). | Article | PubMed | ISI | ChemPort |
- Sperandio, V., Mellies, J.L., Nguyen, W., Shin, S. & Kaper, J.B. Proc. Natl. Acad. Sci. USA 96, 15196−15201 (1999). | Article | PubMed | ChemPort |
- DeLisa, M.P., Wu, C.F., Wang, L., Valdes, J.J. & Bentley, W.E. J. Bacteriol. 183, 5239−5247 (2001). | Article | PubMed | ISI | ChemPort |
- Sperandio, V., Torres, A.G., Giron, J.A. & Kaper, J.B. J. Bacteriol. 183, 5187−5197 (2001). | Article | PubMed | ISI | ChemPort |
- Chung, W.O., Park, Y., Lamont, R.J., McNab, R., Barbieri, B. & Demuth, D.R. J. Bacteriol. 183, 3903−3909 (2001). | Article | PubMed | ISI | ChemPort |
- Day, W.A., Jr. & Maurelli, A.T. Infect. Immun. 69, 15−23 (2001). | Article | PubMed | ISI | ChemPort |
- Lyon, W.R., Madden, J.C., Levin, J.C., Stein, J.L. & Caparon, M.G. Mol. Microbiol. 42, 145−157 (2001). | Article | PubMed | ISI | ChemPort |
|
