2006bassler

Upload: swapnil-shinde

Post on 29-May-2018

218 views

Category:

Documents


0 download

TRANSCRIPT

  • 8/9/2019 2006Bassler

    1/10

  • 8/9/2019 2006Bassler

    2/10

    238 Cell 125, April 21, 2006 2006 Elsevier Inc.

    How does quorum sensing work?

    As a population o quorum-sens-

    ing bacteria grows, a proportional

    increase in the extracellular con-

    centration o the signaling molecule

    occurs. When a threshold concen-

    tration is reached, the group detects

    the signaling molecule and responds

    to it with a population-wide alteration

    in gene expression (Figure 2A). Pro-

    cesses controlled by quorum sensing

    are usually ones that are unproduc-

    tive when undertaken by an individual

    bacterium but become eective when

    undertaken by the group. For exam-

    ple, in addition to competence and

    bioluminescence, quorum sensing

    controls virulence actor secretion,biolm ormation, and sporulation.

    Thus, quorum sensing is a mecha-

    nism that allows bacteria to unction

    as multicellular organisms and to reap

    benets that they could never obtain

    i they always acted as loners.

    A chemical vocabulary has been

    established in which Gram-negative

    quorum-sensing bacteria such as Vib-

    rio communicate with AHLs, which are

    the products o LuxI-type autoinducer

    synthases. These small molecules are

    detected by cognate cytoplasmic LuxRproteins that, upon binding their part-

    ner autoinducer, bind DNA and activate

    transcription o target quorum-sensing

    genes. By contrast, Gram-positive quo-

    rum-sensing bacteria, such as Strep-

    tococcus and Bacillus, predominantly

    communicate with short peptides that

    oten contain chemical modications.

    For example, the signaling molecule or

    genetic competence in B. subtilis ComX

    is a 6 amino acid peptide whose trypto-

    phan residue has been modied by the

    attachment o a geranyl group (Figure 1;

    Okada et al., 2005). Signaling peptides

    such as ComX are recognized by mem-

    brane bound two-component sensor

    histidine kinases. Signal transduction

    occurs by phosphorylation cascades

    that ultimately impinge on DNA bind-

    ing transcription actors responsible or

    regulation o target genes. In general,

    bacteria keep their AHL and peptide

    quorum-sensing conversations private

    by each speacies o bacteria produc-

    ing and detecting a unique AHL (AHLs

    dier in their acyl side-chain moieties),

    peptide, or combination thereo.

    AHLs and peptides represent the

    two major classes o known bacterial

    cell-cell signaling molecules. However,

    our appreciation o the complexity o

    the chemical lexicon is increasing as

    new molecules are discovered that

    convey inormation between cells. For

    example, a amily o molecules generi-

    cally termed autoinducer-2 (AI-2) has

    been ound to be widespread in the

    bacterial world and to acilitate inter-

    species communication. AI-2s are all

    derived rom a common precursor, 4,5-

    dihydroxy-2,3 pentanedione (DPD), the

    product o the LuxS enzyme (Figure

    1). DPD undergoes spontaneous rear-

    rangements to produce a collection o

    interconverting molecules, some (andperhaps all) o which encode inorma-

    tion (Xavier and Bassler, 2005). Pre-

    sumably, AI-2 interconversions allow

    bacteria to respond to endogenously

    produced AI-2 and also to AI-2 pro-

    duced by other bacterial species in

    the vicinity, giving rise to the idea that

    AI-2 represents a universal language:

    a Bacterial Esperanto. AI-2, oten in

    conjunction with an AHL or oligopep-

    tide autoinducer, controls a variety o

    traits in dierent bacteria ranging rom

    bioluminescence in V. harveyi to growthin Bacillus anthracis to virulence in Vib-

    rio cholerae and many other clinically

    relevant pathogens.

    The streptomycetes, common soil-

    dwelling Gram-positive bacteria, use

    -butyrolactones (Figure 1) to con-

    trol morphological dierentiation and

    secondary metabolite production.

    The best studied o these signals, A-

    actor o Streptomyces griseus (one

    o the earliest recognized signaling

    molecules in bacteria), antagonizes

    a DNA binding repressor protein,

    ArpA, thereby promoting the orma-

    tion o hair-like projections known as

    aerial hyphae and the production o

    the antibiotic streptomycin (Khokhlov

    et al., 1967; Onaka et al., 1995). Inter-

    estingly, -butyrolactones are struc-

    tural analogs o AHLs; however, no

    crossrecognition o signals has been

    cited to date. Myxococcus xanthus,

    another soil bacterium, employs a

    mixture o amino acids derived rom

    extracellular proteolysis as signal-

    ing molecules (Kuspa et al., 1992).

    M. xanthus monitors the environment

    Fiur 1. Rprsaiv Bacria

    Mcus Ivvd i Ircuar

    IracisA Vibrio fscheri AHL autoinducer structureis shown as an example; however, a vari-ety o acyl side chains exist in this amily oautoinducers. Likewise, the Bacillus subti-lis ComX peptide autoinducer is shown asa representative o the variety o peptidesused by Gram-positive bacteria as auto-inducers. The ComX and SapB structureswere redrawn rom images kindly providedby D. Dubnau and J. Willey, respectively.

  • 8/9/2019 2006Bassler

    3/10

    Cell 125, April 21, 2006 2006 Elsevier Inc. 239

    or the simultaneous presence o starvation conditions

    and trace amounts o certain amino acids (a mixture

    o tryptophan, proline, tyrosine, phenylalanine, leucine,

    and isoleucine is especially potent) prior to initiating the

    quorum-sensing cascade that culminates in a spore-

    lled ruiting body (discussed urther below). Other mol-

    ecules, including 3-OH palmitic acid methyl ester, cyclic

    dipeptides, and quinolones (see below), also have roles

    in bacterial cell-cell signaling (or review, see Waters and

    Bassler, 2005).

    Sending Mixed MessagesSome quorum-sensing molecules, such as the quino-

    lone signal (2-heptyl-3-hydroxy-4-quinolone) o Pseu-

    domonas aeruginosa (PQS or Pseudomonasquinolone

    signal, Figure 1), are extremely hydrophobic (Pesci et

    al., 1999). This is problematic because PQS must travel

    rom cell to cell in an aqueous environment. How does

    P. aeruginosa manage to disperse a water-insoluble

    signaling molecule in water? A solution to this mystery

    was recently reported (Mashburn and Whiteley, 2005).

    In a process reminiscent o eukaryotic packaging o

    cargo into vesicles that are tracked between organ-

    elles, Gram-negative bacteria, such as Pseudomonas,

    pinch o 0.5 m-sized vesicles rom their outer mem-

    branes, and oten these vesicles transport various kinds

    o macromolecules. Mashburn and Whiteley nd that

    Pseudomonas packages PQS into vesicles derived rom

    the bacterial membrane and the vesicles deliver the qui-

    nolones to neighboring cells (Figure 2B). Remarkably,

    PQS mediates its own packaging; mutants blocked in

    quinolone production ail to produce membrane vesicles

    but regain the capacity to do so when chemically syn-

    thesized PQS is supplied exogenously (Mashburn and

    Whiteley, 2005). In addition to PQS, the Pseudomonas

    vesicles contain other quinolones that unction as anti-

    biotics to kill other species o bacteria, such as Staphy-

    lococcus epidermidis. Thus, P. aeruginosa sends mixedmessages: To its own kind, it emits signals that oster

    group behavior, but to other kinds o bacteria, the mes-

    sage delivered is atal (Figure 2B).

    Static in the Line

    Recent discoveries o quorum-quenching mechanisms

    suggest that biological tit-or-tat matches have evolved

    or counteracting quorum-sensing bacteria. Presum-

    ably, anti-quorum-sensing strategies are deployed so

    that one species o bacteria can outcompete another

    quorum-sensing species or so that a eukaryote can end

    o a quorum-sensing bacterial invader (Figure 2C). One

    such activity was discovered in soil samples assayed

    or intererence with AHL detection and response in a

    Fiur 2. Bacria Cmmuicai vr Disacs(A) Quorum sensing. Quorum-sensing bacteria produce and respond to the extracellular accumulation o signal molecules called autoinducers(depicted as green spheres).(B) Mixed messaging. Membrane vesicles trac the Pseudomonas aeruginosa quinolone signal (PQS; depicted as green triangles) between cells.PQS acilitates group behavior when delivered to other P. aeruginosa cells, but other quinolones (X), also contained in the vesicles, are antibioticsthat kill other bacterial species.(C) Quorum quenching. Quorum-sensing bacteria are vulnerable to a variety o quorum-quenching mechanisms such as the enzymatic inactivationo autoinducers or the presence o autoinducer antagonists (yellow spheres), molecules with structure similar to autoinducers that prevent autoin-ducer detection and response.(D) Conversations across kingdoms. Enterococcus aecalis produces a cytolysin composed o two subunits. The small cytolysin subunit acts as anautoinducer that monitors the environment or other E. aecalis cells. The large cytolysin subunit monitors the vicinity or eukaryotic host cells. Let:I no host cells are present, the two subunits orm an inactive complex, and production o the subunits is held at a low basal level. Right: I host cellsare present, the large subunit binds to the surace o the host cells, leaving the small subunit ree to induce high-level production o both subunits,which together bind to target cells and cause them to lyse.

  • 8/9/2019 2006Bassler

    4/10

    240 Cell 125, April 21, 2006 2006 Elsevier Inc.

    promiscuous AHL quorum-sensing-responsive reporter

    strain. The inhibitory activity was dened as a lactonase

    enzyme, AiiA, which cleaves the acyl moiety rom the

    lactone rings o AHLs (Dong et al., 2000). AiiA is espe-

    cially nonspecic with regard to acyl side chains, so it

    is believed to inactivate many AHL autoinducers. AiiA

    production is attributed to a variety o Bacillus species,

    which is noteworthy because Bacillus spp. are Gram-

    positive bacteria and use peptide autoinducers to com-

    municate. Thus, Bacillus renders the Gram-negative

    bacterial community mute while managing to continue its

    own conversations uninterrupted. Analogous strategies

    have since been discovered. In a particularly insidious

    scheme, Variovorax paradoxus destroys AHLs with an

    AHL-degrading enzyme that unctions by ring opening;

    ater disabling quorum sensing in its oes, V. paradoxus

    consumes the linearized product and uses it to acquire

    carbon and nitrogen (Leadbetter and Greenberg, 2000).It remains to be established whether AHL degradation

    has signicant consequences or bacterial signaling in

    the natural environment.

    Intererence with peptide signaling is also well known.

    Staphylococcus aureus relies on peptide quorum sensing

    or virulence, and strains o this pathogen are grouped

    according to the autoinducing peptide that they produce.

    Each autoinducing peptide, while activating the quorum-

    sensing cascade o the group that produces it, crossin-

    hibits quorum sensing in the other groups (Figure 2C). The

    autoinducing peptides are extremely similar in structure,

    and crossinhibition occurs because the dierent peptides

    compete or binding to the autoinducer receptors (Lyon etal., 2002). Presumably, in mixed inections, the S. aureus

    group that is rst to successully establish its quorum-

    sensing cascade shuts down quorum sensing in the other

    groups and becomes the predominant group to colo-

    nize the host. This has been borne out in in vivo mouse

    abscess models in which mice injected with a particular

    S. aureus group are susceptible to inection, but not i the

    S. aureus group is coinjected with the autoinducing pep-

    tide rom another S. aureus group (Wright et al., 2005).

    Quorum-sensing intererence has been reported or

    AI-2-mediated communication. Some bacteria, such as

    Escherichia coli and Salmonella typhimurium, possess

    AI-2-specic importers that are induced in response

    to high levels o AI-2 in the environment. AI-2 import

    eliminates the signal rom the extracellular environment.

    In mixed-species consortia, because AI-2 molecules

    spontaneously interconvert, bacteria with AI-2 import-

    ers can consume both their own AI-2 and that produced

    by other species in the vicinity. This capability enables

    AI-2-consuming bacteria to interere with other species

    ability to accurately count the cell density o the popula-

    tion and to appropriately respond to changes in it (Xavier

    and Bassler, 2005).

    Eukaryotes too appear to possess armaments that

    are specically aimed at quorum-sensing bacteria.

    The seaweed Delisea pulchra produces halogenated

    uranones that are structural analogs o AHLs. Thesemolecules bind to LuxR-type transcription actors and

    cause their proteolysis (Maneeld et al., 2002). Reactive

    oxygen and nitrogen intermediates produced by NADPH

    oxidase, an important constituent o the mammalian

    immune system, inactivate S. aureus autoinducing pep-

    tides in a mouse virulence model system (Rothork et al.,

    2004). Paraoxonase enzymes hydrolyze esters and, in

    humans, are considered to have important antiathero-

    genic unctions. The human paraoxonase (PON) amily

    has three members: PON1, PON2, and PON3. Although

    their endogenous substrates have not been dened, it is

    clear that they are lactonases. At least PON2 is capable

    o inactivating a variety o bacterial AHL autoinducers,suggesting that humans possess anti-quorum-sensing

    capabilities (Draganov et al., 2005). However, a direct

    in vivo link between PON2 and AHL inactivation awaits

    experimental verication.

    Intimate Conversations

    Some bacteria, as we have discussed, converse over

    a distance by exchanging diusible signals that allow

    members o a community o cells to communicate (quo-

    Fiur 3. Iima Bacria Cvrsais(A) Reciprocal C signaling in Myxococcus xanthus. The surace-displayed C signal protein interacts with a hypothetical receptor on an adjacent cellto transmit a signal that promotes ruiting-body ormation.(B) Crisscross signaling in Bacillus subtilis. Sporulation takes place in a two-compartment sporangium in which three intercellular signals, the se-creted signaling proteins SpoIIR (IIR) and SpoIVB (IVB) and an unknown signal (?), are exchanged back and orth. For simplicity, the illustration doesnot convey that the second and third signaling events take place ater the smaller cell has been enguled by the larger cell.(C) Contact-dependent inhibition in Escherichia coli. Cells producing the proteins CdiAB orm aggregates with, and inhibit the growth o, cells lack-ing the genes or the growth-inhibiting proteins.

  • 8/9/2019 2006Bassler

    5/10

    Cell 125, April 21, 2006 2006 Elsevier Inc. 241

    rum sensing). At the opposite extreme are short-range

    signals that require direct contact between individual

    cells or inormation exchange. One classic example o

    this is C signaling, which is critical or multicellular ruit-

    ing-body ormation by M. xanthus (Figure 3A). Cells o

    this social bacterium move on suraces by gliding motil-

    ity in a manner in which individual cells oten reverse

    direction. C signaling promotes a coordinated gliding

    behavior known as streaming, in which reversals are

    suppressed. Streaming culminates in the ormation o

    a mound o cells within which spore ormation takes

    place. The C signal is a 17 kDa protein that is derived

    by proteolysis rom a larger precursor (Kim and Kaiser,

    1990a, 1990c; Lobedanz and Sogaard-Andersen, 2003).

    Elegant experiments in which cells were articially

    orced into alignment in tiny grooves on an agar plate

    showed that C signaling requires contact between cells.

    It has been hypothesized that the unction o C signalingis to report on cell alignment, which takes place during

    the aggregation phase o ruiting-body ormation (Julien

    et al., 2000; Kim and Kaiser, 1990b). C signaling also

    governs the expression o numerous genes required

    or spore ormation. A cell-surace receptor (perhaps

    located at the cell poles) is presumed to be required

    or C signaling, but the putative receptor and the down-

    stream signal-transduction system remain unknown.

    The concept that M. xanthus cells are capable o inti-

    mate interaction is reinorced by the recent dramatic

    demonstration that particular motility-associated, GFP-

    tagged outer-membrane lipoproteins readily exchange

    between M. xanthus cells (Nudleman et al., 2005). Thisnding suggests that M. xanthus cells are, at a minimum,

    capable o at least briefy using their outer membranes

    and sharing outer-membrane proteins.

    Exactly how C signaling suppresses gliding reversals

    remains largely unknown, but an important insight comes

    rom recent cytological studies on the protein FrzS,

    which is required or directed movement (Mignot et al.,

    2005). Gliding is mediated by ber-like pili that extend

    rom one end o the cell and use their tips like a harpoon

    to latch on to the substratum or another cell. Movement

    is subsequently achieved by retraction o the pili. The

    bacteria reverse direction by disassembling pili at one

    cell pole and reassembling pili at the other pole. FrzS

    oscillates rom pole to pole, evidently traveling along an

    undened, cytoskeletal track. Conceivably, FrzS is part

    o a pilus-assembly complex that alternately governs

    pilus construction at one pole and then the other. I so, a

    downstream consequence o C signaling might be sup-

    pression o this pole-to-pole oscillation.

    An extreme example o intimate cell-to-cell signal-

    ing occurs during the process o spore ormation in

    Bacillus subtilis. Spores are ormed in a two-chamber

    sporangium that consists o a orespore that ultimately

    becomes the spore and a mother cell that nurtures the

    developing spore. Early in sporulation, the orespore and

    mother cell lie side by side, but later in development, the

    mother cell wholly enguls the orespore to create a cell

    within a cell. Thus, during sporulation, the orespore and

    the mother cell are in intimate contact across the mem-

    branes where the cells abut each other. Each o these

    cells ollows its own distinctive program o gene expres-

    sion, but the two lines o gene expression are not inde-

    pendent o one another. Rather, they are linked in criss-

    cross ashion by three intercellular signaling pathways

    (Figure 3B and Figure 4A; or reviews, see Losick and

    Stragier, 1992; Rudner and Losick, 2001). A secreted

    signaling protein (SpoIIR) produced in the orespore

    under the control o the orespore transcription actor

    F triggers the appearance o E in the mother cell. Next,

    E sets in motion a poorly understood chain o events

    that activates G in the orespore. Finally, G directs the

    synthesis o SpoIVB, a secreted signaling protein that

    triggers the appearance o K in the mother cell.

    A case o an in your ace conversation is contact-

    mediated growth inhibition in E. coli (Aoki et al., 2005).

    Certain wild strains o E. coli inhibit the growth o stan-

    Fiur 4. Ircuar Siai iB. subtilis ad Cac-

    Dpd Ihibii byE. coli(A) A fuorescence micrograph o a eld o sporulating B. subtilis cellsharboringgp used to a promoter controlled by E. Cell membraneswere stained with TMA-DPH and were alse colored red. Green fuo-rescence is observed in the mother-cell compartment o sporangia inwhich E was activated by a signal emanating rom the smaller, adja-cent orespore compartment (see text and Figure 3B). The experimentwas carried out and kindly provided by T. Doan and D. Rudner.(B) A fuorescence micrograph o an aggregate o E. coli inhibitor cellslabeled green with GFP and noninhibitor (sensitive) cells labeled redwith DsRed. The inhibitor cells produce the contact-dependent inhibi-tor proteins CdiA and CdiB (see text and Figure 3C). The experimentwas carried out and kindly provided by S. Aoki and D. Low.

  • 8/9/2019 2006Bassler

    6/10

    242 Cell 125, April 21, 2006 2006 Elsevier Inc.

    dard laboratory strains (E. coli K12) by a mechanism

    known as contact-dependent inhibition. Inhibition is

    mediated by a pair o proteins called CdiA and CdiB (or

    contact-dependent inhibitors A and B) that resemble

    two-partner secretion proteins that are exported to the

    cell surace by a pathway involving proteolytic process-

    ing. When the genes or CdiA and CdiB are introduced

    into E. coli K12, the laboratory strain acquires the capac-

    ity to cause growth inhibition in a manner that requires

    cell-to-cell contact (Figure 3C and Figure 4B). Contact

    dependence was demonstrated in an elegant experi-

    ment in which the inhibitory cells were separated rom

    the sensitive cells by a porous membrane. Pores o 0.4

    m prevented growth inhibition, whereas pores o 0.8

    m (large enough to allow cells to slip through) did not.

    Furthermore, inhibitory cells were shown to orm aggre-

    gates with sensitive cells in a manner that depended on

    the Cdi proteins. Importantly, sensitive cells displayingsurace pili are resistant to growth inhibition, presum-

    ably because the pili keep the inhibitory cells at a sae

    distance. The physiological signicance o contact-

    dependent inhibition is unclear, but an attractive pos-

    sibility is that it unctions in the regulation o growth o

    specic cells in complex communities o bacteria.

    Communal Living

    Most o the examples we have considered so ar involve

    interactions among large numbers o bacteria in popula-

    tions existing in liquid environments. Many bacteria are

    also capable o coordinating their behavior to orm ses-

    sile communities consisting o large numbers o denselypacked cells. These architecturally complex communities,

    called biolms, orm on suraces or at air-liquid interaces.

    Cells in biolms are typically held together by an extracel-

    lular matrix composed o polysaccharides, protein, and

    oten DNA. As in communes, all o the members o these

    bacterial communities cooperate in the construction o

    the biolm by contributing matrix components.

    An interesting twist on communal living has recently been

    reported in Vibrio cholerae, the causative agent o cholera

    (Meibom et al., 2005). V. cholerae exists both ree swim-

    ming in the ocean and also as a constituent o biolms,

    requently on chitin-containing exoskeletons. V. cholerae

    eeds on this solid polymer o N-acetylglucosamine, during

    which time it acquires DNA by natural transormation (DNA

    is known to be present in biolms at concentrations above

    100g/ml) . DNA uptake rom the environment is postulated

    to allow V. cholerae to obtain new genes, thereby diversiy-

    ing its genome. Genes encoding type IV pili and homologs

    o genes required or genetic competence in Gram-positive

    bacteria are induced in the presence o chitin and are nec-

    essary or V. cholerae transormation. An intact quorum-

    sensing cascade is also a prerequisite or transormation

    on chitin. These studies demonstrate or the rst time that

    V. cholerae has natural competence, that bacterial evolu-

    tion could be occurring on interaces in which bacteria are

    in direct contact with suraces, and nally that this process

    is driven by cell-cell signaling.

    Another ascinating example o communal living is

    involved in aerial mycelium ormation in streptomycetes,

    ungus-like bacteria that undergo a complex process o

    morphological dierentiation. During this process, an

    aerial mycelium is erected that consists o hair-like la-

    ments that project rom the surace o the colony. The

    colony is composed o a branching network o hyphae

    known as the substrate mycelium. Contributing to the

    erection o the aerial hyphae are cell-surace proteins

    and, interestingly in the present context, lantibiotic-like

    peptides known as SapB (in Streptomyces coelicolor;

    Figure 1) and SapT (in Streptomyces tendae ). SapB

    and SapT, which contain the dening thioester bridge

    o lanthionines, accumulate extracellularly during aerial

    mycelium ormation, serving as suractants to acilitate

    the release o nascent aerial hyphae rom the substrate

    mycelium (Kodani et al., 2004; Willey et al., 2006). Thus,

    as in the case o quorum-sensing molecules, the accu-mulation o SapB and SapT refects the cooperative

    activity o the bacterial community. Unlike quorum-sens-

    ing molecules, SapB and SapT are not signals; rather,

    they are morphogenetic peptides that play a mechanical

    role in development.

    Fratricide

    Since the discovery o penicillin by Sir Alexander

    Fleming in 1928, it has become widely recognized that

    microorganisms engage in chemical warare in which

    one species wards o other species through the pro-

    duction and release o antibiotics and other antimicro-

    bial agents. Sometimes, as in the case o the bacte-riocins, bacteria produce agents that kill other strains

    o the same species. Recently, however, two radical

    examples o ratricide have come to light in which

    some members o a genetically identical population o

    cells kill other members (siblings) o the same popu-

    lation: cannibalism in Bacillus subtilis and allolysis in

    Streptococcus pneumoniae.

    Spore ormation by B. subtilis, which is triggered by

    nutrient limitation, is an elaborate developmental pro-

    cess that takes place over the course o 7 to 10 hours

    and involves the conversion o a growing cell into a dor-

    mant cell type (a spore) that can remain inert or many

    years. Thereore, it comes as no surprise that the deci-

    sion to orm a spore is a lie-or-death one. A B. subtilis

    cell could be at a considerable disadvantage i it com-

    mits to spore ormation in response to what turns out

    to be a brie fuctuation in nutrient availability. To guard

    against this possibility, the bacterium deploys a sys-

    tem o cannibalism in which cells that have entered the

    sporulation pathway orestall the absolute commitment

    to spore ormation (Gonzalez-Pastor et al., 2003).

    At the heart o the cannibalism system is a bistable

    switch that governs the activation o the master regulator

    or entry into sporulation, Spo0A. In response to nutri-

    ent limitation, about hal o the B. subtilis cells activate

    Spo0A and enter the pathway leading to sporulation,

    and the other (Spo0A-inactive) cells do not. Cells that

  • 8/9/2019 2006Bassler

    7/10

    Cell 125, April 21, 2006 2006 Elsevier Inc. 243

    have activated Spo0A produce and export a killing actor

    and a protein toxin that together kill nonsporulating sib-

    lings. Their deaths result in the release o nutrients that,

    in turn, delay or reverse progression into sporulation by

    the cells that have activated Spo0A. When no siblings

    remain to be cannibalized and no other sources o nutri-

    ents become available, development progresses to the

    point that spore ormation becomes irreversible.

    Interestingly, and pertinent to the theme o this review,

    the response to the cannibalism toxin involves an inter-

    cellular chemical signaling system o unusual simplicity

    (Ellermeier et al., 2006). To avoid suicide, toxin-produc-

    ing cells (i.e., Spo0A-expressing cells) simultaneously

    produce a membrane bound immunity protein that neu-

    tralizes the toxin in the membrane. The immunity protein

    dually unctions in signal transduction and in protection

    rom the toxin. The immunity protein is encoded by a

    two-gene operon that also contains the gene or anautorepressor. When bound to the toxin, the immunity

    protein sequesters the autorepressor, thereby dere-

    pressing the operon and initiating synthesis o increased

    immunity protein and autorepressor. Thus, accumulation

    o immunity protein that is not bound to the toxin leads to

    unsequestered repressor, which unctions to downregu-

    late expression o the operon and reset the system.

    The second example o ratricide is the allolysis behav-

    ior o the pathogen S. pneumoniae (Guiral et al., 2005;

    Hvarstein et al., 2006). This Gram-positive bacterium is

    commonly ound in the nasopharynx but also causes a

    variety o invasive diseases including pneumonia, bac-

    teremia, otitis media, meningitis, and sinusitis. As dis-cussed above, S. pneumoniae is known or its ability to

    enter into a state o genetic competence under condi-

    tions o high cell population density in response to a

    secreted signaling peptide. Analogous to the case o B.

    subtilis sporulation, only a raction o the S. pneumoniae

    cells in the population become competent in response

    to the peptide autoinducer. Those that do so elaborate a

    bacteriocin that causes the lysis o noncompetent cells

    in the population. What is the purpose o preying on

    genetically identical siblings? Claverys and coworkers

    (Guiral et al., 2005; Hvarstein et al., 2006) report that

    the lysed cells release not only transorming DNA and

    nutrients but also pneumolysin and other actors impor-

    tant or virulence. Thus, rather than relying on sel or

    secretion o virulence actors, S. pneumoniae sacrices

    some its relatives or this purpose which acilitates inva-

    sion o its host.

    Conversations across the Divide

    Not only do bacteria sense one anothers presence and

    use chemical signals to communicate, there is new evi-

    dence suggesting that some bacteria also detect their

    eukaryotic hosts. Enterococcus aecalis is a Gram-posi-

    tive bacterium that is a commensal o the human intes-

    tine. It is also an opportunistic pathogen involved in

    urinary tract inections, bacteremia, and inective endo-

    carditis. Pathogenesis depends on a cytolysin com-

    posed o two nonidentical peptides (the small subunit,

    CylLS, and the large subunit, CylL

    L) that are posttrans-

    lationally modied, secreted, and activated. Enterococ-

    cal cytolysin is related to lantibiotics, and it is lethal to a

    broad range o prokaryotic and eukaryotic cells. Inter-

    estingly, the smaller o the two peptides (CylLS ) making

    up the cytolysin doubles as a peptide autoinducer in the

    quorum-sensing induction o the operon encoding the

    cytolysin and the genes required to modiy it (Haas et

    al., 2002). Two proteins, CylR1 (a predicted membrane

    protein) and CylR2 (a predicted DNA binding protein)

    repress transcription o the cytolysin genes, and dere-

    pression occurs in response to the cell-density-depen-

    dent extracellular accumulation o the CylLS

    peptide.

    Surprisingly, although the small cytolysin subunit is

    used to monitor bacterial cell density, the large cytolysin

    subunit appears to be used to monitor the environment

    or host cells (Figure 2D). The basis or this discoverywas the observation that cytolysin activity is dependent

    upon the presence o target cells, such as erythrocytes.

    Gilmore and colleagues report that in the absence o

    target cells, CylLL

    and CylLS

    orm a stable complex that

    is inactive or autoinduction and or lysis o target cells

    (Coburn et al., 2004). Importantly, however, CylLL

    has a

    higher anity or target cells than or CylLS. It is specu-

    lated that in the presence o host cells, CylLL

    dieren-

    tially adsorbs to host cells. Titration o CylLL

    by the host

    cells leaves CylLS

    ree to accumulate and to act as an

    autoinducer that promotes high-level synthesis o CylLS

    and CylLL, which, in turn, cause the target cells to lyse,

    presumably by the ormation o a pore.In addition to bacteria detecting host cells, there are

    examples o interkingdom chemical communication in

    which the host perceives the presence o the bacterial

    cells. As discussed above, some eukaryotes perceive

    hostile quorum-sensing bacteria and take measures

    to conound them. Another hostile eukaryotic-pro-

    karyotic dialogue occurs between dicotyledonous

    plants and the pathogenic bacterium Agrobacterium

    tumeaciens. In this case, the bacterium relies on

    host- and sel-produced cues. In response to host-

    produced signals, the bacterium transers a ragment

    o DNA called T-DNA to host cells. The T-DNA orces

    the host to commence production o opines that, in

    turn, eed the bacteria. Bacterially produced quorum-

    sensing signals induce plasmid transer between the

    bacterial community, increasing its overall inectivity

    (Zhu et al., 2000).

    Eukaryotes in symbiotic associations with bacteria

    also participate in the conversation, but they use riendly

    verbiage. For example, in the initial stages o symbio-

    sis between Rhizobium meliloti and its plant host, the

    earliest signals come rom the plant. Plant roots release

    favonoid molecules that Rhizobium detects in a com-

    partment called the rhizosphere (Peters et al., 1986). In

    response to these plant signals, Rhizobium activates

    transcription o nod genes that are under control o

    the NodD DNA binding protein. The nod genes encode

  • 8/9/2019 2006Bassler

    8/10

    244 Cell 125, April 21, 2006 2006 Elsevier Inc.

    enzymes that produce Nod actor, a bacterial signal (Fig-

    ure 1) that induces additional plant processes required

    or the development o nodules (Mulligan and Long,

    1985). Nodules provide the bacteria with a location that

    has conditions appropriate or nitrogen xation, which,

    in turn, benet the plant. Presumably, similar two-way

    conversations occur in

    many other bacterial-

    eukaryotic encounters,

    both aable and antag-

    onistic.

    I Understand You

    The signaling mecha-

    nisms we have con-

    sidered thus ar are

    apparently unique to

    bacteria. Remarkably,the Gram-negative bac-

    terium Providencia stu-

    artii provides an exam-

    ple o an intercellular

    signaling system that

    is used by both pro-

    karyotes and eukary-

    otes. P. stuartii is a quorum-sensing pathogen that

    causes nosocomial inections in humans. The structure

    o the P. stuartii signaling molecule is not known, but it

    appears to be a peptide. A genetic screen or P. stuar-

    tii mutants incapable o extracellular autoinducer pro-

    duction revealed the geneaarA. AarA has homology toRhomboid (RHO), which has been primarily studied in

    Drosophila melanogaster(Rather et al., 1999). In fies,

    RHO, a serine protease, is necessary or the intra-

    membrane cleavage, extracellular release, and activa-

    tion o ligands or the epidermal growth actor (EGF)

    receptor. EGF receptor signaling, in turn, is required

    or specication o cell ate. Consistent with this, D.

    melanogaster rho mutants are deective in several

    developmental processes including the ormation o

    the correct number o veins and their proper position-

    ing in wings and the organization o the compound

    eye. Evidence that AarA and RHO have a shared unc-

    tion in extracellular signaling comes rom the extraor-

    dinary nding that expression o P. stuartii aarA in a D.

    melanogaster rho mutant rescued wing-vein develop-

    ment, and, in the reciprocal experiment, introduction

    o D. melanogaster rho into a P. stuartii aarA mutant

    restored quorum-sensing signal production (Gallio et

    al., 2002). In an extension o the above analysis, eight

    prokaryotic Rhomboids were tested or the ability to

    cleave three D. melanogaster EGF receptor ligands;

    ve o these trials proved successul. The dierent

    rhomboids tested shared only minimal sequence

    homology; however, all contained the putative serine

    catalytic triad residues required or activity, and, in

    every case, amino acid substitutions at these residues

    destroyed the proteins unction in signal production

    (Urban et al., 2002). There are over a hundred known

    rhomboids in archaea, bacteria, and eukaryotes, but

    none o the microbial unctions are known. The above

    results suggest that these diverse proteins have a

    conserved enzymatic unction, as well as a conserved

    role in cell-cell communication.

    Conclusions: A Continuing

    Conversation

    Alexander Tomasz early

    impressions that bacterial

    cell-cell signaling could be an

    organizing principle underly-

    ing multicellular behavior were

    amazingly accurate. What he

    could not have imagined, and

    what we are only now begin-

    ning to appreciate, is the tre-mendous complexity in the

    extracellular chemical milieu

    that bacteria experience and

    interpret to garner inorma-

    tion about their growth status

    and potential; their cell num-

    bers; those o their neigh-

    bors; and the presence or absence o eukaryotes, both

    riend and oe. We now understand that there can be

    one-way, two-way, and multi-way chemical dialogues,

    as well as the spread o misinormation and the deliv-

    ery o deadly messages. Chemical conversations span

    species and kingdoms, and the particular moleculesused to convey particular pieces o inormation appear

    optimized or their specic purpose: long- versus short-

    range messaging, contact-mediated or -inhibited signal-

    ing, or intra- or extracellular delivery. Increasingly, the

    molecular mechanisms underpinning bacterial cell-cell

    signaling begin to resemble those employed by eukary-

    otes or temporal and spatial patterning, suggesting a

    shared evolutionary history.

    Looking ahead, two considerations lead us to believe

    that our understanding o the bacterial lexicon is primi-

    tive. First, bacteria are known to produce an extraordi-

    narily diverse array o small organic molecules with anti-

    bacterial, antiungal, antiviral, and immunomodulatory

    activities. It is generally believed that these agents serve

    as armamentaria in the internecine warare that bacteria

    wage against each other and against other microbes in

    the environment. But recent studies show that at subin-

    hibitory doses, antibiotics such as riampicin and eryth-

    romycin cause global changes in gene-expression pat-

    terns in bacteria, and many genes o unknown unction

    are aected (Goh et al., 2002; Tsui et al., 2004). Based on

    this, Davies and colleagues have suggested that many

    bacterially produced antibiotics may not be released

    exclusively or purposes o chemical warare (Goh et

    al., 2002; Tsui et al., 2004). Rather, these chemicals may

    act as signals in habitats where their local concentra-

    tions are well below those required to inhibit growth. Our

    Since the activatora cell-produced

    chemicalseems to impose a high

    degree of physiological homogeneity in

    a pneumococcal population with respect

    to competence, one is forced to conclude

    that in this case [a] bacterial population

    can behave as a biological unit with

    considerable coordination among its

    members.One wonders whether this kind of control

    may not be operative in some other

    microbial phenomena also.

    Alexander Tomasz, 1965

  • 8/9/2019 2006Bassler

    9/10

    Cell 125, April 21, 2006 2006 Elsevier Inc. 245

    second consideration stems rom estimates suggesting

    that only a minute raction o the Earths bacteria have

    been identied, much less cultivated in the laboratory.

    Indeed, bacteria and their viruses ar and away repre-

    sent the greatest repository o genetic diversity on the

    planet. Provocative new approaches, collectively known

    as metagenomics, in which DNA rom various environ-

    ments is directly cloned into surrogate hosts is providing

    access to this rich vein o genetic inormation (Brady et

    al., 2004; Handelsman et al., 1998; Rondon et al., 2000).

    With it come completely novel biosynthetic pathways or

    previously unknown natural products, many o which are

    small organic molecules. Some molecules discovered

    through metagenomics exhibit signaling activity. By

    extrapolation, it seems reasonable to suppose that vast

    numbers o microbially produced, small organic mole-

    cules with intercellular signaling activity await discovery.

    I so, we may have just begun to eavesdrop on a micro-bial agora: a bustling polyglot o chemical languages,

    each with its own biological story to tell.

    ACknowleDgMentS

    The authors thank Drs. B. Hammer, W. Chai, A. Hitchcock, C. Waters,and N. Wingreen or critical readings o the manuscript and D. Zus-man and L. Kroos or helpul discussions. This work was supportedby The Howard Hughes Medical Institute and NIH grants R01 GM065859 and R01 AI 054442 to B.L.B. and NIH grant R01 GM18568,NSF grant MCB-01 10090, and a Visiting Proessorship rom the MillerInstitute o the University o Caliornia at Berkeley to R.L.

    ReFeRenCeS

    Aoki, S.K., Pamma, R., Hernday, A.D., Bickham, J.E., Braaten, B.A.,and Low, D.A. (2005). Contact-dependent inhibition o growth inEscherichia coli. Science 309, 12451248.

    Brady, S.F., Chao, C.J., and Clardy, J. (2004). Long-chain N-acylty-rosine synthases rom environmental DNA. Appl. Environ. Microbiol.70, 68656870.

    Coburn, P.S., Pillar, C.M., Jett, B.D., Haas, W., and Gilmore, M.S.(2004). Enterococcus aecalis senses target cells and in response ex-presses cytolysin. Science 306, 22702272.

    Dong, Y.H., Xu, J.L., Li, X.Z., and Zhang, L.H. (2000). AiiA, an enzymethat inactivates the acylhomoserine lactone quorum-sensing signaland attenuates the virulence o Erwinia carotovora. Proc. Natl. Acad.Sci. USA 97, 35263531.

    Draganov, D.I., Teiber, J.F., Speelman, A., Osawa, Y., Sunahara, R.,and La Du, B.N. (2005). Human paraoxonases (PON1, PON2, andPON3) are lactonases with overlapping and distinct substrate speci-cities. J. Lipid Res.46, 12391247.

    Eberhard, A., Burlingame, A.L., Eberhard, C., Kenyon, G.L., Neal-son, K.H., and Oppenheimer, N.J. (1981). Structural identication oautoinducer o Photobacterium scheri lucierase. Biochemistry 20,24442449.

    Ellermeier, C., Hobbs, E.C., Gonzalez-Pastor, J.E., and Losick, R.(2006). A three-protein signal transduction pathway governing the re-sponse to a cannibalism toxin. Cell 124, 549559.

    Gallio, M., Sturgill, G., Rather, P., and Kylsten, P. (2002). A conservedmechanism or extracellular signaling in eukaryotes and prokaryotes.Proc. Natl. Acad. Sci. USA 99, 1220812213.

    Goh, E.B., Yim, G., Tsui, W., McClure, J., Surette, M.G., and Davies,

    J. (2002). Transcriptional modulation o bacterial gene expression bysubinhibitory concentrations o antibiotics. Proc. Natl. Acad. Sci. USA99, 1702517030.

    Gonzalez-Pastor, J.E., Hobbs, E.C., and Losick, R. (2003). Cannibal-

    ism by sporulating bacteria. Science 301, 510513.

    Guiral, S., Mitchell, T.J., Martin, B., and Claverys, J.P. (2005). Com-petence-programmed predation o noncompetent cells in the humanpathogen Streptococcus pneumoniae: genetic requirements. Proc.Natl. Acad. Sci. USA 102, 87108715.

    Haas, W., Shepard, B.D., and Gilmore, M.S. (2002). Two-componentregulator o Enterococcus aecalis cytolysin responds to quorum-sensing autoinduction. Nature415, 8487.

    Handelsman, J., Rondon, M.R., Brady, S.F., Clardy, J., and Goodman,R.M. (1998). Molecular biological access to the chemistry o unknownsoil microbes: a new rontier or natural products. Chem. Biol. 5,R245R249.

    Hvarstein, L.S., Martin, B., Johnsborg, O., Granadel, C., and Clav-erys, J.P. (2006). New insights into the pneumococcal ratricide: rela-

    tionship to clumping and identication o a novel immunity actor. Mol.Microbiol. 59, 12971307.

    Julien, B., Kaiser, A.D., and Garza, A. (2000). Spatial control o celldierentiation in Myxococcus xanthus. Proc. Natl. Acad. Sci. USA 97,90989103.

    Khokhlov, A.S., Tovarova, I.I., Borisova, L.N., Pliner, S.A., Schevchen-ko, L.A., Kornitskaya, E.Y., Ivkina, N.S., and Rappoport, I.A. (1967).A-actor responsible or the biosynthesis o streptomycin by a mutantstrain o Actinomyces streptomycini. Dokl. Akad. Nauk SSSR 177,232235.

    Kim, S.K., and Kaiser, D. (1990a). C-actor: a cell-cell signaling pro-tein required or ruiting body morphogenesis o M. xanthus. Cell 61,1926.

    Kim, S.K., and Kaiser, D. (1990b). Cell alignment required in dieren-

    tiation o Myxococcus xanthus. Science249, 926928.

    Kim, S.K., and Kaiser, D. (1990c). Purication and properties o Myxo-coccus xanthus C-actor, an intercellular signaling protein. Proc. Natl.Acad. Sci. USA 87, 36353639.

    Kodani, S., Hudson, M.E., Durrant, M.C., Buttner, M.J., Nodwell, J.R.,and Willey, J.M. (2004). The SapB morphogen is a lantibiotic-likepeptide derived rom the product o the developmental gene ramSin Streptomyces coelicolor. Proc. Natl. Acad. Sci. USA 101, 1144811453.

    Kuspa, A., Plamann, L., and Kaiser, D. (1992). Identication o heat-stable A-actor rom Myxococcus xanthus. J. Bacteriol. 174, 33193326.

    Leadbetter, J.R., and Greenberg, E.P. (2000). Metabolism o acyl-ho-moserine lactone quorum-sensing signals by Variovorax paradoxus.J. Bacteriol. 182, 69216926.

    Lobedanz, S., and Sogaard-Andersen, L. (2003). Identication o theC-signal, a contact-dependent morphogen coordinating multipledevelopmental responses in Myxococcus xanthus. Genes Dev. 17,21512161.

    Losick, R., and Stragier, P. (1992). Crisscross regulation o cell-type-specic gene expression during development in B. subtilis. Nature355, 601604.

    Lyon, G.J., Wright, J.S., Christopoulos, A., Novick, R.P., and Muir, T.W.(2002). Reversible and specic extracellular antagonism o receptor-histidine kinase signaling. J. Biol. Chem.277, 62476253.

    Maneeld, M., Rasmussen, T.B., Henzter, M., Andersen, J.B., Stein-berg, P., Kjelleberg, S., and Givskov, M. (2002). Halogenated ura-

    nones inhibit quorum sensing through accelerated LuxR turnover.Microbiol. 148, 11191127.

  • 8/9/2019 2006Bassler

    10/10

    246 Cell 125 April 21 2006 2006 Elsevier Inc

    Mashburn, L.M., and Whiteley, M. (2005). Membrane vesicles tra-c signals and acilitate group activities in a prokaryote. Nature437,422425.

    Meibom, K.L., Blokesch, M., Dolganov, N.A., Wu, C.Y., and School-

    nik, G.K. (2005). Chitin induces natural competence in Vibrio cholerae.Science 310, 18241827.

    Mignot, T., Merlie, J.P., Jr., and Zusman, D.R. (2005). Regulated pole-to-pole oscillations o a bacterial gliding motility protein. Science 310,855857.

    Mulligan, J.T., and Long, S.R. (1985). Induction o Rhizobium melilotinodC expression by plant exudate requires nodD. Proc. Natl. Acad.Sci. USA 82, 66096613.

    Nealson, K.H., Platt, T., and Hastings, J.W. (1970). Cellular control othe synthesis and activity o the bacterial luminescent system. J. Bac-teriol. 104, 313322.

    Nudleman, E., Wall, D., and Kaiser, D. (2005). Cell-to-cell transer obacterial outer membrane lipoproteins. Science 309, 125127.

    Okada, M., Sato, I., Cho, S.J., Iwata, H., Nishio, T., Dubnau, D., andSakagami, Y. (2005). Structure o the Bacillus subtilis quorum-sensingpeptide phereomone ComX. Nat. Chem. Biol. 1, 2324.

    Onaka, H., Ando, N., Nihira, T., Yamada, Y., Beppu, T., and Horinouchi,S. (1995). Cloning and characterization o the A-actor receptor generom Streptomyces griseus. J. Bacteriol. 177, 60836092.

    Pesci, E.C., Milbank, J.B., Pearson, J.P., McKnight, S., Kende, A.S.,Greenberg, E.P., and Iglewski, B.H. (1999). Quinolone signaling inthe cell-to-cell communication system o Pseudomonas aeruginosa.Proc. Natl. Acad. Sci. USA 96, 1122911234.

    Peters, N.K., Frost, J.W., and Long, S.R. (1986). A plant favone, luteo-lin, induces expression o Rhizobium meliloti nodulation genes. Sci-ence233, 977980.

    Rather, P.N., Ding, X., Baca-DeLancey, R.R., and Siddiqui, S. (1999).

    Providencia stuartii genes activated by cell-to-cell signaling and iden-tication o a gene required or production or activity o an extracel-lular actor. J. Bacteriol. 181, 71857191.

    Rondon, M.R., August, P.R., Bettermann, A.D., Brady, S.F., Grossman,T.H., Liles, M.R., Loiacono, K.A., Lynch, B.A., MacNeil, I.A., Minor, C.,et al. (2000). Cloning the soil metagenome: a strategy or accessingthe genetic and unctional diversity o uncultured microorganisms.Appl. Environ. Microbiol. 66, 25412547.

    Rothork, J.M., Timmins, G.S., Harris, M.N., Chen, X., Lusis, A.J.,Otto, M., Cheung, A.L., and Gresham, H.D. (2004). Inactivation o abacterial virulence pheromone by phagocyte-derived oxidants: newrole or the NADPH oxidase in host deense. Proc. Natl. Acad. Sci.USA 101, 1386713872.

    Rudner, D.Z., and Losick, R. (2001). Morphological coupling in devel-opment: lessons rom prokaryotes. Dev. Cell 1, 733742.

    Tomasz, A. (1965). Control o the competent state in Pneumococcusby a hormone-like cell product: an example or a new type o regula-tory mechanism in bacteria. Nature208, 155159.

    Tsui, W.H., Yim, G., Wang, H.H., McClure, J.E., Surette, M.G., and Da-vies, J. (2004). Dual eects o MLS antibiotics: transcriptional modula-tion and interactions on the ribosome. Chem. Biol. 11, 13071316.

    Urban, S., Schlieper, D., and Freeman, M. (2002). Conservation o in-tramembrane proteolytic activity and substrate specicity in prokary-otic and eukaryotic rhomboids. Curr. Biol. 12, 15071512.

    Waters, C.M., and Bassler, B.L. (2005). Quorum sensing: cell-to-cellcommunication in bacteria. Annu. Rev. Cell Dev. Biol.21, 319346.

    Willey, J.M., Willems, A., Kodani, S., and Nodwell, J.R. (2006). Mor-phogenetic suractants and their role in the ormation o aerial hyphaein Streptomyces coelicolor. Mol. Microbiol. 59, 731742.

    Wright, J.S., 3rd, Jin, R., and Novick, R.P. (2005). Transient interer-ence with staphylococcal quorum sensing blocks abscess ormation.Proc. Natl. Acad. Sci. USA 102, 16911696.

    Xavier, K.B., and Bassler, B.L. (2005). Intererence with AI-2-mediatedbacterial cell-cell communication. Nature437, 750753.

    Zhu, J., Oger, P.M., Schrammeijer, B., Hooykaas, P.J., Farrand, S.K.,and Winans, S.C. (2000). The bases o crown gall tumorigenesis. J.Bacteriol. 182, 38853895.