ABSTRACT The term ‘quorum sensing’ (QS) is generally used to describe the phenomenon that bacteria release and perceive signal molecules to coordinate cooperative behaviour in response to

their population size. QS-based communication has therefore been considered a social trait. Here we show that QS signals (_N_-acyl-homoserine lactones, AHLs) are stochastically produced in

young biofilms of _Pseudomonas putida_ and act mainly as self-regulatory signals rather than inducing neighbouring cells. We demonstrate that QS induces the expression of putisolvin

biosurfactants that are not public goods, thereby triggering asocial motility of induced cells out of microcolonies. Phenotypic heterogeneity is most prominent in the early stages of biofilm

COORDINATION OF SIDEROPHORE GENE EXPRESSION AMONG CLONAL CELLS OF THE BACTERIUM _PSEUDOMONAS AERUGINOSA_ Article Open access 06 June 2022 _IN SILICO_ BACTERIA EVOLVE ROBUST COOPERATION VIA

COMPLEX QUORUM-SENSING STRATEGIES Article Open access 25 May 2020 SPATIAL-TEMPORAL DYNAMICS OF A MICROBIAL COOPERATIVE BEHAVIOR RESISTANT TO CHEATING Article Open access 07 February 2022

INTRODUCTION The term quorum sensing (QS) is used to describe the phenomenon that bacteria are capable of perceiving and responding to self-generated signal molecules to coordinate their

behaviour at the group level1. The general consensus is that bacteria trigger the QS response only when their cell density has reached a certain threshold (the ‘quorum’), on which the

expression of target genes is either activated or repressed. Among the various QS signal molecules identified to date, _N_-acyl-homoserine lactones (AHLs) have been investigated to the

greatest extent2,3 and have been shown to control the expression of a large variety of traits, including bioluminescence, virulence, symbiosis, different forms of motility, biofilm

formation, production of antibiotics and toxins, and conjugation4,5. Many AHL-controlled traits represent cooperative behaviours that can generate benefits to other cells in the local

community6. Consequently, it has been suggested that QS has evolved to restrict the expression of costly cooperative behaviours to conditions, in which they are most beneficial, which is the

case at high cell density7,8,9,10,11,12. However, recent studies in _Pseudomonas aeruginosa_ have demonstrated that QS is more complex, because it also controls expression of a few cellular

enzymes (that is, private goods)10,13,14,15. It has been suggested that co-regulation of public and private goods stabilizes cooperation, because it negates the selective advantage of

cheating mutants, which exploit public goods without contributing to them13. Here we studied the role of QS in biofilm formation in _P. putida_ IsoF, a strain that has been isolated from the

rhizosphere of a tomato plant16,17. In this strain, we have previously identified an AHL-dependent QS system, which is located on a genomic island, encoding PpuI, which directs the

biosynthesis of the two AHLs 3-oxo-C10 and 3-oxo-C12 as major products; PpuR, the AHL receptor; and RsaL a repressor of _ppul_17. For a closely related strain (PCL1445), it is has been shown

that the _ppu_ system controls expression of a large non-ribosomal peptide synthethase (encoded by _psoA_, _psoB_ and _psoC_), which directs the biosynthesis of the two cyclic lipopeptide

biosurfactants putisolvin I and II18. The putisolvins were found to not only inhibit biofilm formation of _P. putida_ PCL1445 but also to break down existing _P. aeruginosa_ biofilms19.

Previous work has shown that _P. putida_ IsoF forms a flat and homogenous biofilm, whereas a _ppuI_ mutant forms a structured biofilm with characteristic microcolonies and water-filled

channel17. Although putisolvin production has not been demonstrated for _P. putida_ IsoF, it has been hypothesized that QS-dependent expression of these biosurfactants could also affect

biofilm structural development of this strain18,20. In this study, we visualize AHL production at the single cell level to clarify the link between AHL-mediated QS, putisolvin production and

biofilm development in _P. putida_ IsoF. We show that at early stages of biofilm development, QS induces putisolvin production, which gives rise to a fraction of motile cells that leave the

microcolony on their own. This asocial motility is possible because: (a) AHL signal production is stochastically expressed in only a fraction of the cells in young biofilms; (b) AHL

production in one cell does not induce AHL production in its neighbouring cells; and (c) putisolvins cannot be used by other cells, and therefore do not represent public goods that can be

shared among cells. At a later stage of biofilm development, the AHL expression pattern is more compatible with the expected cross-induction of cells within microcolonies, which results in a

mass movement of cells and a concomitant collapse of microcolonies, giving rise to an unstructured biofilm. RESULTS QS TRIGGERS ASOCIAL MOTILITY IN YOUNG BIOFILMS OF _P. PUTIDA_ To

investigate the role of AHL-mediated QS in biofilm formation of _P. putida_ IsoF at the single cell level, we transferred the green fluorescent protein (GFP)-based AHL sensor plasmid pRP4las

(with stringently controlled copy number: two to three per cell) into the wild-type strain. To ensure that none of the cells is AHL-induced before inoculation into the flow-through

chamber21, we grew the strains for at least five generations at a low cell density (OD600<0.3) in minimal medium supplemented with citrate as carbon source. Following inoculation into the

flow-through chambers, we monitored the spatial and temporal production of AHLs during biofilm development under defined conditions22. In our setup, small microcolonies were formed within

6–8 h; yet, at this point no fluorescent cells were detectable. When microcolonies reached a population size of 23.3±18.6 bacteria, after 11.9±1.3 h, a small fraction of cells turned green

fluorescent, indicating that they had triggered the production of AHL signal molecules (Fig. 1a). Surprisingly, however, these induced cells did not seem to stimulate AHL production in

neighbouring cells within the colony as one would expect according to the generally accepted paradigm that QS is a regulatory mechanism that co-ordinates behaviour at the group level. To

rule out artefacts potentially associated with the use of a plasmid-based AHL reporter, we integrated the AHL reporter cassette into the chromosome of _P. putida_ IsoF. Using this

single-copy AHL biosensor, we quantified the number of induced cells within and outside of microcolonies. This analysis revealed that free cells outside colonies were significantly more

often induced than cells within colonies (Fig. 1b, linear mixed model (LMM): _t_111=5.68, _P_<0.0001). Although the frequency of induced cells significantly increased over time (LMM:

_t_111=11.73, _P_<0.0001), it increased similarly among free and colony cells (LMM, no significant interaction between time and cell status (free versus colony): _t_111=1.14, _P_=0.26),

showing that QS induction level was consistently higher among free cells. This pattern is compatible with a two-step non-coordinated process starting with stochastic expression of AHL,

followed by induced cells becoming motile and independently leaving the microcolonies. Indeed, we observed that induced cells left the colonies and were either removed by the nutrient flow

or re-attached to the glass surface in the void spaces between the microcolonies (Fig. 1c). To further elucidate the heterogeneity in QS induction, we exposed early-stage biofilms to a

saturating concentration of 3-oxo-C10-HSL (0.5 μM). We observed that the timing of QS induction was slightly advanced (1.5 h), but that the heterogeneity in induction remained (Supplementary

Fig. 1). This supports the idea that there are two distinct subpopulations of QS-responsive and non-responsive cells, whereby cross-induction between the two subpopulations is limited, at

least in early-stage biofilms. Our observation is reminiscent of previous findings by Pradhan and Chatterjee23, who demonstrated the presence of stable subpopulations of QS-responsive and

non-responsive cells in _Pseudomonas syringae_ and _Xanthomonas campestris_. At later stages of biofilm growth (Fig. 1a), we noticed an increase of fluorescent cell clusters within the

microcolonies, which might be the result of AHL-mediated cross-stimulation. These clusters continuously increased in size until the large majority of cells of the microcolonies showed green

fluorescence (Fig. 1a; usually in >30-h-old biofilms). At this point the microcolonies suddenly collapsed as a consequence of a mass movement of cells. The resulting biofilm was

unstructured and uniformly covered the glass surface as has been reported previously for mature IsoF biofilms17. PUTISOLVIN IS REQUIRED FOR MOTILITY AND BIOFILM COLLAPSE In a next step, we

aimed to better understand the mechanistic link between QS-heterogeneity and asocial motility in our system. As previous work had revealed a strong relationship between the production of

putisolvin biosurfactants and biofilm formation in _P. putida_ PCL1445 (refs 18, 19), we examined whether strain IsoF also harbours the putisolvin biosynthetic gene cluster, and whether

asocial motility and biofilm collapse is linked to putisolvin production. By using a PCR approach combined with sequencing, we found that the entire _pso_ gene cluster is present in strain

IsoF and shows >99% DNA sequence identity with the _pso_ locus of strain PCL1445. To investigate whether putisolvins act as a biosurfactant in the IsoF strain, we constructed the defined

_psoA_ knockout mutant PL11, as well as the conditional mutant PL2, in which the native promoter region of _psoA_ has been replaced with the rhamnose-inducible _P__rhaB_ promoter

(Supplementary Fig. 2). In the absence of rhamnose, both mutant strains showed no surfactant activity in a simple drop-collapse assay (Fig. 2a). However, drop-collapsing activity of strain

PL2 but not of PL11 could be restored when the medium was amended with at least 0.5% rhamnose. The Du Nouy ring method was used to measure surface tension of spent culture supernatants along

the growth curve. Surface tension was found to reach a minimum when the cultures had an OD600 of ~2.0. Importantly, surface tension was found to be significantly reduced on rhamnose-induced

putisolvin production (Fig. 2b). Next, we tested whether putisolvin is essential for swarming motility, as has been found to be the case for other biosurfactants in other bacterial

species24. Indeed, when tested on swarming plates containing citrate as carbon source, the wild-type IsoF colonized the entire plate within 3 days, whereas no surface migration was observed

for mutants PL2 and PL11 (Fig. 2c). However, in the case of PL2 swarming could be restored by amending the medium with 0.5% rhamnose (Fig. 2d), whereby the migration speed of the swarm

colony was found to be proportional to the rhamnose concentration. These results demonstrate that putisolvin acts as a biosurfactant in _P. putida_ IsoF and is essential for swarming. To

test whether putisolvin is involved in biofilm collapse, we compared biofilm formation of the wild-type IsoF with the conditional _psoA_ mutant PL2 in flow-through cells using AB minimal

medium supplemented with 1 mM citrate. After 3 days of growth, the wild-type had formed a flat and unstructured biofilm with a low volume/area ratio, while the biofilm of mutant PL2 was

dominated by large microcolonies, characterized by a threefold higher volume/area ratio and with only few cells colonizing the void space (Fig. 3). Addition of 0.2% rhamnose to the medium

recovered the flat wild-type biofilm structure. These experiments demonstrate that putisolvins promote the colonization of the substratum by facilitating the movement of cells out of

microcolonies. PUTISOLVIN PRODUCTION IS QS REGULATED We further investigated whether AHL-mediated QS controls putisolvin production, thereby leading to the asocial motility phenotype and

biofilm collapse observed in our single-cell experiments. Indeed, Dubern _et al_.18 have demonstrated that production of putisolvins in strain PCL1445 is regulated by the _ppuI-rsaL-ppuR_ QS

system. In agreement with this study, we found that a _ppuI_ mutant of strain IsoF, named F117, was unable to swarm and spent culture supernatants of F117 cultures lacked surfactant

activity (Fig. 4a,b). Both defects could be rescued by the addition of 5 μM 3-oxo-C10-AHL. These results strongly suggest that QS deficiency in F117 results in the abolishment of putisolvin

production. To obtain more direct evidence for the link between QS and putisolvin production, we constructed a transcriptional fusion of the _psoA_ promoter (triggering putisolvin synthesis)

with _gfp_ and transferred the resulting plasmid, pLUM1, into the wild-type IsoF, the _ppuI_ mutant F117 (defective for AHL signal production) and the _ppuR_ mutant GC3 (defective for

responding to AHL signals). Measurements of GFP fluorescence revealed that _psoA_ expression was strongly decreased in F117 and GC3, but could be restored to wild-type level for F117, but

not for GC3, when the medium was supplemented with 5 μM 3-oxo-C10-HSL (Fig. 4c). THE ROLE OF FLAGELLA FOR BIOFILM DEVELOPMENT OF _P. PUTIDA_ To investigate whether, in addition to

putisolvins, flagella may be required for cell migration out of microcolonies, we constructed a _fliM_ mutant of _P. putida_ IsoF, designated GC25, which no longer produces flagella and

therefore is unable to swim. The microcolonies that strain GC25 formed were much more compact than the ones of the wild-type strain, suggesting that flagella are important for the

positioning of cells within the aggregates (Fig. 5). Moreover, putisolvin producers migrated only at a very low rate out of the microcolonies when compared with the wild-type. These results

show that the dissociation of cells from microcolonies is dependent on both flagella-driven motility and the production of putisolvin biosurfactants. SPATIAL EXPRESSION OF _PSOA_ IN BIOFILMS

We followed the temporal and spatial expression of putisolvins within biofilms of _P. putida_ IsoF using the _P__psoA_-_gfp_ transcriptional fusion. Similar to our previous results (Fig.

1), we found that fluorescent cells were mainly located at the periphery or outside of microcolonies. In contrast, when the same transcriptional fusion was present in the

putisolvin-defective mutant PL11 or the non-motile mutant GC25, we observed both a higher proportion of QS-induced cells and a more homogenous induction across cells, especially in GC25

(Fig. 5). These findings provide evidence that asocial cell movement out of the colony restricts cross-induction in the wild-type strain, while cross-induction becomes increasingly possible

when cells are forced to stay together. PUTISOLVINS ARE PRIVATE AND NOT PUBLIC GOODS A recent study in _P. aeruginosa_ revealed that biosurfactants can represent public goods, which allow

biosurfactant-defective mutants to swarm along with biosurfactant-producing wild-type cells25. In contrast, our single-cell analyses suggest that biosurfactants in _P. putida_ IsoF remain

associated with the bacterial cell surface, thereby triggering individual-based and not group-based swarming. To test this hypothesis, we mixed a putisolvin-deficient or a flagella-deficient

mutant with the wild-type on swarming plates. In support of our hypothesis, we found that the mutants stayed close to the inoculation point, forming a small colony, whereas the wild-type

swarmed over the entire plate (Fig. 6), with swarming behaviour being unaffected by the presence of another strain (Supplementary Fig. 3). This result held across a wide range of strain

mixing ratios (Supplementary Figs 3 and 4). These findings demonstrate that putisolvins are private rather than public goods, which mostly adhere to the producing cells, and can therefore

not be used by other cells for swarming. DISCUSSION Our work breaks with the central assumption that QS primarily represents a regulatory mechanism to coordinate cooperative behaviours among

cells at high population density. Instead, we show that QS can also do the opposite: trigger uncoordinated self-directed behaviour at low cell density. Specifically, we found that in the

early stages of biofilm development, AHL production occurred stochastically in only a fraction of cells. AHL production in these cells triggered the synthesis of putisolvins, biosurfactants

that remain associated with the producer’s cell surface, thereby resulting in cells individually moving out of the microcolony. As this asocial motility removes individuals with the highest

AHL production from the consortium, it exerts a negative feedback on cells left behind by delaying AHL cross-induction within the microcolony. It is important to note that we were only able

to discover the lack of cross-induction and asocial motility, because AHLs were stochastically expressed at low population density. If all cells had started expressing AHLs at the same time,

we would have erroneously concluded that putisolvin serves as a public good, coordinately expressed in the consortium to allow cooperative motility. This highlights that the mere

observation of individual cells doing the same thing at the same time does not necessarily mean that coordination through communication and the sharing of pubic goods are involved26. There

are at least three reasons why stochasticity can arise in our system. First, Kaplan and Greenberg27 showed that AHL-dependent QS can be an extremely sensitive system, as demonstrated in

_Vibrio fischeri_, where one to two AHL molecules per cell are sufficient to trigger autoinduction, suggesting that at very low AHL concentrations QS is intrinsically stochastic. Second,

although the classic QS model assumes that AHLs are diffusing into the cell from the surroundings, such that the population density determines induction levels, we suggest that at the onset

of QS the signal molecules are not released from the producing bacterium but directly bind to their cognate cytoplasmic receptors, which, as a consequence, results in self-induction of the

cell’s QS cascade, and not in cross-induction. This possibility is especially probable in our study system, as the _P. putida_ IsoF AHLs contain relatively long fatty acid chains, which

often require transporters to be actively released from the cell28,29,30. Third, physiological differences between cells, particularly when grown as a biofilm, may exist that affect AHL

production or the sensitivity of the QS response. Moreover, cross-induction seems to be additionally impeded in our system, because induced cells leave the consortium, which presumably

results in reduced local AHL concentrations. Although previous work has revealed heterogeneity in QS both at low and high cell densities, the situation described in this study is unique, as

it is the first example that QS heterogeneity serves as a mechanism to trigger a self-directed behaviour of individual cells. At low cell density, heterogeneity in the initiation of QS has

been observed in _P. aeruginosa_ when single cells were confined in small volumes in a microfluidic device31. In this study, it is not only shown that QS induction is highly variable but

also that low numbers of cells, even single cells, are able to initiate QS, supporting the idea of QS self-induction. At high cell density, meanwhile, Anetzberger _et al_.32 showed that the

expression of QS-regulated bioluminescence in _Vibrio harveyi_ is heterogenous. In a subsequent study, a working model was presented, in which the combination of the different _V. harveyi_

signal molecules available (this organism produces three structurally unrelated signal molecules), rather than cell density _per se_, determines the timing of QS-regulated traits in this

species33. Likewise, induction of AHL-controlled bioluminescence in individual _V. fischeri_ cells was found to differ widely in time scale and in the overall intensity, suggesting that QS

has relatively imprecise control over the response of an individual cell34. Although heterogeneity in QS seems frequent, a remaining key question is whether the heterogeneity is simply an

inevitable outcome of the regulatory mechanism controlling AHL production, or whether it has an adaptive function. Although speculative at this stage, it seems plausible that in our case of

asocial motility, leaving the microcolony can be beneficial under nutrient depletion. Even more so in mixed biofilms, where natural selection could favour individual-based early dispersal to

reach new resources faster than the competitors35. Although our data indicate that both AHL and putisolvin production is self-directed and thereby asocial during the early stages of biofilm

formation, the pattern conceivably changed in older biofilms. Although AHL-producing cells first leave the microcolonies by themselves, we hypothesize that over time AHLs accumulate in the

microcolonies, eventually leading to cross-induction of neighbouring cells. This idea is supported by the observation that _psoA_ expression is more homogenous and increases faster in

mutants impaired in motility when compared with expression of _psoA_ in the wild-type background (Fig. 5). Furthermore, typical AHL cross-induction of cells has been observed in _P. putida_

IsoF microcolonies grown in a microfluidic setup, in which motility of cells was restricted by coating the substratum with polylysin36. Thus, at later stages of biofilm development, AHL

signalling eventually becomes a social trait, which then results in all cells producing putisolvins. This, in turn, leads to the hallmark of biofilm structural development in our setup,

which was the sudden collapse of microcolonies at high cell densities. As nutrients are probably limiting at this point of biofilm development, even self-directed motility can become a

social trait, because putisolvin-mediated biofilm dispersal will prevent overcrowding and allow the left behind population to resume growth. Indeed, the putisolvin-mediated microcolony

collapse is reminiscent to the role of QS-controlled rhamnolipid production in detachment of cells from mature _P. aeruginosa_ biofilms, which was suggested to release the stress arising

from nutrient limitation at high population density37. Our experiment showing that putisolvin-deficient mutants are unable to swarm with the wild-type strain demonstrates that putisolvins do

not represent public goods (Fig. 6). In support of this, it has been shown that putisolvins preferentially adhere to the cell surface of the producing cell19. Interestingly, adhesion to

bacterial cell surfaces has also been shown for other biosurfactants38,39,40, which may therefore represent a more general phenomenon. In analogy to our findings, Burch _et al_.41 showed

that the biosurfactant syringafactin, which is produced by the plant epiphyte _P. syringae_ pv. syringae B728a, is adsorbed to the adjacent waxy cuticle or retained on the bacterial cell

surface. It is shown that its production primarily benefits the producer both by attracting moisture and facilitating access to nutrients. Our results relate to recent work on QS in _P.

aeruginosa_, where it has been shown that QS not only coordinates the expression of public goods at the group level, but also directs the expression of metabolically important enzymes at the

cellular level13,15. The QS regulatory control over both social and self-directed traits has been interpreted as an adaptation to prevent invasion of cheating mutants. The idea is that the

fitness increase a QS-deficient mutant gains by exploiting a QS wild-type strain, is cancelled by the fitness loss these mutants face, because they lack an important cellular enzyme. The

situation is clearly different in our study system where the risk of cheating is reduced, because both the AHL-signal and putisolvins are not or only partially available to others. Taken

together, our insights highlight that QS is much more complex than previously thought, as the traits being induced by QS can cover the entire continuum from a cooperative public good trait

that generates benefits to others (for example, elastase production11), to extracellular traits that mostly generate self-directed benefits (for example, putisolvin production), to entirely

intracellular traits that solely provide benefits to the producer. METHODS STRAINS AND CULTURE CONDITIONS Bacterial strains and plasmids used in this study are listed in Supplementary Table

S1. _Escherichia coli_ strains used for recombinant manipulations were propagated in Luria–Bertani medium at 37 °C. Plasmids were delivered to _P. putida_ by triparental mating22. Briefly,

donor, recipient and helper strain, _E. coli_ HB101(pRK600), were harvested from overnight cultures, mixed and spot-inoculated on Luria–Bertani plates. After overnight incubation at 37 °C,

transconjugants were isolated on Pseudomonas Isolation Agar (PIA) at 30 °C. _P. putida_ strains were grown in modified AB medium supplemented with 10 mM sodium citrate42 (referred to as ABC

medium). When required, media were supplemented with antibiotics at the following concentrations. For _E. coli_: 50 μg ml−1 ampicilin, 50 μg ml−1 kanamycin, 10 μg ml−1 gentamycin, 10 μg ml−1

tetracycline and 50 μg ml−1 trimethoprim. For _P. putida_: 100 μg ml−1 kanamycin, 20 μg ml−1 gentamycin and 100 μg ml−1 tetracycline. CONSTRUCTION OF _P. PUTIDA_ ISOF MUTANTS The _psoA_

mutant PL11 was generated as follows: an internal _psoA_ fragment was PCR amplified using the primers psoAF (5′- ctgatggtgtcgttcgaagagg -3′) and psoAR (5′- gctcgtcgagcacgtacaactg -3′). The

amplicon was digested with SmaI and cloned into the gene replacement vector pEX18Gm cut with the same enzyme. The resulting plasmid, pEX18_psoA_, was mobilized into _P. putida_ IsoF by

triparental mating and gene replacement mutants were selected on PIA medium containing 20 μg ml−1 gentamycin. The _fliM_ mutant GC25 was constructed by amplifying a _fliM_ internal fragment

using primers FliMF (5′- gccatggccgggttgaytc -3′) and FliMR (5′- gaygaygggctggtrcagac -3′), and cloning the PCR product blunt-ended into the Stul-digested gene replacement vector pSHAFT2Gm.

The resulting plasmid, pSHAFT2_fliM_, was used to construct a _fliM_ mutant as described for the _psoA_ mutant. A _pso_ conditional mutant was constructed as follows: first, the gentamycin

resistance cassette from pBBR1MCS-5 was amplified using the primers genF (5′- gcagcaacgatgttacgcag -3′) and genR (5′- ttggtaccccgatctcggcttgaacg -3′), the amplicon was digested with XbaI and

KpnI (restriction site underlined), and cloned into plasmid pSC200 cut with the same enzymes, yielding plasmid pSC200Gm. Next, a 590-bp fragment beginning at the start codon of _psoA_ was

amplified using primers pos2F (5′- tgcctgccgccgaaacctt -3′) and pos590R (5′- atctagagccagccaataatcgcggtc -3′), and the resulting DNA fragment was blunt ended with Klenow fragment and cloned

into the filled-in NdeI site of pSC200Gm. This plasmid was mobilized from _E. coli_ CC118 into _P. putida_ IsoF by conjugation and the conditional mutant was selected on PIA medium

supplemented with 20 μg ml−1 gentamycin. The genetic structures of all mutants constructed were confirmed by PCR and sequence analysis. The following primers were used: pSHAFT2F (5′-

CGCTCTCGCGGCTTACGTTC -3′), pSHAFT2R (5′- AAGCCAGGGATGTAACGCACTG -3′), peX_F (5′- CACCGACAAACAACAGATAA -3′), peX_R (5′- CCCCAGGCTTTACACTTT -3′) pSC200end (5′- GTCATACTGGCCTCCTGATGTCGT -3′).

CONSTRUCTION OF TRANSCRIPTIONAL FUSIONS The pUT/mini-Tn_5_Km-based plasmid pPLlas21 was used to integrate the GFP-based AHL sensor into the chromosome of _P. putida_ IsoF. Three independent

mutants with different insertion positions were purified and used for flow cell experiments. To construct a _P__psoA_-_gfp_ transcriptional fusion, the _psoA_ promoter region was PCR

amplified using the primers p-psoAF (5′- aggatccgattctaagctttgcggcg -3′) and p-psoAR (5′- tggatccgctcagggcaaaggtttcg -3′). PCR fragments were cloned as BamHI fragments (restriction sites are

underlined) into the respective site of the promoter–probe vector pGA-G1, generating the plasmids pPLM1 (_P__psoA_-_gfp_). A P_psoA__-cfp_ fusion was generated by cloning the PCR product

containing the _ecfp_ gene from pBK-mini-Tn_7_ into the TOPO vector. Then, the _psoA_ promoter region from PLM1 was inserted as a BamHI fragment upstream of the _ecfp_ gene in this plasmid.

Finally, the the _P__psoA__-cfp_ cassette was excised as a SacI fragment and inserted into the same site of plasmid pBBR1MCS-3, yielding pLUM3 (P_psoA__-cfp_). The plasmids were mobilized

from _E. coli_ CC118 to _P. putida_ strains by conjugation and selected on PIA medium supplemented with 50 μg ml−1 gentamycin (pPLM1) or 100 μg ml−1 tetracycline (pLUM3). MEASUREMENT OF

PROMOTER ACTIVITIES _P. putida_ strains harbouring pPLM1 were grown in 10 mM ABC medium for 24 h at 30 °C with continuous shaking. When required, 3-oxo-C10 homoserine lactone was added to

the medium at a final concentration of 5 μM. Green fluorescence was measured using 200 μl samples in a microtitre plate reader (SynergyTM HT, MWG Biotech, Germany) with an excitation

wavelength of 485 nm and emission detection at 528 nm. The data were corrected for autofluorescence and processed with the KC4 software (BioTek Instruments). Specific fluorescence was

calculated by normalizing relative fluorescence to OD600, which was simultaneously measured in the instrument. SWARMING MOTILITY ASSAYS Swarming motility was determined on ABC agar plates

supplemented with 0.1% casamino acids and solidified with 0.4% (wt/vol) agar as described previously43. Briefly, overnight cultures were adjusted to an OD600 of 0.1, and 2 μl samples were

inoculated on swarming plates, which were incubated for 3 days at 30 °C. The swarming plates were supplemented with 0.5%, 1% or 2% (wt/vol) rhamnose when appropriate. BIOSURFACTANT

PRODUCTION Semi-quantitative measurement of biosurfactant activity was done by using the drop-collapsing assay, in which the reduction of surface tension causes a collapse of the droplet

placed on a hydrophobic surface. To quantify biosurfactant production, the decrease of surface tension between culture medium and air was determined with a Du Nouy ring18. CULTIVATION AND

ANALYSIS OF BIOFILMS Biofilms were grown in flow cells supplied with ABC medium. The flow system was assembled and prepared as described previously22. Briefly, the flow channels were

inoculated with _P. putida_ cultures grown for at least five generations at a low cell density (OD600<0.3) in minimal medium supplemented with citrate as the carbon source. The medium

flow was kept at a constant rate of 0.2 mm s–1 by a Watson–Marlow 205S peristaltic pump. The incubation temperature was 30 °C. Microscopic inspection and image acquisition were performed

using a confocal laser scanning microscope (DM5500Q; Leica) equipped with a × 40/1.3 or a × 63/1.4 oil objective. Captured images were analysed with the Leica Application Suite (Mannheim,

Germany) and the Imaris software package (Bitplane, Switzerland). Images were prepared for publication using CorelDraw (Corel Corporation) and PowerPoint (Microsoft) software. To quantify

the proportion of AHL-induced and non-induced cells, five independent experiments were conducted. In each experiment, five random positions were chosen on the flow chamber glass surface and

surveyed every 30 min for 10 h, starting 6 h post inoculation. Aliquots of a low-cell-density inoculum (OD600=0.01) were used to initiate the flow cell biofilms to allow single cell

analysis. To distinguish between free and colony-associated cells, an aggregate size of eight cells was defined as threshold, below which cells were considered as free. Using this threshold,

the average aggregate size for free cells was found to be 3.5±0.5. As transmitted light was used to obtain the total cell number, only microcolonies with few cell layers were used for

quantification. When analysing older biofilms consisting of multiple cell layers, we either used strains marked with mCherry or stained cells with SYTO 62 (Life Technologies). STATISTICAL

ANALYSIS An LMM was used to test whether the proportion of AHL-induced cells differs between free and colony-associated cells, and whether the induction pattern changes over time. Position

identity within experiments was introduced into the model as a random factor to account for the nested approach (that is, five positions within five experiments). Prism (GraphPad Software)

was used for one-way analysis of variance. If the analysis of variance yielded significant differences between factor levels, the Bonferroni method was applied for pairwise comparisons

between factor levels. ADDITIONAL INFORMATION HOW TO CITE THIS ARTICLE: Cárcamo-Oyarce, G. _et al_. Quorum sensing triggers the stochastic escape of individual cells from _Pseudomonas

putida_ biofilms. _Nat. Commun._ 6:5945 doi: 10.1038/ncomms6945 (2015). REFERENCES * Fuqua, W. C., Winans, S. C. & Greenberg, E. P. Quorum sensing in bacteria: the LuxR-LuxI family of

cell density-responsive transcriptional regulators. _J. Bacteriol._ 176, 269–275 (1994). Article  CAS  Google Scholar  * Whitehead, N. A., Barnard, A. M. L., Slater, H., Simpson, N. J. L.

& Salmond, G. P. C. Quorum-sensing in Gram-negative bacteria. _FEMS Microbiol. Rev._ 25, 365–404 (2001). Article  CAS  Google Scholar  * Williams, P. Quorum sensing, communication and

cross-kingdom signalling in the bacterial world. _Microbiology_ 153, 3923–3938 (2007). Article  CAS  Google Scholar  * Bassler, B. L. & Losick, R. Bacterially speaking. _Cell_ 125,

237–246 (2006). Article  CAS  Google Scholar  * Williams, P., Winzer, K., Chan, W. C. & Camara, M. Look who’s talking: communication and quorum sensing in the bacterial world. _Philos.

Trans. R. Soc. Lond. B Biol. Sci._ 362, 1119–1134 (2007). Article  CAS  Google Scholar  * West, S. A., Diggle, S. P., Buckling, A., Gardner, A. & Griffins, A. S. The social lives of

microbes. _Annu. Rev. Ecol. Syst._ 38, 53–77 (2007). Article  Google Scholar  * Parsek, M. R. & Greenberg, E. P. Sociomicrobiology: the connections between quorum sensing and biofilms.

_Trends Microbiol._ 13, 27–33 (2005). Article  CAS  Google Scholar  * Keller, L. & Surette, M. G. Communication in bacteria: an ecological and evolutionary perspective. _Nat. Rev.

Microbiol._ 249–258 (2006). * Diggle, S., Crusz, S. & Camara, M. Quorum sensing. _Curr. Biol._ 17, R907–R910 (2007). Article  CAS  Google Scholar  * Darch, S. E., West, S. A., Winzer, K.

& Diggle, S. P. Density-dependent fitness benefits in quorum-sensing bacterial populations. _Proc. Natl Acad. Sci._ 109, 8259–8263 (2012). Article  CAS  ADS  Google Scholar  * Diggle,

S. P., Griffin, A. S., Campbell, G. S. & West, S. A. Cooperation and conflict in quorum-sensing bacterial populations. _Nature_ 411–414 (2007). * West, S. A., Winzer, K., Gardner, A.

& Diggle, S. P. Quorum sensing and the confusion about diffusion. _Trends Microbiol._ 20, 586–594 (2012). Article  CAS  Google Scholar  * Dandekar, A. A., Chugani, S. & Greenberg, E.

P. Bacterial quorum sensing and metabolic incentives to cooperate. _Science_ 338, 264–266 (2012). Article  CAS  ADS  Google Scholar  * Mellbye, B. & Schuster, M. The sociomicrobiology

of antivirulence drug resistance: a proof of concept. _MBio_ 2, doi:10.1128/mBio.00131-11 (2011). * Heurlier, K., Dénervaud, V. & Haas, D. Impact of quorum sensing on fitness of

_Pseudomonas aeruginosa_. _Int. J. Med. Microbiol._ 296, 93–102 (2006). Article  CAS  Google Scholar  * Steidle, A. et al. Visualization of N-acylhomoserine lactone-mediated cell-cell

communication between bacteria colonizing the tomato rhizosphere. _Appl. Environ. Microbiol._ 67, 5761–5770 (2001). Article  CAS  Google Scholar  * Steidle, A. et al. Identification and

characterization of an N-acylhomoserine lactone-dependent quorum-sensing system in _Pseudomonas putida_ strain IsoF. _Appl. Environ. Microbiol._ 68, 6371–6382 (2002). Article  CAS  Google

Scholar  * Dubern, J.-F., Lugtenberg, B. J. J. & Bloemberg, G. V. The ppuI-rsaL-ppuR quorum-sensing system regulates biofilm formation of _Pseudomonas putida_ PCL1445 by controlling

biosynthesis of the cyclic lipopeptides putisolvins I and II. _J. Bacteriol._ 188, 2898–2906 (2006). Article  CAS  Google Scholar  * Kuiper, I. et al. Characterization of two _Pseudomonas

putida_ lipopeptide biosurfactants, putisolvin I and II, which inhibit biofilm formation and break down existing biofilms. _Mol. Microbiol._ 51, 97–113 (2004). Article  CAS  Google Scholar 

* Dubern, J.-F., Coppoolse, E. R., Stiekema, W. J. & Bloemberg, G. V. Genetic and functional characterization of the gene cluster directing the biosynthesis of putisolvin I and II in

_Pseudomonas putida_ strain PCL1445. _Microbiology_ 154, 2070–2083 (2008). Article  CAS  Google Scholar  * Lumjiaktase, P., Aguilar, C., Battin, T., Riedel, K. & Eberl, L. Construction

of self-transmissible green fluorescent protein-based biosensor plasmids and their use for identification of N-acyl homoserine-producing bacteria in lake sediments. _Appl. Environ.

Microbiol._ 76, 6119–6127 (2010). Article  CAS  Google Scholar  * Christensen, B. B. et al. Molecular tools for study of biofilm physiology. _Methods Enzymol._ 310, 20–42 (1999). Article 

CAS  Google Scholar  * Pradhan, B. B. & Chatterjee, S. Reversible nongenetic phenotypic heterogeneity in bacterial quorum sensing. _Mol. Microbiol._ 92, 557–569 (2014). Article  CAS 

Google Scholar  * Kearns, D. B. A field guide to bacterial swarming motility. _Nat. Rev. Microbiol._ 8, 634–644 (2010). Article  CAS  Google Scholar  * Xavier, J. B., Kim, W. & Foster,

K. R. A molecular mechanism that stabilizes cooperative secretions in _Pseudomonas aeruginosa_. _Mol. Microbiol._ 79, 166–179 (2011). Article  CAS  Google Scholar  * Ross-Gillespie, A. &

Kümmerli, R. Collective decision-making in microbes. _Front. Microbiol._ 5, 54 (2014). Article  Google Scholar  * Kaplan, H. B. & Greenberg, E. P. Diffusion of autoinducer is involved

in regulation of the _Vibrio fischeri_ luminescence system. _J. Bacteriol._ 163, 1210–1214 (1985). CAS  PubMed  PubMed Central  Google Scholar  * Pearson, J. P., Van Delden, C. &

Iglewski, B. H. Active efflux and diffusion are involved in transport of _Pseudomonas aeruginosa_ cell-to-cell signals. _J. Bacteriol._ 181, 1203–1210 (1999). CAS  PubMed  PubMed Central 

Google Scholar  * Chan, Y. Y. et al. Control of quorum sensing by a _Burkholderia pseudomallei_ multidrug efflux pump. _J. Bacteriol._ 189, 4320–4324 (2007). Article  CAS  Google Scholar  *

Buroni, S. et al. Assessment of three resistance-nodulation-cell division drug efflux transporters of _Burkholderia cenocepacia_ in intrinsic antibiotic resistance. _BMC Microbiol._ 9, 200

(2009). Article  Google Scholar  * Boedicker, J. Q., Vincent, M. E. & Ismagilov, R. F. Microfluidic confinement of single cells of bacteria in small volumes initiates high-density

behavior of quorum sensing and growth and reveals its variability. _Angew. Chem. Int. Ed. Engl._ 48, 5908–5911 (2009). Article  CAS  Google Scholar  * Anetzberger, C., Pirch, T. & Jung,

K. Heterogeneity in quorum sensing-regulated bioluminescence of _Vibrio harveyi_. _Mol. Microbiol._ 73, 267–277 (2009). Article  CAS  Google Scholar  * Anetzberger, C., Schell, U. &

Jung, K. Single cell analysis of _Vibrio harveyi_ uncovers functional heterogeneity in response to quorum sensing signals. _BMC Microbiol._ 12, 209 (2012). Article  CAS  Google Scholar  *

Perez, P. D. & Hagen, S. J. Heterogeneous response to a quorum-sensing signal in the luminescence of individual _Vibrio fischeri_. _PLoS One_ 5, e15473 (2010). Article  ADS  Google

Scholar  * Rendueles, O. & Ghigo, J.-M. Multi-species biofilms: how to avoid unfriendly neighbors. _FEMS Microbiol. Rev._ 36, 972–989 (2012). Article  CAS  Google Scholar  * Meyer, A. et

al. Dynamics of AHL mediated quorum sensing under flow and non-flow conditions. _Phys. Biol._ 9, 026007 (2012). Article  ADS  Google Scholar  * Boles, B. R., Thoendel, M. & Singh, P. K.

Rhamnolipids mediate detachment of _Pseudomonas aeruginosa_ from biofilms. _Mol. Microbiol._ 57, 1210–1223 (2005). Article  CAS  Google Scholar  * Neu, T. R. Significance of bacterial

surface-active compounds in interaction of bacteria with interfaces. _Microbiol. Rev._ 60, 151–166 (1996). CAS  PubMed  PubMed Central  Google Scholar  * Zhong, H. et al. Adsorption of

dirhamnolipid on four microorganisms and the effect on cell surface hydrophobicity. _Appl. Microbiol. Biotechnol._ 77, 447–455 (2007). Article  CAS  Google Scholar  * Yuan, X. et al.

Adsorption of surfactants on a _Pseudomonas aeruginosa_ strain and the effect on cell surface lypohydrophilic property. _Appl. Microbiol. Biotechnol._ 76, 1189–1198 (2007). Article  CAS 

Google Scholar  * Burch, A. Y., Zeisler, V., Yokota, K., Schreiber, L. & Lindow, S. E. The hygroscopic biosurfactant syringafactin produced by _Pseudomonas syringae_ enhances fitness on

leaf surfaces during fluctuating humidity. _Environ. Microbiol._ 16, 2086–2098 (2014). Article  CAS  Google Scholar  * Heydorn, A. et al. Experimental reproducibility in flow-chamber

biofilms. _Microbiology_ 146, 2409–2415 (2000). Article  CAS  Google Scholar  * Eberl, L. et al. Involvement of N-acyl-L-hormoserine lactone autoinducers in controlling the multicellular

behaviour of _Serratia liquefaciens_. _Mol. Microbiol._ 20, 127–136 (1996). Article  CAS  Google Scholar  Download references ACKNOWLEDGEMENTS We thank Stefanie Heller for technical support.

Work on QS and biofilm formation has been supported by the Swiss National Science Foundation (Project 31003A-143773). G.C. is supported by a CONICYT Bicentennial Becas-Chile Scholarship.

R.K. is supported by the Swiss National Science Foundation (PP00P3-139164). AUTHOR INFORMATION Author notes * Putthapoom Lumjiaktase & Rolf Kümmerli Present address: Present Address:

Faculty of Medicine, Department of Pathology, Ramathibodi Hospital, Mahidol University, Bangkok, Thailand, * Gerardo Cárcamo-Oyarce and Putthapoom Lumjiaktase: These authors contributed

equally to this work AUTHORS AND AFFILIATIONS * Department of Microbiology, Institute of Plant Biology, University of Zürich, Zollikerstrasse 107, Zürich, CH-8008, Switzerland Gerardo

Cárcamo-Oyarce, Putthapoom Lumjiaktase & Leo Eberl * Department of Microbial Evolutionary Ecology, Institute of Plant Biology, University of Zürich, Winterthurerstrasse 190, Zürich,

CH-8057, Switzerland Rolf Kümmerli Authors * Gerardo Cárcamo-Oyarce View author publications You can also search for this author inPubMed Google Scholar * Putthapoom Lumjiaktase View author

publications You can also search for this author inPubMed Google Scholar * Rolf Kümmerli View author publications You can also search for this author inPubMed Google Scholar * Leo Eberl View

author publications You can also search for this author inPubMed Google Scholar CONTRIBUTIONS L.E. designed the research; G.C.-O. and P.L. performed the experiments and contributed equally

to the study. All authors analysed the data and contributed to the writing of the paper. CORRESPONDING AUTHOR Correspondence to Leo Eberl. ETHICS DECLARATIONS COMPETING INTERESTS The authors

declare no competing financial interests. SUPPLEMENTARY INFORMATION SUPPLEMENTARY INFORMATION Supplementary Figures 1-4, Supplementary Table 1, and Supplementary References (PDF 282 kb)

RIGHTS AND PERMISSIONS This work is licensed under a Creative Commons Attribution 4.0 International License. The images or other third party material in this article are included in the

article’s Creative Commons license, unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license, users will need to obtain permission

from the license holder to reproduce the material. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/ Reprints and permissions ABOUT THIS ARTICLE CITE THIS

ARTICLE Cárcamo-Oyarce, G., Lumjiaktase, P., Kümmerli, R. _et al._ Quorum sensing triggers the stochastic escape of individual cells from _Pseudomonas putida_ biofilms. _Nat Commun_ 6, 5945

(2015). https://doi.org/10.1038/ncomms6945 Download citation * Received: 26 May 2014 * Accepted: 24 November 2014 * Published: 16 January 2015 * DOI: https://doi.org/10.1038/ncomms6945 SHARE

THIS ARTICLE Anyone you share the following link with will be able to read this content: Get shareable link Sorry, a shareable link is not currently available for this article. Copy to

clipboard Provided by the Springer Nature SharedIt content-sharing initiative