A New Class of Quorum Quenching Molecules from Staphylococcus Species Affects Communication and Growth of Gram-Negative Bacteria

The knowledge that many pathogens rely on cell-to-cell communication mechanisms known as quorum sensing, opens a new disease control strategy: quorum quenching. Here we report on one of the rare examples where Gram-positive bacteria, the ‘Staphylococcus intermedius group’ of zoonotic pathogens, excrete two compounds in millimolar concentrations that suppress the quorum sensing signaling and inhibit the growth of a broad spectrum of Gram-negative beta- and gamma-proteobacteria. These compounds were isolated from Staphylococcus delphini. They represent a new class of quorum quenchers with the chemical formula N-[2-(1H-indol-3-yl)ethyl]-urea and N-(2-phenethyl)-urea, which we named yayurea A and B, respectively. In vitro studies with the N-acyl homoserine lactone (AHL) responding receptor LuxN of V. harveyi indicated that both compounds caused opposite effects on phosphorylation to those caused by AHL. This explains the quorum quenching activity. Staphylococcal strains producing yayurea A and B clearly benefit from an increased competitiveness in a mixed community.


Introduction
In many bacteria, pathogenicity is controlled and coordinated by an inter-cellular communication process named quorum sensing (QS). QS is based on the synthesis and secretion of small hormone-like molecules termed autoinducers that bind to cognate receptors. The signal-activated receptor controls directly or indirectly expression of target genes. Since the concentration of signaling molecules in liquid culture is proportional to cell density in the culture, gene expression is coordinated in response to the bacterial population density [1,2]. In V. harveyi, the QS system consists of three autoinducers and three cognate receptors functioning in parallel to channel information into a shared regulatory pathway [2]. Similar to other Gram-negative bacteria, V. harveyi produces an AHL signal termed HAI-1, 3-hydroxy-C4homoserine lactone [3], which binds to the membrane-bound sensor histidine kinase (LuxN). The second molecule is AI-2, a furanosyl borate diester, which binds to the periplasmic protein LuxP. The LuxP-AI-2 complex interacts with another membranebound sensor histidine kinase, LuxQ. The third molecule is termed CAI-1 (for cholera autoinducer-1), a long chain ketone [4,5], which is recognized by the membrane-bound sensor histidine kinase, CqsS [6]. At low cell density, in the absence of appreciable amounts of autoinducers, the three sensors (LuxN, LuxQ, and CqsS) act as autophosphorylating kinases that subsequently transfer the phosphate to the cytoplasmic protein LuxU, which passes the phosphate to the DNA-binding response regulator protein LuxO [7,8]. Phosphorylated LuxO represses the master regulator of QS, LuxR, via sigma factor s 54 and regulatory small RNAs [9,10].
QS system is a promising target for anti-virulence therapy [1,18]. In contrast to classic antibiotics, quorum-quenching compounds are inhibitors of bacterial virulence, rather than of bacterial growth [19]. Since the first studies on QS inhibitors, the halogenated furanones [20], more compounds have been identified [21,22].
In contrast to Gram-negative bacteria, many Gram-positive bacteria communicate using modified oligopeptides as signals and ''two-component''-type membrane-bound sensor histidine kinases as receptors. The well-studied QS system in Staphylococcus is the agr QS system [23]. The excreted signal is a thiolactone-or lactonebased peptide [24] (AIP, autoinducer peptide) that mediates communication with other staphylococci in a cell density dependent way [25,26].
While studying the potential interaction of staphylococci with Gram-negative bacteria [27,28], we came across another communication system in a Staphylococcus species group, named 'intermedius group'. This group consists of closely related mainly coagulase-positive bacterial species including S. delphini, S. intermedius, S. lutrae, S. pseudintermedius, and S. schleiferi. They are all phylogenetically related, are zoonotic pathogens, and only rarely occur in human infections. We found that these species excrete two low molecular compounds that inhibit the expression of QScontrolled toxins and other QS-regulated compounds in Gramnegative bacteria. The excreted compounds, which we named yayurea A and B, were isolated from S. delphini and structurally characterized. Yayurea A and B represent new bacterial products, and were able to quench the QS regulation in a wide spectrum of Gram-negative bacteria. Furthermore, growth of yayurea A and B producing S. delphini is not suppressed by respiratory toxins when co-cultured with P. aeruginosa. This suggests that the quorum quenchers have a function in self-protection and competitiveness in natural environments shared with Gram-negatives. Here we show an example of inter-phylum interference between Firmicutes (Gram-positive) and the Gram-negative beta-and gammaproteobacteria.

Staphylococcus delphini suppresses production of QSregulated phenotypes in various Gram-negative bacteria
Our aim was to find out if some staphylococcal species are able to suppress the QS controlled phenotypes in Gram-negative bacteria. To investigate this, we tested the ability of several staphylococcal species to inhibit pyocyanin production of P. aeruginosa in a co-cultivation assay, as pyocyanin production is QS controlled. We found that S. delphini DSMZ 20771 completely inhibited pyocyanin production over 24 h co-cultivation with P. aeruginosa PAO1, while Staphylococcus aureus showed no such activity (Fig. 1A). This led us to investigate if S. delphini could also suppress QS-controlled phenotypes in other Gram-negative bacteria such as the QS-regulated prodigiosin production in Serratia marcescens [29]; bioluminescence in Vibrio harveyi [7,30]; or violacein production in Chromobacterium subtsugae, a pathogen of potato beetles [31]. Indeed, in co-cultivation studies with these Gram-negative bacteria, S. delphini did also suppress prodigiosin and violacein production as well as bioluminescence, while S. aureus did not (Fig. 1BCD). Sterile filtered culture supernatant of a 24 h S. delphini culture had the same QS-inhibiting effect as the co-culture, indicating that the QS-inhibiting compound(s) were excreted. The supernatants of S. aureus and S. epidermidis caused no effect.
Structural analysis of the QS-inhibiting compounds from S. delphini The QS inhibitors were isolated from the supernatant of an overnight culture of S. delphini DSMZ 20771. Further purification revealed that the supernatant contained two compounds with different retention times (R t ) in HPLC and distinct UV spectra. We named the two compounds yayurea A and B.
Yayurea B (phenethylurea) was obtained as a colorless solid. A preliminary molecular formula, C 8 H 10 N, was deduced from the FT-ICR-MS spectrum which showed an ion at m/z = 121 ([M+H] + ). GC-EI-MS provided a signal at R t = 15.0 min pointing to 2-phenethylamine. The 1 H-NMR again showed a set of aromatic protons and a singlet at d H = 8.54 ppm, while the 13 C-NMR displayed a signal at d H = 170.4 ppm (C = O) and five aromatic and 4 aliphatic protons. In summary, the structure of N-(2-phenethyl)-urea was assigned for yayurea B (Fig. 2B). UV-Absorption maxima emphasized the phenyl-(l = 260 nm) and indole chromophores (l = 225 nm and l = 280 nm respectively) of the yayureas. Comprehensive physicochemical characteristics and mass spectra of yayurea A and B are shown (Table 1 and Figure S1). In the meantime, yayurea A and B could also be chemically synthesized and their identity confirmed by mass spectral and NMR analyses (Wölfle et al. manuscript in preparation). Both the synthesized and the natural compounds revealed the same chemical properties and the same QS-quenching activity in Gram-negative bacteria (data not shown).
Suppression of QS-regulated respiratory toxins protects S. delphini from killing by Pseudomonas aeruginosa P. aeruginosa produces various QS-controlled respiratory toxins such as pyocyanin and hydrogen cyanide [32,33], which kill S. aureus [27]. If yayurea A and B repress the production of QS-regulated pyocyanin or hydrogen cyanide, one would expect that S. delphini survives better in a co-culture with P. aeruginosa than, for example, S. aureus. Indeed, co-

Author Summary
While studying the potential interaction of staphylococci with Gram-negative bacteria, we came across another communication system in a Staphylococcus species group, which consists of closely related coagulase-positive bacterial species that play a role as zoonotic pathogens. We found that these species excrete two small compounds that inhibit both the expression of QS-controlled toxins and other QS-regulated compounds as well as growth in Gram-negative bacteria. The excreted compounds, which we named yayurea A and B, were isolated from S. delphini and structurally characterized. They represent new bacterial products, which quench the QS regulation in a wide spectrum of Gram-negative bacteria by stimulating the LuxN-mediated phosphorylation of LuxU. Furthermore, growth of yayurea A and B producing S. delphini is not suppressed by respiratory toxins when co-cultured with P. aeruginosa. This suggests that the quorum quenchers have a function in self-protection and competitiveness in natural environments shared with Gram-negatives. Here we show one of the rare cases of inter-phylum interference between firmicutes (Gram-positive) and beta-/gammaproteobacteria (Gram-negative).
cultivation studies with S. aureus (non-producer) or S. delphini with P. aeruginosa SH1 revealed that the viability (CFU) of S. aureus significantly decreased in the stationary growth phase, most likely due to the respiratory toxins produced by P. aeruginosa (Fig. 3A), while that of S. delphini was uninfluenced (Fig. 3B). The addition of yayurea A and B to the mixed S. aureus and P. aeruginosa culture protected S. aureus from Figure 1. Quenching of QS-regulated pigments and bioluminescence by S. delphini. P. aeruginosa (A), S. marcescens (B), V. harveyi (C) and C. subtsugae (D) were each co-cultivated with S. aureus (1) or S. delphini (2) for 24 h. Pyocyanin, which is excreted by P. aeruginosa, was determined in the supernatant at its absorption maximum A 520 nm . Prodigiosin, which is cell wall bound in S. marcescens, was ethanol-extracted from the cell pellet and determined at its absorption maximum A 534 nm . Bioluminescence of V. harveyi was intensified by aeration before measuring in a bioluminescence reader. Violacein from C. subtsugae was quantitatively extracted with butanol and determined at its absorption maximum A 585 nm . doi:10.1371/journal.ppat.1003654.g001 killing in the stationary phase (Fig. 3A). Furthermore, the CFU of P. aeruginosa SH1 was unaffected while co-cultured with S. delphini or S. aureus, indicating that none of the two staphylococcal species was able to kill P. aeruginosa. Co-cultivation of P. aeruginosa with S. aureus in the presence of yayureas (100 mg/ml yayurea A and 900 mg/ml yayurea B) had also no effect on viability of P. aeruginosa (Fig. 3C). Other tested Pseudomonas strains PAO1 and DSMZ 50071 showed similar results (data not shown). All in all, these results showed that production of yayurea A and B enables staphylococci to coexist with Gram-negative bacteria in a mixed community.  Yayurea A and B are mainly produced in stationary growth phase of S. delphini We followed the production of yayurea A and B over 24 h in the supernatant S. delphini grown in TSB. The amount of produced yayurea A and B was determined by HPLC-analysis; peak integration was correlated with standard yayurea A and B. The production of both compounds started at the transition of exponential to stationary growth phase (after approximately 4 h) and increased rapidly for the next 5 h; after 24 h little more was produced (Fig. 4). The production kinetics is reminiscent of a QS controlled expression. Both compounds were produced in amazingly high concentrations: yayurea A reached concentration of 120 mg/ml and yayurea B even 900 mg/ml. This high concentration is entirely sufficient to suppress QS-systems in Gram-negative bacteria as can be seen below.

Yayurea A is more active than yayurea B in inhibiting various QS-controlled traits in Gram-negative bacteria
Purified yayurea A and B inhibited QS-regulated factors in Gram-negative bacteria in a dose-dependent manner. We tested prodigiosin production in S. marcescens, bioluminescence in V. harveyi, and pyocyanin production in P. aeruginosa ( Fig. 5 and 6).
Yayurea A and B inhibited production of prodigiosin in S. marcescens in a dose-dependent way. Inhibition already started at low concentrations (15 mg/ml) and increased with increasing concentrations of yayurea A or B. At 250 mg/ml, prodigiosin production was inhibited at 60% (yayurea A) and 40% (yayurea B). At a concentration of 1000 mg/ml, prodigiosin production was completely inhibited by yayurea A, and to approximately 70% by yayurea B (Fig. 5A). At 500 mg/ml, bioluminescence of V. harveyi was inhibited by yayurea A and B by 99% and 76% respectively (Fig. 5B). We also investigated whether yayurea A and B inhibited biofilm formation in P. aeruginosa and S. aureus. Again yayurea A quite efficiently inhibited biofilm formation in P. aeruginosa, while yayurea B was less effective (Fig. 5C). In contrast to gallidermin, a good biofilm inhibitor in staphylococci [34], both yayureas showed no biofilm-inhibiting effect with S. aureus (Fig. 5D).
We noticed that high dose of yayurea A and B (especially yayurea A) inhibit the growth of Gram-negative bacteria. We used P. aeruginosa as an example to verify the effect of growth and QS inhibition by yayurea A and B, and used the antibiotic tetracycline and the well-known QS-inhibitor furanone [35] as controls (Fig. 6). For QS inhibition, around 50% of the yayurea A and B concentration was necessary compared to that needed for growth inhibition (Fig. 6 A, B). Furanone revealed a similar correlation of growth and QS inhibition; the concentration (31 mg/ml) that inhibited 40% of growth inhibited 80% of QS (Fig. 6C). In contrast to yayurea and furanone, tetracycline inhibited growth and QS almost linearly (Fig. 6D). Besides P. aeruginosa, yayureas affect also growth of S. marcescens, V. harveyi, and V. cholerae ( Figure  S2); however, growth of E. coli was not affected. All staphylococcal species tested, such as S. aureus, S. carnosus, S. delphini or S. schleiferi, are resistant to yayurea A and B, independently whether they are producing these compounds or not.

Yayurea A and B are perceived by the AHL-receptor LuxN of Vibrio harveyi
To gain insight into the molecular mechanism behind why yayurea A and B cause a decrease in bioluminescence of V. harveyi, we performed in vitro phosphorylation assays of the corresponding signaling proteins. The full-length hybrid kinases LuxN, LuxQ, CqsS (tagged with 6 histidine residues) were heterologously expressed in an E. coli strain that lacks the F 1 /F o -ATPase (to prevent ATP degradation during phosphorylation experiments), and inverted membrane vesicles prepared from this strain were directly used for the phosphorylation experiments. Then we tested the effect of yayurea A and B on the LuxN, LuxQ [in interplay with LuxP (LuxPQ)], and CqsS-mediated timedependent phosphorylation of the HPt protein LuxU (Fig. 7A). Yayurea A and B had no effect on LuxPQ or CqsS-mediated phosphorylation of LuxU (Fig. 7B). However, they significantly stimulated the LuxN-mediated phosphorylation of LuxU in comparison to the control (Fig. 7C, compare lane 2 with lanes 4 and 10). LuxN is the sensor for the AHL autoinducer N-(3hydroxybutyryl)-homoserine lactone (HAI-1) (Fig. 7A). It is known [36] that the presence of HAI-1 inhibits the autophosphorylation activity of LuxN, thus decreasing the level of phospho-LuxU (Fig. 7C, compare lanes 2 and 6).
Yayurea A and B caused the opposite effect, leading to an increase in the level of phospho-LuxU, which explains the  decrease in bioluminescence when V. harveyi is exposed to these compounds. Under our test conditions, autoinducer HAI-1 was dominant in relation to yayurea A or B (Fig. 7C, compare lane 4 with lanes 8 and 12). Taken together these data show that the receptor kinase LuxN of V. harveyi specifically recognizes yayurea A and B. It is still unclear whether yayurea A and B also bind to the HAI-1 binding site or whether LuxN contains an independent binding site for these new compounds.

Phylogenetic position of yayurea A and B producing species in the Staphylococcus taxa
We tested a number of staphylococcal species representatives (listed in Table 2) by co-cultivation with P. aeruginosa for suppression of pyocyanin production. Only five species exerted such an activity: S. delphini, S. intermedius, S. lutrae, S. pseudintermedius, and S. schleiferi, and they all produced yayurea A and B as determined by HPLC analysis. Based on 16S rRNA-and multilocus sequence typing (MLST) the species are phylogenetically related and summarized in the 'intermedius group' [37] ( Figure  S3). Typically, the species are coagulase positive (with exception of S. schleiferi subsp. schleiferi), and oxidase negative. Interestingly, they all colonize various animals, and many represent zoonotic pathogens [38,39]. None of the other listed staphylococcal species (even not the next related S. hyicus, S. chromogenes or S. muscae) inhibited QS-system in Gram-negative bacteria or produced yayurea A and B.

Discussion
In this study, we found that some staphylococcal species excrete two novel compounds, yayurea A and B that interfere with the QS system of diverse Gram-negative bacteria. We tested 24 staphylococcal species with respect to production of yayurea A and B by co-cultivation with Gram-negative bacteria. The results showed that only five species (S. delphini [40], S. intermedius [41], S. pseudintermedius [42], S. lutrae [43], and S. schleiferi [44]) produced these compounds and were able to inhibit QS-regulated markers of Gram-negative test bacteria. All five species are naturally associated with various animals, where they act as zoonotic pathogens. They are only rarely associated with human infections. It is remarkable that these species are clustered in various phylogenetic trees, based on either 16S rRNA [45,46], on thermonuclease (nuc) sequences [47], on major autolysin (atl) sequences [48], or on multilocus sequence typing (MLST) [37]. The five species form a phylogenetic cluster, termed ''intermedius group'' ( Figure S3).  To gain insight into the mode of QS quenching by yayurea A and B, we performed in vitro phosphorylation assays with the autoinducer receptors LuxN, LuxQ, CqsS (Fig. 7A). Yayurea A and B had no effect on LuxPQ or CqsS-mediated phosphorylation of LuxU. However, they significantly stimulated the LuxNmediated phosphorylation of LuxU (Fig. 7B-C), suggesting that they interact with LuxN and cause LuxU activation. While the LuxN autoinducer HAI-1, an AHL, inhibits the autophosphorylation of LuxN, consequently leading to a decrease in LuxUphosphorylation, yayurea A and B caused the opposite effect by increasing LuxU phosphorylation. Thus, it is suggested that yayurea A and B keep the V. harveyi in a phenotypic state of low cell density, although the cells have grown to high density.
When both AI-1 and yayurea A or B were applied, HAI-1 overruled the effect of yayurea A and B in vitro. Bioluminescence of the wild type strain in vivo significantly decreased after exposure to yayurea A and B, although the effect of the latter compound was smaller. Obviously, HAI-1 did not overrule the effect of yayurea A and B in vivo. This difference can be explained by the fact that bioluminescence was measured in stationary phase grown cells at a time when the impact of HAI-1 on QS induction is low [49]. Furthermore, the contribution of the other receptors (LuxQ and CqsS) and their cognate autoinducers (AI-2 and CAI-1) needs to be considered in vivo [49]. The ratio of the three receptors and the impact of their kinase and phosphatase activities on the output of the phosphorylation cascade is not yet fully understood. This could be another explanation for the strong effect of the quorum quenchers in vivo. Membrane-topology analysis predicts that LuxN is bound to the bacterial inner-membrane by nine transmembrane (TM) spanning domains [50], and periplasmic loop 3 might be the HAI-1 binding site [21]. We believe that yayurea A and B interfere with the AHL quorum sensing response of Gram-negative bacteria; since the tested bacteria on which the effects were observed have at least one AHL-based quorum sensing system (Pseudomonas, Chromobacterium, Vibrio, and Serratia).
Most natural environments harbor a stunningly diverse collection of microbial species. One example, is the marine bacterium Halobacillus salinus, which produces N-(29-phenylethyl)isobutyramide and 2,3-methyl-N-(29-phenylethyl)-butyramide [51]. These compounds are unrelated to yayurea A (N-[2-(1Hindol-3-yl)ethyl]-urea) and B (N-(2-phenethyl)-urea). Within these communities, bacteria compete with their neighbors for space and resources [52]. Zoonotic commensals and pathogens, including yayurea producing staphylococcal species, use animals as a habitat. Whether these animals are also colonized by Gramnegative bacteria, other than in the gut, has barely been investigated. However, one can assume that animals bathe in puddles, lakes, and rivers, suggesting that their mucus, skin, fur, or feathers may easily encounter Gram-negative bacteria, being transient to permanent colonizers. One can also assume, that the 'intermedius group' share the habitat animal with Gram-negative bacteria. S. delphini was isolated from dolphin. Since both V. harveyi and S. delphini are marine bacteria and animal pathogens, it is conceivable that both bacteria share the same habitat and moreover, the same host surface. Our results demonstrated suppression of the quorum sensing regulated bioluminescence of V. harveyi during co-culture with S. delphini (Fig. 1), indicating that yayurea A and B are effective quorum quenchers.  Table 2. Strains tested for Q-quenching activity and yayurea A and B production.

QS-inhibition
The in vitro phosphorylation assay gave first mechanistic insights on how yayurea A and B affect quorum sensing of V. harveyi. We detected a significant and specific effect of yayurea A and B on the AHL-receptor LuxN. The concentrations used for these in vitro assays might not be physiological, especially when the influence of both molecules (HAI-1 and yayurea) was studied. Furthermore, it should be noted that HAI-1 is not constantly produced [49]. Therefore, in the natural habitat there might be times, when yayurea from S. delphini can fully interact with LuxN from V. harveyi in the absence of any competition with HAI-1. It therefore makes sense that the zoonotic staphylococci impair the QS-system of Gram-negative bacteria for competitive reasons. The growth of these staphylococci is not impaired by P. aeruginosa because the excreted yayureas suppress the production of the QS-controlled toxins and are thereby protected from being killed by the toxins (Fig. 3). While biofilm formation appears to be modulated by many regulators and environmental conditions in P. aeruginosa, pyocyanin and cyanide are controlled by the Las-QS system [53], implying that yayurea A and B might compete with 3-oxo-C12-HSL for LasR interaction.
When yayurea A (125 mg/ml) was added, the growth of P. aeruginosa was not affected (Fig. 6A), but quorum sensing and biofilm formation (Figs. 6A and 5C) were inhibited by 20%, which indicates that biofilm inhibition takes place prior to the onset of growth inhibition. In addition to their quorum quenching activity in Gram-negative bacteria, yayureas also inhibited their growth at higher dose (Fig. 2), which is a further advantage in the race for space and resources. For this advantage the zoonotic staphylococci are apparently prepared to pay a certain price, namely the production and excretion of comparatively high amounts of yayurea A and B. However, the benefit in competitiveness appears to prevail the cost disadvantages. As QS controls not only virulence factors but also many metabolic functions important for fitness, we don't know whether the inhibition of growth is a consequence of QS-inhibition or vice versa [54][55][56]. For some antibiotics (azithromycin, ceftazidime, and ciprofloxacin), it has been shown that they decrease the expression of QS-regulated virulence and many other genes [57].
The yayureas are potential candidates for use as antiinfectives. The only disadvantage might be the rather high concentration needed to completely quench QS; on the other hand, preliminary results suggest that they hardly have cytotoxic activities.
An interesting question is why just the 'intermedius group' is producing yayureas. The most likely answer is that this group shares its habitat with Gram-negative bacteria. There must be a benefit to producing yayurea A and B because they are excreted in such high amounts certainly costing energy. In addition, the yayureas show growth inhibition only at quantities that exceed the physiological concentrations. It is a clever arrangement, that S. delphini produces just enough yayurea A and B to almost completely silence the expression of the studied QS-regulated compounds or biofilm formation in diverse Gram-negative bacteria. In mixed cultures, the yayurea-producing staphylococci arrest Gram-negative bacteria in a pheno-and genotypic state of low cell density, although the cells have grown to high density. The advantage for the staphylococci is twofold, on one hand, they are protected from QS-controlled toxins and on the other hand, yayurea A and B affect stationary growth of particularly those Gram-negative bacteria with prominent QS-control systems. This is one of the rare cases of inter-phylum interference between firmicutes (Gram-positive) and beta-/gammaproteobacteria (Gram-negative). A schematic presentation of the interference is shown in Fig. 8.

Bacterial strains and growth conditions
Staphylococcus strains were grown in Tryptic Soy Broth (TSB, Sigma) medium at 37uC, Vibrio cholerae SP27459 (O1 El Tor), Serratia marcescens, Chromobacterium subtsugae DSMZ 17043 and Pseudomonas aeruginosa were grown in lysogenic broth (LB) medium, and Vibrio harveyi DSMZ6904 in marine broth (MB; 5 g peptone, 3 g yeast extract and 75% sea water per liter of deionized water) at 28uC. The Staphylococcus strains tested in this work are listed in Table 2. Mixed cultivation of Staphylococcus sp. with P. aeruginosa In co-culture experiments, we inoculated TSB with S. aureus (OD 578 0.005) and P. aeruginosa SH1 (OD 578 0.005). For the protection test, yayurea A (100 mg/ml), yayurea B (900 mg/ml), or water (control) were added to the mixed culture of S. aureus and P. aeruginosa after 3 h incubation. For co-cultivation of S. delphini with P. aeruginosa, each of the strains was inoculated with OD 578 0.001, because doubling time of S. delphini was shorter than that of S. aureus. Co-cultures were aerobically grown at 37uC. For CFU determination, samples were diluted and plated on Chapman agar (selective medium for staphylococci, while growth of P. aeruginosa is retarded) and BM agar (1% tryptone, 0.5% yeast extract, 0.5% NaCl, 0.1% K 2 HPO 4 , 0.1% glucose) on which P. aeruginosa grows better than the staphylococci. Colonies were counted after incubating for 24 h at 37uC; staphylococcal and pseudomonas colonies could be easily distinguished by colony shape and pigmentation.
Purification and structural analysis of the QS-inhibiting compounds from S. delphini The QS inhibitors were isolated from aerobically-grown overnight culture of S. delphini (vigorous shaking at 37uC, 20 h). Cells were centrifuged and the supernatant was applied to an Amberilite XAD 16 resin (Rhom& Haas, Germany). The column was first washed with water, then with 40% and 60% methanol and finally eluted with solvent A (80% methanol containing 5% acetic acid). Yayureas were significantly enriched in the last fraction. In the second purification step, the eluate was evaporated in vacuo, suspended in water, and further separated on an Amberilite IRC 50 cation exchange resin (Serva, Heidelberg, Germany) after the pH was adjusted to 7.0 with 1 M NaOH. The column was washed with water, then 70% ethanol and eluted with solvent A. The eluate was concentrated in vacuo, diluted, and adjusted to a final concentration of 50 mM sodium phosphate buffer with a pH of 4.2. In the third purification step, the eluate was separated on a SP Sepharose cation exchange column (GE Healthcare, Germany) with a linear 1 M NaCl gradient in 50 mM sodium phosphate buffer on a Ä kta FPLC (GE Healthcare, Germany). The two active compounds were eluted separately at 120 mM and 280 mM NaCl concentration respectively. The final purification and desalting of each peak was carried out by reversed phase preparative HPLC (RP-HPLC) (Bischoff, Leonberg, Germany) on a nucleosil 100 C-18, 86250 mm column (Machery Nagel, Düren, Germany) with a linear water acetonitrile (containing 0.1% TFA) gradient of 0% to 60% in 25 min. Purified compounds were lyophilized and stored at 220uC. Qualitative analysis was carried out on an Agilent 1200 HPLC system (Agilent technologies, Waldbronn, Germany) and a RP-HPLC Waters xBridge C18, 5 mm, 4.66150 mm column. Compounds were eluted with a 15 min linear gradient of 0.1% phosphoric acid to acetonitrile at a flow rate of 1.5 ml/min. For structure elucidation, mass spectrometry was carried out on GC-MS and FT-ICR MS (Bruker, ApexII). NMR spectra were measured in d 4  Assessing QS-regulated compounds in S. marcescens, P. aeruginosa, C. subtsugae, and bioluminescence in V. harveyi in co-cultivation with Staphylococcus sp.
Overnight cultures of S. aureus and S. delphini cells were diluted in LB medium containing 0.3% glucose to an OD 578 value of 1.0 (we also tried 0.1, 0.5, but OD 1.0 results were most pronounced), incubated for 4 h at 37uC, then co-cultivated at 30uC with S. marcescens or P. aeruginosa; each strain was inoculated with a starting OD 600 of 0.01. After 24 h of cultivation, prodigiosin of S. marcescens was extracted from the cell pellets by ethanol acidified with 4% of 1 M hydrochloric acid and then quantified by A 534 nm determination. Pyocyanin in the supernatant was determined by its absorption maximum at 520 nm [58]. For bioluminescence tests, overnight-cultured V. harveyi was diluted to an OD 578 of 1.0 and cocultivated with an equal amount of S. aureus or S. delphini cells in LB-MB (LB medium containing 0.1% glucose mixed with equal volumes of MB) for 18 h. Overnight cultures of S. aureus and S. delphini cells were diluted to an OD 578 of 0.1, incubated for 4 h at 37uC, then co-cultivated at 30uC with C. subtsugae, which was diluted to a final OD 600 value of 2; this high OD was necessary to pronounce violacein production, whose expression is dependent on AHL at higher cell density. After 24 h of incubation, cell pellets were collected and resuspended in water. Cells were lysed by 10% sodium dodecyl sulfate and incubated for 5 min at room temperature. Violacein was quantitatively extracted from the cell by adding watersaturated butanol. The butanol phase containing the violacein was collected and determined at its absorption maximum at 585 nm.
Production of yayurea A and B in S. delphini and their effect on bacterial growth S. delphini was inoculated to a starting OD 578 of 0.1 and incubated at 37uC. During the 24 h incubation, cell density was followed by OD 578 and the active compounds in the supernatants were determined by RP-HPLC and phosphoric acid -acetonitrile gradient as described above. The Gram-negative representatives, E. coli, P. aeruginosa, S. marcescens, V. cholerae and V. harveyi were inoculated to OD 600 of 0.1 and grown in LB for 2 h. After that time, 1 mg/ml yayurea A or B (as solution in H 2 O) or an equal volume of water as negative control was added. Cells were either incubated at 37uC (S. aureus or S. delphini) or 30uC (E. coli, S. marcescens, V. harveyi, V. cholerae and P. aeruginosa). Cell density was followed for 24 h.
Activity assays for QS-regulated compounds, bioluminescence, and biofilm formation in Gramnegative bacteria Pyocyanin production. An overnight culture of P. aeruginosa PAO1 was diluted with LB medium to an OD 600 of 0.1. The (Z-)-4bromo-5-(bromomethylene)-2(5H)-furanone (Sigma), yayureas and tetracycline were dissolved in 10% DMSO. 190 ml diluted PAO1 was cultured in 96-well plates with 10 ml different concentrations of pure yayurea A, B, furanone, or 10% DMSO (negative control) at 30uC. After 24 h of incubation, cell density was measured at 600 nm and pyocyanin was isolated as described previously [58]. To measure pyocyanin production, the supernatants were subjected to HPLC separation on a Nucleosil 100, C-18 column and a 0-100% Water-ACN Gradient (water containing 0.1% phosphoric acid) in 15 min at a flow rate of 1.5 ml/min. The content of pyocyanin was determined at its absorption maximum at 520 nm [58]. The relative pyocyanin production was calculated as the ratio of the amount of pyocyanin content from the HPLC measurements to cell density.
Prodigiosin production. S. marcescens cells were inoculated from an overnight culture to a starting OD 600 of 0.1. Then 100 ml of the diluted culture was mixed with 100 ml of different amounts of the purified compounds or water in wells of 96-well microtiter plates and incubated for 24 h at 28uC. Cell density was measured at 600 nm. Prodigiosin was extracted as described above and the relative prodigiosin production was calculated as the ratio between the amount of prodigiosin at its absorption maximum at 534 nm and cell density values at 600 nm. The prodigiosin suppressing effect of yayurea A and B was determined (H 2 O was used as control).
Biofilm formation assay. Biofilm formation assays were performed as described previously [34]. A full-grown culture of P. aeruginosa PAO1 or S. aureus was diluted with LB medium containing 0.3% (w/v) glucose to an OD 600 of 0.1. 100 ml of the dilution was mixed with 100 ml of the yayurea solutions with the indicated concentrations, and incubated in 96-wellplates at 37uC for 24 h. To determine the biofilm formation, planktonic cells were discarded and the plates gently washed with PBS and air-dried for 30 min. The wells were stained with 200 ml of a 0.1% crystal violet solution at room temperature for 30 min. The stained biofilm was rinsed with distilled water followed by the addition of 200 ml DMSO. Absorbance of crystal violet as indicator for biofilm-forming bacteria was measured at 590 nm. The effects of the compounds were evaluated for which the sample with water was set to 100%.
Bioluminescence. An overnight culture of V. harveyi was diluted to an OD 600 of 0.1 and grown with yayurea A, B, or water for 24 h at 28uC. Bioluminescence was determined in a Tecan infinite M200 plate reader (Tecan, Groedig, Austria). Relative luminescence was normalized by the OD 600 values. The effects of the compounds were calculated by the relative luminescence, for which the sample with water was set to 100%.

Investigation of the QS-target(s) of yayureas in vitro
Production of QS-receptors and QS-molecules. Inverted membrane vesicles were prepared using E. coli strainTKR2000 expressing plasmids pNKN, pNKQ, and pNKS encoding wild type LuxN, LuxQ, and CqsS (each with a C-terminal His-tag), respectively. Inverted membrane vesicles were prepared as described by [36]. LuxP was produced in, and purified from, E. coli MDAI-2 transformed with the plasmid pGEX_LuxP as described before [59]. LuxU was produced and purified as described before, using E. coli JM109 transformed with plasmid pQE30LuxU-6His [36]. Synthetic autoinducer, HAI-1, (from the University of Nottingham) was dissolved in a minimal volume of acetonitrile [10% (v/v)], diluted with water to a concentration of 100 mM and stored at 220uC.
Phosphorylation assays. Each QS-receptor kinase was tested as full-length membrane integrated proteins in inverted membrane vesicles. The buffer used for phosphorylation reactions was as follows: 50 mM Tris/HCl pH 8.0, 10% (v/v) glycerol, 500 mM KCl, 2 mM DTT. Each phosphorylation reaction contained equimoloar receptor concentrations [the total protein concentration were accordingly adapted: 3.2 mg/ml (LuxN), 5.5 mg/ml (LuxQ), and 5.2 mg/ml (CqsS) membrane proteins], and purified LuxU at 0.1 mg/ml. For phosphorylation of LuxQ, LuxP was added at a concentration of 0.4 mg/ml. LuxP integration into LuxQ-bearing membrane vesicles was triggered by several cycles of thawing and freezing. When indicated, yayurea A or B were added to the reaction mix at a final concentration of 220 mg/ml, and HAI-1 at 20 mM. The reaction was started by adding radio labeled Mg 2+ -ATP, typically 100 mM [c-32 P]ATP (specific radioactivity of 0.94 Ci/mmol; Perkin Elmer) and 110 mM MgCl 2 . The reaction was stopped after 1 and 10 minutes by adding SDS Laemmli-loading buffer followed by separation of the proteins on SDS-polyacrylamide gels. Gels were dried at 80uC on filter paper, exposed to a phosphoscreen for at least 24 h, and subsequently scanned using a PhosphorImager SI (GE Healthcare). The gels presented are representative of at least three independent sets of experiments.  Figure S3 Phylogenetic tree among Staphylococcus species. The tree is based on 16S rRNA relationships according to [45]. The phylogenetic position of the 'intermedius group' composed of S. intermedius, S. pseudintermedius, and S. delphini is marked in bold. Based on a combinatin of 16S rRNA and multilocus data, the group was recently complemented by the next related species S. lutrae, S. schleiferi subsp. schleiferi and S. schleiferi subsp. coagulans [37]. All members of this group produce yayurea A and B. (TIF)