Pseudomonas aeruginosa quorum sensing inhibition by clinical isolate Delftia tsuruhatensis 11304: involvement of N-octadecanoylhomoserine lactones

Pseudomonas aeruginosa is one of the most common opportunistic pathogens that use quorum sensing (QS) system to regulate virulence factors expression and biofilm development. Delftia sp. 11304 was selected among 663 Gram-negative clinical isolates based on its QS inhibitory activity against P. aeruginosa MMA83 clinical isolate. Whole genome sequencing identified this isolate as D. tsuruhatensis and revealed genetic armamentarium of virulence factors and antibiotic resistance determinants. Ethyl acetate extract of D. tsuruhatensis 11304 culture supernatant (QSI extract) prevented biofilm formation of P. aeruginosa MMA83, but was unable to cause biofilm decomposition. QSI extract showed a synergistic effect in combination with meropenem and gentamycin, against P. aeruginosa MMA83. A dose-dependent reduction of the virulence factors: elastase, rhamnolipid and pyocyanin production by P. aeruginosa MMA83 and significant downregulation of lasI, lasR, rhlI, rhlR, pqs and mvfR expression were observed. Matrix-assisted Laser Desorption Ionization (MALDI) mass spectrometry of D. tsuruhatensis 11304 QSI extract revealed the presence of N-acyl homoserine lactones (AHL) with chain lengths of C12 to C18. The main ion peak was identified as N-octadecanoylhomoserine lactone (C18-HSL). Commercial C18-HSL (20 µM) reduced pyocyanin production as well as mRNA level of the lasI gene. A novel AHL species, dihydroxy-N-octadecanoylhomoserine lactone, was also described.


Results clinical isolate Delftia sp. 11304 produces QSI molecule(s). Among 633 clinical isolates from a
Laboratory collection screened for quorum sensing inhibitory (QSI) activity (Supplementary Table 1) 19 strains (belonging to five genera) were selected as positive (Supplementary Table 2). Delftia sp. 11304 was selected as one of the most promising candidate with QSI activity. Growth phase dependence of the Delftia sp. 11304 QSI phenotype was determined, as the QSI phenotype could not be detected before 4 hours of growth, while the peak of QSI activity was observed after16 hours (corresponding to the stationary phase). High production of QSI molecule(s) was retained after up to 30 hours of growth (Fig. 1). As a control QSI activity of previously characterized QSI molecules produced by Delftia tsuruhatensis: bis (2-ethylhexyl) phthalate and diisooctyl phthalate were tested in parallel. Neither of these compounds showed QSI activity at the used concentrations (ranging from 4 µg/ml to 1024 µg/ml) when tested in a colorimetric agar well diffusion assay using Chromobacterium violaceum CV026 as an indicator strain ( Supplementary Fig. S1). Additionally, no measurable effects of bis (2-ethylhexyl) phthalate and diisooctyl phthalate neither on the viability of C. violaceum CV026 and P. aeruginosa MMA83 (Supplementary Table S3) nor to P. aeruginosa MMA83 biofilm production were detected ( Supplementary  Fig. S2).
Whole genome sequencing. Genomic DNA of the strain Delftia sp. 11304 was sequenced using Illumina HiSeq. 2500 platform (MicrobesNG, University of Birmingham, United Kingdom). A total of 6,616,336 sequences were generated from the genome. Assembly of reads resulted in 199 contigs with the largest contig 337425 and an average GC content of 66.78%. The contig dataset was used to determine the functional analysis. According www.nature.com/scientificreports www.nature.com/scientificreports/ to the genomic sequence analysis Delftia sp. 11304 was identified as a Delftia tsuruhatensis species. A repertoire of different virulence factors and antibiotic resistance determinants were detected in D. tsuruhatensis 11304 genome (Supplementary Tables 4 and 5). Among the virulence factors presence of the genes encoding for different siderophores, pyoverdine, pyochelin and ornitobactin, as well as capsule and alginate could be pointed out. Additionally, within the D. tsuruhatensis 11304 resistome antibiotic resistance determinants such as genes encoding for efflux pumps of Resistance-Nodulation-Division (RND) superfamily (including: mexF, mexY, mexB, mexD, mexI, mexA, smeB, smeE, amrB and mdtB), beta-lactamases genes bla OXA-258 and bla LRA-13, as well as spiramycin resistance gene were identified as the most significant.

Stability of the QSI molecule(s).
In order to determine whether D. tsuruhatensis 11304 QSI activity was based on proteinaceous or non-proteinaceous molecule activity, thermostability and resistance to proteinase were tested. Results showed that both overnight culture and cell-free supernatant retained QSI activity after the heat and proteinase K treatments (Fig. 2). Obtained results suggested that the QSI activity of D. tsuruhatensis 11304 was based on non-enzymatic quorum sensing inactivation mechanism. ethyl acetate D. tsuruhatensis 11304 extract exhibits QSI activity and inhibits biofilm formation of P. aeruginosa MMA83. Organic solvents with different polarity were used to optimize the extraction of the QSI molecule(s) from D. tsuruhatensis 11304 overnight culture supernatant. Only the ethyl acetate extract resuspended in DMSO was shown to be QSI active (Fig. 2), while methanol, chloroform and hexane extracts did not exhibit any QSI activity.
To exclude that the QSI effect of extract was a result of a bactericidal effect on the P. aeruginosa MMA83, we evaluated the antimicrobial activity of the QSI extract by microdilution method. As shown in Fig. 3, D. tsuruhatensis 11304 QSI ethyl acetate extract had no statistically significant effect on the P. aeruginosa MMA83 growth.
Additionally, we have demonstrated that the inhibition of P. aeruginosa MMA83 biofilm formation by D. tsuruhatensis 11304 was dose-dependent using crystal violet staining of biofilm biomass (Fig. 3). Indeed, biofilm formation decreased by 88.5% in the presence of 5 mg/ml QSI extract comparing to the control (***p < 0.001). About 50% inhibition of biofilm formation was observed with a 0.156 mg/ml QSI extract (***p < 0.001). The lowest tested concentration of QSI extract (0.038 mg/ml) showed decreasing activity of 21.9% in regard to positive control (*p < 0.05). Treatment of P. aeruginosa MMA83 with the highest dose of DMSO used in experiment (0.5% v/v) had no statistically significant impact on bacterial growth and biofilm formation. D. tsuruhatensis 11304 QSI extract prevents biofilm formation of P. aeruginosa MMA83, but it is unable to induce biofilm decomposition. Fluorescence microscopy has been used to visualize the effect of the QSI extract on P. aeruginosa MMA83 biofilm formation as well as on the decomposition of preformed biofilm. The captured images showed that P. aeruginosa MMA83 that grew in the absence of D. tsuruhatensis 11304 QSI extract formed a compact biofilm. In the presence of D. tsuruhatensis 11304 QSI extract biofilm was highly dispersed, with a clear decrease in the surface coverage and density of bacteria. Indeed, as shown in Fig. 4, minimal attachment of P. aeruginosa MMA83 cells could be observed in the treatment with a 5 mg/ml QSI extract (after 24 hours of incubation) (Fig. 4b).  (1) and cell-free supernatant (2). The heat treatment at 100 °C for 30 min (3, 4) and 60 min (5,6). D. tsuruhatensis 11304 overnight culture (3,5) and cell-free supernatant (4, 6) retained QSI activity and shown to be thermostable. Proteinase K treatment did not affect QSI activity of D. tsuruhatensis 11304 overnight culture (7) and cell-free supernatant (8). Ethyl acetate extract of D. tsuruhatensis 11304 exhibits QSI activity (9). Proteinase K and LB growth medium with 0.5% DMSO (10,11) are used as negative controls. (b) Quantitative data of these experiments obtained by measurement of inhibition zones of violacein production by Chromobacterium violaceum CV026.
Interestingly, co-incubation of preformed P. aeruginosa MMA83 biofilm with the QSI extract (5 mg/ml) did not result in decomposition of the biofilm (Fig. 4d), although the biofilm architecture was less consistent and slightly different compared to the positive control (Fig. 4c), there were still visible living bacterial cells resembling the architecture observed in the positive control. Numbers on x-axis represent two-fold serial dilutions of the D. tsuruhatensis 11304 QSI extract with the unit of measure being mg/ml. DMSO -bacterial culture with 0.5% v/v of DMSO that is the highest concentration of DMSO in experiment (used when 5 mg/ml of QSI extract was added). Data analysis by Student's t-test demonstrates significant difference (*p < 0.05, **p < 0.01, ***p < 0.001) between the treated biofilms and the control (DMSO treated control). No significant difference in the cell growth between treated and untreated controls was observed. and decomposition of preformed biofilms (c,d), captured with fluorescence microscopy. P. aeruginosa MMA83 was coincubated for 24 hours with 5 mg/ml D. tsuruhatensis 11304 QSI extract in order to visualize its effects on biofilm forming ability. The effect of the D. tsuruhatensis 11304 QSI extract on decomposition of 24 hours old preformed biofilm was evaluated after 24 hours of treatment with QSI extract. Fluorescent dyes SYTO9 (green) and PI (red) were used for the visualization of biofilms. Scale bar represents 10 µm.
tigate the clinical relevance of the QSI extract and its effectiveness in combination with clinically used drugs, the checkerboard method was assessed. Results indicate that the test strain P. aeruginosa MMA83 was susceptible to meropenem and gentamycin at very high used concentrations (MIC values 0.512 mg/ml and 4.096 mg/ml, respectively). MIC value of the D. tsuruhatensis QSI extract was 20 mg/ml. However, their combined application with the QSI extract showed a synergistic outcome for meropenem (∑FIC = 0.125) and for gentamycin (∑FIC = 0.047) ( Table 1). DMSO (MIC value 145.45 mg/ml) had indifferent outcome for both of used antibiotics.

D. tsuruhatensis 11304 QSI extract inhibits the virulence factors production of P. aeruginosa
MMA83. The impact of the QSI extract on the QS regulated virulence factors (elastase, rhamnolipid and pyocyanin) production in P. aeruginosa MMA83 was next investigated. A dose-dependent decrease in the production of all the analyzed virulence factors was observed (Fig. 5). The lack of the production of elastase, rhamnolipid and pyocyanin after treatment of the P. aeruginosa MMA83 strain with a 5 mg/ml D. tsuruhatensis 11304 QSI extract has been noticed. The treatment with a 0.156 mg/ml QSI extract resulted in about 40% inhibition of rhamnolipid and 50% inhibition of elastase and pyocyanin production compared to the positive control.

D. tsuruhatensis 11304 QSI extract inhibits the expression of P. aeruginosa MMA83 QS systems.
In order to confirm that the above described effects of D. tsuruhatensis 11304 QSI extract on P. aeruginosa MMA83 virulence were caused by transcriptional inhibition of QS systems genes belonging to three P. aeruginosa QS networks (las -lasI, lasR; rhl -rhlI, rhlR; PQS -pqs, mvfR) were selected for RT-qPCR study. The obtained results revealed that the treatment of P. aeruginosa MMA83 with a 5 mg/ml QSI extract significantly decreased the mRNA levels (***p < 0.001) (Fig. 6). Relative expression of the genes coding for inducer synthases lasI and rhlI was 2.2 times lower, while the expression of transcriptional regulators lasR and rhlR was even more reduced (7.6 and 4.1 times lower, respectively) compared to the control. The most significant downregulation of transcription was observed in the case of the pqs gene (15 times lower), while the transcription of transcriptional regulator mvfR was 5.7 times lower compared to the control.  aeruginosa MMA83 virulence factors production. Black bars -production of elastase, white bars -production of rhamnolipid and gray bars -production of pyocyanin. Statistical significance was evaluated by Student's ttest against control (0) without QSI extract (*p < 0.05, **p < 0.01, ***p < 0.001).

Determination of D. tsuruhatensis 11304 QSI extract composition by MALDI mass spectrometry.
To define the nature of the molecules composing the D. tsuruhatensis 11304 QSI extract, an aliquot of the sample was preliminarily purified on a thin-layer chromatography (TLC) silica gel. Separated spots were visualized, and then scratched off. The silica gel was put into a cotton-plugged column, and the sample was eluted from the gel by ethyl acetate, dried and then investigated by MALDI mass spectrometry. Several solvents were employed to dilute and analyze the QSI extract; nevertheless, the best resolved spectra were obtained by diluting the QSI extract directly in the matrix solution (see Methods section for details). The positive ion MALDI mass spectrum, recorded in reflectron mode, is reported in Fig. 7a. The spectrum clearly indicated the presence of several [M + H] + ions attributed to N-acyl homoserine lactones (AHLs) with chain lengths of C12 to C18. In particular, the main ion peak at m/z 368.3 was identified as N-octadecanoylhomoserine lactone (C 18 -HSL). This was proven by the observation in the related positive ion MS 2 spectrum (Supplementary Fig. S3) of (i) the typical intense ion at m/z 102.06, common to all AHLs detected, relative to the protonated α-amino-γ-butyrolactone deriving from the cleavage of the carboxamide linkage between the acyl chain and the homoserine group; (ii) the counterpart acylium ion arising from the same cleavage (m/z 267.27) as well as (iii) the occurrence of a fragment originated from the loss of a water molecule from the [M + H] + molecular ion (m/z 350.29) 32,33 . Similarly, other AHLs species with a shorter acyl chain have been identified at m/z 284.2 (C 12 -HSL), m/z 312.3 (C 14 -HSL) and m/z 340.3 (C 16 -HSL). In traces, Figure 6. The changes of relative mRNA levels of QS genes without or with the treatment of P. aeruginosa MMA83 with a 5 mg/ml D. tsuruhatensis 11304 QSI extract. Genes examined in this study belong to three P. aeruginosa QS pathways (las -lasI, lasR; rhl -rhlI, rhlR; PQS -pqs, mvfR). RT-qPCR data were normalized against the ribosomal gene rpsL as an internal control. Student's t-test was used to compare the differences between the control and experimental groups (***p < 0.001). www.nature.com/scientificreports www.nature.com/scientificreports/ N-3-oxo-octadecanoylhomoserine lactone (oxo-C 18 -HSL) has also been detected at m/z 382.3 (Fig. 7a). In addition, a clearly intense peak at m/z 400.3 was noted and isolated as MS 2 precursor ion furnishing important structural data. The corresponding MALDI MS 2 spectrum (Fig. 7b), revealed the occurrence of the common ion at m/z 102.06 but also of an ion derived from the loss of water from the protonated molecular ion (m/z 382.29). c 18 -HSL reduces pyocyanin production and quorum sensing lasI gene expression in P. aeruginosa MMA83. In order to investigate QSI potential of C 18 -HSL its effect on expression of quorum sensing genes and pyocyanin production in P. aeruginosa MMA83 was analyzed. Small but statistically significant decrease in pyocyanin production by MMA83 grown in presence of 20 µM C 18 -HSL was observed (Fig. 8b). In addition, the same concentration of C 18 -HSL statistically significant reduced transcription of the lasI gene in MMA83, although the fold-change was small (Fig. 8a).

Discussion
Our understanding of Delftia tsuruhatensis has experienced a complete turnaround from the promising plant growth-promoting bacteria to emerging human pathogen predominantly isolated from respiratory specimens, blood and urine 27,30 . There is a disparity between the growing importance of Delftia spp. isolates and genomic data from these species. Noticeably, D. tsuruhatensis genomic data are scarce since only the genome of D. tsuruhatensis MTQ3 is publicly available on NCBI. Although we could not speculate about the origin of the D. tsuruhatensis 11304 strain, which was characterized in this study (was it intrahospital, or an environmental strain introduced to the hospital by a patient), the presence of genetic determinants for the production of siderophores, capsule and alginate is a good indication of a possible virulent phenotype. Along with the genetic determinants of antibiotic resistance that were found within the D. tsuruhatensis 11304 genome, like beta-lactamases and RND pumps, that could contribute to the overall pathogenicity of this particular strain and could provide insight into the genetic armory of D. tsuruhatensis in general.
The reason why we included Delftia spp. in testing was our initial presumption that bacteria, like D. tsuruhatensis 11304, which colonize the same tissues during infection of a host as P. aeruginosa, probably compete with P. aeruginosa by employing diverse mechanisms, including the interference of cell-to-cell signaling, i.e. quenching of the quorum sensing systems. Diisooctyl ester of 1,2 benzenedicarboxylic acid, a compound that is called by several different names, was previously identified as a QSI molecule, produced by Delftia tsuruhatensis, active against P. aeruginosa PAO1 QS 31 . It is worth noticing that a contamination of laboratory samples by this compound was found when plastic tubes were used, since it is widely used as a plasticizer 34,35 . In addition, our data showed that within the range of concentrations used in this study, bis (2-ethylhexyl) phthalate and diisooctyl phthalate were unable to interfere with C. violaceum CV026 quorum sensing system or with P. aeruginosa biofilm formation. Thus, our main aim was to characterize the QSI potential of D. tsuruhatensis 11304 against P. aeruginosa by using C. violaceum and a multidrug-resistant clinical isolate P. aeruginosa MMA83 as model systems. The D. tsuruhatensis 11304 ethyl acetate extract showed a remarkable ability to attenuate P. aeruginosa MMA83 QS, and to suppress its virulent phenotype by disabling its intrinsic ability to form a biofilm or to produce virulence factors such as elastase, rhamnolipid and pyocyanin.
Biofilm formation ability of P. aeruginosa MMA83 was significantly impaired by the D. tsuruhatensis 11304 extract, without affecting overall bacterial cell growth. This reduction in biofilm forming ability occurred in a dose-dependent manner, similarly as previously reported for other QSI extracts 36,37 . However, the disruption of preformed biofilm was not achieved, which might result from biofilm matrix impermeability or possibility that it acts through QS inhibition. The production of extracellular virulence factors is of crucial importance for the invasion of host tissues, initial phases of biofilm formation and promoting virulence expression and it was shown RT-qPCR data were normalized against the ribosomal gene rpsL as an internal control. (b) The changes in pyocyanin production after the treatment of P. aeruginosa MMA83 with a commercial C 18 -HSL. "C" on both graphs refers to the DMSO control. Student's t-test was used to compare the differences between the control and experimental groups (*p < 0.05, **p < 0.01, ***p < 0.001). www.nature.com/scientificreports www.nature.com/scientificreports/ to be controlled by las and rhl QS systems 38,39 . Our study demonstrated that the D. tsuruhatensis 11304 extract was highly capable of attenuating P. aeruginosa virulence, and thus possibly infection, by significantly reducing the production of these virulence factors. This anti-virulent effect of the D. tsuruhatensis 11304 extract was based on the inhibition on transcriptional level of las, rhl and pqs QS systems, indicating a possible therapeutic potential of the QSI molecule(s) produced by strain 11304.
Considering that clinical isolates of P. aeruginosa often possess MDR or even XDR resistant phenotype, novel strategy for clinical treatment of infections caused by such isolates could be the combination of antibiotics with non-antibiotic bioactive compounds. It has been reported that different QS inhibitors caused a decrease in antibiotic resistance in P. aeruginosa PAO1 through synergistic effects with clinically used drugs 40,41 . In our study, susceptibility of the MDR clinical isolate P. aeruginosa MMA83 to meropenem and gentamycin was enhanced by synergistic interactions of the D. tsuruhatensis 11304 QSI extract with antibiotics.
MALDI MS data showed that AHLs with chain lengths of C12 to C18 were present in the D. tsuruhatensis 11304 QSI extract, with the main ion peak attributed to N-octadecanoylhomoserine lactone (C 18 -HSL). Interestingly, peak at m/z 400.3 was noted and isolated as MS 2 precursor ion furnishing important structural data. The corresponding MALDI MS 2 spectrum (Fig. 7b), revealed the occurrence of the common ion at m/z 102.06 but also of an ion derived from the loss of water from the protonated molecular ion (m/z 382.29). Two peaks at m/z 281.25 and 263.24, matching with the loss of one and two water molecules respectively from the acyl chain, suggested the occurrence of two hydroxyl groups decorating the acyl chain. None of the fragments helped in understanding the position of the two hydroxyl moieties; nevertheless, according to literature data, it could be assumed that a single hydroxyl group is placed at position 3 of the acyl chain whereas the position of the second remains to be defined. Therefore, a novel AHL species could be identified as composing the D. tsuruhatensis 11304 QSI extract, specifically a dihydroxy-N-octadecanoylhomoserine lactone (Fig. 7b). Although C 18 -HSLs were previously described 42 , to our best knowledge this is the first report of a naturally occurring dihydroxy-N-octadecanoylhomoserine lactone. Finding that commercial C 18 -HSL reduces lasI quorum sensing gene expression and pyocyanin production indicates its involvement in virulent potential of P. aeruginosa MMA83. This finding is of importance since it is known that, in certain bacterial species, long-chain AHLs interfere with the short-chain AHLs-mediated QS signaling 17,18 . For example, C 10 -HSL fails to activate the CviR-dependent transcription in C. violaceum CV026, and yet functions as an antagonist in the presence of a native autoinducer C 6 -HSL. Additionally, it was shown that the lengthening of the AHLs' acyl-tails in the CV026 model system reduces agonism and enhances antagonism due to the promotion of an inactive conformation of the CviR:ligand complex 43 . It is tempting to speculate that the long-chain AHLs in the D. tsuruhatensis 11304 extract, among which a predominantly novel dihydroxy-N-octadecanoylhomoserine lactone is present, could be responsible for the interference of the P. aeruginosa MMA83 quorum sensing system and mitigation of its virulence. An analogy could be drawn to the CV026 quorum sensing system since P. aeruginosa doesn't produce AHLs with side chains longer than C12 44 . Our findings could be also supported with observations made by Zhu et al. 45 , who showed that one of the strongest antagonists of TraR, which is a 3-oxo-C 8 -HSL-responsive transcriptional activator in Agrobacterium tumefaciens, had also hydroxyl residues placed at position 3 of the acyl chain. The authors of mentioned study also claimed that the length of the side acyl chain was one of key determinants affecting antagonistic properties of synthetic AHLs used in their model system. We should take into account that the observed QSI activity could not be attributed solely to C 18 -HSL but to an orchestrated activity of several molecules from the extract, due to limited effect of C 18 -HSL comparing to the effect of entire 11304 QSI extract. Thus, our further work will be focused on the separation and identification of all the components in D. tsuruhatensis 11304 ethyl acetate extract, and testing of their QSI activities independently.

Methods
Bacterial strains and cultivation conditions. Strain Delftia sp. 11304 from the collection of Laboratory for Molecular Microbiology (LMM), Institute of Molecular Genetics and Genetic Engineering (IMGGE), University of Belgrade, used in this study was isolated in a co-culture with Achromobacter xylosoxidans 11304 from a cough swab 46 in a tertiary type hospital in Belgrade, Serbia. Quorum sensing inhibition (QSI) activity of 633 clinical isolates from LMM collection was tested by using AHL biosensor strain Chromobacterium violaceum 026 (CV026), a mini-Tn5 mutant deficient in the AHLs synthase cvil 17 . Pseudomonas aeruginosa MMA83, a New Delhi metallo-beta-lactamase-producing clinical strain from the laboratory collection (LMM, IMGGE) was used as a test strain 47,48 . The bacteria were cultured aerobically in either Luria-Bertani (LB) broth medium or Mueller-Hinton medium (MH) at 37 °C for all isolates except C. violaceum CV026 which was grown at 30 °C. M9 medium (10 x M9 salts -Na 2 HPO 4 x7H 2 O, KH 2 PO 4, NaCl, NH 4 Cl; 20% glucose; 1 M MgSO 4 ; 1 M CaCl 2 ; dissolved in miliQ water) supplemented with 0.24% pyruvate was used for bacterial cultivation on a large scale for extraction of molecule(s) that interfere with quorum sensing system. QSI detection assay. Initial screening for Delftia sp. 11304 QSI activity was performed with colorimetric agar well diffusion assay using Chromobacterium violaceum CV026 as an indicator strain. Overnight culture of C. violaceum CV026 (0.5% v/v) was inoculated in Luria-Bertani soft-agar (0.5% v/v) with addition of 5 μM N-(hexanoyl)-l-homoserine lactone -HHL (Sigma-Aldrich, Missouri, USA) and overlaid on LB agar (17% v/v). Wells were made in lawn of the solidified LB soft agar. Wells were filled with 50 μl of Delftia sp. 11304 culture and incubated for 24 h at 30 °C in order to test QSI activity 49 . The absence of the violet color around the wells was considered as evidence of the inhibition of violacein pigment synthesis, and demonstrative of positive QSI activity.
Chemical compounds bis (2-ethylhexyl) phthalate and diisooctyl phthalate that possess the same molecular formula C 24 H 38 O 4 (molecular weight 390.564 g/mol) and the same structural formula (diisooctyl phthalate) as a previously described quorum sensing inhibitor compound 31 were used in order to confirm their QSI and antibacterial activity. Two-fold serial dilutions (started from 1024 µg/ml to 4 µg/ml) of bis (2-ethylhexyl) phthalate and diisooctyl phthalate solutions (Sigma-Aldrich) dissolved in ethanol (1:1 ratio)-were tested for antibacterial activity on C. violaceum CV026 and P. aeruginosa MMA83, QSI activity (C. violaceum CV026) as well as biofilm forming effect (P. aeruginosa MMA83) using previously described methods.
Whole genome sequencing and genome analyses. Genomic DNA of Delftia sp. 11304 was sequenced using Illumina HiSeq by MicrobesNG service (MicrobesNG, IMI-School of Biosciences, University of Birmingham, Birmingham, UK). The quality of each sequencing library was assessed using FastQC 50 . IDBA-UD with multi k-mer mode outperformed the assembly using De Bruijn Graph methods 51 . In addition, unassembled reads were collected and assembled by Celera Assembler with the Best Overlap Graph-CABOG. Finally, raw reads were mapped to assembled scaffolds with Burrows Wheeler Aligner-BWA 52 . Delftia sp. 11304 was identified using the genome sequence and EzBioCloud 16 S rRNA gene database available online at https://help.ezbiocloud. net/ezbiocloud-16s-database/ 53 .
Identification of genes coding virulence factors in the genome sequence was performed using virulence factor database (VFDB), an online resource for curating information about virulence factors of bacterial pathogens (http://www.mgc.ac.cn/VFs/main.htm) 54 and Linux command line. Presence of the antibiotic resistance genetic determinants in the sequenced genome was determined by the publicly available database, The Comprehensive Antibiotic Resistance Database (CARD, https://card.mcmaster.ca/) 55 using Linux command line. Draft genome sequence of Delftia tsuruhatensis 11304 has been deposited at the NCBI GenBank database under accession number SMMJ00000000.

Stability of D. tsuruhatensis 11304 QSI molecule(s).
In order to test the thermostability of D. tsuruhatensis 11304 QSI activity, overnight culture and cell-free supernatant were incubated at 100 °C for 30 and 60 min. Additionally, a potential proteinaceous nature of QSI activity was tested by treatment of D. tsuruhatensis 11304 overnight culture and cell-free supernatant with proteinase K (500 μg/ml) at 37 °C for 3 hours. Residual QSI activities after the heat or proteinase K treatments were evaluated by colorimetric agar well diffusion assay.
Preparation of the D. tsuruhatensis 11304 QSI extract. Bacterial culture D. tsuruhatensis 11304 (five liters) was cultivated in M9 medium supplemented with 0.24% pyruvate for 24 h, at 37 °C, aerobically. After centrifugation at 13,680 x g, at 4 °C, for 30 min, the collected supernatant was divided into four glass flasks and each of them was extracted with an equal volume of methanol, chloroform, ethyl acetate and hexane (Sigma-Aldrich) with vigorous shaking for 30 min at room temperature. Extracts were evaporated to dry using a vacuum rotary evaporator at 50 °C for ethyl acetate and 30 °C for methanol, chloroform and hexane (Buchi Rotavapor, R200, Fisher Scientific, Hampton, New Hampshire, US) and the dry mass was dissolved in dimethyl sulfoxide (DMSO) for further analyses 56,57 . In order to avoid contamination with phthalates from plasticware all experiments were done in glassware.
The effect of the D. tsuruhatensis 11304 QSI extract on P. aeruginosa MMA83 biofilm formation ability. In order to determine the minimal inhibitory concentration (MIC) of ethyl acetate QSI extract dissolved in DMSO on P. aeruginosa MMA83 planktonic cells, a 5 mg/ml QSI extract (following two-fold serial dilutions) was inoculated with 2 × 10 5 CFU/ml of the test strain using a 96-wells microdilution method. The MIC value was defined as the lowest concentration which inhibited bacterial growth. Controls were incubated with a concentration of DMSO equal to that used in the treatment (0.5% v/v). Number of CFU/ml was determined by plating serial tenfold dilutions after 24 h of incubation at 37 °C. The experiments were performed in sextuplicate and repeated three times. Afterwards, wells were washed to remove planktonic bacterial cells and stained with 0.1% (v/v) crystal violet (HiMedia Labs Pvt. Ltd., India) 58 . Biofilm formation was quantified by recording the absorbance at 595 nm using Plate Reader Infinite 200 pro (MTX Lab Systems, Austria).
The effects of the QSI extract on biofilm formation or decomposition of preformed biofilm were additionally visualized using fluorescent dye SYTO9 (TermoFisher Scientific, Massachusetts, USA) and propidium iodide (PI) (Sigma-Aldrich) 59 . For visualization of biofilms, P. aeruginosa MMA83 (2 × 10 5 CFU/ml) was cocultivated with a 5 mg/ml QSI extract in 24-well plates (Tissue Culture Plate, Sarstedt, Germany) which contained microscopic cover glass. After incubation (24 h, at 37 °C), the cells were washed 3 times with PBS and stained with SYTO9 (2.5 µM) and PI (2.5 µM), green and red fluorescent dyes, respectively. In order to test the decomposition of preformed biofilm in the presence of the QSI extract, P. aeruginosa MMA83 was cultivated for 24 hours, at 37 °C in 24-well plates, washed 3 times with PBS and, subsequently, treated with 5 mg/ml of QSI extract. After next 24 hours (48 hours in total), cells were washed and stained as previously described. were supplemented with two-fold diluted concentrations of the QSI extract (starting with 5 mg/ml, down to 0.038 mg/ml) and then mixed with Elastin-Congo red (Sigma-Aldrich) at a final concentration of 2 mg/ml. After 24 h of incubation at 37 °C, with shaking (180 rpm), the mixtures were centrifuged at 15,700 x g for 15 min, after which the elastase activity was measured at 495 nm using Plate Reader Infinite 200 pro (MTX Lab Systems, Austria) 61 . Supernatant of P. aeruginosa MMA83 culture without the QSI extract treatment was used as a positive control.
Rhamnolipid assay. Rhamnolipid production was examined by acidification of supernatants with HCl (to pH 2) of the previously cultivated test strain (co-incubated with two-fold serial dilutions of the QSI extract or without the QSI extract) 62 . The absorbance was monitored spectrofotometrically at 570 nm by Plate Reader Infinite 200 pro (MTX Lab Systems, Austria).
Pyocyanin assay. P. aeruginosa MMA83 overnight culture (with two-fold dilutions of the QSI extract or without the QSI extract) was centrifuged at 15,700 x g for 15 min and pyocyanin was extracted from the supernatant using chloroform in a 1:2 ratio following the re-extraction of chloroform phase by 0.2 N HCl (3:1 ratio) according to the previously described method 63,64 . The concentration of pyocyanin was evaluated by measuring the absorbance of the red top layer at 520 nm using Plate Reader Infinite 200 pro (MTX Lab Systems, Austria). Concentrations, expressed as micrograms of pyocyanin produced per milliliter of culture supernatant, were determined by multiplying the optical density at 520 nm (OD 520 ) by 17.072.

Quantification of P. aeruginosa MMA83 quorum sensing genes expression by RT-qPCR.
Reverse transcription-quantitative PCR (RT-qPCR) was performed to determine the expression levels of regulatory genes involved in quorum sensing. Primers used for the amplification of selected genes are listed in the Table 2. The total RNA was isolated from the test strain P. aeruginosa MMA83 (grown for 10 hours, at 37 °C) supplemented with the QSI extract (5 mg/ml) or without it by RNeasy Mini Kit (Qiagen, Germany). The total RNA was then treated with DNase using an Ambion DNA-free TM Kit (ThermoFisher, MA, US) and reverse transcribed by a Rever-tAid RT Reverse transcription Kit (ThermoFisher) according to the manufacturer's instructions. Isolation of QSI molecules by TLC. An aliquot (1 mg) of the D. tsuruhatensis 11304 QSI extract was dissolved in 100 µl of ethyl acetate and was applied to TLC (Silica gel 60, F 254 , Aluminium sheets, Merck, Darmstadt, Germany) followed by developing with a mixture of chloroform and ethyl acetate in the ratio 5:4.
Separated spots were scratched off from the plate and the silica gel was put into a cotton-plugged column. The sample was then eluted from the silica gel with ethyl acetate.  www.nature.com/scientificreports www.nature.com/scientificreports/ MALDI MS and MS 2 of the D. tsuruhatensis 11304 QSI extract. The MS structural analysis was performed on an ABSCIEX TOF/TOF TM 5800 Applied Biosystems mass spectrometer equipped with an Nd:YLF laser (λ = 345 nm), with a pulse length of <500 ps and a repetition rate of up to 1000 Hz. The dried sample obtained after TLC separation was redissolved either in acetonitrile, ethyl acetate, or directly in an α-cyano-4-hydroxycinnamic acid matrix solution (5 mg/ml) in acetonitrile/TFA 0.1% (70:30) 32 . The prepared QSI extract (0.5 µl) and matrix solution (0.5 µl), or 1 µl of the mixture were deposited on a stainless-steel plate and left to dry at room temperature. Each spectrum, acquired in positive ion mode, was a result of the accumulation of 2,000 laser shots, whereas 2,500 shots were summed for the MS 2 data acquisitions.
The effect of the commercial C 18 -HSL on the pyocyanin production and quantification of quorum sensing genes expression in P. aeruginosa MMA83. Commercial N-octadecanoylhomoserine lactone (C 18 -HSL) was purchased from Cayman Chemical (Tallin, Estonia) and resuspended in DMSO (0.5 mg/ ml) according to manufacturer instructions. Concentrations of 5, 20 and 50 µM C 18 -HSL were used for evaluation of pyocyanin production and quantification of quorum sensing genes (listed in Table 2) expression in the presence of C 18 -HSL. Each concentration of C 18 -HSL was compared with control containing the exact volume of DMSO without C 18 -HSL to eliminate biological effect of DMSO in the experiment. Pyocyanin productions was expressed in µg/ml as described previously 63,64 . Experiments were done as described above, in triplicate.
Statistical analyses. The statistical analyses and visualization were performed using GraphPad Prism software and SPSS 20.0 for Windows. The results are shown as means ± standard errors. The differences between control and experimental groups were compared using Student's t-test. A p value less than 0.05 was considered to be statistically significant.