Surfing motility is a complex adaptation dependent on the stringent stress response in Pseudomonas aeruginosa LESB58

Cystic fibrosis (CF) is a genetic disease that affects mucin-producing body organs such as the lungs. Characteristic of CF is the production of thick, viscous mucus, containing the glycoprotein mucin, that can lead to progressive airway obstruction. Recently, we demonstrated that the presence of mucin induced a rapid surface adaptation in motile bacteria termed surfing motility, which data presented here indicates is very different from swarming motility. Pseudomonas aeruginosa, the main colonizing pathogen in CF, employs several stress coping mechanisms to survive the highly viscous environment of the CF lung. We used motility-based assays and RNA-Seq to study the stringent stress response in the hypervirulent CF isolate LESB58 (Liverpool Epidemic Strain). Motility experiments revealed that an LESB58 stringent response mutant (ΔrelAΔspoT) was unable to surf. Transcriptional profiling of ΔrelAΔspoT mutant cells from surfing agar plates, when compared to wild-type cells from the surfing edge, revealed 2,584 dysregulated genes. Gene Ontology and KEGG enrichment analysis revealed effects of the stringent response on amino acid, nucleic acid and fatty acid metabolism, TCA cycle and glycolysis, type VI secretion, as well as chemotaxis, cell communication, iron transport, nitrogen metabolic processes and cyclic-di-GMP signalling. Screening of the ordered PA14 transposon library revealed 224 mutants unable to surf and very limited overlap with genes required for swarming. Mutants affecting surfing included two downstream effector genes of the stringent stress response, the copper regulator cueR and the quinolone synthase pqsH. Both the cueR and pqsH cloned genes complemented the surfing deficiency of ΔrelAΔspoT. Our study revealed insights into stringent stress dependency in LESB58 and showed that surfing motility is stringently-controlled via the expression of cueR and pqsH. Downstream factors of the stringent stress response are important to investigate in order to fully understand its ability to colonize and persist in the CF lung.


Introduction
Pseudomonas aeruginosa is a ubiquitous Gram-negative pathogen that causes opportunistic and difficult-to-treat infections. It is strongly associated with chronic debilitating lung infections in individuals with cystic fibrosis (CF) [1]. A hallmark of pathogenic P. aeruginosa is their evolutionary adaption to challenging host conditions such as those found in the CF lung [2]. A particularly aggressive CF isolate is the transmissible Liverpool Epidemic Strain (LES), originally isolated from the sputum of a CF patient [3]. LES isolates are associated with enhanced virulence and greater patient morbidity [4], predominance in intra-species competition [5], stronger persistence in the bronchial lumen [6], increased biofilm formation, decreased motility, and elevated levels of antibiotic resistance [2]. P. aeruginosa possesses several virulence factors that facilitate invasion of the host, compromise immune defences, and ensure its survival and colonization in the host environment. To establish an infection, it has been proposed that Pseudomonas type IV pili are involved in binding to the apical epithelial surface, while flagellar-mediated motility is proposed to be important for the colonization of the basolateral surface of airway epithelium [7]. The type of motility can vary substantially depending on the viscosity and composition of the environment [8]. Recently, Abdullah et al. [9] showed that the reduced water content of the CF airways can trigger the misfolding of the MUC5B mucin protein that further contributes to the thick, viscous gel-like properties of the mucus in the CF lung. At the viscosity encountered in the mucous layer of the CF lung, Pseudomonas likely adopts complex surface growth and motility behaviours, such as swarming and biofilm formation using flagella and type-IV pili [10] or flagella-mediated surfing in the presence of the glycoprotein mucin [11].
P. aeruginosa employs a variety of adaptation mechanisms to cope with environmental stresses [12]. The stringent stress response signaling pathway, mediated by the second messenger molecule guanosine (penta)tetraphosphate (p)ppGpp, is an evolutionarily conserved response that is activated in various cellular stress conditions that restrict bacterial growth (e.g. nutrient or iron limitation). In P. aeruginosa, (p)ppGpp is synthesized through the two enzymes RelA and SpoT. Both enzymes can transfer pyrophosphate from ATP onto GTP to form guanosine pentaphosphate, which is rapidly converted to ppGpp. The accumulation of ppGpp remodels the RNA polymerase that leads to a switch from energetically-costly processes to energy-saving and stress coping mechanisms [13]. The stringent stress response has been previously studied in various Pseudomonas human and plant isolates and shown to be associated with twitching and swarming motility, biofilm formation, hemolytic capacity and abscess formation as well as growth in a murine skin infection model [6,[14][15][16][17][18].
The CF lung is considered a stressful environment due to increased oxidative stress levels [19], chronic activation of the innate immune system [20], and bacterial competition upon antibiotic exposure [12]. Therefore, here we aimed to characterize the role of the stringent stress response in relation to motility of the hypervirulent CF epidemic isolate LESB58. We present evidence that the stringent stress response is required for P. aeruginosa to adapt to environmentally challenging conditions. This stress response was shown to link the complex adaptive motile behaviour, surfing, with the strong involvement of the stringently-controlled copper regulator cueR and quinolone synthase pqsH.

Results
The P. aeruginosa LESB58 stringent response mutant was impaired in surfing We confirmed literature reports [14,15,21] showing that the stringent response was required for swarming, adherence and biofilm formation, as well as pyoverdine and pyocyanin production, in the clinical isolate P. aeruginosa LESB58 (S1 Text, S1 Fig). Vogt et al. [14] demonstrated that ppGpp was not required for swimming in P. aeruginosa PAO1, as indeed we confirmed (S2 Fig), and despite the rather poor swimming motility of strain LESB58, ppGpp did not appear to be required for the weak swimming motility observed for this strain either (S1 Text, S2 Fig). To determine the role of the stringent stress response in other adaptive behaviours, we further investigated surfing motility. Surfing is a conserved form of surface motility that occurs when mucin, a high molecular weight glycoprotein present in mucus, is added to the medium [22]. It exhibits clear differences from swarming surface motility in that it is dependent on flagella but not pili or rhamnolipids (required for P. aeruginosa swarming), and can occur at a wide range of substrate viscosities under both nutrient-limiting or rich conditions (restricted in the case of swarming), and with certain other lubricating agents in place of mucin [11,22,23]. At the molecular level, surfing cells exhibit differential expression of more than a thousand genes, including many regulators, when compared to swimming, swarming and sessile cells [23].
Intriguingly, LESB58 showed excellent surfing capability (Fig 1). In contrast, a double deletion mutant of relA and spoT in the LESB58 background, ΔrelAΔspoT, was defective in surfing.

PLOS PATHOGENS
Surfing motility is stringently controlled in Pseudomonas aeruginosa This could be partially complemented with either the spoT or relA cloned genes, although the spoT complement was much more highly pigmented (Fig 1). There was no basic defect in flagella production due to any of these mutations, since we observed similar swimming zone sizes after 24 h for WT and all mutants in both the strains LESB58 and PAO1 (although the swimming colony morphology for strain LESB58 was somewhat unusual). Furthermore, all mutants appeared to show similar swimming ability when viewed in aqueous suspension using a light microscope. Nevertheless, to investigate if the loss of surfing involved flagella loss, we examined cell morphology of mutant cells on surfing agar plates and compared them to wildtype cells using transmission electron microscopy (TEM). The LESB58 wild-type showed two distinct morphologies with electron-dense thin rods and less dense enlarged cells at the edge and center, while the stringent response mutant only exhibited enlarged cells that were both longer and wider than wild-type cells (S3 Fig). Flagella could only be observed on cells at the centre of surfing plates, but not at the edge for WT, as previously described [11]. We were unable to detect flagella in the spots formed by the ΔrelAΔspoT double mutant (S4 Fig), although the large amount of extracellular material under these circumstances made observations difficult.
Since surface motility adaptations (i.e., swim, swarm, surf) utilize flagella for colony propagation, we examined the expression of some genes involved in flagella synthesis, chemotaxis, and quorum-sensing using qRT-PCR (Table 1). RNA was isolated from P. aeruginosa LESB58 stringent response mutant cells and compared to wild-type cells from the motility edge. Under swarming conditions, downregulation in the ΔrelAΔspoT stringent response mutant was observed for the chemotaxis gene cheY by 3.8-fold, the two-component response regulator of flagella synthesis fleR by 3.5-fold, and the rhamnosyltransferase rhlB (which is involved in the synthesis of rhamnolipid, a surfactant required for swarming) by 5.8-fold. We also observed downregulation of the quorum-sensing regulators lasR and rhlR by 4.3-and 3.5-fold respectively. In contrast under surfing conditions, most of these genes were relatively upregulated or unaltered in expression in the double mutant ( Table 1).

The LESB58 stringent stress response regulated surfing motility
To further understand the surfing deficiency of the stringent response mutant, the transcriptional profile of the LESB58 ΔrelAΔspoT (non-surfing) mutant cells from surfing agar plates was compared to wild-type cells isolated from the edge of the surfing colony. This revealed the

PLOS PATHOGENS
Surfing motility is stringently controlled in Pseudomonas aeruginosa

PLOS PATHOGENS
Surfing motility is stringently controlled in Pseudomonas aeruginosa

Identification of genes required for surfing motility and stringent dependency
To further characterize the genes that might underlie the effects of the stringent response double mutant, a group of genes that were required for surfing motility was determined by screening the ordered PA14 transposon library [24] under surfing conditions. This screen revealed 224 transposon mutants, including 25 regulatory mutants, that were either defective in surfing or had a non-surfing phenotype (S1 Table). Intriguingly there was very little overlap with mutants identified from an analogous screen performed to identify mutants involved in swarming [25].
To identify downstream mediators of the stringent stress response involved in surfing in LESB58, we focused on regulatory genes in the ppGpp-deficient ΔrelAΔspoT mutant. To extend this analysis, qRT-PCR was utilized to examine gene expression of 11 of the regulators for which mutants had a surfing defect ( Table 2). Dysregulation of 9 of these by more than 2-fold was demonstrated and of these, four were demonstrated to be downregulated, including the copper-transport regulator cueR, a positive regulator of copper metabolism [26], that was strongly down regulated by 4.4-fold (Table 2).
Interestingly, we observed upregulation in the stringent response double mutant of the two quorum sensing regulators rhlR and pqsR, which when mutated lead to surfing deficiency [22,23] (S1 Table) and also mediate stress responses [12,27]. Since both are involved in the regulation of Pseudomonas quinolone signal (PQS) synthesis during stress [27], downstream PQS genes were further investigated and it was shown that the pqsABCDE operon, pqsH and phnAB were all downregulated (by -1.6 to -3.5-fold) in the mutant (S5 Fig). The gene expressing FADdependent monooxygenase pqsH, involved in the last step of PQS synthesis, was confirmed by qRT-PCR to be 2.8-fold downregulated in the ΔrelAΔspoT mutant.
To determine the involvement of cueR and pqsH in stringent regulation of surfing motility in strain LESB58, the ΔrelAΔspoT mutant was complemented with each cloned gene. Intriguingly, overexpression of both cueR and pqsH separately restored surfing motility in the surfing-deficient ΔrelAΔspoT mutant (Fig 3).

Discussion
Microorganisms move on moist surfaces to enable them to survey new nearby ecological habitats and favourable environmental conditions or to avoid deleterious situations [28]. In Table 2. Relative fold-changes of LESB58 ΔrelA/ΔspoT mRNA expression compared to wild-type levels of expression. 11 regulators were investigated for which PA14 transposon mutant variants exhibited surfing deficiency (no motility or a different form of surfing). The stringent response double mutant was compared to the edges of wild-type motility zones. All experiments were performed on KB agar plates.

PLOS PATHOGENS
Surfing motility is stringently controlled in Pseudomonas aeruginosa response to specific chemical and physical cues sensed in the environment, microbes can adapt through physiological and morphological differentiation. The stringent stress response is a powerful set of mechanisms that enable bacteria to rapidly adjust to unfavourable environmental conditions and stressful circumstances. Here we demonstrated that the loss of the stringent response in the epidemic P. aeruginosa CF isolate LESB58 inhibited a mucin-induced rapid form of surface motility known as surfing. We investigated in more detail the role of ppGpp in surfing motility; a recently identified novel form of motility that requires the presence of mucin that abounds in the lungs of CF patients, contributes to broad-spectrum antibiotic resistance, and has been shown to be conserved in several bacterial species [11,22]. To investigate the mechanisms underlying surfing impairment in the ΔrelAΔspoT stringent response mutant, the transcriptional profile of this non-surfing mutant was compared to that of wild-type cells isolated from the edge of a surfing colony. Intriguingly, flagella are required for surfing [11] and most flagellar biosynthesis and chemotaxis genes were upregulated in the non-surfing ΔrelAΔspoT mutant. This upregulation could be partially due to protein degradation and/or lack of surface flagella. To further identify a mechanism that might explain surfing deficiency, we screened the comprehensive ordered P. aeruginosa strain PA14 transposon mutant library, which revealed 224 surfing-deficient mutants. Interestingly, of the 233 mutants affecting swarming motility that were revealed in an analogous screen of the same library [25] there were only 28 mutants that overlapped between the two screens (PA066, rhlE, PA0475, epd, aphB, gacS, fliL, hcnC, gacA, pqsH (S9 Fig), PA2685, minD, rhlR, tli5, PA3628, PA3631, PA3749, PA4137, PA4144, PA4168, PA4398, PA4616, cbrAB, cueR (S9 Fig), pstC, cbcV, and  rmd). This together with the limited overlap in genes with altered expression between swarming [8] and surfing edge [23] cells, indicates that these two surface motility adaptations, swarming and surfing, are very different.
Further investigation of the expression, in the LESB58 ΔrelAΔspoT stringent response mutant, of regulatory genes required for surfing revealed that the copper-responsive transcriptional regulator cueR was strongly downregulated. Copper is an essential trace element inside bacterial cells, acting as a reactive center for many enzymatic reactions that are important for numerous vital biological processes including respiration, iron uptake and transport, and superoxide detoxification; however too much copper is toxic [29]. In P. aeruginosa, CueR has been shown to directly bind to and activate five promoters controlling the expression of 11 genes, namely PA3515 to PA3519, PA3520, efflux operon mexPQ-opmE, non-coding RNA

PLOS PATHOGENS
Surfing motility is stringently controlled in Pseudomonas aeruginosa PA3574.1, and a copper homeostasis P-type ATPase cueA (PA3920), the last of which was shown to be a virulence factor in a murine model and involved in dissemination to the liver [30]. Of these, PA3518, PA3519 and PA3520 were all upregulated during surfing [23]. Critically we showed here that when cueR was overexpressed in a ΔrelAΔspoT ppGpp-deficient strain, it restored surfing motility. The copper-responsive regulator cueR is therefore a new downstream effector of the stringent stress response required for surfing, as well as swarming motility [25].
Transcriptomic analysis also revealed a dysregulation of cell communication, which prompted us to look further into quorum-sensing (QS) signalling pathways. Our previous studies revealed the dependence of surfing motility on all three major QS pathways in P. aeruginosa [11,22]. This was also confirmed by our mutant screen since we identified suppressed surfing motility in mutants in lasI, rhlR, rhlI, pqsA-E, pqsR, pqsH, phzA1, and hcnC. Interestingly, the expression of the transcriptional regulator pqsR and the acyl-homoserine-lactone responsive regulator rhlR was 2-fold upregulated in the ΔrelAΔspoT mutant, but other QSdepended genes, including the pqsABCDE operon that includes the quinolone signal response protein gene pqsE, pqsH, phnAB, as well as QS-regulated virulence factors lasAB, aprI, aprA/ DEFX, phzA1/A2/H/M/S, hcnABC, and ambA/CDE were all downregulated in the stringent response mutant, by around 2.5-22-fold.
The Pseudomonas quinolone signal (PQS, 2-heptyl-3-hydroxy-4-quinolone) pathway depends on PqsR [31], which regulates the expression of the pqsA-E operon once activated by 2-heptyl-4-quinolone (HHQ). Further expression of pqsE upregulates virulence factors (e.g., phz, hcn) [31], which is in accordance with our study where we observed a downregulation of all those genes. PQS and pqsE have both been suggested to be required for Lectin A production in P. aeruginosa [32], however under surfing conditions lectin production appears to depend exclusively on the rhl QS system [12] as we found a 7-fold upregulation of lecA in the ΔrelAΔ-spoT. Together with the upregulation of the rhl system we found that two out of three AHLacylases, hacB and pvdQ, both hydrolyse long-chain AHLs to prevent excessive signal build-up [33], were downregulated in the stringent response mutant by 1.7-and 3-fold, respectively. Schafhauser et al. [34] and Nguyen et al. [35] reported that the stringent response negatively regulates 4-hydroxy-2-alkylquinolines (HAQs) by reducing the expression of the rhl system, which leads to increased expression of the HAQ biosynthetic gene pqsA, the monooxygenase pqsH, and the positive regulator pqsR. Here, we observed an opposite regulation under surfing conditions, where the rhl system was upregulated and HAQ-associated genes were downregulated.
We further focused more closely on the effector functions (signalling molecules) and investigated the quinolone synthase pqsH gene that encodes the enzyme involved in the last step of synthesis to produce PQS from HHQ. Overexpression of this gene complemented the surfing deficiency of the ΔrelAΔspoT stringent stress response mutant. Our previous studies [22] demonstrated that a mutation in the pqs operon was surfing deficient and this could be complemented by adding exogenous PQS signalling molecule. This then indicates that the loss of signalling through PQS is one reason why the ppGpp-deficient double mutant ΔrelAΔspoT is surfing deficient. Thus, our data supports the conclusion that quorum sensing is stringentlycontrolled to influence surfing motility in LESB58.
Another intriguing feature of the data provided here is the apparent importance of cyclic di-GMP in regulating surfing. We identified four enzymes mediating the regulation of this messenger as being required for surfing motility including the diguanylate cyclases TpbB and SadC and phosphodiesterase FimX and DipA as well as one cyclic-di-GMP binding regulator FleQ (S1 Table). These genes separately regulate other complex adaptive processes including virulence, biofilm formation/dispersion and persistence, a form of adaptive resistance to

PLOS PATHOGENS
Surfing motility is stringently controlled in Pseudomonas aeruginosa antimicrobials, thus, contributing to a relatively large regulon [36][37][38]. Accordingly, several genes involved in the cyclic di-GMP pathway were found to be dysregulated in the stringent response mutant, including dipA which was shown to be essential in mediating surfing motility. Thus, the stringent response and surfing motility in general have a very substantial influence on second messenger activity in Pseudomonas.
In summary, we demonstrated that the ΔrelAΔspoT stringent stress response mutant of the P. aeruginosa clinical CF isolate LESB58 had pronounced deficiencies in adapting to environmental stress conditions, suggesting a potential role of the stringent stress response during pathogenesis. We identified two novel downstream effector genes, cueR and pqsH, both involved in stringent response regulation of surfing motility, and believe it is important to further identify other downstream factors of the stringent stress response, to fully understand its ability to colonize and persist in the CF lung.

Bacterial strains, media, and growth conditions
Bacterial strains used in this study are listed in Table 3 and primers in S2 Table. Pseudomonas strains were cultured at 37˚C in LB (Difco), 2xYT (Sigma), King's B medium (KB), or synthetic cystic fibrosis medium (SCFM) [39]. Bacterial growth was monitored at an optical density at 600 nm (OD 600 ) using either a spectrophotometer or a 96-well microtiter plate reader (Synergy H1; BioTek). Cultures harboring individual plasmids were maintained by supplementation with 15 μg/ml gentamicin (Gm) for E. coli or 500 μg/ml Gm for P. aeruginosa LESB58.

Swarming motility assay
Swarming motility of P. aeruginosa strain LESB58 was examined as previously described [40], with slight modifications. Strains were scraped from overnight grown plates and suspended in sterile demineralized water to an OD 600 of 0.1. Ten μl of a bacterial cell suspension was applied onto KB plates containing 0.35% agar. Plates were then incubated at 37˚C for 48 h. Experiments were repeated at least three times with three replicates for each strain.

Swimming and surfing motility assay
Swimming motility of P. aeruginosa LESB58 strains was examined on KB and SCFM plates containing 0.3% agar, while surfing motility was investigated on the same plates supplemented with 0.4% mucin from porcine stomach (Type II; Sigma-Aldrich). Briefly, P. aeruginosa strains Table 3. Pseudomonas aeruginosa strains used in this study.

LESB58
Liverpool Epidemic Strain isolate [3] LESB58 ΔrelAΔspoT Double deletion mutant of relA and spoT [18] LESB58 ΔrelAΔspoT/relA + Double deletion mutant of relA and spoT chromosomally complemented with the native relA gene including its promoter region; gentamicin r [18] LESB58 ΔrelAΔspoT/ spoT + Double deletion mutant of relA and spoT chromosomally complemented with the native spoT gene including the rpoZ and the rpoZ-spoT promoter region; gentamicin r [18] LESB58

PLOS PATHOGENS
Surfing motility is stringently controlled in Pseudomonas aeruginosa were grown to mid-log phase (OD 600 = 0.5), and subsequently spotted (1 μl) onto the respective agar plate, further incubated at 37˚C, and motility zones visually inspected after 15-48 h. The experiments were performed at least three times. The P. aeruginosa PA14 transposon library [24] was screened on SCFM plates containing 0.3% agar and 0.4% mucin. Surfing deficiency was identified either as an inability to spread past the point of inoculation, >70% reduced surface spreading, or a switch to in agar growth/motility which is characteristic of swimming rather than surfing which occurs on the surface.

Growth curves experiments and pyoverdine production
LESB58 strains were grown overnight (16-18 h) at 37˚C with shaking (250 rpm). Bacteria were pelleted (5000 g, 3 min) and suspended in the respective growth media: KB, dYT, MHB, or SCFM to an OD 600 of 0.1. Then 200 μl was transferred to a flat bottom 96-well polystyrene microtiter plate (Corning) and incubated at 37˚C with continuous fast linear shaking at 567 cycles per minute (cpm) in a microplate reader (Synergy H1; BioTek). OD 600 and fluorescence [assessing pyoverdine production [41] at an excitation wavelength of 400 nm and emission wavelength of 460 nm] readings were taken every hour over a 24-hour period. Experiments were performed three times with at least three technical replicates.

Pyocyanin measurement
Overnight cultures of the LESB58 wild-type, ΔrelAΔspoT stringent response mutant, and its according relA and spoT complemented strains, were grown in LB medium, washed in SCFM and resuspended in SCFM at an adjusted OD 600 of 0.1 and further cultivated at 37˚C with aeration (250 rpm) for 22 h. Pyocyanin was extracted from filter-sterilized supernatants and measured as previously described [42]. Briefly, 2.88 ml chloroform was added to 4.8 ml of culture supernatant and vortexed 10 times for 2 seconds. After centrifugation (14000 rpm, 8 min), 2.4 ml of the chloroform layer was transferred to a fresh tube and 1.2 ml of 0.2 N HCl was added and the mixture vortexed 10 times for 2 seconds. After centrifugation (14000 rpm, 2 min), 1 ml of the top layer (containing pyocyanin) was removed and its absorbance measured at 520 nm (OD 520 ). The OD 520 values were multiplied by the extinction coefficient of 17.072 [42] to obtain the pyocyanin concentrations (μg/ml) of the bacterial supernatants. Experiments were performed three times with at least two technical replicates.

Adherence experiments
Strains were streaked onto KB agar plates and grown overnight at 37˚C. Bacteria were scraped from the plates and resuspended in KB medium to an OD 600 of 0.5. One hundred μl of a bacterial suspension was added into polystyrene round bottom 96-well microtiter plates (Falcon) and incubated at room temperature for 1 h. Then each well was washed three times with water, and adhered cells subsequently stained by adding 105 μl of 0.03% crystal violet. Staining was performed on a table top shaker (100 rpm) for 20 minutes at room temperature. Next, plates were washed three times with water and adhered crystal violet dissolved in 110 μl 70% ethanol at room temperature for at least 20 minutes. Absorbance was measured at 595 nm with a BioTek plate reader. Data analysis was performed to calculate the mean and standard deviation, after removal of outliers that were more than one standard deviation from the mean. Data was further normalized to the wild type. Experiments were performed at least three times with up to six technical replicates.

PLOS PATHOGENS
Surfing motility is stringently controlled in Pseudomonas aeruginosa

Biofilm formation under flow conditions and flow-cell imaging
Overnight grown cultures of the wild-type, ΔrelAΔspoT mutant, and complemented strains were cultivated for biofilm formation under flow conditions in chambers as described previously [43]. Briefly, flow-cells were inoculated with P. aeruginosa LESB58 strains at an OD 600 of approximately 0.005 in dYT broth and kept under static conditions for 3 h to enable adherence to the glass chambers after which continuous flow was applied for 3 days at 37˚C. For imaging, the cells were stained with 1 μM SYTO 9, and images captured using a confocal laser scanning microscope (Zeiss, LSM 800) and analyzed using the Zeiss Zen software (v2.3, blue edition). Experiments were performed at least twice.

Antibiotic susceptibility
The MIC of antibiotics for P. aeruginosa LESB58 and stringent response mutants was determined at 37˚C in Mueller-Hinton broth (MHB; Difco) using the broth microdilution assay in 96-well plates [44]. All tests were performed at least three times following the Clinical and Laboratory Standards Institute recommendations. Bacterial growth was examined by visual inspection after 24-48 h of incubation. The MIC was defined as the lowest concentration of a compound that completely prevented visible cell growth.

Construction of cueR and pqsH overexpression plasmids
A 501-bp fragment containing the cueR gene including its upstream promoter region was PCR amplified using the primers cueR_oe_fwd-Kpn/cueR_oe_rev-Hin. An 1,834-bp fragment containing the pqsH gene including its upstream promoter region was PCR amplified using the primers pqsH_oe_fwd-Apa/pqsH_oe_rev-Hin. Each PCR fragment was gel purified and cloned into KpnI/HindIII or ApaI/HindIII restriction sites of plasmid pBBR1MCS-5 [45], yielding pBBR5.cueR and pBBR5.pqsH with additional ability to be expressed from the lac promoter. All constructs were sequenced before transformation into P. aeruginosa LESB58 ΔrelAΔspoT mutant as previously described [40].

RNA isolation, quantitative real-time (qRT)-PCR, and RNA-Seq
RNA was isolated from wild-type cells (edge of surfing colony) and the stringent response double mutant (total colony) as previously described [18,23]. Bacterial cells were collected (six biological replicates) and resuspended in a sterile water mixture with RNAprotect Bacteria Reagent (QIAGEN), isolated using the RNeasy Mini Kit (QIAGEN), and the RNA obtained was DNAse-treated (Ambion/Life Technologies). RNA was quantified using the Synergy H1 microplate reader (BioTek) and RNA integrity determined using the Agilent Bioanalyzer. Three biological replicates (S10 Fig) with the highest RIN score were selected for further analysis. Quantitative Real-Time (qRT-)PCR was performed as described previously [18] at least three times independently using rpoD for normalizing transcript levels.
Ribosomal RNA was depleted using the RiboZero Bacteria Kit (Illumina), and cDNA libraries were constructed using the KAPA Stranded Total RNA Kit (KAPA Biosystems). Sequencing was performed by the University of British Columbia Sequencing Consortium using an Illumina HiSeq-2500, generating single end reads (1×100 bp). The read quality of the sequencing samples was checked using FastQC v0.11.6 [46] and MultiQC v1.6 [47]. Alignment of transcriptomic reads to the LESB58 reference genome (obtained from the Pseudomonas Genome Database, ww.pseudomonas.com) was performed using STAR v2.6 [48]. Counts were generated using HTSeq v0.11.2 [49]. Differentially expressed (DE) genes between the double mutant

Functional enrichment of DE genes
Enrichment of GO terms were performed using GofuncR, testing the DE genes against a custom set of GO annotations downloaded from the Pseudomonas Genome Database. The full list of 2,584 DE genes was split into up and down regulated, with GO enrichment being performed independently on each of these sets. Results were filtered using a significance threshold of family-wise error rate (FWER) � 0.1. Enrichment of KEGG Pathways was done using Gage v2.3.0 on the full list of 2,584 DE genes. Results were filtered for significance based on qvalue � 0.2.
Enrichment of cellular functions, based on manually curated lists was performed on the full list of 2,584 DE genes using Fisher's Exact Test, implemented via a custom script in R v3.6.0 (R Core Team, 2019). Multiple test correction was performed using the Benjamin-Hochberg method and filtered on a significance of � 0.05.

Transmission electron microscopy (TEM)
Bacterial cells were picked from SCFM surfing plates and re-suspended in 10 μl dH 2 O. Formvar/carbon TEM grids (200 mesh, copper; Ted Pella Inc.) were placed on top of the suspension for 30 s to allow for cell adherence. Excess liquid was removed using filter paper and grids were subsequently stained with 5 μl of 2% aqueous uranyl acetate for 30 s and then washed for 5 s in 10 μl dH 2 O. Images from multiple grid sections were taken with a Hitachi H-7600 transmission electron microscope (UBC Bioimaging facility).

Statistical analysis
Statistical evaluations were performed using GraphPad Prism 7.0 (GraphPad Software, La Jolla). P-values were calculated using one-way ANOVA, Kruskal-Wallis multiple-comparison test followed by the Dunn procedure. Data was considered significant when p-values were below 0.05.
Supporting information S1 Text. The P. aeruginosa LESB58 stringent response mutant was impaired in swimming, swarming, adherence, and biofilm formation, as well as being more susceptible to antibiotics. (DOCX) S1

PLOS PATHOGENS
Surfing motility is stringently controlled in Pseudomonas aeruginosa the wild-type absorbance. � indicates p-value < 0.05 compared to wild-type. Experiments were performed at least three times. Error bars indicate ± standard error. D) Biofilm formation under flow-cell conditions in dYT broth. Cells were stained after three days for one hour with 1 μM SYTO-9 and subsequently imaged using a Zeiss LSM800 confocal microscope. E) Pyoverdine production (fluorescence: excitation 400 nm; emission 460 nm) after 20 h incubation in KB broth in a 96-well plate under shaking conditions (567 cpm). F) Pyocyanin production after 22 h incubation in SCFM broth under shaking conditions (220 rpm). E, F) Analysis was performed using One-way ANOVA with Dunn correction. �� , indicate p-value < 0.01 compared to wild-type. D-F) Experiments were performed 2-3 times. Error bars indicate ± standard deviation.