The stringent response promotes biofilm dispersal in Pseudomonas putida

Biofilm dispersal is a genetically programmed response enabling bacterial cells to exit the biofilm in response to particular physiological or environmental conditions. In Pseudomonas putida biofilms, nutrient starvation triggers c-di-GMP hydrolysis by phosphodiesterase BifA, releasing inhibition of protease LapG by the c-di-GMP effector protein LapD, and resulting in proteolysis of the adhesin LapA and the subsequent release of biofilm cells. Here we demonstrate that the stringent response, a ubiquitous bacterial stress response, is accountable for relaying the nutrient stress signal to the biofilm dispersal machinery. Mutants lacking elements of the stringent response – (p)ppGpp sythetases [RelA and SpoT] and/or DksA – were defective in biofilm dispersal. Ectopic (p)ppGpp synthesis restored biofilm dispersal in a ∆relA ∆spoT mutant. In vivo gene expression analysis showed that (p)ppGpp positively regulates transcription of bifA, and negatively regulates transcription of lapA and the lapBC, and lapE operons, encoding a LapA-specific secretion system. Further in vivo and in vitro characterization revealed that the PbifA promoter is dependent on the flagellar σ factor FliA, and positively regulated by ppGpp and DksA. Our results indicate that the stringent response stimulates biofilm dispersal under nutrient limitation by coordinately promoting LapA proteolysis and preventing de novo LapA synthesis and secretion.


Results
Stringent response mutants are defective in starvation-induced dispersal. We previously showed 33 that a P. putida mutant bearing a transposon insertion in dksA showed a delayed dispersal response in biofilms formed on microtiter plate wells. P. putida dksA (PP_4693) is located upstream and in the same orientation to PP_4694, encoding glutamyl-Q tRNA(Asp) synthetase. Operon prediction by the DOOR algorythm 34 , as well as the lack of transcription start sites mapped by differential RNA sequencing upstream from PP_4694 35 strongly suggest that both genes are co-transcribed. Thus, to prevent polar effects on the downstream gene, we used allele replacement to construct MRB46, a derivative of the reference strain P. putida KT2440 bearing an unmarked in-frame deletion of dksA. In addition, we also generated the same ∆dksA mutation in PP1922, a KT2440 derivative unable to synthesize (p)ppGpp (henceforth designated ppGpp 0 ), due to deletions of the (p) ppGpp sythetase-encoding genes relA and spoT 32 . To analyse the ability of this mutant set to form and disperse biofilms in a microtiter dish setting, we performed dilution series-based growth curves, a method in which a dilution series is used to recapitulate the time-course of planktonic and biofilm growth 33 . To this end, the wildtype (KT2440), ∆dksA (MRB46), ∆relA (PP1437), ppGpp 0 (PP1922) and ∆dksA ppGpp 0 (MRB53) strains were serially diluted in LB, inoculated in microtiter plate wells and planktonic and biofilm growth was monitored after 20 hour incubation (Fig. 1).
Consistent with the behaviour of the insertion mutant, the ∆dksA strain showed somewhat slower planktonic and biofilm growth relative to the wild-type (Fig. 1a). In addition, this mutant failed to disperse the biofilm at the onset of stationary phase. Instead, biofilm biomass was retained until well into stationary phase and then a delayed dispersal response was observed. The ppGpp 0 and ∆dksA ppGpp 0 mutants exhibited similar behaviour: after a prolonged lag phase, planktonic and biofilm growth occurred similarly to that of the wild-type, but the biofilm was not dispersed at the onset of stationary phase (Fig. 1b,c). Rather, biofilm biomass accumulated to peak at levels 2-to 3-fold greater than those in the wild-type and, even though biofilm levels decreased somewhat during late stationary phase, a full dispersal response did not occur. Extended incubation of the ∆dksA ppGpp 0 mutant dilution set to 26 hours resulted in a similar increase in biofilm accumulation, while dispersal was still not observed ( Supplementary Fig. S1). These results strongly suggest that the lack of the components of the stringent response is associated with a permanent defect in biofilm dispersal. Interestingly, a ∆relA mutant (PP1437) showed a phenotype indistinguishable from that of the wild-type (Fig. 1d), suggesting that the (p)ppGpp synthesis activity of SpoT is sufficient to support a normal biofilm growth cycle under our experimental conditions.
A similar set of dilution-based growth curves was performed in K10T-1, a medium containing tryptone and glycerol that is commonly used for biofilm studies in Pseudomonas fluorescens 36 (Supplementary Fig. S2). The general pattern of planktonic growth and biofilm accumulation and dispersal was similar to that observed in LB, but the dispersal defect of the ∆dksA mutant was more severe, as biofilm biomass was not diminished in late stationary phase. Complementation assays confirmed that the lack of DksA is responsible for these phenotypes, as insertion of a miniTn7BB-Gm derivative expressing dksA from its own promoter rescued both the growth and biofilm dispersal SCIeNtIfIC RepoRTs | (2017) 7:18055 | DOI:10.1038/s41598-017-18518-0 defects of the ∆dksA strain ( Supplementary Fig. S3). Taken together, our results strongly suggest that an active stringent response involving DksA and (p)ppGpp synthesis is required for starvation-induced biofilm dispersal.
The ability of our mutant set to form a biofilm at the medium-air interphase (pellicle) was also assessed using shaking culture tubes containing K10T-1 medium (Fig. 2). While the wild-type strain showed little or no pellicle biomass, the KT2442-derived ∆bifA strain MRB32 (used here as a positive control) displayed a thick pellicle, as previously described 23 . The ∆relA and ∆dksA mutants displayed behaviour similar to that of the wild-type, while Planktonic (left axes, open symbols) or biofilm growth (right axes, closed symbols) is plotted against the initial A 600 of each dilution. Blue circles represent the wild-type KT2440 strain and red squares represent the ∆dksA mutant MRB46 (a), the ppGpp 0 mutant PP1922 (b), the ∆dksA ppGpp 0 mutant MRB53 (c), or the ∆relA mutant PP1437 (d) Plates were incubated for 20 hours prior to measurement. Plots display one representative experiment of at least three biological replicates. Error bars represent the standard deviation of the six technical replicates. the ppGpp 0 and ∆dksA ppGpp 0 mutants showed pellicle levels comparable to those in the ∆bifA mutant. The correlation between the inability to synthesize (p)ppGpp, high levels of biofilm and pellicle formation suggests that (p)ppGpp is a negative regulator of cell adhesion and biofilm formation. Ectopic (p)ppGpp synthesis restores biofilm dispersal in the ppGpp 0 mutant. To further characterize the role of (p)ppGpp in the biofilm growth cycle, we determined the effect of ectopic (p)ppGpp synthesis on planktonic and biofilm growth of the wild-type and ppGpp 0 strains. To this end, plasmids pMRB160 and pMRB162 were constructed to produce the C-terminally truncated derivatives of the Escherichia coli RelA protein RelA ∆456-743 , displaying constitutive (p)ppGpp synthetic activity, and RelA ∆332-743 , which is inactive 37,38 , from the nahR-Psal expression system 39 . The expressing plasmids were transferred to the wild-type KT2440 and ppGpp 0 PP1922 strains and biofilm and planktonic growth was assessed by means of serial dilution-based growth curves. Even though the Psal promoter is inducible by salicylate, induction of RelA ∆456-743 synthesis was inhibitory to growth, and therefore assays were only performed in the absence of salicylate (Fig. 3).
The planktonic and biofilm growth curves of the wild-type strain bearing the RelA ∆456-743 and RelA ∆332-743 -producing plasmids (Fig. 3a) were similar to that observed in the absence of plasmid (Fig. 1a-d). When tested in the ppGpp 0 background, the inactive RelA ∆332-743 -producing strain also showed a behaviour similar to that of the plasmid-free ppGpp 0 strain (Fig. 1b). In contrast, production of the constitutive RelA ∆456-743 derivative rescued both the planktonic growth and biofilm dispersal defects associated to the lack of (p)ppGpp (Fig. 3b), suggesting that basal transcription from Psal provides sufficient amount of RelA ∆456-743 to promote biofilm dispersal, while not being detrimental to growth. It could be argued that ectopic (p)ppGpp production may promote non-specific biofilm detachment by mechanisms unrelated to the starvation-induced biofilm dispersal signalling pathway. To test this hypothesis, the effect of RelA ∆456-743 and RelA ∆332-743 production was also tested in pMRB160 and pMRB162-bearing P. putida strain MRB1 33 . This mutant lacks the periplasmic protease LapG, the effector element of starvation-induced dispersal response, required for LapA cleavage. Although production of the constitutive RelA ∆456-743 derivative provoked a considerable delay in planktonic growth of MRB1 (Fig. 3c), both plasmid-bearing strains displayed high levels of biofilm that failed to reproduce the dispersal response upon entry in stationary phase, much like our previous observations with plasmid-free MRB1 33 . These results are consistent with the hypothesis that (p)ppGpp induces a dispersal response functionally similar to the previously characterized starvation-induced dispersal response.
The flagellar σ factor FliA and the stringent response coordinately regulate bifA transcription. Our recent work revealed that BifA is the PDE responsible for c-di-GMP depletion during biofilm dispersal. On the other hand, our results above establish that presence of the elements of the SR is required for biofilm dispersal. In P. putida, bifA transcription is dependent on the flagellar σ-factor FliA 40 . In addition, it has been shown that P. putida RNAP utilizing FliA is co-ordinately regulated by DksA and (p)ppGpp to elevate the expression of the aer-2 gene, encoding an oxygen-and metabolism-sensor protein 41 . Therefore, we reasoned that, similarly to aer-2, bifA transcription may be subjected to dual regulation by FliA and the SR. To test this possibility, plasmid pMRB68, harbouring a PbifA-gfp-lacZ fusion was transferred to KT2440 and its SR and fliA − mutant derivatives and expression was monitored across the growth curve by means of β-galactosidase assays (Fig. 4).
Expression of the PbifA promoter in the wild-type strain increased 3-fold as the culture progressed from exponential to stationary phase. A similar behaviour was observed in the ∆dksA mutant, suggesting that DksA does not greatly influence bifA expression in these conditions. In contrast, low, basal expression levels were observed in the fliA − , ppGpp 0 and ∆dksA ppGpp 0 mutants. These results confirm the FliA-dependence of PbifA transcription and indicate that (p)ppGpp synthesis is required for high level expression of bifA.
To determine whether the effect of FliA and the SR on bifA is exerted at the level of PbifA transcription initiation, single-round in vitro transcription assays were performed. These assays used plasmid pVI2407, containing the PbifA promoter, as a template, purified P. putida core RNA polymerase, FliA, and DksA, and ppGpp. As a control, similar reactions were performed using plasmid pVI1011, which bears the previously analysed FliA-, DksA-and ppGpp-regulated P. putida Paer2 promoter 41 .
In the presence of FliA-RNAP but in the absence of ppGpp and DksA the levels of PbifA transcript were low. Transcript levels were increased in the presence of DksA, to reach a maximum 6-fold stimulation at 4 µM DksA (Fig. 5a). While ppGpp did not substantially affect PbifA transcription in the absence of DksA (Figs 5c and S4a), simultaneous addition of ppGpp and DksA resulted in the synergistic stimulation of PbifA: while 1 µM DksA alone provoked a 3-fold increase in transcription relative to the condition with no protein or effector added, stimulation was increased up to 5-to 9-fold when 1 µM DksA was combined with 100, 200, 400 or 600 µM ppGpp (Figs 5c and S4a). Conversely, 100-600 µM ppGpp did not stimulate transcription noticeably in the absence of DksA, but provoked a 2-to 3-fold increase in the transcript levels in the presence of 1 µM DksA relative to those obtained with 1 µM DksA in the absence of ppGpp. Consistent with previous results 41 , only a comparatively modest regulation of the Paer2 promoter, amounting to a maximum 2-fold stimulation by DksA alone (Fig. 5b) and a 3-fold combined effect of 1 µM DksA plus 200-600 µM ppGpp (Figs 5d and S4b), was observed. Taken together these results confirm that the PbifA promoter is FliA-dependent and performance of FliA-RNAP at this promoter is directly stimulated by ppGpp and DksA in vitro. Furthermore, the high extent of the stimulation of the PbifA promoter indicates that the global effectors of the SR -ppGpp and DksA -can strongly influence transcription by FliA-RNAP.
(p)ppGpp negatively regulates LapA synthesis and secretion. In addition to BifA, other elements involved in the dispersal response are LapD, a c-di-GMP sensor that regulates the proteolytic activity of the LapG, and LapA, a high molecular weight outer membrane adhesin. On the other hand, LapA localization at the outer membrane is strictly dependent on a ABC-type secretion system, comprised of LapB, LapC and LapE 42 . In Planktonic (left axes, open symbols) or biofilm growth (right axes, closed symbols) is plotted against the initial A 600 of each dilution. Panel (a): blue circles and red squares represent the wild-type strain KT2440 bearing the constitutive RelA ∆456-743 -producing plasmid pMRB160 and the inactive RelA ∆332-743 -producing plasmid pMRB162, respectively. Panel (b): blue circles and red squares represent the ppGpp 0 strain PP1922 bearing the constitutive RelA ∆456-743 -producing plasmid pMRB160 and the inactive RelA ∆332-743 -producing plasmid pMRB162, respectively. Panels (c): blue circles and red squares represent the lapG − strain MRB1 bearing the constitutive RelA ∆456-743 -producing plasmid pMRB160 and the inactive RelA ∆332-743 -producing plasmid pMRB162, respectively. The growth medium was LB. Plates were incubated for 20 hours prior to measurement. Plots display one representative experiment of at least three biological replicates. Error bars represent the standard deviation of the six technical replicates.
SCIeNtIfIC RepoRTs | (2017) 7:18055 | DOI:10.1038/s41598-017-18518-0 order to identify possible additional biofilm dispersal-related targets of SR response signaling, plasmids pMRB67, pMRB66, pMRB240 and pMRB241, bearing transcriptional gfp-lacZ fusions to the PlapA and PlapBC, PlapE, and PlapGD promoters, respectively, were introduced into the wild-type KT2440 and the ppGpp 0 mutant PP1922, and expression was assessed by online monitoring of GFP fluorescence during the growth curve. Differential rates of GFP fluorescence accumulation during exponential phase were calculated from the slopes of the fluorescence vs. absorbance plots (Fig. 6a-d). Expression from the PlapA, PlapBC and PlapE promoters was increased 4-, 2-and 3-fold respectively in the ppGpp 0 relative to the wild-type strain, while, PlapGD expression was not significantly altered in these conditions (Fig. 6e). These results indicate that, in addition to stimulating the synthesis of the c-di-GMP-depleting PDE BifA, (p)ppGpp acts directly or indirectly by diminishing the synthesis and secretion of the high molecular weight biofilm adhesin LapA.

Discussion
Programmed biofilm dispersal is a mechanism described in multiple bacteria that enables them to actively exit the biofilm structure and resume a planktonic lifestyle in response to environmental and/or physiological cues. Here we describe the involvement of the components of the SR, the nucleotide alarmone (p)ppGpp and the protein DksA, in the induction of biofilm dispersal in the Gram-negative soil bacterium P. putida KT2440. The SR and (p) ppGpp have previously shown to influence biofilm development in multiple bacterial species, including Listeria monocytogenes, Streptococcus mutans, Enterococcus faecalis, Vibrio cholerae and uropathogenic E. coli 29,[43][44][45][46][47] . In contrast to our findings, induction of the SR stimulated biofilm formation in these organisms, and this positive effect has been used as an argument for the development of SR inhibitors as therapeutic agents against biofilm infections [48][49][50] . However, our results indicate that this is not always the case and, depending on the bacterial species, the result of treatment with SR antagonists may be exactly the opposite -i.e., stimulation of biofilm formation. Interestingly, (p)ppGpp is a negative regulator of biofilm growth in Francisella novicida 51,52 , although the mechanisms involved have not been explored further.
Our phenotypic analysis of ∆dksA, ppGpp 0 and ∆dksA ppGpp 0 strains revealed defects in starvation-induced biofilm dispersal. Strains unable to synthesize (p)ppGpp (ppGpp 0 and ∆dksA ppGpp 0 ) were strongly impaired in biofilm dispersal, at least within the 20-26 h time frame of our experiments ( Fig. 1; Supplementary Figs S1 and S2), and displayed elevated biofilm levels, both in a microtiter plate submerged biofilm model ( Fig. 1; Supplementary Figs S1 and S2), and in a medium-air interphase (Fig. 2) setting. Similarly elevated biofilm levels have also been observed in other P. putida mutants lacking elements involved in biofilm dispersal, such as LapG 7,33 , or BifA 24,33 . In contrast to the (p)ppGpp 0 strain, a wild-type dispersal response was observed in the ∆relA mutant (Fig. 1), indicating that SpoT-dependent (p)ppGpp synthesis is sufficient to trigger dispersal under our culture conditions. An analogous situation has been reported for Salmonella virulence in mice, which is attenuated in a ppGpp 0 , but not in a ∆relA mutant [53][54][55] . On the other hand, the defects observed in the ∆dksA strain were comparatively small relative to those observed in the ppGpp 0 and ∆dksA ppGpp 0 strains -biofilm dispersal was delayed but not abolished, and biofilm overgrowth was not observed (Figs 1 and 2; Supplementary Figs S1 and S2). This relatively mild in vivo phenotype, and the fact that dksA inactivation did not have an additive effect in a ppGpp 0 background may be attributed to only a partial dependence of the dispersal response on DksA, or alternatively, to the presence of redundant functions that may partially substitute for DksA in its absence (also see below).
We traced the effect of the stringent response on the starvation-induced dispersal response to the transcriptional control of bifA, which encodes a c-di-GMP phosphodiesterase responsible for the drop in c-di-GMP concentration that triggers biofilm dispersal in P. putida 24 , and lapA, lapBC and lapE, encoding the high molecular weight adhesin and the LapA-specific secretion system 42 . During revision of the present manuscript, a paper showing the involvement of (p)ppGpp in the regulation of biofilm formation in P. putida KT2440 was published 56 . These authors identified biofilm-related targets for positive (lapA and the exopolysaccharide synthesis clusters peb and bcs) and negative regulation (lapF, encoding a second adhesin and the exopolysaccharide synthesis cluster pea). These results are complementary to our own observations, and reinforce the notion that the stringent response is an integral component of the regulatory network of the biofilm cycle, acting directly or indirectly on multiple factors.
Further characterization of the PbifA promoter revealed that bifA transcription is subjected to dual regulation in vivo, by the flagellar σ factor FliA, as previously described 40 , and by (p)ppGpp (Fig. 4). Furthermore, single-round in vitro transcription showed that the PbifA promoter is regulated synergistically by ppGpp and DksA (Fig. 5), thus substantiating the notion that bifA is a target for activation through the SR. Despite the clear in vitro evidence for the involvement of DksA in stimulation of FliA-dependent transcription from PbifA, the absorbance plots of the data collected from one representative replicate of the assays performed with the PlapA (a), PlapBC (b), PlapE (c) and PlapGD (d) promoter fusions. Blue circles denote the wild-type strain KT2440 and red squares denote de ppGpp 0 strain PP1922. Highlighted data points were used for linear regression, which is shown as a hatched black line. R-squared values were always greater than 0.98. (e). Differential rates of GFP accumulation. Columns and error bars represent the averages and standard deviations obtained from from at least three independent biological replicates. Stars denote statistical significance assessed by two-tailed T-tests not assuming equal variance (*p < 0.05; **p < 0.01; ***p < 0.001). effect of dksA inactivation on PbifA stimulation in vivo was negligible (Fig. 4). We can envisage two plausible explanations for this apparent contradiction between the in vivo and in vitro data, both of which involve functional replacement by another family member. Firstly, the P. putida KT2440 genome contains two additional dksA paralogs: PP_2220, which encodes a protein with 68% similarity and 45% identity to P. aeruginosa DksA2, and PP_3037. Given that in P. aeruginosa, DksA2 is able to functionally replace DksA 57 , it is possible that the P. putida PP2220 and/or PP3037 proteins may simply take over the role of DksA. Secondly, DksA belongs to a family bacterial transcription factors that includes the GreA and GreB proteins amongst others, all of which bind and penetrate the secondary channel of RNAP to access the active site. In E. coli, partial redundancy and competition between GreA, GreB and DksA has been thoroughly documented and interplay between these factors can lead to discrepancies between in vivo and in vitro effects 44,[58][59][60] . Solving which of these alternative possibilities is the case for PbifA is the subject of ongoing investigations.
Based on the results presented here and previous work by others and us, we propose a regulatory circuit for starvation-induced biofilm dispersal in P. putida as depicted in Fig. 7. This circuit connects nutrient stress with its ultimate outcome, namely LapA proteolysis that allows the bacterium to exit the biofilm and resume a planktonic lifestyle. In this model, nutrient limitation is signaled via the stringent response RelA and/or SpoT (p)ppGpp synthetases. Although we do not currently know the identity of the signal, our results indicate that SpoT, but not RelA, is required for this response; nevertheless, RelA may also be involved under a different growth regimes. The resulting elevated (p)ppGpp levels, along with the auxiliary protein DksA and the flagellar σ factor FliA then lead to stimulation of bifA transcription and hence increased levels of the phosphodiesterase BifA, which is responsible for depleting the c-di-GMP pool prior to biofilm dispersal 24 . The subsequent decrease in c-di-GMP levels would predictably release LapD inhibition of LapG protease activity, resulting in cleavage of LapA from the cell surface and biofilm dispersal 7,22 . We recently demonstrated that c-di-GMP inhibits FleQ-dependent activation of the flagellar cascade 61 , while stimulating FleQ-regulated synthesis of the biofilm matrix components LapA and cellulose 61,62 . Given those findings, we consider it likely that BifA-dependent c-di-GMP depletion would also result in concurrent cessation of LapA and cellulose synthesis accompanied by the stimulation of flagellar synthesis. In turn, FliA activation during flagellar assembly would probably increase BifA synthesis further, thus completing a positive feedback-loop to tightly couple biofilm dispersal with the de novo synthesis of the flagellar apparatus in preparation for the resumption of a planktonic lifestyle.
The model outlined above places BifA as a central integrator of the nutrient stress and biofilm dispersal responses mediated through the global regulatory networks of (p)ppGpp and c-di-GMP. In addition, we have recently shown that FleQ and FliA influence the synthesis of several other enzymes involved in c-di-GMP metabolism in P. putida 61 . Hence, it appears likely that a complex sub-network of interactions could also act to additionally fine-tune c-di-GMP levels as required for the physiological regulation of the switch between the planktonic and sessile lifestyles and vice versa.
Plasmid and strain construction. Plasmids and oligonucleotides used in this work are summarized in Supplementary Table 1. All DNA manipulations were performed following standard procedures 63 . Restriction and modification enzymes were used according to the manufacturers instructions (Fermentas, Roche and NEB). When required, blunt ends were generated using the Klenow fragment or T4 DNA polymerase. E. coli DH5α was used as a host in cloning procedures. All cloning steps involving PCR were verified by commercial sequencing (Secugen). Plasmid DNA was transferred to E. coli and P. putida strains by transformation 64 , triparental mating 65 or electroporation 66 . Site-specific integration of miniTn7 derivatives in P. putida strains was performed essentially as described 67 .
To construct P. putida strains with an in-frame deletion of the dksA gene, 701 bp and 780 bp from the upstream and downstream chromosomal regions flanking dksA were PCR-amplified with oligonucleotide pairs dksAFupstm/dksARupstm (upstream region) and dksAFdwstm/dksARdwstm (downstream region). The PCR products were cleaved with EcoRI and BamHI or BamHI and HindIII, respectively, and three-way ligated into EcoRI-and HindIII-digested pEX18Tc, yielding pMRB33. A FRT-flanked kanamycin resistance cassette was generated by cloning the BamHI-excised kanamycin resistance gene from pUTminiTn5-Km into EcoRV-digested pPS854 to generate plasmid pMPO284. The BamHI-excised FRT-flanked kanamycin resistance cassette was then cloned into the BamHI site in pMRB33, yielding pMRB38. This plasmid was transferred to P. putida KT2440 by electroporation. Selection of integration, allelic replacement and FLP-mediated excision of the kanamycin resistance marker was performed essentially as described 68,69 . The structure of the deleted dksA locus was verified by PCR and Southern blot. For construction of a ∆dksA derivative of the ppGpp 0 strain PP1922, the kanamycin resistance cassette in pMRB38 was replaced by a streptomycin/spectinomycin resistance cassette from pUTminiTn5-Sm/ Sp, which was first excised with BamHI and cloned at the EcoRV site between the FRT sites of pPS854 to yield pMRB97. The FRT-flanked Str r /Spec r cassette was then excised with BamHI and cloned into BamHI-linearized pMRB33 to produce pMRB98, which was subsequently used to generate the dksA deletion in PP1922 as described above for pMRB33 in KT2440. P. putida strains with deletions within relA and spoT were generated as previously described 32 using suicide plasmids pVI681 and pVI682, and double site recombination, with a KT2440 derivative harbouring miniTn5-Tel as the recipient.
The truncated versions of E. coli relA, encoding RelA ∆456-743 and RelA ∆332-743 were excised from pALS13 and pALS14, respectively, by EcoRI and HindIII digestion and cloned into pBBR1MCS-4 digested with the same enzymes to yield pMRB153 and pMRB154. Subsequently, the 1075 bp nahR-Psal expression cassette was obtained from XbaI-and SpeI-digested pMRB120 and cloned into the single XbaI site in pMRB153 and pMRB154 in the same orientation as the relA derivatives to yield pMRB160 and pMRB162.
PCR fragments containing the PlapE and PlapGD promoter regions were inserted in the directional TOPO ® cloning vector pENTR TM /D-TOPO ® and subsequently transferred using the Gateway ® recombination technology (Thermo Fisher Scientific) into the Gateway ® gfpmut3-lacZ fusion vector pMRB3, to produce pMRB240 and pMRB241, as previously described 61 . A 217 bp synthetic DNA fragment (GenScript) containing the putative PbifA promoter region (positions −198 to +3) and flanked by EcoRI and BamHI sites was digested with these enzymes and cloned into EcoRI-and BamHI-cleaved pTE103 to yield the PbifA in vitro transcription template plasmid pVI2407.
SCIeNtIfIC RepoRTs | (2017) 7:18055 | DOI:10.1038/s41598-017-18518-0 Biofilm growth and quantification. For most procedures involving biofilm growth, overnight cultures grown in LB or K10T-1 broth 34 were diluted in the same medium to an A 600 of 0.1 and 150 μl were dispensed into wells of Costar 96 microtiter polystyrene plates (Corning). The plates were incubated at 25 °C with moderate shaking (150 rpm) for the desired period of time and processed for planktonic and biofilm growth quantification, essentially as described 70 . Serial dilution-based growth curves were performed as described 33 . For each experiment, at least 3 biological replicates were assayed in sextuplicate.
For biofilm growth at the medium-air interphase (pellicle), fresh colonies were inoculated in glass tubes containing 5 ml of K10T-1 broth. Cultures were incubated overnight at 30 °C with shaking, after which the tubes were placed on a rack for 10 minutes and documented by digital photography.
In vivo gene expression assays. β-galactosidase assays were used to examine the expression of the PbifA promoter in P. putida KT2440 and derivatives. Pre-inocula of bacterial strains harbouring the fusion plasmid pMRB67 were grown to the stationary phase in LB with carbenicillin. Cultures were diluted 100-fold in the same medium and incubated in the same conditions. Samples were withdrawn from the cultures at 1-hour intervals and β-galactosidase activity was determined from SDS-and chloroform-permeabilized cells as previously described 71 .
Alternatively, the rates of GFP accumulation during exponential growth were used to measure gene expression in cultures bearing gfp-lacZ fusions to different promoters. Overnight LB cultures of the fusion-bearing strains were diluted in 1/10 strength LB broth, and dispensed into the wells of a Costar 96 microtiter polystyrene plate (Corning). The plate was incubated in a Spark microtiter plate reader/incubator at 30 °C with 510 rpm shaking to mid-exponential phase. Cultures were then diluted in the same medium and incubated in the same conditions for 23 hours. A 600 and GFP fluorescence (485 nm excitation, 535 nm emission) were monitored in 15-minute intervals during incubation. Differential rates of GFP fluorescence accumulation were calculated as the slopes of the linear regression (R 2 > 0.98) performed on the exponential growth fluorescence vs. absorbance plots (Fig. 6).
In vitro transcription assays. Nucleotides and [α-32 P]-UTP were purchased from Roche Molecular Biochemicals and Perkin Elmer, respectively, while ppGpp was synthesized and purified as previously described 72 . Purification of P. putida-derived native core RNA polymerase, N-terminally His-tagged DksA (His-DksA) and C-terminally His-tagged FliA (FliA-His) was as previously described 41,73,74 . Protein concentration was determined using a BSA ™ Protein Assay Kit (Pierce) with bovine serum albumin (BSA) as a standard.
Transcription assays were performed at 30 °C essentially as described previously 40 , using 10 nM supercoiled pTE103-based plasmids bearing the PbifA promoter (pVI2407) or the Paer2 promoter (pVI1011) as the DNA template. Assays of a final volume of 20 µl were performed in T-buffer (35 mM Tris-acetic acid, pH 7.9; 70 mM potassium acetate; 5 mM magnessium acetate; 20 mM ammonium acetate; 1 mM dithiothreitol and 0.275 mg ml −1 BSA). Core RNA polymerase (10 nM) and FliA-His (40 nM) were pre-mixed and incubated for at least 5 min to allow holoenzyme formation. When required, ppGpp, His-DksA and/or His-DksA storage buffer were added to holoenzyme mixes and incubated for 5 min prior to addition of template DNA. Reactions were incubated for 20 minutes to allow open-complex formation. Single-round transcription was initiated in the presence of anti-RNase (Ambion) by the addition of NTPs (final concentration: ATP, 500 µM; GTP and CTP, 200 µM each; UTP, 80 µM and [α-32 P]-UTP, 5 mCi at >3000 Ci mmol −1 ) and heparin (0.1 mg ml −1 ) to prevent re-initiation. After a further 10 minutes at 30 °C, the reactions were terminated by adding 5 µl of a stop/load mix [150 mM EDTA, 1 M NaCl, 14 M urea, 3% glycerol, 0.075% (w/v) xylene cyanol, 0.075% (w/v) bromophenol blue]. Transcripts were analysed on 7 M urea/5% (w/v) polyacrylamide sequencing gels and quantified using a Molecular Dynamics Phosphorimager. Data availability statement. All data generated or analyzed during this study are included in this published article (and its Supplementary Information files).