Arginine as an environmental and metabolic cue for cyclic diguanylate signalling and biofilm formation in Pseudomonas putida

Cyclic diguanylate (c-di-GMP) is a broadly conserved intracellular second messenger that influences different bacterial processes, including virulence, stress tolerance or social behaviours and biofilm development. Although in most cases the environmental cue that initiates the signal transduction cascade leading to changes in cellular c-di-GMP levels remains unknown, certain l- and d-amino acids have been described to modulate c-di-GMP turnover in some bacteria. In this work, we have analysed the influence of l-amino acids on c-di-GMP levels in the plant-beneficial bacterium Pseudomonas putida KT2440, identifying l-arginine as the main one causing a significant increase in c-di-GMP. Both exogenous (environmental) and endogenous (biosynthetic) l-arginine influence biofilm formation by P. putida through changes in c-di-GMP content and altered expression of structural elements of the biofilm extracellular matrix. The contribution of periplasmic binding proteins forming part of amino acid transport systems to the response to environmental l-arginine was also studied. Contrary to what has been described in other bacteria, in P. putida these proteins seem not to be directly responsible for signal transduction. Rather, their contribution to global l-arginine pools appears to determine changes in c-di-GMP turnover. We propose that arginine plays a connecting role between cellular metabolism and c-di-GMP signalling in P. putida.


Results
l-Arginine increases c-di-GMP levels and promotes biofilm formation in P. Putida.. To expand our previous observations connecting arginine biosynthesis and c-di-GMP levels, deletion mutants in argG and argH, previously constructed and confirmed to be auxotrophs for l-arginine 10 , were analysed in terms of second messenger contents by introducing the c-di-GMP biosensor plasmid pCdrA::gfp C11 and measuring fluorescence during growth in diluted LB. As observed earlier with transposon insertion mutants in these genes 7 , both the ΔargG and ΔargH strains showed significantly less fluorescence than the parental strain despite having similar growth patterns, indicative of reduced intracellular c-di-GMP contents ( Supplementary Fig. S1), and lost the crinkly colony morphology associated to the plasmid harbouring cfcR in multicopy unless supplied with l-arginine ( Supplementary Fig. S2). Addition of increasing concentrations of l-arginine to the growth medium strongly enhanced fluorescence in the wild type and restored it to a limited extent in the ΔargG and ΔargH mutants (Fig. 1). As expected, fluorescence was severely reduced in a ΔcfcR mutant even in the presence of l-arginine, although a certain dose-dependent response was still detectable (Fig. 1), suggesting that the raise of c-di-GMP levels due to the amino acid is mostly but not exclusively through the DGC activity of CfcR.
To define if the stimulatory effect of l-arginine on c-di-GMP contents was specific of this amino acid, fluorescence of KT2440 harbouring pCdrA::gfp C was tested during growth in rich and minimal medium supplied with each of the 20 proteinogenic l-amino acids at 5 or 15 mM during 24 h. As shown in Fig. 2, l-arginine was the only amino acid causing a relevant, concentration-dependent increase in relative fluorescence in both media (between 1.5-and threefold) throughout culture growth. Addition of 15 mM l-tryptophan also resulted in a relevant increase (nearly twofold) in rich medium but not in minimal medium, which could suggest the need for additional molecules for the response to l-tryptophan. Further analysis revealed a synergistic effect of l-arginine and l-tryptophan: addition of l-tryptophan in minimal medium had a minor influence on c-di-GMP levels, but the combination of both amino acids caused a significantly higher response than addition of l-arginine alone ( Supplementary Fig. S3). Statistically significant, yet quantitatively less relevant increases were also observed with other l-amino acids. Negative effects could also be detected in some cases, particularly with proline, which caused a 35% reduction in relative fluorescence at 15 mM in both media (Fig. 2).
In many bacteria, including P. putida KT2440, c-di-GMP levels directly correlate with biofilm development. We therefore tested if increasing concentrations of l-arginine enhanced attachment and biofilm formation. Assays were done in polystyrene multiwell plates under static conditions in minimal medium with glucose as carbon source. The presence of l-arginine did not influence planktonic growth in these conditions (Fig. 3a), but increased the amount of attached biomass (Fig. 3b).
Arginine biosynthesis modulates the expression of biofilm structural elements. The crinkly colony phenotype associated to high levels of c-di-GMP in P. putida requires the species-specific exopolysaccharide (EPS) Pea 5 . On the other hand, the large adhesins LapA and LapF are essential for the development of mature biofilms in P. putida, a process in which the four EPS present in this bacterium would contribute differently depending on the environmental conditions [12][13][14] . These elements are differentially modulated by the c-di-GMP dependent regulator FleQ [14][15][16] . All these facts prompted us to investigate if expression of any of those structural elements of the biofilm matrix was affected in the ΔargG and ΔargH mutants. Plasmids harbouring transcriptional fusions of lapA, lapF and the first gene in each EPS cluster with the reporter gene lacZ devoid of its own promoter [16][17][18] were introduced in P. putida KT2440 and the two mutants and β-galactosidase activity was followed during growth in LB. Results are summarized in Fig. 4. A significant reduction in expression was observed in stationary phase for the lapF::lacZ and the pea::lacZ fusions (Fig. 4b,c) in both arginine biosynthesis mutants compared to the wild type. In contrast, the other four fusions showed only minor differences between strains.
Interestingly, the appearance of the crinkly colony phenotype associated to high levels of c-di-GMP does not take place in P. putida KT2440 harbouring cfcR in multicopy when grown on M9 minimal medium agar plates with glucose as the only carbon source, unless l-arginine is added ( Supplementary Fig. S2). We therefore tested if expression of pea, required for this phenotype, also responded to exogenous l-arginine. As shown in Fig. 5, expression of the pea::lacZ fusion was enhanced with increasing concentrations of the amino acid. This effect was more evident upon entry in stationary phase. and lapF is under the control of the stationary phase sigma factor RpoS 5,17 , and the same has been recently reported for pea 19 , a result that we have independently confirmed ( Supplementary Fig. S4). Hence, we considered the possibility that the differences in expression observed for these two genes in the arginine biosynthesis mutants could reflect an influence of arginine availability on expression of rpoS. To test this hypothesis, a translational rpoS'-'lacZ fusion, harboured in pMAMV21 5 , was introduced in P. putida KT2440 and the ΔargG and ΔargH mutants, and β-galactosidase activity was measured during growth in LB. Results in Fig. 6a, indicate that a functional arginine biosynthesis pathway is required for full expression of rpoS expression, since β-galactosidase activity was reduced in both mutants in stationary phase. Addition 5 or 15 mM of l-arginine increased β-galactosidase activity in the ΔargG and ΔargH mutants (Fig. 6b), whereas addition of 25 mM of the amino acid caused a less stimulatory effect compared with 15 mM in the case of the ΔargG mutant and had no significant effect in the ΔargH strain. In the wild type, significantly increased rpoS expression was only observed with 5 mM l-arginine. Activity of a transcriptional cfcR::lacZ fusion, harboured in pMIR200 6 , was also tested in the ΔargG and ΔargH mutants. As shown in Fig. 6c, the expression pattern of cfcR was similar to that observed for rpoS, with the mutants having lower activity upon entry into stationary phase.
Substrate binding proteins participate in the response to environmental l-arginine. To explore in more detail the response of P. putida to l-arginine, we carried out a similarity search analysis to identify proteins that could be analogous to ArtI, the l-arginine binding protein involved in c-di-GMP signalling in S. enterica serovar Typhimurium 9 . Two proteins, corresponding to loci PP_0282 and PP_4486, present around 40% identical residues with ArtI of S. enterica, and a third one, encoded by PP_3593, shows 36% identity. The three are periplasmic substrate binding proteins sharing around 25% identical residues, amino acids likely involved in arginine binding are conserved, and the corresponding genes are located in clusters encoding predicted amino acid ABC transporters ( Supplementary Fig. S5). PP_0282 is annotated in the Pseudomonas genome database (https ://www.pseud omona s.com; 20 ) as ArtJ (l-arginine ABC transporter substrate- www.nature.com/scientificreports/ binding subunit) and PP_4486 as ArgT (lysine/arginine/ornithine ABC transporter substrate-binding protein); PP_3593 has no specific annotation, but the protein is 72% identical to the octopine-binding protein OccT of Pseudomonas protegens CHA0. Hereafter, this nomenclature is followed.  www.nature.com/scientificreports/ To define the potential role of these substrate-binding proteins in l-arginine transport, deletion mutants were constructed in each of the corresponding genes, as well as a double ΔargTΔartJ mutant, and their growth was tested in M8 minimal medium with glucose as cabon source and l-arginine as nitrogen source. Results presented in Fig. 7a indicate that ArgT is the main contributor to l-arginine uptake, given the long lag phase and extended doubling time of the ΔargT mutant. The ΔoccT and ΔartJ mutants were not affected in growth, whereas the double ΔargTΔartJ mutation caused a much stronger effect on growth than the single ΔargT mutation, suggesting that in the absence of ArgT, ArtJ also plays a relevant role in l-arginine uptake. When l-arginine was supplied as the sole carbon and energy source in M9 minimal medium, all mutants showed a slight delay in growth, being greater in the double mutant (Fig. 7b). Experiments with other basic amino acids as nitrogen or carbon sources revealed only minor differences between strains ( Supplementary Fig. S6), except in the case of the ΔoccT mutant, which was unable to grow in l-lysine as carbon and energy source.
To test the involvement of these binding proteins in arginine-dependent c-di-GMP signalling, the biosensor pCdrA::gfp C was introduced in the ΔargT, ΔartJ, ΔoccT and ΔargTΔartJ mutants, and fluorescence was analysed during growth in M9 minimal medium with glucose as carbon source and in the presence of increasing concentrations of l-arginine. As shown in Fig. 8, no difference in relative fluorescence was observed between the wild type and the ΔargT, ΔartJ, and ΔargTΔartJ mutants in the absence of l-arginine (Fig. 8a). However, the dose-dependent response to l-arginine was significantly reduced in all these mutants, a cumulative effect being observed in the double ΔargTΔartJ mutant in these conditions (Fig. 8b,c). Surprisingly, the ΔoccT mutant showed increased fluorescence with respect to KT2440 in the absence of the amino acid ( Fig. 8a) and maintained the dose-dependent response to l-arginine, reaching higher fluorescence levels than the wild type at the different concentrations of amino acid tested (Fig. 8b,c). www.nature.com/scientificreports/ Role of arginine binding proteins in biofilm development associated to l-arginine. We tested if the above results correlated with changes in biofilm formation between the wild type and the different mutants in the presence or absence of exogenous l-arginine. Results are shown in Fig. 9. As previously observed, addition

Discussion
In recent years, evidence has been accumulating that connects bacterial social behaviours with the presence in the environment of certain amino acids. d-amino acids prevent biofilm formation in Staphylococcus aureus and P. aeruginosa 21 . On the other hand, several l-amino acids have been described to hamper swarming motility and stimulate biofilm formation in P. aeruginosa PA14 8 ; among them, arginine caused a significant increase in c-di-GMP content, even though it was not the most relevant in terms of enhancing biofilm formation 8 . However, the positive effect of arginine on biofilm formation was only observed in cultures grown with the amino acid as the only carbon and nitrogen source, but not when both arginine and glucose were present 8 . In contrast, our results show that l-arginine increases c-di-GMP levels and promotes biofilm formation in P. putida regardless of the presence of other carbon and nitrogen sources. Furthermore, in this bacterium l-arginine appears to function both as a metabolic signal and as an environmental signal: mutants deficient in arginine biosynthesis show low c-di-GMP levels, partly restored by exogenous l-arginine, while mutants limited in arginine transport present reduced response to the presence of the amino acid in the growth medium, in terms of c-di-GMP levels and biofilm formation. We have also confirmed previous observations on the importance of l-arginine in the development of crinkly colony morphology 7 , a phenotype associated to high levels of c-di-GMP in P. putida KT2440 which requires of the species-specific EPS Pea 4 . Addition of l-arginine is required for this phenotype to develop in minimal medium, and the amino acid specifically restores this phenotype in mutants deficient in arginine biosynthesis, indicating that l-arginine plays a relevant role in Pea production. Accordingly, expression of Pea is significantly reduced in ΔargG and ΔargH mutants and increased by addition of l-arginine to the growth medium. EPS production dependent on the presence of l-asparagine in the culture medium has been reported in Bacillus 22 . Interestingly, amino acid-decorated EPSs have been identified in Vibrio 23,24 . Whether arginine is a component of the EPS Pea in P. putida remains unknown. The effect of exogenous l-arginine on motility and biofilm development can vary in different bacteria depending on the concentration of amino acid. For instance, in Streptococcus gordonii, low l-arginine concentrations (between 0.5 and 500 µM) enhance biofilm development and promote the establishment of structured biofilms, while high concentrations (≥ 50 mM) alter biofilm architecture, biomass and thickness 25 . In the case of P. aeruginosa PAO1, l-arginine concentrations above 250 mM inhibit swimming motility, whereas lower concentrations (100 mM) favour this type of motility 26 . In this work we have observed positive effects on biofilm formation and www.nature.com/scientificreports/ c-di-GMP levels with l-arginine concentrations ranging between 5 and 25 mM, but we have also seen different responses in terms of rpoS expression depending on l-arginine concentration. Other l-amino acids also seem to have a positive or negative influence on c-di-GMP levels in KT2440, but to a lesser and in some cases variable extent depending on the growth medium. Among them, l-tryptophan causes a significant increase in c-di-GMP levels in rich medium but not in minimal medium, and our data indicate the existence of a synergistic effect of l-arginine and l-tryptophan. This amino acid has been described to positively impact biofilm development in S. enterica serovar Typhimurium 27 , and genes related to tryptophan biosynthesis are upregulated during early biofilm formation in E. coli 28,29 . However, no effect of tryptophan alone was reported in P. aeruginosa PA14 8 . The connection between arginine and tryptophan signalling will deserve further detailed exploration.
In these experiments, the negative effect of l-aspartic acid, previously shown to reduce c-di-GMP levels in KT2440 7 , was not so evident. It should be noted that the data presented in Fig. 2 correspond to the area below the curve for the relative fluorescence data over 24 h, in order to assess overall differences. Changes in c-di-GMP levels with l-aspartic acid were only evident at late times of growth 7 , and are therefore likely underscored when data throughout 24 h of bacterial culture are compiled together. The negative effect of l-aspartic acid was further evidenced by the reduction of the crinkle colony morphology of KT2440 harbouring cfcR in multicopy when grown in the presence of increasing concentrations of the amino acid ( Supplementary Fig. S2). It could also explain why arginine supplementation did not fully restore c-di-GMP levels in the ΔargG and ΔargH mutants (Fig. 1), since these mutants are bound to accumulate aspartic acid 7 .
There is still limited information about the mechanisms of action of amino acids that lead to changes in the turnover of the second messenger. Our results indicate that in P. putida KT2440, the response regulator with DGC activity CfcR, the chief contributor to c-di-GMP levels in stationary phase 6 , is the main element in the increase in c-di-GMP levels caused by exogenous l-arginine, despite its lack of amino acid-binding or protein-protein interaction domains. Still, a ΔcfcR mutant retains some response to l-arginine, suggesting additional protein(s) with DGC activity yet to be identified also participate in the process. In P. aeruginosa PAO1, SadC and RoeA, www.nature.com/scientificreports/ two of the most important DGCs implicated in biofilm formation, are necessary for the l-arginine response 8 . In addition, a multidomain transmembrane protein with PDE activity encoded by locus PA0575 (RmcA) binds l-arginine in its N-terminal domain, and a mutant in this gene shows increased c-di-GMP levels in response to the amino acid 30 . Homologs of SadC or RoeA are missing in P. putida KT2440, but a homolog of RmcA can be found (PP_0386). It will be worth exploring its potential contribution to the arginine response, although it www.nature.com/scientificreports/ would be expected to correlate with a decrease rather than an increase in the levels of c-di-GMP, based on its role in P. aeruginosa. Arginine has also been found to induce the synthesis of c-di-GMP in Salmonella enterica serovar Thyphimurium. Although the mechanism is not fully characterized, the substrate binding subunit ArtI of the arginine transporter and the diguanylate cyclase STM1987, containing a periplasmic sensing domain, are required for the response to the amino acid 9 . P. putida KT2440 does not appear to have an equivalent of STM1987, but our results show that substrate binding proteins associated to amino acid transport systems are important for the response to l-arginine: mutants lacking ArgT and/or ArtJ, both of which participate in arginine transport, partially lose the increase in c-di-GMP levels observed in P. putida KT2440 in the presence of the amino acid. In contrast, deletion of a third substrate binding protein, OccT, limits lysine utilization as carbon source ( Supplementary  Fig. S6) but has little influence on arginine transport and causes an increase in c-di-GMP levels. The transport systems associated to ArgT and ArtJ had been previously described to participate in l-lysine transport 31 . It has been reported that in KT2440, two active metabolic pathways are required for utilization of l-lysine as the sole carbon source: the aminovalerate pathway and the aminoadipate pathway 32 . The second one requires conversion of l-lysine to d-lysine by a periplasmic racemase 33 . Since the occT mutant can use l-lysine as nitrogen source but not as carbon source, it seems plausible that this transport system is in fact required for d-lysine uptake, and that d-lysine, as described for other d-amino acids in different bacteria, reduces c-di-GMP levels in P. putida. This would be consistent with the increased second messenger levels detected in the occT mutant. Such idea is further supported by the fact that the occT gene is in the same genomic context as genes related to d-lysine catabolism 33 , but additional work will be required to confirm it.
We hypothesize all these data, along with those obtained with arginine biosynthesis mutants, indicate that cellular arginine pools are sensed and transduced into c-di-GMP turnover and signalling in P. putida. This notion, rather than direct interaction between a substrate binding protein and a DGC, is compatible with the changes in expression of RpoS and elements under its control (lapF, pea, cfcR) in the ΔargG and ΔargH mutants, and the influence of exogenous l-arginine through periplasmic binding proteins associated to different amino acid transport systems. It is possible that two independent signalling pathways exist for intracellular and extracellular arginine. Yet, if our hypothesis is correct, the results presented here open the way to further exploring a still poorly developed area of research, namely how central metabolism and second messenger turnover are connected in bacteria. The underlying molecular mechanisms will be analysed in future work.

Methods
Bacterial strains, culture media and growth conditions. Strains used in this work are listed in Table 1. Pseudomonas putida KT2440 is a plasmid-free derivative of P. putida mt-2, which was isolated from a vegetable orchard in Japan and whose genome is completely sequenced 34,35 . Pseudomonas strains were routinely grown at 30 °C in Luria-Bertani (LB) medium 36 . Where indicated, M9 37 or modified FAB 38 minimal media supplied with glucose (20 mM) as carbon source were used. Escherichia coli strains were grown at 37 °C in LB. When Figure 9. Influence of periplasmic substrate binding proteins on biofilm formation by P. putida. KT2440 and the ΔargT, ΔartJ, ΔargTΔartJ, and ΔoccT mutants were grown in FAB minimal medium with glucose and different l-arginine concentrations (shown as increasing intensity colour bars). Attached biomass was analysed after 10 h of growth. Values correspond to absorbance (A 595 ) after staining with crystal violet and subsequent solubilisation of the dye, normalized with respect to culture growth (OD 660 ). Data are averages and standard errors from two independent experiments with four technical replicates each. Asterisks indicate statistically significant differences between the wild type and the corresponding mutant in each condition (Student's t test; *p ≤ 0.05; **p ≤ 0.01). Molecular biology techniques. DNA preparation, digestion with restriction enzymes, plasmid dephosphorylation, adenylation, ligation and cell transformations were carried out using standard protocols 39,40 . PCR amplifications were done using Phusion High-Fidelity DNA polymerase (Thermo Fisher Scientific). Plasmid purification and gel extraction from agarose gels were done with appropriate kits, following manufacturers´ instructions (NZYTech and QIAgen, respectively). Transfer of plasmids to Pseudomonas cells was performed by electrotransformation or triparental conjugation as previously described 10,41 . construction of null mutants. Null mutants were obtained by gene replacement of the wild type allele with a null allele via homologous recombination, without inserting any antibiotic resistance marker. The strategy designed to obtain the mutants consisted of the amplification of the upstream and downstream fragments surrounding the gene to be replaced by overlapping PCR using Phusion High-Fidelity DNA polymerase (Thermo Fisher Scientific). Oligonucleotides used are detailed in Supplementary Table S1. PCR reactions were carried out in two steps. Firstly, flanking regions of the gene to be removed were amplified separately using primers with NotI restriction site on one end and a complementary tail on the other end. Secondly, overlapping upstream and downstream regions were used as template for the second PCR, obtaining a single amplicon flanked with NotI www.nature.com/scientificreports/ restriction sites. PCR product was cloned into pCR2.1-TOPO vector after its adenylation, transferred to E. coli DH5α by heat shock transformation, and sequenced to ensure the absence of mutations. The fragment was then subcloned into the NotI site of the suicide vector pKNG101, which is unable to replicate in Pseudomonas and allows the generation and selection of double recombination events 42 . Each pKNG101 derivative containing the mutation was mobilized from E. coli CC118λpir to P. putida KT2440 by triparental conjugation 10 . Merodiploid exconjugants were selected in M9 minimal medium with citrate as carbon source and streptomycin. One of them was selected to obtain clones in which a double recombination event had taken place after growth in LB medium supplied with 14% sucrose. Resulting mutants were sucrose-resistant and streptomycin-sensitive. Null mutants were checked by PCR, followed by sequencing of the corresponding genome region.
Growth curves. To analyse the growth of P. putida KT2440 and its mutant derivatives in basic l-amino acid transport, overnight cultures grown on glucose-M9-plates at 30 °C were scrapped out in 1 ml of M9 salts and washed two times in the same medium. Inocula were adjusted to a final optical density at 660 nm (OD 660 ) of 0.02 in M9 salts and distributed in 100-well plates (150 µL/well). l-arginine, l-lysine, l-histidine or l-ornithine were added as carbon and energy sources at a final concentration of 10 mM. Alternatively, M8 minimal medium with glucose as carbon and energy source and the amino acids supplied as nitrogen source, was used. Plates were incubated at 30 °C with continuous shaking (200 r.p.m.) and growth of the cultures was monitored at 30 min intervals for 24 h in an automated BioScreen C MBR apparatus equipped with a wide band filter (420-580 nm).
Biofilm assays. Biofilm formation assays were performed in 96-well polystyrene microtiter plates as previously described 43 , using modified FAB medium with glucose as carbon source, based on the presence of calcium in its composition, which is important for adhesion 44,45 . Briefly, overnight cultures were diluted to an OD 660 of 0.02 and 150 µL were added to each well. Where indicated, l-arginine was added at final concentrations of 5 or 15 mM. Plates were incubated at 30 °C in static conditions. At the indicated times, growth of the cultures (OD 660 ) was measured, liquid was removed and wells were washed twice with distilled water. Biomass attached to the surface was stained with crystal violet (0.4%) for 15 min and quantified after dye solubilisation with glacial acetic acid (30% v/v) by measuring absorbance at 595 nm in a Tecan Sunrise plate reader.
Measurement of β-galactosidase activity. β-galactosidase activity was assayed during growth in LB as described 46 . Alternatively, where indicated, M9 minimal medium with glucose, with or without l-arginine was used. Overnight cultures were diluted to an optical density of 0.05 in fresh medium. After 1 h of growth at 30 °C and 200 rpm, cultures were diluted 1:10 to ensure proper dilution of β-galactosidase that might have accumulated after overnight growth; this step was omitted when experiments were done in minimal medium. Incubation was continued in the same conditions, collecting samples at the indicated times. The results are expressed in Miller units and correspond to averages and standard deviations of at least two independent experiments with three technical replicas per sample.
Comparative analysis of c-di-GMP levels based on a bioreporter. The bioreporter plasmid pCdrA::gfp C was used for quantitative analysis of c-di-GMP levels based on fluorescence. This plasmid carries a fusion of gfp to the promoter of the P. aeruginosa gene cdrA, which responds to c-di-GMP via the transcriptional regulator FleQ 11 . Overnight cultures were diluted in fresh medium (LB diluted 1:3 or M9 with glucose) to a final OD 600 of 0.02 and distributed into suitable 96-well plates (Greiner or Nunc Flat Bottom Black Polystyrol 96-well plates). Where indicated, l-amino acids were added at final concentrations of 5, 15 or 25 mM. Plates were incubated at 30 °C in static conditions and growth (OD 660 ) and fluorescence (excitation: 485 nm, emission: 535 nm) were monitored every 30 min for 24 h using microplate fluorescence readers equipped with shaking and temperature control (TECAN Infinite 200, Synergy Neo2 Biotek, and Varioskan Lux). Data are presented as fluorescence/OD 600 . In the case of Fig. 2, data correspond to the calculated area below the curve for all the relative fluorescence with respect to growth.