P. fluorescens and Escherichia coli strains were grown in Luria-Bertani (LB) medium at 28 and 37°C, respectively. M9 medium was supplemented with acetate, glucose, or histidine at final concentrations of 2.5–20 mM. When required, antibiotics were added at the following concentrations: ampicillin (Ap), 100 μg ml−1; gentamicin (Gm), 25 μg ml−1; kanamycin (Km), 30 μg ml−1. For selection against pK18mobSacB, LB agar plates were supplemented with 10% sucrose. Genes cloned in pSRKgm were induced by the addition of 100 μM of isopropyl-β-D-thiogalactopyranoside (IPTG).

Pre-cultures for assays were prepared in a volume of 5 ml as described previously (Zhang et al., 2015). Growth curves were performed using a Synergy H4 Hybrid microplate reader with the Gen5 software (BioTek, Winooski, VT) as described previously (Zhang et al., 2015).

Bacterial strains and plasmids used in this study are listed in Table 1. All molecular techniques were conducted following to standard protocols (Sambrook and Russell, 2001). Escherichia coli DH5α was used for general cloning, E. coli DH5αλpir was used for cloning in pK18mobSacB, and E. coli Bl21 DE3 was used for the heterologous expression of proteins. P. fluorescens was transformed by electroporation as described elsewhere (Artiguenave et al., 2006). All constructs were verified by sequencing. Primers used in this study are listed in Table 2.

Plasmids pSRKgm-acs, pSRKgm-crbR, and pET30-crbR were built by cloning the ORF of each gene into the NdeI/BamHI sites of plasmids pSRKgm and pET30.

Derivatives of pK18mobSacB used for allelic exchange were obtained by cloning in the plasmid a PCR product containing the gene to be deleted, together with 500 bp to 1 kb of upstream and downstream sequence. Then the targeted gene was amplified out by total PCR of the plasmid with phosphorylated divergent primers. The resulting PCR product was ligated, and the religated plasmid was recovered by transformation.

To construct plasmid pUC18T-mini-Tn7T-Gm-GW-crbS, the crbS coding region with its putative promoter region was PCR-amplified and cloned into the pEntr221 vector from Invitrogen using the Gateway BP Clonase II enzyme mix (Invitrogen). The crbS region was then transferred to vector pUC18T-mini-Tn7T-Gm-GW, using the Gateway LR Clonase II enzyme mix (Invitrogen). To construct plasmids pUC18T-mini-Tn7T-Gm-GW-crbS ΔSTAC and pUC18T-mini-Tn7T-Gm-GW-cbrA ΔSTAC, the STAC domain was amplified from plasmids pEntr221-crbS and pCR8-cbrA respectively, using phosphorylated divergent primers. The resulting PCR product was ligated, and the religated plasmid was recovered by transformation. The crbS ΔSTAC and cbrA ΔSTAC regions were then transferred to vector pUC18T-mini-Tn7T-Gm-GW, using the Gateway LR Clonase II enzyme mix (Invitrogen).

Chimeric constructs PrcrbS-SLC5(CrbS)/HK(CbrA) and PrcbrA-SLC5(CbrA)/HK(CrbS) were obtained by overlapping PCR of a synthetic DNA fragment and two PCR products as illustrated in Figure S1 and in Supplementary File 1. The resulting product was cloned in pUC18T-mini-Tn7T-Gm-GW with pEntr221 as an intermediary using gateway technology as described earlier in this section.

Plasmid pBBGFP was constructed by amplifying out the uidA gene of pBB53Gus with divergent primers. The resulting PCR was ligated with a phosphorylated PCR product of the GFP gene from pET3a[GFP(LVA); Ferris et al., 2012]. Transcriptional fusions were constructed by cloning a PCR product of each putative promoter region in the EcoRI/KpnI sites of plasmid pBB53Gus or pBB53GFP. Mutagenesis of the acs promoter transcriptional fusions was achieved by the quick-change strategy (Agilent) using plasmid pBBR53PracsGUS or pBBR53acsGFP as a template. Transcriptional fusions used to map the acs promoter were built by amplifying plasmid pBBR53PracsGUS with divergent primers carrying an EcoRI site or a KpnI site. The resulting PCR product was digested with the corresponding restriction enzyme, ligated, and transformed for the recovery of the reconstituted plasmid.

Strains ES01, ES02, ES03, ES07, ES08, ES09, and ES10 were constructed by allelic exchange selected by sucrose sensitivity using the corresponding derivative of plasmid pK18mobSacB (Schäfer et al., 1994). The same strategy was used for the double and triple mutants in the following combinations. Strains SES11 and ES12 were derived from strain ES07, strains ES13, and ES14 were obtained from strain ES08, strain ES15 was constructed from strain ES09, strain ES16 was generated from strain ES13, and strain ES17 from strain PBR809.

Strains ES04 and ES21 were built through Tn7 mediated integration by transforming strain ES01 with pUX-BF13 and the corresponding derivative of vector pUC18T-mini-Tn7T-Gm-GW as describe elsewhere (Choi et al., 2005; Choi and Schweizer, 2006). The same strategy was used for strains ES18 and ES19, which were derived from strain ES17, and for strains ES20 and ES21 derived from strain PBR809 and strain ES01, respectively.

Five-milliliter cultures of strains sbw25, ES01, and ES03 were inoculated from pre-cultures to an A₆₂₀ of 0.5 in M9 medium supplemented with acetate and incubated at 30°C. After 8 h cultures were pelleted, frozen with liquid nitrogen and delivered to the Proteome Center of the University of Tübingen for processing and analysis. After tryptic digestion and methylation, samples were examined by liquid chromatography-tandem mass spectrometryLC-MS/MS analysis on a Proxeon Easy-nLC1200 coupled to a QExactive HF (130 min gradient, HCD, Top12). Processing of data was done using MaxQuant software (vs. 1.5.2.8). The spectra were searched against a UniProt- Pseudomonas fluorescens SBW25 database (7,063 entries). Down-stream analysis was done using Perseus software (vs. 1.5.0.15) to perform significanceB calculation of normalized protein group ratios (P-value threshold 0.01). Dataset is available as Supplementary File 2.

Five-milliliter cultures of the appropriate strains were inoculated from pre-cultures to an A₆₂₀ of 0.5 in M9 medium supplemented with glucose or acetate and incubated at 30°C for 8 h. β-Glucuronidase activity expressed in modified Miller units was quantified as described previously (Jefferson et al., 1986). Photometric measurements were performed with the Synergy H4 Hybrid microplate reader.

Five-milliliter cultures of the appropriate strains were inoculated from pre-cultures to an A₆₂₀ of 0.5 in LB medium, and heterologous expression of CrbR was induced with 500 μM IPTG (isopropyl-β-D-thiogalactopyranoside). After 4 h of incubation at 37°C, 200 μl of each culture were transferred to a 96-well plate, and GFP fluorescence (485/530 nm), and cell density A₆₂₀ were measured using the Synergy H4 Hybrid microplate reader.

RNA was isolated using the Rneasy Mini kit (Qiagen) from a 5 ml culture of P. fluorescens SBW25 (O.D.600 0.5) incubated under acetate-utilizing conditions for 8 h. cDNA synthesis was performed with the RevertAid First-Strand cDNA Synthesis Kit (ThermoScientific) using primers ACS1RACE1, ActPRACE1, 0110RACE1, and 5625RACE1 for acs, pflu1813/actP, pflu0110, and pflu5625, respectively. A poly-ATP tail was added to the 5′ end of cDNAs using Terminal Transferase (ThermoScientific) as instructed by the manufacturer and the tailed cDNA sample was subjected to PCR amplification using kit's oligo(dT)₁₈ primer and primers ACS1RACE2, ActPRACE2, 0110RACE2, and 5625RACE2. Only PCR reactions for acs and pflu1813/actP generated products of the expected size. Finally, the PCR products were sequenced using primers ACS1RACEseq and ActPRACEseq.

The Microbial Genomic Context Viewer (Overmars et al., 2013) was used to search for homologs of the CrbS/R regulon in Gammaproteobacteria and to retrieve their putative promoter region. The search was performed in all the genomes available in the platform, but when several strains of the same species were available, only one was used. MEME (Bailey et al., 2009) and FIMO (Grant et al., 2011) searches were run with the default options searching for a motif between 6 and 30 nucleotides long. Sequence analyses were performed in the MPI bioinformatics toolkit (Alva et al., 2016; https://toolkit.tuebingen.mpg.de) using Quick2D, HHpred (Hildebrand et al., 2009), PCOILS (Lupas et al., 1991), and MARCOIL (Delorenzi and Speed, 2002). Transmembrane segments were predicted with FMHMM (Sonnhammer et al., 1998) and Phobius (Käll et al., 2004). For transcriptional fusions, comparisons between samples were performed by ANOVA and Tukey's multiple comparison test using SPSS version 23 (IBM).

In P. fluorescens sbw25, CrbS is encoded by the gene pflu4471, CrbR by the gene pflu1195, and the acetyl-CoA synthase by the gene pflu4766. Although, regulation of acs by CrbR has been shown in several species of Gammaproteobacteria, there is little knowledge about other possible genes under the control of the CrbS/R system. First, we verified that the behavior of the CrbS/R system in P. fluorescens is as reported in other bacteria. To do so, we constructed strains ES01 (ΔcrbS), ES02 (ΔcrbR), and ES03 (Δacs) from P. fluorescens SBW25 and tested them for their ability to grow on acetate. As expected, the three deletion mutants were unable to grow on M9 minimal medium supplemented with acetate, unless complemented with the corresponding gene (Figures 2A–C). Next, we performed a quantitative proteome analysis of cultures of strains SBW25, ES01 (ΔcrbS), and ES03 (Δacs) after incubation for 8 h in M9 minimal media supplemented with acetate. This approach permits the direct comparison of protein abundance between strains, allowing us to distinguish those proteins that changed their abundance as a consequence of acetate utilization from those that were affected by the deletion of ΔcrbS. We identified four proteins, Pflu0110, Pflu1813, ActP, and Pflu5625 that increased their abundance in the SBW25 and ES03 strains but decreased it in strain ES01. Since both mutants are unable to grow on acetate, differences in protein abundance between the two strains can be directly linked to the deletion of CrbS (Figure 3).

actP has been shown previously to be transcriptionally activated by CrbR (Jacob et al., 2017) and pflu1813 is found in the same operon as actP (pflu1814) throughout Gammaproteobacteria. This supports the notion that CrbR also regulates the other genes identified through our approach. To confirm this hypothesis we built β-glucuronidase transcriptional fusions of the putative promoter region of each of the genes that code for the identified proteins, plus acs, and tested their expression in different genetic backgrounds. Our results show that all five transcriptional fusions are activated in a wt strain when incubated in M9 medium supplemented with acetate, but not with glucose (Figure 4 and Figure S2). This activation did not happen in the ES01 (ΔcrbS) nor the ES02 (ΔcrbR) strains (Figure 4 and Figure S2), demonstrating that the CrbS/R system controls these genes. Interestingly, in strain ES03 (Δacs), all transcriptional fusions showed high levels of transcription when incubated with glucose. This observation is consistent with the results from the proteome analysis. It could be explained by the continuous triggering of the CrbS/R system, present in this strain, because of the accumulation of acetate from the growth medium and the cell metabolism, due to lack of acetyl-CoA synthase activity (Figure 4). Whether acetate or an unknown related molecule is the signal responsible for triggering the CrbS/R system is still pending to be experimentally determined.

So far, we have demonstrated that Pflu0110, Pflu1813, ActP, and Pflu5625 are part of the CrbS/R regulon. Pflu0110 is predicted to be an Acyl-CoA Hydrolase, an enzyme that could hydrolyze acetyl-CoA into CoA and acetate. Pflu1813 is annotated as a putative membrane protein of unknown function that belongs to the DUF485 family. ActP is a cation/acetate symporter, which as mentioned previously, has been shown to be part of the CrbS/R regulon. Seeking to understand the role that these proteins have in acetate utilization we constructed deletion mutants of each of the genes and challenged them to grow on M9 minimal media supplemented with acetate. While strains ES07, ES08, and ES09 (Δpflu1813, ΔactP, and Δpflu5625, respectively) showed a negligible effect, the strain ES10 (Δpflu0110) presented a 5-h delay in reaching stationary phase, which also occurred at a lower optical density (Figure 2D). The minimal impact of pflu1813, actP, and plfu5625 deletions on acetate utilization could be explained by a suppression effect between them. To discard this possibility, we constructed different combinations of double and triple mutants and tested their ability to grow on acetate. Only those in which pflu0110 was deleted showed an impairment of their capacity to utilize acetate on the same scale as strain ES10 (Data not shown).

In the next step, we studied the sequence properties of crbR-responsive promoters that mediate specific activation under acetate utilizing conditions. We used 5′ RACE to determine the transcription initiation site (TIS) from the acs and actP promoters. These are located at 34 and 36 base pairs from the first codon of the ORF, respectively. In both regions, immediately after the transcription initiation site, we found a Crc motif (AAnAAnAA) that suggests catabolite repression (Moreno et al., 2015). Additionally, in the first 36 base pairs upstream of the initiation sites we located putative RpoD promoters (Potvin et al., 2008), an observation congruent with an earlier bioinformatic analysis in P. aeruginosa (Schulz et al., 2015). Although, we did not succeed to identify the transcription initiation site in the case of genes pflu0110, and pflu5625, we also found putative Crc motifs around 30 nucleotides upstream of the beginning of the ORFs.

To determine the minimal region necessary for CrbR-dependent regulation, we built transcriptional fusions from several derivatives of the acs promoter and measured their activity in M9 medium supplemented with acetate in a wt and a ΔcrbR background (Figure 5A). Our results outlined a minimal inducible promoter encompassing the region between the −62 and +1 nucleotides of our original construct (Figure 5 fusion Pracs-62+1). Most notably, the dramatic loss of induction of the Pracs-35 derivative that ends just after the predicted RpoD promoter, compared to Pracs-62, indicated that an element essential for CrbS/R-mediated induction could be located in this region (Figure 5 fusion Pracs-62 and Pracs-35).

CrbR is a member of the LuxR family, whose members are known to recognize and bind inverted repeats (Evangelista-Martínez et al., 2006; Antunes et al., 2008). To identify any conserved motifs required for CrbR-mediated promoter induction, we obtained the sequence of the putative promoter region of acs homologs from 31 Gammaproteobacteria (Table 3) and ran a motif search using the MEME, a tool for the discovery of novel, ungapped motifs in sequences (Bailey et al., 2009). We were able to identify an imperfect inverted repeat toward the 5′ end of the coding strand, which was present in all the sequences analyzed. Particularly, in the acs promoter of P. fluorescens SBW25, it is located between the −52 and −36 bases from the transcription initiation site, immediately before the start of the RpoD promoter (Figure 5A). Not only is this observation is congruent with our results from the mapping of the acs promoter, but the identified inverted repeat is also located in the putative promoter regions of actP, pflu5625, and pflu0110 (Figure 5B). Moreover, we performed a motif search of the putative promoter regions of a total of 112 identified homologs of acs, actP, pflu5625, and pflu0110 in 31 species of Gammaproteobacteria. We found the imperfect inverted repeat in 91 (81%) of the analyzed sequences (Table 3) and determined a consensus sequence using MEME (Figure 5C). Consistent with the absence of the CrbS/R system in members of the family, the identified motif was not found in the vicinity of the homologs of these genes in enterobacteria. All the evidence we gathered pointed to a role of this inverted repeat in the transcriptional activation of the CrbS/R regulon in the presence of acetate, so we named it the ACTU motif (acetate utilization). To test this hypothesis, we constructed transcriptional fusions with two derivatives of the acs promoter with one (Pracs1M) and four (Pracs4M) point mutations, respectively, which break the symmetry of the repeat. When introduced into a wt strain and tested under acetate-utilizing, conditions Pracs1M showed reduced induction levels while fusion Pracs4M show no activity (Figures 5A,B). These results confirm the role of the motif in the acetate-dependent induction of the promoter. Nevertheless, we still needed to establish a direct link between the ACTU motif and the induction of the acs promoter via CrbR, as we could not discard the possibility that other proteins in P. fluorescens were interacting with the motif. To do so, we cloned CrbS in pET30 (pET30-CrbR) and introduced this plasmid in E. coli Bl21 DE3 strains carrying (I) a GFP transcriptional fusion to the acs promoter (Pracs), (II) the acs promoter derivative with the mutagenized palindrome (Pracs4M-GFP), or (III) a promoterless GFP reporter plasmid. When CrbS expression was induced with IPTG the Pracs-GFP transcriptional fusion, but not the Pracs4M-GFP, was activated. Without the addition of IPTG, or in strains carrying the empty pET30, neither fusion was induced (Figure 6). These results demonstrate that expression of the acs promoter requires the presence of Crb and that the activation of the acs promoter by CrbR is contingent on the presence of the ACTU motif. Activation was independent of the carbon source, as the reporter levels were measured in complex medium. This suggests that CrbR is activated in E. coli through an unknown mechanism. Unspecific phosphorylation is a common occurrence in members of the OmpR/LuxR family (Barbieri et al., 2013; Huynh et al., 2015). The fact that the acs promoter is transcribed in E. coli is not surprising, considering that RpoD from Pseudomonas and its promoter consensus sequence are highly similar to those of E. coli (Potvin et al., 2008; Schulz et al., 2015).

Finally, we used the FIMO module of the MEME suite, that finds individual motif occurrences in a sequence database (Grant et al., 2011), to prospect for additional genes regulated by the CrbR/S system in P. fluorescens SBW25. Besides the genes identified in this work, the program predicted new occurrences of the palindromic sequence in the upstream region of several genes (Table 4). These predictions should be approached cautiously as the q-value suggests they may not be significant. However, this was the case for pflu0110, which we have experimentally proved to be regulated by CrbR. Therefore, experimental verification should be obtained for the most interesting predictions, for example, genes pflu3808 and pflu4952, which code for Isocitrate dehydrogenase and Fumarase, respectively. Both enzymes are part of the TCA cycle, and particularly the Isocitrate dehydrogenase is located at the branching point of the glyoxylate cycle that cells activate when growing on acetate (Wolfe, 2005).

In bacteria, there is mounting evidence for the existence of multi-protein complexes that link transport and signaling. For example, several signal transduction systems use permeases, ABC-transporters, or soluble substrate-binding proteins as co-sensors (Tetsch and Jung, 2009; Västermark and Saier, 2014). In this context, CrbS and CbrA represent a unique type of sensor, in which the signaling and transport domains are connected within the same polypeptide. In the case of CbrA, previous data suggest that the protein not only senses histidine, but is also capable of internalizing it, and that this process is dependent on signaling. Physical coupling between the N-terminal SLC5 domain and the C-terminal Histidine kinase of CbrA is required for function (Zhang et al., 2015). These observations raise the possibility that transport of the substrate through the SLC5 domain elicits a signal, which could trigger or modulate the activity of the histidine kinase. To test this hypothesis, we generated strain ES17, a double ΔcbrA ΔcrbS deletion mutant. When tested, this strain was not able to grow on acetate as a carbon source nor in histidine as a carbon and nitrogen source (Figures 7A–D). Next, using overlapping PCR, we built two chimeric constructs, from CbrA and CrbS, in which the promoter and the SLC5 domain regions of each gene were exchanged. A conserved arginine located immediately before the STAC domain was used as the crossover point (Figure 1 and Figure S1 and Supplementary File 1). These constructs were used independently to complement strain ES17, giving rise to strains ES18 (SLC5-CrbS/HK-CbrA) and ES19 (SLC5-CbrA/HK-CrbS; Figure S1). Both strains were then challenged to grow in M9 media supplemented with histidine or acetate or a combination of both. Strain ES18 was able to thrive in any media containing histidine (Figures 7A,B). The presence of acetate in the medium had no effect, as long as histidine was also added (Figures 7B,D) but the strain was not able to grow solely on it (Figure 7C). These results show that, in CbrA, substrate transport is not needed for the induction of the signal. The previously observed lack of activity of the histidine kinase when not physically coupled to the SLC5 domain can be explained then by increased instability, misfolding, or failure to dimerize. Strain ES19 was not able to grow in any of the conditions tested (Figures 7A–D). Two factors may explain this. (I) The chimera is expressed from the cbrA promoter, which may not be active under acetate-utilizing conditions; (II) CrbS may have additional structural limitations.

Since the STAC domain has been described recently, there are no data yet on its function. To gain an insight into the role of the STAC domain, we constructed strains ES20 (CbrAΔSTAC) and ES21 (CrbSΔSTAC), in which the STAC domain was removed. The limits of the deleted region were carefully designed to avoid disruption of the secondary structure of the protein (Figure 1; boxed residues). Surprisingly, when tested for their ability to utilize the corresponding substrate, there was no obvious difference in growth when compared to the wt strain. Moreover, the activity of the Pracs transcriptional fusion in strain ES21 showed no significant difference under acetate utilizing conditions, when compared to the wt strain (Data not shown).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.