All bacterial strains and the general transducing phage used in the study are listed in Table 1. E. coli strains were grown at 37°C while Serratia and Pectobacterium strains were grown at 30°C in Lysogeny Broth (LB). Overnight liquid cultures were grown in 5 mL LB in a 30 mL plastic universal tube, inoculated with a single bacterial colony and aerated on a rotary wheel for 14–16 h. Bacterial growth was determined by measuring the optical density of the liquid culture at 600 nm wavelength using a Unicam Heλios spectrophotometer (Thermo Scientific™).

Random transposon mutagenesis was carried out using a plasmid-transposon (plasposon) hybrid system, pKRCPN1, as previously described (Monson et al., 2015; Lee et al., 2017). The plasmid-transposon hybrid vector pKRCPN1 was delivered into NWA19 recipient cells by conjugation using the DAPA auxotroph E. coli β2163 as donor strain. Conjugations were performed by mixing overnight cultures that had been normalized to an OD₆₀₀ of 1.0 in a 1:2 or 1:3 donor to recipient ratio. Putative transconjugants were plated onto selective agar (LB + Kn). The absence of DAPA in the plates allows counterselection of the donor while kanamycin ensured selection of successful NWA19 transconjugants. Resuspended cells from donor-and recipient-only spots were also plated out onto selective agar as controls.

Putative transconjugants were visually screened for any notable difference in colony opacity, compared to NWA19, as a proxy for gas vesicle production. Colonies that were noticeably more translucent, slightly less opaque, hyper opaque or that had an unusual colony morphology were selected putative gas vesicle mutants. The putative mutants were passaged thrice on selective agar to confirm the observed colony phenotypes. The transposon insertion location in the mutants was determined by random primed PCR (RP-PCR) (Jacobs et al., 2003) and subsequent DNA sequencing across the transposon junction using oligos oPF109 and oMAMV2. Prior to characterization, the transposon insertion was transferred into a clean S39006 LacA background by transduction using the generalized transducing phage ϕOT8 (Evans et al., 2010).

Phenotypic plate assays for carbapenem, cellulase, and swimming and swarming motilities were performed as described previously (Slater et al., 2003; Andro et al., 1984; Lee et al., 2017). Overnight cultures of the test strains were normalized to an OD₆₀₀ of 1.0 before spotting 10 μL (or 5 μL for swimming and swarming assays) onto appropriate agar plates. Indicator plates for carbapenem production were prepared using LB agar overlaid with 0.75% top agar seeded with E. coli ESS. Antibiotic production was indicated by inhibition zones around test colonies (Slater et al., 2003). Carboxymethylcellulose plates for cellulase assays were developed by flooding with 0.2% (w/v) Congo red for 20 min, bleached with 1 M NaCl then stained with 1 M HCl for 5 min (Andro et al., 1984). Cellulase activity was indicated by a halo surrounding the test strain. Swimming and swarming plate assays were assessed visually as previously described (Lee et al., 2017).

Potato rotting assays were performed as described by Walker et al. (1994) with some modifications. Maris Piper potatoes were surface sterilized by dipping the tubers in 1% (w/v) Virkon™ solution for 10 min, rinsed with sterile deionized water and air-dried at RT. Overnight cultures of the test strains were normalized to contain approximately 10⁷ CFU/mL based on the OD₆₀₀. Using a sterile 200-μL pipette tip, holes of the same depth were created on opposite sides of a potato tuber and inoculated with 10 μL (10⁵ cells) of the WT and mutant strains. The infection sites were sealed with sterile vacuum grease and the tuber was wrapped in several layers of wet paper towel then finally with cling film. The inoculated tubers were incubated at 30°C for 5 days to allow infection. After incubation, the tubers were cross sectioned, and the rotten tissue was scraped off and weighed. To determine the bacterial load, 0.1 g of rotten tissue samples were resuspended in 1 mL of 1X M9 solution by vortexing. A 100 μL aliquot of the resuspended cells were serially diluted and plated onto LB plates to determine the number of viable cells. Viable cell count for each strain was expressed as CFU per 0.1 g of rotten tissue.

C. elegans killing assays were performed based on the method described by Kurz et al. (2003), using the DH26 strain. The nematodes were maintained on Nematode Growth Medium (NGM) agar covered with an E. coli OP50 lawn. Prior to each killing assay, the nematodes were synchronized to ensure all worms used were at the same development stage. Fifty (50) L4 stage nematodes were used for each test strain in the killing assay. Ten nematodes were placed onto each NGM plate containing a 50 μL spot of the test strain (grown overnight at 30°C prior to the worm transfer). The plates were then incubated at 25°C and scored for dead nematodes every 24 h until the population was eradicated. Nematodes that no longer responded to touch were considered dead and live worms were transferred onto fresh plates every day.

β-glucuronidase (β-glu) activity was determined throughout growth as described by Ramsay et al. (2011). Aliquots (100 μL) from growth experiments of S39006 strains containing uidA gene fusions were taken and frozen immediately at −80°C. The samples were thawed at room temperature and 10 μL of each was transferred to 96-well microtiter plates. The samples were mixed with 100 μL of PBS containing 20 mg/mL lysozyme and 250 μg/mL 4′-methylumbelliferyl-β-D-glucuronide (MUG). Fluorescence emission was measured using the Gemini XPS plate reader (Molecular Devices) under the following settings: excitation 360 nm, emission 450 nm, cut-off 435 nm, 8 reads per well measured every 30 s for 30 min. The β-glu activity of each sample was expressed as Relative Fluorescence Units (RFU)/OD₆₀₀.

The β-galactosidase (β-gal) activity of E. coli DH5α strains carrying lacZ promoter fusion plasmids was measured in the same manner except the substrate used was 4′-methylumbelliferyl-β-D-galactoside.

For Phase Contrast Microscopy (PCM), wet mounts of liquid culture or colonies resuspended in LB were used for phase contrast imaging. Samples were visualized using an Olympus BX-51 microscope with 100X oil immersion lens and images were captured with QICAM Fast 1,394 digital camera.

Transmission Electron Microscopy (TEM) was performed at the Cambridge Advanced Imaging Center, University of Cambridge. Bacterial samples were attached to a carbon-coated glow discharge grid for 10 min and washed thrice with dH₂O. The cells were stained with 2% phosphotungstic acid (pH 7.0) for 5 min and viewed under a FEI Tecnai G2 TEM.

Plasmids pQE80-iclR, pQE80Cm-iclR and pBAD30-iclR for genetic complementation were constructed using standard restriction cloning protocols. Oligonucleotides used in PCR amplification and plasmids used in this study are listed in Tables 2, 3. The coding sequence of iclR was amplified with primer pairs oCMCS14/oCMCS15 and oCMCS16/oCMCS15 for cloning into pQE80-oriT and pBAD30, respectively. To construct pQE80-iclR and pQE80Cm-iclR, the PCR product was digested with BamHI (NEB) and HindIII (NEB) and ligated into compatibly digested pQE80-oriT or pQE80-oriT-Cm with T4 DNA ligase (Thermo Scientific), according to manufacturer’s instructions. Plasmid pBAD30-iclR was likewise constructed as described above except the PCR amplicon and pBAD30 were digested with SphI and HindIII. The promoter fusion plasmid pRW50-yagE_prom was constructed by Gibson cloning using a prepared reaction mix (Supplementary Table 1). The plasmid backbone was amplified by long PCR using PrimeSTAR® GXL DNA Polymerase (Takara) and the 150 bp fragment containing the yagE promoter region was amplified using Phusion DNA polymerase. Equimolar amounts of the DNA fragments to be assembled were mixed used for the reaction. Gibson assembly was carried out by mixing 15 μL of the reaction buffer with 5 μL of DNA fragment mixture and incubated at 50°C for 1 h. All plasmids were subjected to sequencing (Eurofins Genomics or DNA Core Sequencing Facility, Department of Biochemistry, University of Cambridge) to confirm the sequence.

IclR was overexpressed and purified in S39006 to preserve the nativity of the protein. Plasmid pQE80-iclR was transferred into S39006 by conjugation using E. coli β2163. IclR expression was induced with 1 mM IPTG at mid-logarithmic phase of growth and the IclR protein was purified using Ni-NTA agarose (Qiagen) according to manufacturer’s instructions. The protein content of the purified sample was measured by Bradford assay using BSA as standard. The identity of the purified his-tagged protein was confirmed by SDS-PAGE and Western blot analysis against the 6xHis tag.

Preparation of DNA probes and EMSA analyses were carried out according to the Labeled Universal Electrophoretic Gel Shift Oligonucleotide (LUEGO) method with minor modifications (Jullien and Herman, 2011). The fluorescently labeled Cy5-LUEGO was used to generate the different probes together with the designed “short” and “long” oligonucleotides (Table 2). Labeled probes were prepared by mixing the oligonucleotides in a 5:5:1 (LUEGO:short:long) ratio and annealed using the Veriti Thermo Cycler (Applied Biosystems) with the following profile: 2 min at 95°C then a 100% ramp down to 70°C followed by slow cooling at 0.1% ramp rate to 18°C. The resulting probe was diluted to 5 nM and titrated with increasing concentrations of the purified test protein. The protein-probe mixture was incubated in 1X EMSA reaction buffer (10 mM Tris–HCl pH 8, 5 mM DTT, 50 mM KCl, 1 mM MgCl₂, 10% (v/v) glycerol, 0.1% Triton X-100) for 30 min at 20°C to allow binding. The reaction mixture was loaded onto a native resolving gel and separated in 1X TGE (0.5 mM Tris–HCl pH 8, 19.2 mM glycine, 0.2 mM EDTA) at 100 V. Gels were imaged using the Bio-Rad ChemiDoc™ MP imaging system at the excitation wavelength of Cy5 (650 nm).

The protocol for the preparation of cell lysates for proteomic analysis was largely as described by Dolan et al. (2020). Briefly, 25 mL of LB was inoculated with overnight cultures of the strain of interest to starting OD₆₀₀ of 0.05. Strains were grown for 14 h in 250 mL flasks at 30°C and 215 rpm. These cultures were then normalized to an OD₆₀₀ of 2.0 and 25 mL of cells was harvested by centrifugation (2,739 g, 10 min, 4°C). The cell pellets were resuspended in 500 μL lysis buffer (100 mM Tris–HCl, 50 mM NaCl, 10% (v/v) glycerol, 1 mM Tris (2-carboxyethyl) phosphine, 1 mM cOmplete Mini protease inhibitor cocktail (Roche), pH 7.5) followed by three rounds of sonication (3 × 10 s), with 30 s rest between each round. Samples were then centrifuged (21,130 g, 15 min, 4°C) to remove cell debris. Protein content was determined using a Bradford assay (Bio-Rad) and samples were normalized to 2 mg/mL. At this stage, an SDS-PAGE analysis of the protein extracts and a western blot for GvpC were performed as quality controls. Cell lysates containing 100 μg protein were then sent to the Cambridge Center for Proteomics (CCP) for quantitative proteomic analysis.

The statistical significance of any differences seen in RFU or antibiotic production over the course of a growth curve was determined using a repeated measures ANOVA where p-values of less than 0.05 were considered significant. Differences in the weight of rotten tissues from potato tubers between test strains was evaluated using a paired t-test. Viable cell counts from rotten tissues were likewise compared using a paired t-test. Finally, differences in the virulence of C. elegans were assessed using a Mantel-Cox log-rank test. Asterisks in figure legends indicate p-values: *, p < 0.05; **, p < 0.01; ***, p < 0.001; ****, p < 0.0001; NS, not significant.

S39006 colonies are opaque due to the ability of gas vesicles to refract light. Therefore, changes in colony opacity can be used as a simple and rapid visual screen for gas vesicle regulatory mutants. However, the opacity of S39006 colonies can be confounded in certain mutants where production of the red pigment, prodigiosin, is also altered. To overcome this complication, transposon mutagenesis was carried out in the non-pigmented S39006 strain, NWA19. This strain has an in-frame deletion of pigC which encodes the condensing enzyme necessary for prodigiosin production. Random transposon mutagenesis was carried out by conjugation of the DAPA auxotroph E. coli β2163 harboring plasposon pKRCPN1 with NWA19 cells. The pKRCPN1 plasmid contains a Tn5 transposon that is fused to a promoterless lacZ gene, an oriR6K replication origin and a kanamycin resistance cassette. In addition, the plasmid backbone contains a transposase gene to facilitate transposition and a tetracycline resistance cassette. Since the plasmid requires the pir-encoded π protein for replication which is present in the donor strain but not in NWA19, the Tn5 transposon is forced to integrate randomly into the recipient’s chromosome. By screening transconjugants for kanamycin resistance and tetracycline sensitivity, successful integration of only the transposon and not the entire plasmid can be easily confirmed. Lastly, the absence of DAPA in the screening plates was used to select against E. coli β2163.

A total of 30,072 transposon mutants were screened from 25 independent conjugation patches. The insertion mutants were visually screened based on differences in colony opacity or morphology relative to the recipient strain NWA19 and were classified into four distinct categories: translucent, intermediate, bull’s-eye and hyper opaque (Hill, 2021). From this mutant bank, nine (9) novel gas vesicle regulatory mutants (rbsB, orf1875, draG, cysB, sdhC, orf5465, ubiJ, orf6410, and fliD) were identified. The corresponding transposon insertions were transferred into a LacA (wild-type) background prior to further characterization. Here, we present the characterization of a mutant, CS109, with a transposon inserted in orf6410 which is predicted to encode an IclR-type transcriptional regulator.

The transposon was inserted 296 bp from the start codon and in the same direction of transcription of orf6410, resulting in a lacZ transcriptional fusion (Figure 1A). Interrogation of the region surrounding orf6410 showed that there are no predicted genes immediately upstream or downstream with which it could be co-transcribed. Since orf6410 is predicted to be monocistronic and is divergently transcribed from the genes immediately downstream, the transposon insertion should not have any polar effects which had been confirmed by genetic complementation assays (Supplementary Figure 1). The orf6410 gene product contains an N-terminal helix-turn-helix (HTH) and effector binding domains (Figure 1B) characteristic of IclR family transcriptional regulators (Molina-Henares et al., 2006). The transposon insertion is located within the effector binding domain of Orf6410. Based on the similarities, orf6410 will now be designated as iclR.

A BLASTP search revealed that the amino acid sequence of Orf6410 had the highest degree of similarity and identity to IclR family transcriptional regulators from several species of Brenneria and Pectobacterium (Supplementary Table 2). Notably, the predicted protein was also similar to XynR transcriptional regulators from Enterobacter hormaechei (66.4% identity, 82.0% similarity) and E. coli (65.6% identity, 80.8% similarity). In E. coli K12, XynR (formerly YagI) is encoded on the CP4-6 cryptic prophage and has been shown to regulate the bidirectional transcription units: yagA, encoding an uncharacterised transcription factor and yagEF, encoding a xylonate dehydratase and a 2-keto-3-deoxygluconate aldolase, respectively (Shimada et al., 2017). Comparison of this genomic region in E. coli K12 to S39006, E. hormaechei and P. parmentieri (Figure 1C) revealed a similar gene organization across the strains. E. coli xynR, yagE, yagF and yagG (encoding a D-xylonate transporter) have homologous genes in all three strains analyzed but the putative S39006 yagG is located in a different genomic region. Homologs of E. coli yagH (encoding a xylosidase/arabinosidase) are found in E. hormaechei and P. parmentieri but not in S39006. Furthermore, it was predicted that S39006 iclR may play a role in the regulation of a two-gene operon composed of orf6395 (encoding a putative dihydrodipicolinate synthase family protein) and orf6400 (YjhG/YagF family D-xylonate dehydratase) similar to XynR’s predicted regulation of yagEF in E. coli K12 (Figure 1D) (Shimada et al., 2017).

The iclR mutant was classified as an intermediate gas vesicle producer because the individual colonies appeared less opaque than the wild type on agar plates. Further, the culture spot of the transposon mutant appeared less opaque than the wild type but not translucent like the GV-defective GPA1 strain (Figure 2). The flotation ability of the mutants was also slightly decreased compared to the wild type, as seen in the static cultures where the air-liquid interface was more thinly populated by cells (Figure 2). PCM images of cells from culture spots and flotation assays further confirmed the gas vesicle phenotype of these mutants. Cultures of the intermediate gas vesicle producers were composed of a mixture of cells with and without gas vacuoles when viewed under PCM.

The GV gene cluster of S39006 is 16.6 kb long and composed of 19 open reading frames (ORFs) responsible for gas vesicle morphogenesis arranged in two non-overlapping operons, gvpA1-gvpY and gvrA-gvrC (Ramsay et al., 2011). The impact of the disruption of iclR on the transcription of the GV cluster was also investigated using gvpA1::uidA and gvrA::uidA chromosomal fusions. The activity of β-glucuronidase (the product of the uidA gene) was used as a proxy for transcription of each operon and monitored over the course of a growth curve under aerobic or microaerophilic conditions. Under aerobic conditions, the expression of the gvpA1-gvpY operon in the iclR mutant was 2.12-fold, 1.66-fold, and 1.65-fold lower compared to the wild type at 10, 12, and 14 h of growth, respectively. Similarly, the expression of the gvrA-gvrC operon was 1.54-fold, 1.47-fold, and 1.12-fold lower at the same time points (Figures 3A,B). A similar trend was observed under microaerophilic conditions; although the difference in the gvpA1-gvpY transcription between the wild type and the transposon mutant was not statistically significant (Figure 3C).

A series of phenotypic plate assays were performed with the iclR mutant to evaluate the pleiotropic effects of the mutation. Carbapenem production was greatly reduced in the mutant both in the plate spot tests and growth curve assay (Figure 4A). Results confirmed the significant reduction in carbapenem production in mutant with around a 2-fold decrease in the levels of antibiotic at the peak of its production (10 h) (Figure 4B). Production of prodigiosin and cellulase was also significantly reduced in the mutant (Figures 4C,D) while flagellar motility (swimming and swarming) was increased compared to the wild type (Figures 4E,F).

The impacts of the disruption of iclR on the virulence of S39006 in planta and in a C. elegans model were also established. Plant virulence of S39006 was assessed using potato rotting assays (Figure 5A). The mutant exhibited a significant reduction in virulence in planta with about a 3-fold decrease in the amount of rotten tissue after 5 days, compared to the wild type (Figure 5B). In addition, there was no significant difference in the viable cell counts recovered from the rotten tissues in both strains indicating that growth of the mutant is not impaired in potato tubers (Figure 5C). This reduction in plant virulence in the iclR mutant may be related to the decrease in its production of cellulase. Similarly, the mutant exhibited attenuated virulence against C. elegans with half the worms killed after 4 days and the entire population eliminated after 7 days. On average, worms fed with the mutant survived 1 day longer than with wild type which represents a slight, but statistically significant, reduction in virulence (Figure 5D).

To determine any further impacts caused by the disruption of iclR, we performed a quantitative proteomic analysis comparing the intracellular proteomes of wild type S39006 and the transposon mutant. Moreover, information on proteins with altered abundances in the iclR mutant was used to identify potential mechanisms through which various phenotypes were affected in S39006. Analysis of the differentially expressed proteins focused on those with adjusted p-values of less than 0.01 and a log₂FC cut-off of ±0.50. Using these parameters, 200 proteins in the iclR transposon mutant showed significantly altered abundances compared to the wild type. Of these, 81 proteins were downregulated (40.5%) while 119 proteins were upregulated (59.5%). Around 1,350 other proteins had a significant adjusted p-value (p < 0.01) but with log₂FC values not within the cut-off. While such smaller changes in expression in some proteins (e.g., transcriptional regulators) may still have important physiological impacts, only those with highly altered abundances were analyzed for the purpose of this study. A volcano plot showing the protein abundances (log₂FC) in the iclR mutant relative to the wild type as a function of statistical significance (−log of adjusted p-value) is presented in Figure 6.

Among the most downregulated proteins in the iclR mutant are those involved in gas vesicle and carbapenem biosynthesis, two key phenotypes that were attenuated in the mutant. Seventeen (17) of the 19 GV proteins were detected in the proteomic analysis, with log₂FC ranging from −1.19 (GvpC) to −0.30 (GvrB) (Supplementary Figure 2A). The gas vesicle structural components GvpA1 (log₂FC -0.97), GvpA3 (−1.05) and GvpC (−1.19) showed the most dramatic changes among the proteins encoded in the GV cluster. This is consistent with the decreased colony opacity and flotation ability observed in the iclR mutant. Furthermore, this also corroborated the observed reduction in transcription of the gvpA1-gvpY operon under aerobic conditions (Figure 3A). Expression of GV proteins from the gvrA-gvrC operon showed smaller, but still significant, changes ranging from −0.50 (GvpH) to −0.30 (GvrB). Unsurprisingly, the proteins involved in carbapenem synthesis (except for CarE which was not detected) and intrinsic resistance were all significantly less abundant in the mutant, with log₂FC between −2.4 and − 1.2 (Supplementary Figure 2B). The reduced expression of the carbapenem biosynthetic enzymes CarA (−2.2), CarB (−2.4), CarC (−2.2) and CarD (−1.4) explains the impaired antibiotic activity of the iclR mutant. Moreover, the proteins responsible for providing intrinsic resistance to the carbapenem, CarF (−1.5) and CarG (−1.5), were also less abundant in the mutant. Despite this, the mutant was still resistant to exogenous carbapenem (Supplementary Figure 3) implying that low levels of CarF and CarG are still sufficient to confer carbapenem resistance in S39006. The abundances of the prodigiosin biosynthetic enzymes were also lower in the transposon mutant compared to the wild type (Supplementary Figure 2C) as expected. Of the 14 proteins encoded in the prodigiosin operon, 12 were detected in the proteomics analysis. Of these, 11 biosynthetic enzymes were downregulated (log₂FC between −0.4 to −0.04) while one protein with unknown function was upregulated (PigK, log₂FC = 1.36). Although the log₂FC in the abundance of these enzymes are over the threshold of −0.5, they are still statistically significant except for PigA. The endoglucanase Orf13495 was also less abundant in the mutant by a log₂FC of −0.53. This corroborates the observed reduction of cellulase production in the plate assay (Figure 4D) and may explain the attenuated phytopathogenicity of the iclR mutant demonstrated in the potato rotting assays (Figure 5A).

In addition to proteins that were expected to be downregulated in the iclR mutant based on its confirmed phenotypes – GV, carbapenem, prodigiosin and cellulase production, several other proteins were identified to be less abundant in the analysis. The 30 proteins with the most reduced abundance (p < 0.01) in the mutant, other than those involved in GV and carbapenem production, are summarized in Supplementary Table 3. Two uncharacterised proteins encoded by orf12725 (−1.46) and orf12730 (−1.30) were also heavily downregulated in the iclR mutant. Other notable proteins that were downregulated in the mutant are the heat shock proteins IbpA (−0.73) and IbpB (−1.16) and the dimethylsulfoxide reductase components DmsA (−0.95) and DmsB (−0.90).

On the contrary, the most abundant proteins in the iclR mutant are involved in chemotaxis and flagellar assembly (Supplementary Table 4) which supports the increased swimming and swarming motilities of the mutant observed in the plate assays (Figures 4E,F). The 30 most abundant proteins apart from chemotaxis or flagellar proteins are listed in Supplementary Table 5. Three proteins were extremely upregulated in the mutant – a dihydrodipicolinate synthase family protein YagE (Orf6395, log₂FC 5.30), a D-xylonate dehydratase YagF (Orf6400, log₂FC 5.26) and a gluconate permease YjhF (Orf6405, log₂FC 7.38). As mentioned earlier, the two-gene yagEF operon is under the regulation of XynR in E. coli K12 (Figure 1D) (Shimada et al., 2017). A homologous operon is present in S39006 with a third gene, yjhF, located immediately downstream (Figure 1C). The significant increase in the abundance of YagE, YagF and YjhF in the iclR mutant provides strong evidence that yagEF-yjhF could be a single transcriptional unit under the repression of IclR.

Since IclR is a DNA-binding transcriptional regulator, it can also indirectly modulate expression of various genes via other regulators. The abundance of known regulators in the iclR mutant compared to the wild type are presented in Table 4. Several known regulators of gas vesicle and secondary metabolite production had significantly altered abundances including Rap (log₂FC − 0.55), RpoS (−0.54), PstA (−0.37), PigX (−0.31), SmaR (−0.25), RbsR (−0.25), RsmA (−0.23), PhoB (−0.11), PigP (0.19), PigZ (0.32), PigT (0.32) and PigU (0.35). Although these regulators modulate either by activation or repression of expression, the overall effect in the iclR mutant was that GV and carbapenem production was diminished while prodigiosin production was slightly, but still significantly, reduced. Among the known activators of carbapenem production, only Rap was downregulated while PigP, PigZ and PigU were all upregulated. On the other hand, the repressors of carbapenem RpoS, SmaR, RsmA and PhoB were all less abundant in the mutant. Similarly, Rap was the only activator of prodigiosin synthesis that was less abundant while PigP, PigT and PigU were all upregulated. The repressors of prodigiosin synthesis RpoS, PstA, PigX and RsmA were upregulated and only PigZ was downregulated. Lastly, the GV activators Rap, RsmA and RbsR were both less abundant in the iclR mutant compared to the wild type which could have contributed to the decreased expression of the GV cluster. RpoS represses swimming motility, hence its downregulation could explain the increased swimming ability of the mutant. Consistently, the swimming motility activators PigZ, PigT and PigU were all upregulated. Swarming motility is known to be repressed by Rap, RpoS, PigS and RsmA which were all less abundant in the iclR mutant while the repressors PigT and RhlA were both upregulated. Several activators of cellulase production also showed altered abundances in the mutant. RpoS and RbsR were downregulated and PigP and PigZ were only slightly upregulated in the mutant.

Differential proteomic analysis revealed that the cognate proteins of yagE, yagF and yjhF were extremely upregulated in the iclR transposon mutant. This strongly suggested that yagEF-yhjF locus could be an operon possibly under the control of IclR, similar to XynR in E. coli K12 (Shimada et al., 2017). The target site of XynR had been identified to be the palindromic sequence TGTTCAacattatAGAACA in E. coli K12 (Shimada et al., 2017). The Finding Individual Motif Occurrences (FIMO) program of the MEME Suite (Bailey et al., 2009; Grant et al., 2011) was used to scan for high-scoring motif occurrences of the consensus XynR box in the upstream sequences in the S39006 genome. Putative XynR binding sites were found in the upstream sequences of 76 genes (Supplementary Table 6). Some of the relevant putative XynR binding sequences are summarized in Figure 7A. Unsurprisingly, a XynR box was found upstream of yagE which supports the proposed role of XynR as a repressor of the putative yagEF-yjhF operon in S39006. A possible XynR binding site is also located upstream orf17610, a yagE homolog, which also encodes a dihydrodipicolinate synthase family protein (Figure 7B). Furthermore, a putative binding site was also found in the open reading frame of iclR which suggests that IclR might regulate its own synthesis (Figure 8A). Autoregulation is a typical feature of IclR transcriptional regulators in bacteria (Molina-Henares et al., 2006). XynR boxes were also found upstream of a wide variety of genes encoding methyl-accepting chemotaxis proteins (orf5750 and orf11345), transporters (orf5180, orf6375, orf10785, orf11825, orf13740, orf15415 and orf16295) and enzymes (hrpA, parC, orf2665, orf5700, orf9200, orf10320 and orf19225).

Further analysis of the sequence directly upstream of yagE confirmed the presence of promoter elements and a XynR box overlapping with the −10 site (Figure 7C). To investigate if IclR represses yagEF-yjhF transcription by binding the yagE promoter (yagE_prom), a 150-bp fragment containing the yagE_prom was cloned into the β-gal transcriptional fusion vector pRW50, creating pRW50-yagE_prom. DH5α cells were transformed with either pRW50 or pRW50-yagE_prom then the resulting strains were subsequently transformed with pQE80-oriT or pQE80-iclR. The constructed strains were grown in LB supplemented with the appropriate antibiotics and 0.1 mM IPTG and β-gal activity was monitored over the course of a growth curve (Figure 8A). Results showed that the yagE_prom is able to drive expression of the promoter-less lacZ gene as evidenced by the β-gal activity detected in the DH5α (pRW50-yagE_prom) (pQE80-oriT). However, when IclR was expressed in trans in the DH5α (pRW50-yagE_prom) (pQE80-oriT-xynR), the enzyme activity significantly dropped during exponential phase and was no longer detected throughout growth. This indicates that IclR could be directly repressing yagE transcription by binding to the XynR box within the promoter region.

Finally, to confirm that IclR binds the yagE promoter and investigate if it can bind its own ORF, gel shift assays were performed following the LUEGO method (Jullien and Herman, 2011). Cyanine-5 (Cy-5)-LUEGO-labeled yagE or iclR probes were mixed with increasing amounts of purified IclR and separated on a polyacrylamide gel. The mobility of the yagE probe on the gel was reduced and a complete band shift was achieved with 0.8 μM IclR (protein to DNA molar ratio of 32:1) (Figure 8B). This confirms that IclR represses transcription of the yagEF-yjhF operon by directly binding the yagE promoter. On the contrary, the mobility of the iclR probe was unchanged across the protein concentrations tested (Figure 8C) indicating that iclR transcription is unlikely to be negatively autoregulated.