ABSTRACT

   The monothiol glutaredoxin GrxD plays an essential role in the
   biosynthesis of the antibiotic 2,4-diacetylphloroglucinol (2,4-DAPG)
   and the biocontrol capacity of the soil bacterium Pseudomonas
   fluorescens 2P24. However, the detailed mechanism underlying
   GrxD-mediated activation of the production of 2,4-DAPG remains unclear.
   Here, we found that GrxD directly interacted with IbaG, a BolA protein
   family member. The mutation of ibaG significantly decreased 2,4-DAPG
   production. Furthermore, expressing ibaG restored the production of
   2,4-DAPG in the grxD ibaG double mutant to wild-type levels in the
   presence of dithiothreitol, suggesting that IbaG was required for
   GrxD-mediated regulation of 2,4-DAPG production. Transcriptome
   sequencing analyses revealed that IbaG plays a global role in gene
   regulation by affecting the expression of numerous genes throughout the
   genome. We also demonstrated that IbaG is an important regulator of
   several cellular processes, including swarming motility, biofilm
   formation, siderophore production, and acid resistance. Altogether, our
   data suggest that IbaG has an essential role in 2,4-DAPG production,
   motility, and biofilm formation. We also propose a regulatory mechanism
   linking GrxD to 2,4-DAPG production via IbaG.

IMPORTANCE

   The production of 2,4-diacetylphloroglucinol (2,4-DAPG) is positively
   influenced by the monothiol glutaredoxin GrxD in Pseudomonas
   fluorescens 2P24. However, the regulatory mechanism underlying
   GrxD-mediated regulation of 2,4-DAPG biosynthesis is mostly
   uncharacterized. Here, we show the function of the BolA-like protein
   IbaG in 2,4-DAPG biosynthesis. We also demonstrate that GrxD directly
   interacts with IbaG and influences the redox state of IbaG. Altogether,
   this work provides new insights into the role of the highly conserved
   IbaG protein in regulating 2,4-DAPG synthesis, biofilm formation, and
   other biocontrol traits of P. fluorescens.

   KEYWORDS: Pseudomonas fluorescens; 2,4-DAPG; monothiol glutaredoxin
   GrxD; BolA; IbaG

INTRODUCTION

   Pseudomonas fluorescens belongs to a group of ubiquitous beneficial
   rhizobacteria that play important roles in inhibiting the growth of
   phytopathogens, promoting plant growth, and eliciting plant immune
   responses ([32]1). The production of secondary metabolites with
   antibiotic activity is crucial for P. fluorescens to exert its
   biocontrol functions. Among these secondary metabolites,
   2,4-diacetylphloroglucinol (2,4-DAPG) has received special attention
   because of its broad-spectrum antimicrobial activity ([33]2). The gene
   cluster involved in 2,4-DAPG biosynthesis comprises the phlACBD operon.
   Specifically, the phlD gene encodes a type III polyketide synthase,
   which catalyzes the formation of phloroglucinol (PG) from malonyl-CoA.
   Monoacetylphloroglucinol (MAPG) acetyltransferase, which is encoded by
   phlACBD, acetylates PG to MAPG and 2,4-DAPG ([34]3).

   The biosynthesis of 2,4-DAPG is regulated by biotic and abiotic factors
   in response to various environmental cues. The pathway-specific
   regulators PhlF and PhlH are TetR-like repressors that inhibit the
   transcription of phlACBD and phlG, respectively. They bind directly to
   their target promoter region to regulate 2,4-DAPG metabolism ([35]2,
   [36]4). Additionally, the Gac/Rsm signaling cascade plays a vital role
   in regulating 2,4-DAPG biosynthesis in P. fluorescens at the
   post-transcriptional level ([37]5). The GacS/GacA two-component system
   induces the transcription of small regulatory RNAs, e.g., RsmX, RsmX1,
   RsmY, and RsmZ, which have high affinity for the CsrA protein family
   ([38]6). In Pseudomonas spp., the members of the CsrA family, RsmA and
   RsmE, interact directly with their target mRNAs to affect mRNA
   stability or alter translational efficiency. It has been hypothesized
   that the distribution of a conserved 5′-CANGGAYG-3′ sequence motif,
   which overlaps with the ribosome binding site, is essential for
   RsmA/CsrA proteins to bind their target mRNA and regulate translation
   ([39]7). For example, a previous study has shown that RsmA binds
   directly to phlA mRNA and influences its expression ([40]6). Finally,
   various carbohydrate compounds, low temperatures, and metabolite
   exudates released by fungal hyphae and plant roots affect the
   production of 2,4-DAPG. In P. protegens Pf-5, the production of
   2,4-DAPG is co-regulated with the biosynthesis of another antibiotic
   called pyoluteorin by PG ([41]8).

   P. fluorescens 2P24 was originally isolated from take-all decline soil.
   It suppresses a variety of soilborne diseases caused by plant
   pathogens, including Rhizoctonia solani, Gaeumannomyces graminis var.
   tririci, and Ralstonia solanacearum ([42]9). Genetic analyses showed
   that 2,4-DAPG is the key biocontrol component of strain 2P24.
   Additionally, strain 2P24 is an efficient root colonizer that can
   durably colonize plant rhizosphere. Root colonization is a complicated
   process involving interactions between plant growth-promoting
   rhizobacteria, phytopathogens, and host plants ([43]10). Reactive
   oxygen species produced by plant cells negatively affect rhizosphere
   colonization by P. fluorescens. Recently, we showed that the monothiol
   glutaredoxin GrxD is responsible for 2,4-DAPG production, oxidative
   stress tolerance, and other biocontrol traits of strain 2P24 ([44]11).
   GrxD is involved in the biosynthesis of iron-sulfur (Fe-S) clusters and
   the reduction of thiol-disulfide exchange reaction in a
   glutathione-dependent manner. The conserved monothiol motif (CGFS) is
   essential for the function of GrxD ([45]12). Furthermore, evidence
   shows that GrxD directly interacts with the proteins MiaB, Aft1, Php4,
   HapX, and BolA to influence redox biology and iron metabolism ([46]13).
   However, the detailed mechanism through which GrxD controls 2,4-DAPG
   production remains unclear.

   BolA protein is widely distributed in prokaryotes and eukaryotes and
   has been associated with a range of cellular processes, including
   biofilm formation, oxidative stress tolerance, bacterial motility, and
   membrane permeability ([47]14). BolA mediates alterations in bacterial
   permeability by directly binding to the promoter regions of dacA, dacC,
   and mreB, which are involved in the regulation of OmpF/OmpC balance
   ([48]15). Transcriptomic analyses have indicated that BolA recognizes
   and binds its target genes, which have the 5′-YYGCCAGH-3′ consensus
   sequence. Escherichia coli encodes two proteins of the BolA family,
   i.e., BolA and IbaG. Notably, BolA and IbaG have different functions:
   BolA is thought to affect cell shape, whereas IbaG does not ([49]16).
   IbaG also plays an important role in bacterial growth, cell morphology,
   and acidic tolerance response. The BolA protein family has been well
   characterized in various bacterial pathogens, such as E. coli,
   Klebsiella pneumoniae, Salmonella enterica serovar Typhimurium, and
   Vibrio cholerae ([50]17 [51]– [52]19), and studies have shown that the
   BolA protein family regulates bacterial pathogenicity. However, the
   role of BolA in the plant beneficial of the P. fluorescens group
   remains to be investigated.

   Here, we show that the BolA-like protein IbaG plays a critical role in
   coordinating the expression of plant-beneficial traits in P.
   fluorescens 2P24. We found that IbaG directly interacted with GrxD and
   controlled 2,4-DAPG production. Moreover, we demonstrated the
   pleiotropic role of IbaG in regulating cell motility, siderophore
   production, and acid tolerance. Altogether, our data provide novel
   insights into the function of IbaG in regulating biocontrol traits of
   the plant-beneficial bacterium P. fluorescens 2P24.

RESULTS

Identification of BolA homolog in P. fluorescens

   We have previously shown that the monothiol glutaredoxin GrxD
   positively regulates 2,4-DAPG biosynthesis, and bioinformatic analyses
   unraveled a potential link between monothiol glutaredoxin and BolA-like
   proteins ([53]11, [54]20). To assess the possible role of BolA in
   regulating the production of 2,4-DAPG mediated by GrxD, we used the
   amino acid sequence E. coli K-12 (GenBank accession no. [55]APC50728.1)
   to search for BolA homologs in P. fluorescens using the psi-BLAST
   algorithm. We found two proteins homologous to BolA of E. coli.
   Sequence analyses revealed that C0J56_08545 and C0J56_05010 had 42% and
   15% sequence identity, respectively, with BolA of E. coli. Moreover,
   they exhibited 26% and 32% sequence identity, respectively, with IbaG
   of E. coli. Thus, we named C0J56_08545 and C0J56_05010 of P.
   fluorescens BolA and IbaG, respectively, based on their homologs in E.
   coli. Structural predications of P. fluorescens BolA and IbaG based on
   E. coli BolA protein (PBD ID 2DHM) indicated 100% coverage and 99%
   confidence ([56]Fig. 1; Fig. S1). The predicted structures of BolA and
   IbaG had a α1β1β2α2α3β3α4 topology, which is characteristic of the BolA
   protein family ([57]Fig. 1; Fig. S1).

Fig 1.

   [58]Fig 1
   [59]Open in a new tab

   Sequence comparison and secondary structure prediction of BolA-like
   proteins. (A) Clustal omega multiple sequence alignment of P.
   fluorescens IbaG protein with BolA-like proteins from other bacteria.
   Identical residues (asterisks), conserved residues (colons), and
   semi-conserved residues (dots) are indicated. The secondary structure
   of the P. fluorescens IbaG protein is shown above the alignment. (B)
   Computer model of the 3D structure predicted for P. fluorescens IbaG
   using the I-TASSER server.

IbaG is required for 2,4-DAPG biosynthesis in P. fluorescens

   To determine whether BolA and IbaG proteins are involved in 2,4-DAPG
   biosynthesis in P. fluorescens, we constructed bolA and ibaG mutants.
   High-performance liquid chromatography (HPLC) assays showed that the
   production of 2,4-DAPG significantly decreased in the ibaG mutant
   compared with that in wild-type P. fluorescens. Complementation with
   the wild-type ibaG gene restored 2,4-DAPG production to the wild-type
   level ([60]Fig. 2A; Fig. S2). In contrast, mutation of the bolA gene
   did not affect 2,4-DAPG biosynthesis ([61]Fig. 2A). Furthermore, the
   ibaG mutant did not inhibit mycelial growth of R. solani, whereas the
   bolA mutant retained an inhibitory activity similar to that of
   wild-type P. fluorescens ([62]Fig. 2B). Next, we examined the effects
   of BolA and IbaG on PhlA protein expression. PhlA protein levels were
   significantly decreased in the ibaG mutant compared with those in the
   wild-type strain. However, deletion of bolA did not affect PhlA protein
   expression, which was similar to that of the wild-type strain 2P24
   ([63]Fig. 2C). In addition, the growth of strain 2P24 and its
   derivatives was analyzed in KBG broth, and the growth characteristics
   of the bolA and ibaG mutants were indistinguishable from those of the
   wild-type strain (Fig. S3). Altogether, these data suggested that ibaG,
   but not bolA, was required for 2,4-DAPG biosynthesis of P. fluorescens.

Fig 2.

   [64]Fig 2
   [65]Open in a new tab

   The role of IbaG in the production of 2,4-DAPG in P. fluorescens. (A)
   Quantification of 2,4-DAPG by HPLC analysis in strain 2P24 and its
   derivatives. (B) Inhibition of mycelial growth of Rhizoctoia solani by
   2P24 (pRK415), the ibaG mutant (pRK415), the ibaG mutant (p415-ibaG),
   and the bolA mutant (pRK415). (C) Western blot analysis was performed
   to detect the levels of PhlA-FLAG in strain 2P24, the bolA mutant, and
   the ibaG mutant. The relative intensities of the PhlA-FLAG bands were
   marked below the bolt. All experiments were performed in triplicate,
   and * represents significant difference (P < 0.05).

IbaG interacts with the GrxD

   Grx-BolA interactions occur in the model plant Arabidopsis thaliana,
   and these interactions can influence the redox state of cells ([66]21).
   To investigate whether IbaG interacted with GrxD in P. fluorescens, we
   used the bacterial adenylate cyclase two-hybrid system. E. coli BTH101
   co-expressing GrxD and IbaG fusion proteins exhibited significantly
   higher β-galactosidase activity than the negative control, suggesting
   an interaction between IbaG and GrxD ([67]Fig. 3A). A similar activity
   was observed when BolA was used as bait and GrxD as prey ([68]Fig. 3A).

Fig 3.

   [69]Fig 3
   [70]Open in a new tab

   Bacterial adenylate cyclase-based two-hybrid analysis of the
   interactions between BolA/IbaG and GrxD. (A) Quantitative
   β-galactosidase assays of E. coli BTH101 harboring both pKT25- and
   pUT18C-derived plasmids. Cells were cultured at 30°C for 8 h and
   β-galactosidase activity was determined as Miller units. (B)
   Mutagenesis of the CGFS motif of GrxD or the conserved C26 of IbaG
   alters the interactions between IbaG and GrxD. All experiments were
   performed in triplicate, and * represents significant difference (P <
   0.05). Zip indicates the positive interaction control.

   GrxD catalyzes the direct reduction of protein disulfides in a
   thiol-disulfide exchange reaction ([71]12). Sequence analyses suggested
   that the C26 residue of IbaG was mostly conserved among different
   Pseudomonas spp. strains ([72]Fig. 1). Therefore, we hypothesized that
   the C26 residue of IbaG is important for the GrxD-IbaG interaction. To
   test this hypothesis, we changed the C26 residue to serine and used the
   mutant version of the protein in the interaction assay. The mutation of
   the C26 residue of IbaG into serine abolished the interaction between
   GrxD and IbaG ([73]Fig. 3B). Nuclear magnetic resonance spectroscopy
   and X-ray crystallography analyses of GrxD ([74]22) have revealed that
   the CGFS motif is required for the function of GrxD. We found that the
   CGFS motif of GrxD was necessary for GrxD-IbaG interaction ([75]Fig.
   3B). Altogether, our data suggested that the GrxD-IbaG interaction was
   specific, and the C26 residue of IbaG and the CGFS motif of GrxD were
   critical for this interaction.

IbaG is required for GrxD-mediated downregulation of 2,4-DAPG production

   Given that GrxD negatively regulates 2,4-DAPG production in P.
   fluorescens, we assessed whether IbaG was involved in this function of
   GrxD. We constructed a grxD ibaG double mutant and measured the
   production of 2,4-DAPG. The grxD ibaG double mutant, as well as the
   grxD mutant or the ibaG mutant, produced low levels of 2,4-DAPG
   ([76]Fig. 2A). We then transformed the grxD ibaG double mutant with the
   wild-type ibaG or grxD gene and tested whether any of these two genes
   restored antibiotic production. HPLC assays showed that complementation
   of the grxD mutant with wild-type grxD restored 2,4-DAPG production to
   the wild-type level. Neither ibaG nor grxD restored 2,4-DAPG production
   in the ibaG grxD double mutant ([77]Fig. 2A). Previous studies have
   shown that GrxD plays an important role in thiol-disulfide homeostasis.
   Thus, we hypothesized that GrxD affected the redox state of IbaG, which
   then regulated 2,4-DAPG production. To test this hypothesis, we
   cultured bacterial cells in LB medium containing 5 mM DTT
   (dithiothreitol), which is known to affect the redox state of proteins
   ([78]12). Expression of IbaG, but not the mutant version IbaG^C26S, in
   the ibaG grxD double mutant restored the production of 2,4-DAPG to
   levels observed in strain 2P24 in the presence of DTT ([79]Fig. 2A).

   We have previously shown that RsmA and RsmE negatively regulate
   2,4-DAPG production ([80]6). Thus, we assessed the effects of ibaG on
   the expression of rsmA and rsmE. Translational fusion assays and
   real-time quantitative PCR (RT-qPCR) analyses showed that deletion of
   ibaG significantly increased the expression levels of rsmA and rsmE
   ([81]Fig. 4; Fig. S4). These results were consistent with our previous
   data, showing that GrxD negatively regulates RsmA and RsmE ([82]11).
   Therefore, IbaG likely regulated 2,4-DAPG production by influencing
   rsmA/rsmE expression in strain 2P24. Altogether, these findings showed
   that IbaG was required for GrxD-mediated downregulation of RsmA and
   RsmE, and 2,4-DAPG production. Moreover, GrxD likely regulated the
   function of IbaG by influencing its redox states.

Fig 4.

   [83]Fig 4
   [84]Open in a new tab

   Effect of IbaG on the translational expression of rsmA and rsmE in P.
   fluorescens 2P24. Expression of rsmA (A) and rsmE (B) was measured in
   2P24 (pRK415), the ibaG mutant (pRK415), and the ibaG mutant
   (p415-ibaG). The strains were inoculated on the LB medium plates
   supplemented with X-gal. The β-galactosidase activity was determined as
   Miller units. All experiments were performed in triplicate, and *
   represents significant difference (P < 0.05).

Identification of IbaG-regulated genes in P. fluorescens

   To identify genes, whose expression depended on IbaG, RNA sequencing
   analyses (RNA-seq) were performed to characterize the transcriptome of
   strain 2P24 and the ibaG mutant. A total of 705 genes were
   differentially expressed in the ibaG mutant compared with the wild-type
   strain 2P24 (366 genes were downregulated and 339 were upregulated)
   ([85]Fig. 5). All differentially expressed genes (DEGs) with known
   functions are summarized in [86]Table S1. The functional classification
   of the identified genes suggested that the DEGs were mainly involved in
   fatty acid biosynthesis, amino acid metabolism, and carbon metabolism
   ([87]Fig. 5B and C). The genes responsible for phage assembly
   (C0J56_06120, C0J56_06140, C0J56_06105, and C0J56_06165) were all
   upregulated in the ibaG mutant. Consistent with these results, the
   transcript levels of C0J56_06120, C0J56_06140, C0J56_06105, and
   C0J56_06165 were significantly higher in the ibaG mutant than those in
   the wild-type strain (Fig. S4). These data suggested that IbaG silenced
   prophage genes, thus preventing the activation of and lysis by
   prophages.

Fig 5.

   [88]Fig 5
   [89]Open in a new tab

   Comparative transcriptomics of the ibaG mutant normalized to those of
   strain 2P24. (A) The volcano plot of the DEGs was analyzed in strain
   2P24 and its ibaG mutant by RNA-seq. Red dots mean up-regulated. Green
   dots mean down-regulated. (B) IbaG regulon categorized by the Gene
   ontology (GO) functional enrichment analysis. (C) Kyoto Encyclopedia of
   Genes and Genomes (KEGG) enrichment analysis of detected DEGs.

   The BolA protein family is involved in stress response and bacterial
   motility. Our transcriptomic results also showed that several stress
   response-associated genes and genes encoding stress proteins, including
   C0J56_16570, C0J56_24075, and C0J56_19920, were significantly
   downregulated in the ibaG mutant. Genes responsible for flagellar
   assembly (FlgF and FlgC) and the chemotaxis response regulator CheY
   were also downregulated in the ibaG mutant. Additionally, the ibaG
   regulon includes genes involved in Fe-S cluster biosynthesis and
   assembly, secretion system, amino acid metabolism, transporter, signal
   transduction, and other functions.

   To verify the RNA-seq data, we randomly selected four genes (cheY,
   iscR, gabP, and purU) from the identified DEGs. Transcriptional fusion
   assays indicated that the expression of cheY and iscR significantly
   decreased, whereas that of gabP and purU significantly increased in the
   ibaG mutant compared with that in the wild-type strain, thus validating
   the RNA-seq data (Fig. S5).

   Altogether, our transcriptome analysis suggested that IbaG plays a
   global role in transcriptional regulation and can influence different
   cellular processes, including Fe-S cluster biosynthesis, flagella
   assembly, motility, and stress response.

Effect of IbaG on acid resistance and siderophore production in P.
fluorescens

   Given that IbaG of E. coli influences acid tolerance ([90]16), we
   explored whether the absence of ibaG impaired P. fluorescens survival
   following exposure to acidic stress. Our data showed that the survival
   of the ibaG mutant was significantly lower than that of the wild-type
   and complemented strains ([91]Fig. 6A). Notably, the mutation of bolA
   did not affect survival under an acid stress condition ([92]Fig. 6A).
   These data suggested the important role of ibaG in growth under acidic
   conditions.

Fig 6.

   [93]Fig 6
   [94]Open in a new tab

   Effect of IbaG on the acid tolerance and siderophore production of P.
   fluorescens. (A) Acid tolerance was determined by calculating the
   proportion of cells that survived during growth in acidic conditions
   (LB medium at pH 5.5) versus in normal conditions (LB medium at pH
   7.0). (B) The production of siderophore of strain 2P24 and its
   derivatives was determined by measuring the diameter of the orange
   circle on the chrome azurol S (CAS) plates. The experiments were
   performed in triplicate, and * represents significant difference (P <
   0.05).

   Siderophore production by strain 2P24 is partially responsible for
   promoting plant growth ([95]9). A previous study has shown the
   downregulation of siderophore production in the grxD mutant ([96]11).
   Therefore, we compared siderophore production in the wild-type and ibaG
   mutant strains and found that the ibaG mutant produced less siderophore
   than the wild-type and bolA mutant strains ([97]Fig. 6B). This result
   is consistent with the RNA-seq data that showed that the levels of
   pvdP, pvdE, and C0J56_20220, which encoded proteins involved in the
   synthesis of a primary siderophore, significantly decreased in the ibaG
   mutant ([98]Table S1). We further analyzed the effects of IbaG on pvdP,
   pvdE, and C0J56_20220 expression using RT-qPCR and confirmed a
   significant induction of these genes (Fig. S4). Collectively, these
   results demonstrate that IbaG is involved in siderophore production in
   P. fluorescens.

IbaG is required for swimming motility and root colonization of P.
fluorescens

   Motility is a key bacterial trait for root colonization and is probably
   involved in biocontrol activities. We compared the swimming motility of
   strain 2P24 and ibaG mutant. The swimming motility of the ibaG mutant
   was significantly reduced compared with that of the wild-type strain
   ([99]Fig. 7A). Complementation of the ibaG mutant with p415-ibaG
   restored the swimming motility, suggesting that ibaG positively
   regulated swimming motility. However, the mutation of bolA did not
   influence bacterial swimming motility ([100]Fig. 7A). We then
   investigated whether the loss of swimming motility in the ibaG mutant
   was due to the influence of the growth or cellular morphology of strain
   2P24 and found no significant difference between the ibaG mutant and
   wild-type strain (Fig. S3).

Fig 7.

   [101]Fig 7
   [102]Open in a new tab

   Swimming motility and colonization of wheat rhizosphere by P.
   fluorescens 2P24 and its derivatives. (A) The colony diameter of strain
   2P24 and its derivatives was measured in swimming assay. (B) The
   rhizosphere population of strain 2P24 (pRK415), the ibaG mutant
   (pRK415), the ibaG mutant (p415-ibaG), and the bolA mutant (pRK415)
   were determined at 14 day after inoculation. All experiments were
   performed in triplicate, and * represents significant difference (P <
   0.05).

   To determine the contribution of IbaG to the root colonization of P.
   fluorescens 2P24, we conducted wheat root colonization assays under
   greenhouse conditions. Strain 2P24 and its derivatives were introduced
   into sterile soil. The ability of the ibaG mutant to colonize the root
   rhizosphere was significantly impaired 14 days after inoculation
   compared with that of the wild-type and bolA mutants ([103]Fig. 7B).
   Thus, the ibaG gene is important for rhizosphere colonization of P.
   fluorescens.

DISCUSSION

   P. fluorescens 2P24 is an effective biocontrol agent that can prevent
   several soilborne diseases by directly inhibiting the pathogen’s growth
   through the production of 2,4-DAPG ([104]9). Rhizosphere colonization
   is required for strain 2P24 to protect plants from phytopathogens.
   Colonization by strain 2P24 relies on a complicated interaction between
   the plant, pathogens, and environmental stimuli ([105]23). We recently
   reported that GrxD is required for the production of 2,4-DAPG and
   rhizosphere colonization by strain 2P24 ([106]11), although the
   underlying molecular mechanism remains largely elusive. Here, we report
   that GrxD regulated the production of 2,4-DAPG likely by modulating the
   activity of the BolA family protein IbaG in P. fluorescens 2P24. This
   conclusion is supported by the following results: (i) the ibaG mutant,
   like the grxD and ibaG grxD double mutants, produced undetectable
   amounts of 2,4-DAPG; (ii) GrxD protein directly interacted with IbaG;
   (iii) the C26 of IbaG and the C29 of GrxD were needed for IbaG-GrxD
   interaction; (iv) expression of a functional IbaG, but not GrxD,
   restored 2,4-DAPG production in the ibaG grxD double mutant; (v) DTT,
   which mimics the function of GrxD, was needed for the function of IbaG.
   DTT is a known strong reducing agent that can block the formation of
   disulfide bonds. A previous study has shown that the redox shift
   pattern of At3g11630 of Arabidopsis thaliana changes in the presence of
   DTT ([107]24). In Vibrio cholerae, TcpP forms a dimer through an
   intermolecular disulfide bond, and the dimer formation is abolished
   with 10 mM DTT ([108]25). In Saccharomyces cerevisiae, cytosolic Grx3/4
   influences the maturation of Fe-S clusters by interfering with the
   function of the BolA-like proteins Fra1 and Fra2, which regulate the
   expression of the iron transport system ([109]26). Moreover, the
   function of the transcriptional regulator OxyR is regulated by Grx1
   through the formation of a disulfide bond in E. coli ([110]27).
   Therefore, GrxD directly interacts with IbaG and influences the redox
   state of IbaG which regulates the production of 2,4-DAPG. BolA-like
   proteins have been extensively studied in prokaryotes and eukaryotes
   and were shown to play important roles in different bacterial
   physiological processes, including virulence-related traits. To our
   knowledge, the present study is the first to report that a BolA family
   protein, specifically IbaG of P. fluorescens 2P24, regulates antibiotic
   production of plant-associated beneficial bacteria.

   We investigated further the mechanism underlying the regulation of
   2,4-DAPG production by IbaG. We found that mutation of ibaG increased
   the translation of RsmA and RsmE but decreased the levels of the
   2,4-DAPG biosynthesis-associated gene phlA ([111]Fig. 2C). These
   results are consistent with a previous report showing that RsmA/E
   directly bind to the 5′ end untranslated region of phlA mRNA to block
   its translation ([112]6). Overall, these data suggest that GrxD
   activates IbaG, which downregulates the expression of RsmA/E,
   consequently promoting the translation of genes involved in 2,4-DAPG
   biosynthesis and activating the 2,4-DAPG production. Phosphorylation of
   BolA is critical for the function of BolA in E. coli. Specifically,
   four phosphorylation sites (S26, S45, T81, and S95) are responsible for
   the stability of BolA protein ([113]28). Future work is needed to
   explore whether post-translational modifications of IbaG in P.
   flurescens 2P24 are a regulating IbaG control of the expression of its
   target genes.

   Although there is a significant homology among BolA-like proteins, our
   results showed that the functions of IbaG and BolA in regulating
   biocontrol-related traits of P. fluorescens differ. Protein-protein
   interaction assays showed that BolA and IbaG interacted with GrxD
   ([114]Fig. 3), whereas IbaG, but not BolA, was involved in the
   regulation of biocontrol-related traits of strain 2P24. For example,
   mutation of ibaG reduced the production of 2,4-DAPG, which was not
   different between the bolA mutant and wild-type strains ([115]Fig. 1A).
   Moreover, siderophore production, bacterial swimming motility, and acid
   tolerance were significantly decreased in the ibaG mutant but not in
   the bolA mutant ([116]Fig. 6 and 7). Transcriptome data suggested that
   IbaG direct effects were related to the regulation of the expression of
   flagellar and stress response-associated genes ([117]Table S1).
   Additionally, inducing the expression of BolA from a plasmid failed to
   complement the phenotype of the ibaG mutant (data not shown),
   suggesting that IbaG and BolA proteins have different functions in P.
   fluorescens. Such functional differences among the BolA-like proteins
   have previously been reported in E. coli, V. cholerae, and
   Sinorhizobium meliloti, indicating that BolA and IbaG may be involved
   in different regulatory pathways to allow adaptation to the constantly
   changing environment ([118]16, [119]18, [120]29).

   A previous study showed higher cellular levels of cyclic-di-GMP
   (c-di-GMP) in an E. coli bolA mutant ([121]19). The secondary messenger
   c-di-GMP plays a pleiotropic role in many biological processes. In
   strain 2P24, c-di-GMP was reported to regulate the biocontrol traits,
   including 2,4-DAPG production, swimming motility, and biofilm formation
   ([122]30). However, our transcriptome data did not reveal any
   difference in the expression of genes involved in c-di-GMP metabolism
   between the ibaG mutant and wild-type strain, suggesting that IbaG
   modulated 2,4-DAPG biosynthesis independently from the c-di-GMP
   regulatory pathway ([123]Table S1).

   Overall, our results demonstrated that the ibaG gene is involved in
   2,4-DAPG production mediated by GrxD, the swimming motility, and the
   acid tolerance of P. fluorescens. This finding highlights the
   complexity of the regulation of 2,4-DAPG production and provides new
   insights for fine-tuning the biocontrol traits of P. fluorescens under
   adverse environmental conditions.

MATERIALS AND METHODS

Bacterial strains and culture conditions

   The strains and plasmids used in this study are listed in [124]Table 1.
   P. fluorescens 2P24 and E. coli DH5α strains were routinely grown in
   lysogeny broth (LB), KB, or ABM medium at 28°C. Antibiotics were
   supplemented when required for plasmid maintenance or transformation at
   the following concentrations: ampicillin (50 µg/mL), kanamycin (50
   µg/mL), and tetracycline (20 µg/mL).

TABLE 1.

   Bacterial strains and plasmids used in this study
   Strains, plasmids, or oligonucleotides Relevant characteristics[125] ^a
   Reference or source
   Strains
    P. fluorescens
     2P24 Wild type, Ap^r ([126]23)
     WPM31 In-frame deletion of grxD, Ap^r ([127]11)
     WPM81 In-frame deletion of ibaG, Ap^r This work
     WPM82 In-frame deletion of bolA, Ap^r This work
     WPM83 Double deletion of ibaG and grxD, Ap^r This work
     WPM28 Strain 2P24 with a FLAG epitope sequence tagged to the C
   terminus of PhlA, Ap^r ([128]6)
     WPM84 WPM81 with a FLAG epitope sequence tagged to the C terminus of
   PhlA, Ap^r This work
     WPM85 WPM82 with a FLAG epitope sequence tagged to the C terminus of
   PhlA, Ap^r This work
    Rhizoctonia solani Basidiomycete fungus caused cotton blight Lab stock
    E. coli
     DH5α supE44 lacU169 (ϕ80lacZ M15) hsdR17 recA1 endA1 gyrA96 thi-1
   relA1 ([129]31)
     BTH101 F-, cya-99, araD139, galE15, galK16, rpsL1 (Strr), hsdR2,
   mcrA1, mcrB1 Euromedex
   Plasmids
     p2P24Km Sucrose-based counter-selectable plasmid, Km^r ([130]6)
     p2P24Km-ibaG Plasmid p2P24Km carrying a deleted ibaG gene, Km^r This
   work
     p2P24Km-bolA Plasmid p2P24Km carrying a deleted bolA gene, Km^r This
   work
     pKT25 pSU40 derivative with T25 fragment of CyaA, Km^r Euromedex
     pUT18C pUC19 derivative with T18 fragment of CyaA, C-terminal
   fusions, Ap^r Euromedex
     pKT25-grxD pKT25 containing the grxD gene, Km^r This work
     pUT18C-ibaG pUT18C containing the ibaG gene, Ap^r This work
     pKT25-ibaG pKT25 containing the ibaG gene, Km^r This work
     pUT18C-grxD pUT18C containing the grxD gene, Ap^r This work
     pKT25-grxDC29S pKT25 containing grxD ^C29S, Ap^r This work
     pUT18C-ibaGC26S pKT25 containing ibaG ^C26S, Ap^r This work
     pUT18C-grxDC29S pKT25 containing grxD ^C29S, Ap^r This work
     pKT25-ibaGC26S pKT25 containing ibaG ^C26S, Ap^r This work
     pKT25-bolA pKT25 containing the bolA gene, Km^r This work
     pUT18C-bolA pUT18C containing the bolA gene, Ap^r This work
     pKT25-ZIP pKT25 derivative with the leucine zipper of GCN4, Km^r
   Euromedex
     PUT18C-ZIP pUT18C derivative with the leucine zipper of GCN4, Ap^r
   Euromedex
     p6013-rsmA rsmA′-′lacZ translational fusion, Tet^r ([131]6)
     p6013-rsmE rsmE′-′lacZ translational fusion, Tet^r ([132]6)
     pRK415 Broad-host-range cloning vector, Tet^r ([133]32)
     p415-grxD pRK415 containing the grxD gene, Tet^r ([134]11)
     p415-ibaG pRK415 containing the ibaG gene, Tet^r This work
     p415-ibaGM pRK415 containing ibaG ^C26S, Tet^r This work
   [135]Open in a new tab
   ^^a

   Ap, ampicillin; Km, kanamycin; Tet, tetracycline.

Gene deletion in P. fluorescens

   The markerless P. fluorescens ibaG deletion mutant was constructed as
   follows. The flanking regions of the ibaG gene were amplified by PCR
   (primers are listed in [136]Table 2). The PCR products of the upstream
   and downstream regions were gel extracted and then connected by fusion
   PCR to delete ibaG. The fusion PCR product was digested with EcoRI and
   BamHI and then cloned into the suicide plasmid p2P24Km ([137]6) (Fig.
   S6), yielding p2P24-ibaG. Plasmid p2P24-ibaG was transferred into P.
   fluorescens 2P24 by electroporation. Plasmid p2P24-ibaG was integrated
   into the chromosome of strain 2P24 by the first crossover and selected
   on LB plates with kanamycin. Cells with the second crossover to
   generate the deletion were selected by culture on LB plates containing
   5% sucrose. The ibaG mutant was confirmed by amplified and Sanger
   sequencing. In the same way, the in-frame knockout mutant of bolA was
   constructed using the primers listed in [138]Table 2.

TABLE 2.

   Primers used in this study
   Name Sequence (5′−3′)[139] ^a Comment
   bolA-HindIII-F1 GGCAAGCTTCGGGGTTGAGGCCCGGCGTTTTTTAG Constructing an
   in-frame bolA deletion mutant
   bolA-R1 TCAGTGTTTGCTGCCACCGGCACATTCGATGCGTTGTTGCATGCTCAT
   bolA-F2 ATGAGCATGCAACAACGCATCGAATGTGCCGGTGGCAGCAAACACTGA
   bolA-EcoRI-R2 GGGAATTCGCGAAAGCTGGTGGTCTTGGGGTC
   ibaG-EcoRI-F1 ATGAATTCAAGCAGGACCCAAGCGCGTTCTATG Constructing an
   in-frame ibaG deletion mutant
   ibaG-R1 CTTGGGCTCAGGTGCGCTCGGCCCATACAGCCTGCATGCTCAACCTCAAT
   ibaG-F2 ATTGAGGTTGAGCATGCAGGCTGTATGGGCCGAGCGCACCTGAGCCCAAG
   ibaG-BamHI-R2 ATGGATCCGCTTGCCGTGCATGTTCACTTCG
   pKT25-grxD-F GGGTCGACTGTGGATATCATCGAAACGAT Cloning the coding region of
   grxD into pKT25
   pKT25-grxD-R GGGGATCCTCGGCTTCGGCCTTGTTCGCCG
   pUT18C-bolA-SalI AGGTCGACTATGAGCATGCAACAACGCATC Cloning the coding
   region of bolA into pUT18C
   pUT18C-bolA-BamHI GGGGATCCTCTCAGTGTTTGCTGCCACCGGCAC
   pKT25-bolA-F ACTCTAGAGATGCTGATTCTGACCC Cloning the coding region of
   bolA into pKT25
   pKT25-bolA-R TCGAATTCATAATGGCTTGGTTCGA
   pUT18C-ibaG-PstI ACACTGCAGGATGCAGGCTGTAGAAGTGAAGAG Cloning the coding
   region of ibaG into pUT18C
   pUT18C-ibaG-EcoRI ATGAATTCGATCAGGTGCGCTCGGCCCAGG
   pKT25-ibaG-F CGGGTACCTGGTGCGCTCGGCCCA Cloning the coding region of ibaG
   into pKT25
   pKT25-ibaG-R ACTCTAGAGATGCAGGCTGTAGAAGTG
   ibaG-C26S-F GTTGAAGGCGAAGGCGCCAACTTCCAGTT Site-directed mutagenesis of
   C26 to S of IbaG
   ibaG-C26S-R GCGCCTTCGCCTTCAACTTCCACCTGCGT
   [140]Open in a new tab
   ^^a

   Restriction sites inserted in the primer for the cloning are
   underlined.

   For the complementation of ibaG, the ibaG gene and its native promoter
   region were amplified by PCR using 2P24 genomic DNA as a template
   (primers are listed in [141]Table 2). The PCR product was then ligated
   into the shuttle plasmid pRK415 ([142]32), yielding p415-ibaG. The
   complementation plasmid p415-ibaG was introduced into the corresponding
   ibaG mutant by electroporation.

Bioinformatic analyses

   Amino acid alignment of BolA or IbaG homologs was performed using the
   Clustal Omega multiple sequence alignment tool at the European
   Bioinformatics Institute (EMBL-EBI) website
   ([143]https://www.ebi.ac.uk/Tools/msa/clustalo/). Protein structural
   prediction was performed using the I-TASSER server
   ([144]https://zhanggroup.org/I-TASSER/).

Site-directed mutagenesis

   The site-directed point mutant of ibaG was performed by using the
   QuikChange site-directed mutagenesis kit according to the
   manufacturer’s protocol (Agilent Technologies, USA). Primers are listed
   in [145]Table 2. Substitutions were confirmed by Sanger sequencing.

Immunoblot analyses

   Strain 2P24, the bolA mutant, and the ibaG mutant were grown overnight
   on LB agar and then cultured in liquid LB medium to an OD600 of 1.0
   (ca. 1 × 10^8 CFU/mL) (Fig. S7) and 1 mL samples were taken. Cells were
   collected by centrifugation, resuspended in 1× phosphate-buffered
   saline (PBS) buffer, and lysed by sonication. Samples were separated on
   a 10% SDS-PAGE gel and blotted on a polyvinylidene difluoride membrane
   according to the manufacturer’s instructions (Bio-Rad, USA). The
   blotting membrane was blocked in blocking buffer (5% milk powder) for 1
   h at room temperature and then incubated with an anti-FLAG antibody
   (Sangon Biotech, Shanghai, China). After washing with Tris-buffered
   saline with Tween 20 buffer, the membrane was then incubated with
   horseradish peroxidase-goat anti-mouse IgG secondary antibody, and the
   signal was detected using a chemiluminescence detection kit (Thermo
   Fisher, USA). RNA polymerase beta was used as the reference protein as
   the internal standard (Thermo Fisher, USA).

Bacterial two-hybrid assays

   Bacterial two-hybrid assays were performed as described previously
   ([146]33). The grxD and ibaG genes were amplified and cloned into the
   pUC18C and pKT25 vectors in frame with two gene fragments of the
   adenylate cyclase of Bordetella pertussis. The resulting vectors were
   then cotransformed into E. coli BTH101 and protein interactions were
   assessed by β-galactosidase activity ([147]34).

Quantification of 2,4-DAPG

   Strain 2P24 and its derivatives were grown in 30 mL KBG (KB broth with
   2% glucose) at 30°C for 36 h. The fermentation culture was mixed with
   0.1% (vol/vol) formic acid and then extracted with an equal volume of
   ethyl acetate as previously described ([148]35). The ethyl acetate
   extract was dried and resuspended in 50 µL of methanol. A 10-µL aliquot
   of the extracted sample was then detected by HPLC analysis (Waters
   2489) using the following conditions: C18 reversed-phase column (5 µm,
   4.5 × 250 mm, Agilent) eluted with 10% acetonitrile (vol/vol).
   Quantification of 2,4-DAPG was performed by integrating the area under
   the curve at 300 nm and compared with the standard curve prepared by
   injection of purified 2,4-DAPG.

RNA-seq analysis

   Cells of strain 2P24 and its ibaG mutant were cultured to the
   exponential phase (OD600 = 1.0) in LB medium. Three biological
   replicates were prepared for each strain. The cells (10 mL) were
   collected and centrifuged for 3 min at 9,600 × g. The total RNA was
   extracted using the RNeasy minikit (Qiagen, MD, USA). Extracted RNA was
   treated with Turbo DNase I (Thermo Fisher, USA) to remove genomic DNA
   contaminants. The rRNA was removed with the Ribo-Zero rRNA removal Kit
   (Illumina, USA). The integrity of the total RNA was then assessed with
   the Agilent TapeStation System (Agilent Technologies, UK). The cDNA
   library was constructed using the NEBNext UltraTM II RNA Library Prep
   Kit, which was then sequenced using Illumina HiSeq 4000 in rapid mode
   at a read length of 100 bp paired ends by Sangon Biotech Co., Ltd.
   (Shanghai, China). The RNA-seq raw data were mapped to the genome of P.
   fluorescens 2P24 (GenBank accession no. [149]CP025542) using HISAT
   software. The DESeq2 method was used to identify the DEGs of the ibaG
   mutant compared to the wild type ([150]36), and a log2 fold change of
   ≥1 and a P value (adjusted) of <0.05 were used to establish
   significance. The volcano plots were generated using the R language
   ggplots 2 package. To identify the potential pathways that the DEGs
   were involved, KEGG pathway enrichment analysis was performed using
   DAVID ([151]https://david.cifcrf.gov/). The GO analysis was performed
   with Blast2GO (BioBam)
   ([152]http://bowtie-bio.sourceforge.net/index.shtml).

Real-time quantitative PCR assay

   The validation of the RNA-seq analysis was performed by RT-qPCR. After
   RNA isolation and DNase I treatment, reverse transcription was done
   with cDNA Synthesis SuperMix (TransGen Biotech, Beijing, China).
   RT-qPCR was carried out with a TaqMan PCR master mix (Thermo Fisher,
   USA). Normalization was performed against the 16S rRNA gene using the
   2^−ΔΔCt method.

Siderophore assay

   Siderophore production was determined by CAS assay ([153]37). Briefly,
   strain 2P24 and its derivatives grown overnight were diluted to an
   OD600 of 0.6, and 5 mL of each culture was dropped onto CAS plates.
   Plates were incubated at 30°C for 36 h, and the production of
   siderophore was determined by measuring the diameter of an orange zone.

Swimming motility assays

   Swimming assays were performed as previously described ([154]11).
   Briefly, 2P24 and its derivatives were cultured overnight.
   Five-microliter aliquots of culture were then spotted onto the center
   of LB plate with 0.5% agar. The plates were incubated for 16 h at 30°C.
   The zone of swimming motility was recorded by measuring the diameter of
   the halo.

Rhizosphere colonization

   Rhizosphere colonization was performed as described previously
   ([155]24). Briefly, strain 2P24 and its derivatives were labeled by
   streptomycin sulfate resistance ([156]38). Surface-sterilized wheat
   seeds (Triticum aestivum cultivar Yumai 49) were soaked inside the
   bacterial cultures (10^8 CFU/mL) before sowing into seedling plates
   containing sterile soil. After 10 days of growth, ten plants were
   harvested randomly from each treatment, and bacteria were recovered
   from the rhizosphere by vortexing the root tips (last centimeter of the
   main root) for 2 min in a tube containing 5 mL of PBS buffer. The
   samples were serially diluted and then plated on LB agar plates. The
   experiment was performed three times with three replicates, and
   population data, collected as CFU counts, were log[10] transformed
   before statistical analysis.

Statistical analysis

   Statistical analysis was performed by the least significant difference
   test or the Student’s t test. Data were presented as mean values ±
   standard deviations, and P values less than 0.05 were considered
   statistically significant. All experiments were performed at least
   three times independently to confirm reproducibility.

ACKNOWLEDGMENTS