Sinorhizobium fredii HH103 RirA is required for oxidative stress resistance and efficient symbiosis with Soybean

International Journal of Molecular Sciences Article Sinorhizobium fredii HH103 RirA Is Required for Oxidative Stress Resistance and Efficient Symbiosis with Soybean Juan Carlos Crespo-Rivas 1,†, Pilar Navarro-Gómez 1,†, Cynthia Alias-Villegas 1,†, Jie Shi 2, Tao Zhen 3, Yanbo Niu 3, Virginia Cuéllar 4, Javier Moreno 5, Teresa Cubo 1, JoséMaría Vinardell 1, JoséEnrique Ruiz-Sainz 1, Sebastián Acosta-Jurado 1,* and María JoséSoto 4,* 1Departamento de Microbiología, Facultad de Biología, Universidad de Sevilla, Avda. Reina Mercedes 6, 41012 Sevilla, Spain; jccr[email protected] (J.C.C.-R.); [email protected] (P.N.-G.); [email protected] (C.A.-V.); [email protected] (T.C.); [email protected] (J.M.V.); [email protected] (J.E.R.-S.) 2Daqing Branch of Heilongjiang Academy of Sciences, Daqing 163000, China; [email protected] 3Institute of Microbiology, Heilongjiang Academy of Sciences, Harbin 150001, China; [email protected] (T.Z.); [email protected] (Y.N.) 4Departamento de Microbiología del Suelo y Sistemas Simbióticos, Estación Experimental del Zaidín, CSIC, c/ Profesor Albareda 1, 18008 Granada, Spain; [email protected] 5Departamento de Biología Celular, Facultad de Biología, Universidad de Sevilla, Avda. Reina Mercedes 6, 41012 Sevilla, Spain; [email protected] *Correspondence: [email protected] (S.A.-J.); [email protected] (M.J.S.); Tel.: +34-954-557121 (S.A.-J.); +34-958-181600 (M.J.S.) † These authors contributed equally to this work. Received: 14 January 2019; Accepted: 9 February 2019; Published: 12 February 2019   Abstract: Members of Rhizobiaceae contain a homologue of the iron-responsive regulatory protein RirA. In different bacteria, RirA acts as a repressor of iron uptake systems under iron-replete conditions and contributes to ameliorate cell damage during oxidative stress. In Rhizobium leguminosarum and Sinorhizobium meliloti, mutations in rirA do not impair symbiotic nitrogen fixation. In this study, a rirA mutant of broad host range S. fredii HH103 has been constructed (SVQ780) and its free-living and symbiotic phenotypes evaluated. No production of siderophores could be detected in either the wild-type or SVQ780. The rirA mutant exhibited a growth advantage under iron-deficient conditions and hypersensitivity to hydrogen peroxide in iron-rich medium. Transcription of rirA in HH103 is subject to autoregulation and inactivation of the gene upregulates fbpA, a gene putatively involved in iron transport. The S. fredii rirA mutant was able to nodulate soybean plants, but symbiotic nitrogen fixation was impaired. Nodules induced by the mutant were poorly infected compared to those induced by the wild-type. Genetic complementation reversed the mutant’s hypersensitivity to H 2 O 2 , expression of fbpA, and symbiotic deficiency in soybean plants. This is the first report that demonstrates a role for RirA in the Rhizobium-legume symbiosis. Keywords: iron; Rhizobium; regulation; siderophore; nitrogen-fixation; plant–bacteria interaction 1. Introduction As for almost all living organisms, iron is an essential nutrient for bacteria but in excess it becomes toxic. This active redox metal forms part of the prosthetic group (heme or iron–sulfur clusters) of many proteins that participate in important metabolic processes such as respiration, central metabolism, Int. J. Mol. Sci. 2019,20, 787; doi:10.3390/ijms20030787 www.mdpi.com/journal/ijms Int. J. Mol. Sci. 2019,20, 787 2 of 16 redox stress resistance, or nitrogen fixation. Although it is not a rare element in the Earth’s crust, iron is not readily bioavailable in most environments. Under aerobic conditions and at neutral pH values, the concentration of soluble iron compounds is too low to fulfill the cell needs. Bacteria circumvent this problem with high-affinity uptake systems whose expression is triggered under iron-deficient conditions [ 1 ]. On the other hand, the redox properties of iron are responsible for the generation of highly toxic free radicals via the Fenton reaction. Reactive oxygen species (ROS) such as H 2 O 2 generated during aerobic metabolism in combination with ferrous iron leads to the production of hydroxyl radicals (OH • ), which are extremely reactive oxidants capable of damaging DNA, proteins, and lipids [ 2 ]. During oxidative stress, i.e., an increased ROS production that cannot be balanced by antioxidants, toxicity caused by iron is more critical. Superoxide (O 2•- ) and hydrogen peroxide (H 2 O 2 ) can react with iron present in the prosthetic group of metalloproteins, resulting in enzyme inhibition and the release of free iron, which could then promote the Fenton reaction [ 3 ]. Therefore, bacteria require mechanisms for iron homeostasis that allow them to maintain sufficient levels of the metal to support vital metabolic processes and at the same time prevent toxicity. Iron homeostasis is achieved by controlling the expression of proteins involved in iron uptake, storage, consumption, and export. When intracellular levels of iron are insufficient, many bacteria produce siderophores, i.e., water soluble Fe 3+ chelating compounds, and express several high-affinity uptake systems to acquire iron from the environment [ 4 ]. Under iron sufficient conditions, expression of iron uptake systems is repressed and iron is incorporated into metalloproteins. Excessive iron is accumulated in a non-reactive form in iron storage proteins and/or expelled out of the cell by iron efflux transporters to mitigate cytotoxic effects [ 5 – 7 ]. All of these processes are coordinately controlled through the activity of iron response regulatory proteins. In many bacteria, including Escherichia coli, the major iron response regulator is Fur (Ferric Uptake Regulator) [ 1 , 8 – 10 ]. Fur is a transcriptional regulator that senses and responds to iron by directly binding this metal [ 10 ]. In the presence of certain levels of iron, the Fe 2+ -Fur complex binds to DNA sequences called Fur boxes and represses the transcription of iron uptake systems, preventing iron overload toxicity [ 1 ]. Remarkably, iron-responsive regulation differs from the Fur paradigm in rhizobia, i.e., soil-dwelling bacteria that are known for their ability to establish a symbiotic relationship into nitrogen-fixing nodules on the roots of leguminous plants [ 11 ]. Symbiotic nitrogen fixation is highly dependent on iron because different proteins involved in the process require the metal to be functional. This is the case of the nitrogenase complex responsible for the nitrogen fixation reaction, or the abundant nodule protein leghemoglobin that buffers oxygen levels to avoid inactivation of nitrogenase activity. Although a Fur-like protein is present in most rhizobia, functional studies performed on the fur genes of Sinorhizobium (Ensifer) meliloti and Rhizobium leguminosarum revealed that the encoded proteins are responsive to manganese and not to iron, which is the reason why they were renamed as Mur [ 12 – 14 ]. Instead, iron-responsive regulation in rhizobia is mediated by the Irr and RirA regulatory proteins that sense iron availability indirectly [15–17]. The Irr (Iron Responsive Regulator) protein belongs to the Fur superfamily. Described initially in the soybean symbiont Bradyrhizobium japonicum [ 18 ], Irr occurs in all rhizobia and is the only iron responsive regulator in Bradyrhizobiaceae. Under iron-deficient conditions, Irr usually represses genes encoding proteins that function under iron-sufficient conditions (i.e., proteins containing iron, heme and iron–sulfur cluster biosynthetic proteins). Upon sensing iron indirectly through the status of heme biosynthesis, Irr is either degraded (as occurs in B. japonicum) or becomes inactive (as is the case of R. leguminosarum), which allows transcription of Irr-repressed genes [17]. In rhizobia distinct from Bradyrhizobiaceae, iron homeostasis is achieved with the participation of a second regulatory protein, RirA (Rhizobial Iron Regulator A). This regulator, which is found exclusively in alphaproteobacteria, is present in Rhizobium,Sinorhizobium (Ensifer), and Mesorhizobium spp., but also in the phytopathogen Agrobacterium tumefaciens and in the animal pathogens Brucella and Bartonella [ 14 , 17 ]. RirA is an iron–sulfur protein that belongs to the Rrf2 family of putative transcription regulators. Under iron-replete conditions, this regulator acts as a repressor of many Int. J. Mol. Sci. 2019,20, 787 3 of 16 iron-responsive genes by binding to Iron-Responsive Operator sequences (5’-TGA-(N9)-TCA-3’) known as IRO boxes [ 14 , 19 ]. A recent study has demonstrated that R. leguminosarum RirA contains a [4Fe–4S] cluster and that the holoprotein ([4Fe–4S]RirA) binds to an IRO box sequence [ 20 ]. Interestingly, the [4Fe–4S] cluster is sensitive to iron levels but also to O 2 , with low iron concentrations and aerobic conditions promoting the conversion of [4Fe–4S] cluster to [2Fe–2S] and subsequent cluster loss, which relieves gene repression [ 20 ]. Therefore, RirA seems to act as an iron sensor via iron–sulfur cluster availability. In rhizobia, RirA has been functionally characterized only in R. leguminosarum and S. meliloti. In R. leguminosarum, RirA regulates several iron-responsive genes such as genes for the synthesis and uptake of the siderophore vicibactin (vbs), genes involved in heme uptake (hmu and tonB), genes for the synthesis of iron–sulfur clusters (suf), the irrA gene as well as its own expression [ 21 , 22 ]. In S. meliloti, the RirA regulon comprises several iron-responsive genes, which include genes involved in iron uptake, energy metabolism, exopolysaccharide synthesis or iron storage, but also genes that are not iron-responsive (e.g., suf) [ 23 – 25 ]. Several free-living phenotypes have been associated with rirA mutants. Thus, inactivation of rirA leads to: (i) constitutive expression of siderophores in R. leguminosarum,S. meliloti and in A. tumefaciens, (ii) species-specific growth phenotypes which can be iron-dependent or -independent, (iii) higher sensitivity to oxidants in S. meliloti and in A. tumefaciens, and (iv) decreased ability to cope with cobalt and nickel toxicity in A. tumefaciens [ 21 , 23 , 24 , 26 – 28 ]. Hypersensitivity to peroxides and metals in rirA mutants was explained by the derepression of iron uptake systems and the consequent iron overload in the cell, which promotes the generation of ROS via Fenton chemistry. The role of RirA in the establishment of plant–bacteria interactions has also been investigated. Loss of function of RirA in R. leguminosarum or S. meliloti does not affect symbiotic nitrogen fixation [21,23,24] . In A. tumefaciens, RirA was found to be important for tumor formation on tobacco leaf pieces [ 26 ] but not on potato disks [ 27 ]. The defect in tumorigenesis on tobacco leaves was associated with increased sensitivity to oxidants and reduced virulence gene expression exhibited by the rirA mutant. S. fredii HH103 is a fast-growing rhizobial strain, which was isolated from a Chinese soil from the Hubei Province. HH103 is able to nodulate American and Asiatic varieties of soybean, as well as many other different legumes that form determinate or indeterminate nodules [ 29 , 30 ]. Iron homeostasis in broad host range rhizobia has not been investigated yet. The genome sequence of S. fredii HH103 reveals the presence of putative iron-responsive regulatory genes, including a rirA homologue [ 31 , 32 ]. With the aim of gaining insights into iron-mediated responses in HH103, in this study, we have investigated the role of RirA during free-living growth and in symbiosis with soybean. Our data demonstrate that, in S. fredii HH103, RirA is important for oxidative stress resistance and for effective symbiosis with soybean plants. 2. Results 2.1. The Lack of RirA in S. fredii HH103 Confers a Siderophore-Independent Growth Advantage under Low Iron Conditions S. fredii HH103 encodes homologues of Fur (SFHH103_03131), Irr (SFHH103_03814), and RirA (SFHH103_00388). The chromosomal SFHH103_00388 gene encodes a 154 amino acid-long protein annotated as an iron-responsive regulator of the Rrf2 family of transcriptional regulators. The predicted protein, which is identical to one encoded by S. fredii NGR234, exhibits high sequence similarity to the RirA proteins of S. meliloti,A. tumefaciens and R. leguminosarum bv. viciae (93.5%, 83.4%, and 79.4% identity, respectively), and conserves the three Cys residues predicted to ligate iron–sulfur clusters in Rrf2 family regulators [ 20 ] (Figure S1). The genetic context of the rirA gene is similar in HH103 and in S. meliloti. In the two rhizobia, the gene is flanked downstream by a cluster of dpp genes involved in the uptake of dipeptides and a heme precursor [ 33 ], and upstream by the fbpA gene that potentially codes for a periplasmic ferric binding protein, which is regulated by RirA in S. meliloti [23]. Int. J. Mol. Sci. 2019,20, 787 4 of 16 In order to investigate the role of the rirA homologue in HH103, mutant SVQ780 was constructed by inserting a lacZ ∆ p-Gm R cassette into the coding sequence of this gene (see Material and Methods). A battery of free-living phenotypes that have been associated with rirA mutants were evaluated such as siderophore production, growth in media with different iron content, resistance to H 2 O 2 , and expression of iron-responsive genes. Under our experimental conditions, Chrome Azurol S (CAS) assays performed using cultures of the wild-type strain SVQ269 (HH103-Rif R ) grown in iron-deficient media failed to show any siderophore activity. In fact, CAS assays performed on forty S. fredii strains isolated from different locations in China only showed activity in two of them (strains S2 and S5; Figure S2), suggesting that siderophore production is a rare attribute in Chinese S. fredii strains. RirA loss-of-function leads to overexpression of siderophore biosynthesis and transport genes in R. leguminosarum,S. meliloti and A. tumefaciens [ 22 – 24 , 26 ]. However, no CAS activity was detected upon inactivation of the rirA homologue in HH103 (Figure S2c), indicating that this strain does not acquire iron via CAS-detected siderophore production. Depending on the bacterial species, inactivation of rirA results in different growth phenotypes ranging from iron-dependent or -independent reduced growth to no effect at all [ 21 , 23 , 27 ]. In our study, we evaluated the growth rates of S. fredii wild-type and rirA mutant strains in minimal medium (MM) agar plates supplemented with different concentrations of FeCl 3 as the added iron source, as well as in un-supplemented MM with ambient iron levels. As shown in Figure 1, the rirA mutant exhibited similar growth capacities in un-supplemented MM and in MM supplemented with 0.22, 2.2 and 22 µ M FeCl 3 . The wild-type and the rirA mutant exhibited similar growth rates on tryptone-yeast extract (TY) medium. Likewise, no differences were detected between the two strains on MM plates supplemented with 22 µ M FeCl 3 , which is the optimum concentration of iron for the wild-type strain under the conditions tested. Interestingly, on MM containing lower concentrations of iron (i.e., un-supplemented and supplemented with 0.22 and 2.2 µ M FeCl 3 ) the rirA mutant showed a slight advantage in growth with respect to SVQ269 and the complemented strain SVQ780C. The low iron availability-dependent growth advantage associated with the rirA mutant might be caused by derepression of as yet uncharacterized high-affinity iron uptake systems. In MM with the highest iron content (220 µ M FeCl 3 ), all three strains exhibited similar growth reductions compared with MM containing 22 µM FeCl3, suggesting probable iron toxicity. Int. J. Mol. Sci. 2019,20, 787 5 of 16 Int. J. Mol. Sci. 2018, 19, x FOR PEER REVIEW 4 of 16 In order to investigate the role of the rirA homologue in HH103, mutant SVQ780 was constructed by inserting a lacZΔp-GmR cassette into the coding sequence of this gene (see Material and Methods). A battery of free-living phenotypes that have been associated with rirA mutants were evaluated such as siderophore production, growth in media with different iron content, resistance to H2O2, and expression of iron-responsive genes. Under our experimental conditions, Chrome Azurol S (CAS) assays performed using cultures of the wild-type strain SVQ269 (HH103-RifR) grown in iron-deficient media failed to show any siderophore activity. In fact, CAS assays performed on forty S. fredii strains isolated from different locations in China only showed activity in two of them (strains S2 and S5; Figure S2), suggesting that siderophore production is a rare attribute in Chinese S. fredii strains. RirA loss-of-function leads to overexpression of siderophore biosynthesis and transport genes in R. leguminosarum, S. meliloti and A. tumefaciens [22–24,26]. However, no CAS activity was detected upon inactivation of the rirA homologue in HH103 (Figure S2c), indicating that this strain does not acquire iron via CAS-detected siderophore production. Depending on the bacterial species, inactivation of rirA results in different growth phenotypes ranging from iron-dependent or -independent reduced growth to no effect at all [21,23,27]. In our study, we evaluated the growth rates of S. fredii wild-type and rirA mutant strains in minimal medium (MM) agar plates supplemented with different concentrations of FeCl3 as the added iron source, as well as in un-supplemented MM with ambient iron levels. As shown in Figure 1, the rirA mutant exhibited similar growth capacities in un-supplemented MM and in MM supplemented with 0.22, 2.2 and 22 µM FeCl3. The wild-type and the rirA mutant exhibited similar growth rates on tryptone-yeast extract (TY) medium. Likewise, no differences were detected between the two strains on MM plates supplemented with 22 µM FeCl3, which is the optimum concentration of iron for the wild-type strain under the conditions tested. Interestingly, on MM containing lower concentrations of iron (i.e., un-supplemented and supplemented with 0.22 and 2.2 µM FeCl3) the rirA mutant showed a slight advantage in growth with respect to SVQ269 and the complemented strain SVQ780C. The low iron availability-dependent growth advantage associated with the rirA mutant might be caused by derepression of as yet uncharacterized high-affinity iron uptake systems. In MM with the highest iron content (220 µM FeCl3), all three strains exhibited similar growth reductions compared with MM containing 22 µM FeCl3, suggesting probable iron toxicity. Figure 1. Growth of S. fredii wild-type strain SVQ269 (wt), rirA mutant SVQ780 (rirA) and complemented strain SVQ780C (rirAC) on tryptone-yeast extract (TY) and minimal medium (MM) agar plates in the absence or presence of different concentrations of FeCl 3 as indicated. Tenfold serial dilutions of bacterial cultures (10 9 cells/mL) are marked above each column. A representative example of at least two experiments is shown. 2.2. RirA Plays A Role in Oxidative Stress Resistance in HH103 It has been shown that rirA mutants of S. meliloti and A. tumefaciens are more sensitive to oxidative stress than their corresponding wild-type strains [ 23 , 26 , 27 ]. Therefore, we decided to test whether inactivation of rirA in HH103 alters the bacterial resistance to hydrogen peroxide. During these experiments, we noticed that HH103 was much more sensitive to H 2 O 2 than S. meliloti or A. tumefaciens. Thus, approximately only 1 out of 10 7 cells survived after exposing SVQ269 cultures to 10 mM H 2 O 2 for only 1 h. When SVQ269 cultures were exposed to a 100-fold lower concentration of H 2 O 2 (100 µ M), the survival ratio of the wild-type strain was slightly reduced to 71%. Under the same conditions, the viability of rirA mutant cells was extremely low (only 20 out of every 10 7 cells could survive after treatment with 100 µ M H 2 O 2 ). The higher sensitivity of the rirA mutant to H 2 O 2 compared to the wild-type strain was confirmed by growing serial dilutions of cells on TY plates containing different concentrations of the oxidant (Figure 2). Int. J. Mol. Sci. 2018, 19, x FOR PEER REVIEW 5 of 16 Figure 1. Growth of S. fredii wild-type strain SVQ269 (wt), rirA mutant SVQ780 (rirA) and complemented strain SVQ780C (rirAC) on tryptone-yeast extract (TY) and minimal medium (MM) agar plates in the absence or presence of different concentrations of FeCl3 as indicated. Tenfold serial dilutions of bacterial cultures (109 cells/mL) are marked above each column. A representative example of at least two experiments is shown. 2.2. RirA Plays A Role in Oxidative Stress Resistance in HH103 It has been shown that rirA mutants of S. meliloti and A. tumefaciens are more sensitive to oxidative stress than their corresponding wild-type strains [23,26,27]. Therefore, we decided to test whether inactivation of rirA in HH103 alters the bacterial resistance to hydrogen peroxide. During these experiments, we noticed that HH103 was much more sensitive to H2O2 than S. meliloti or A. tumefaciens. Thus, approximately only 1 out of 107 cells survived after exposing SVQ269 cultures to 10 mM H2O2 for only 1 h. When SVQ269 cultures were exposed to a 100-fold lower concentration of H2O2 (100 µM), the survival ratio of the wild-type strain was slightly reduced to 71%. Under the same conditions, the viability of rirA mutant cells was extremely low (only 20 out of every 107 cells could survive after treatment with 100 µM H2O2). The higher sensitivity of the rirA mutant to H2O2 compared to the wild-type strain was confirmed by growing serial dilutions of cells on TY plates containing different concentrations of the oxidant (Figure 2). Figure 2. Growth of S. fredii wild-type strain SVQ269 (wt), rirA mutant SVQ780 (rirA) and complemented strain SVQ780C (rirAC) on TY in the absence or presence of different concentrations of H2O2. Tenfold serial dilutions of bacterial cultures (109 cells/mL) are marked above each column. A representative example of at least two experiments is shown. No differences in growth were observed for the wild-type and mutant strains on TY plates in the absence of H2O2. While viability of SVQ269 remained unchanged in the presence of 5 µM H2O2, survival of the rirA mutant was already severely affected. The presence of higher concentrations of H2O2 (7.5 and 10 µM) decreased the viability of the wild-type and especially that exhibited by the rirA mutant. The hypersensitive phenotype of SVQ780 could be reversed by introducing a functional rirA gene in cis (strain SVQ780C). These data indicate that inactivation of rirA in S. fredii HH103 results in reduced resistance to oxidative damage. 2.3. Role of RirA in Iron-Responsive Gene Expression Inactivation of rirA in R. leguminosarum, S. meliloti, and A. tumefaciens causes upregulation of several genes whose transcription is normally reduced/repressed under iron-sufficient conditions. These include genes involved in the synthesis and uptake of siderophores, heme uptake, inorganic ferric ion transport, biosynthesis of Fe–S clusters, and the regulatory irr and rirA genes [21–24,26,27]. The expression levels of potentially iron-responsive genes were determined by Reverse Transcription-quantitative Polymerase Chain Reaction (RT-qPCR) in the wild-type strain SVQ269 after growth in iron-deficient or iron-replete media. Based on the bacterial growth shown in Figure 1, MM supplemented with 0.22 and 22 µM FeCl3 were used as the iron-deficient and iron–replete conditions, respectively. The chosen genes comprise those putatively coding for the periplasmic Figure 2. Growth of S. fredii wild-type strain SVQ269 (wt), rirA mutant SVQ780 (rirA) and complemented strain SVQ780C (rirAC) on TY in the absence or presence of different concentrations of H 2 O 2 . Tenfold serial dilutions of bacterial cultures (10 9 cells/mL) are marked above each column. A representative example of at least two experiments is shown. Int. J. Mol. Sci. 2019,20, 787 6 of 16 No differences in growth were observed for the wild-type and mutant strains on TY plates in the absence of H 2 O 2 . While viability of SVQ269 remained unchanged in the presence of 5 µ M H 2 O 2 , survival of the rirA mutant was already severely affected. The presence of higher concentrations of H 2 O 2 (7.5 and 10 µ M) decreased the viability of the wild-type and especially that exhibited by the rirA mutant. The hypersensitive phenotype of SVQ780 could be reversed by introducing a functional rirA gene in cis (strain SVQ780C). These data indicate that inactivation of rirA in S. fredii HH103 results in reduced resistance to oxidative damage. 2.3. Role of RirA in Iron-Responsive Gene Expression Inactivation of rirA in R. leguminosarum,S. meliloti, and A. tumefaciens causes upregulation of several genes whose transcription is normally reduced/repressed under iron-sufficient conditions. These include genes involved in the synthesis and uptake of siderophores, heme uptake, inorganic ferric ion transport, biosynthesis of Fe–S clusters, and the regulatory irr and rirA genes [21–24,26,27]. The expression levels of potentially iron-responsive genes were determined by Reverse Transcription-quantitative Polymerase Chain Reaction (RT-qPCR) in the wild-type strain SVQ269 after growth in iron-deficient or iron-replete media. Based on the bacterial growth shown in Figure 1, MM supplemented with 0.22 and 22 µ M FeCl 3 were used as the iron-deficient and iron–replete conditions, respectively. The chosen genes comprise those putatively coding for the periplasmic binding protein of a ferric type ATP-Binding Cassette (ABC) transporter (fbpA), the ATPase component of a heme ABC transporter (hmuS), a protein involved in Fe-S cluster formation (sufS), the Irr homologue (SFHH103_03814), and rirA. As shown in Figure 3a, all five genes exhibited upregulation in wild-type cells grown in iron-deficient medium compared with cells grown in iron-replete medium. Int. J. Mol. Sci. 2018, 19, x FOR PEER REVIEW 6 of 16 binding protein of a ferric type ATP-Binding Cassette (ABC) transporter (fbpA), the ATPase component of a heme ABC transporter (hmuS), a protein involved in Fe-S cluster formation (sufS), the Irr homologue (SFHH103_03814), and rirA. As shown in Figure 3a, all five genes exhibited upregulation in wild-type cells grown in iron-deficient medium compared with cells grown in iron-replete medium. Figure 3. Gene expression analyses of iron-related genes in S. fredii. (a) Relative expression of rirA, sufB, fbpA, hmuS and irr genes in the wild-type strain SVQ269 of S. fredii as determined by Reverse Transcription-quantitative Polymerase Chain Reaction (RT-qPCR). The relative expression was calculated as the fold change between growth in iron-deficient (0.22 µM FeCl3) and iron-replete (22 µM FeCl3) media; (b) relative expression of iron-related genes in rirA mutant versus wild-type strain after growth in iron-sufficient (22 µM FeCl3) medium as determined by RT-qPCR; (c) transcriptional activity of rirA in the SVQ780 mutant (white bars) and complemented strain (SVQ780C) (black bars) after growth in media with different iron concentration as determined by β-galactosidase activity. Results are averages from at least two independent biological experiments with two technical replicates. Error bars indicate standard error. The asterisk indicates significant difference with respect to data obtained in cells grown in un-supplemented media according to ANOVA test (p < 0.05). To examine the effects of RirA on iron homeostasis, the transcript levels of fbpA, hmuS, sufS and irr were also determined by RT-qPCR in rirA mutant cells after growth in iron-sufficient (22 µM FeCl3) media (Figure 3b). Under these conditions, no differential expression was detected for hmuS between the mutant and the wild-type (1.5 ± 0.3 fold increase in rirA mutant versus wild-type strain). In the case of sufB and irr, only slight decreases in expression were observed in the mutant compared to the wild-type (2 ± 1.3 and 4.4 ± 3.3 fold decrease in rirA mutant versus wild-type strain for sufB and irr, respectively). In contrast, inactivation of rirA caused a clear upregulation of fbpA (95.4 ± 39.6 fold increase in rirA mutant versus wild-type strain). No differential expression for fbpA was detected between the wild-type and the complemented strain. This suggests that RirA represses transcription of fbpA, either directly or indirectly. Since the promoterless lacZ-GmR cassette used to obtain the rirA mutant was inserted in the same transcriptional orientation as that of the mutated gene, rirA expression levels could be determined by measuring β-galactosidase activity in SVQ780 and SVQ780C strains grown in un-supplemented MM or in MM supplemented with different iron concentrations. As shown in Figure 3c, expression of rirA in the SVQ780 mutant increased slightly with increased iron content in Figure 3. Gene expression analyses of iron-related genes in S. fredii. ( a ) Relative expression of rirA,sufB,fbpA,hmuS and irr genes in the wild-type strain SVQ269 of S. fredii as determined by Reverse Transcription-quantitative Polymerase Chain Reaction (RT-qPCR). The relative expression was calculated as the fold change between growth in iron-deficient (0.22 µ M FeCl 3 ) and iron-replete (22 µ M FeCl 3 ) media; ( b ) relative expression of iron-related genes in rirA mutant versus wild-type strain after growth in iron-sufficient (22 µ M FeCl3) medium as determined by RT-qPCR; ( c ) transcriptional activity of rirA in the SVQ780 mutant (white bars) and complemented strain (SVQ780C) (black bars) after growth in media with different iron concentration as determined by β -galactosidase activity. Results are averages from at least two independent biological experiments with two technical replicates. Error bars indicate standard error. The asterisk indicates significant difference with respect to data obtained in cells grown in un-supplemented media according to ANOVA test (p< 0.05). Int. J. Mol. Sci. 2019,20, 787 7 of 16 To examine the effects of RirA on iron homeostasis, the transcript levels of fbpA,hmuS,sufS and irr were also determined by RT-qPCR in rirA mutant cells after growth in iron-sufficient (22 µ M FeCl 3 ) media (Figure 3b). Under these conditions, no differential expression was detected for hmuS between the mutant and the wild-type (1.5 ± 0.3 fold increase in rirA mutant versus wild-type strain). In the case of sufB and irr, only slight decreases in expression were observed in the mutant compared to the wild-type (2 ± 1.3 and 4.4 ± 3.3 fold decrease in rirA mutant versus wild-type strain for sufB and irr, respectively). In contrast, inactivation of rirA caused a clear upregulation of fbpA (95.4 ± 39.6 fold increase in rirA mutant versus wild-type strain). No differential expression for fbpA was detected between the wild-type and the complemented strain. This suggests that RirA represses transcription of fbpA, either directly or indirectly. Since the promoterless lacZ-Gm R cassette used to obtain the rirA mutant was inserted in the same transcriptional orientation as that of the mutated gene, rirA expression levels could be determined by measuring β -galactosidase activity in SVQ780 and SVQ780C strains grown in un-supplemented MM or in MM supplemented with different iron concentrations. As shown in Figure 3c, expression of rirA in the SVQ780 mutant increased slightly with increased iron content in the medium, exhibiting the lowest and highest expression levels in un-supplemented MM and in MM containing 220 µ M FeCl 3 , respectively. These results are in agreement with data obtained in R. leguminosarum and in A. tumefaciens rirA mutants, in which rirA expression was downregulated under iron-deplete conditions [ 21 , 27 ]. In the complemented SVQ780C strain, expression of rirA was low irrespective of the concentration of iron in the medium, indicating an auto-regulatory role similar to that of RirA in R. leguminosarum and in A. tumefaciens [21,27]. 2.4. Nitrogen-Fixing Symbiosis with Soybean Plants Is Impaired in an S. fredii HH103 rirA Mutant Previous studies have shown that mutations in rirA in R. leguminosarum or in S. meliloti do not affect nodulation or symbiotic nitrogen fixation with their respective hosts (peas and vetch for R. leguminosarum, and alfalfa for S. meliloti) [ 21 , 23 , 24 ]. The symbiotic phenotype of the S. fredii rirA mutant was investigated in soybean (Glycine max) cv. Williams, which forms determinate nodules. The mutant was able to induce nodule formation on soybean roots. However, plants inoculated with the rirA mutant exhibited symptoms of impaired symbiotic effectiveness since they were smaller than plants inoculated with the wild-type strain and their leaves showed a chlorotic appearance similar to that exhibited by non-inoculated soybeans (Figure 4). These symptoms were suggestive of nitrogen deficiency. No significant differences were detected in the number of nodules developed in the roots of soybean plants inoculated with the wild-type or the rirA mutant. However, the nodules induced by the mutant exhibited a significant reduction in size and fresh weight and were not as red as those induced by the wild-type strain (Figure 4and Table 1). Moreover, acetylene reduction assays (ARA) revealed that nitrogenase activity of the mutant was 5.6-fold reduced with respect to that of SVQ269 (Table 1). In order to better understand the symbiotic deficiency of the mutant, nodules induced by the wild-type and the rirA mutant were analyzed by light microscopy (Figure 5). Int. J. Mol. Sci. 2019,20, 787 8 of 16 Int. J. Mol. Sci. 2018, 19, x FOR PEER REVIEW 7 of 16 the medium, exhibiting the lowest and highest expression levels in un-supplemented MM and in MM containing 220 µM FeCl3, respectively. These results are in agreement with data obtained in R. leguminosarum and in A. tumefaciens rirA mutants, in which rirA expression was downregulated under iron-deplete conditions [21,27]. In the complemented SVQ780C strain, expression of rirA was low irrespective of the concentration of iron in the medium, indicating an auto-regulatory role similar to that of RirA in R. leguminosarum and in A. tumefaciens [21,27]. 2.4. Nitrogen-Fixing Symbiosis with Soybean Plants Is Impaired in an S. fredii HH103 rirA Mutant Previous studies have shown that mutations in rirA in R. leguminosarum or in S. meliloti do not affect nodulation or symbiotic nitrogen fixation with their respective hosts (peas and vetch for R. leguminosarum, and alfalfa for S. meliloti) [21,23,24]. The symbiotic phenotype of the S. fredii rirA mutant was investigated in soybean (Glycine max) cv. Williams, which forms determinate nodules. The mutant was able to induce nodule formation on soybean roots. However, plants inoculated with the rirA mutant exhibited symptoms of impaired symbiotic effectiveness since they were smaller than plants inoculated with the wild-type strain and their leaves showed a chlorotic appearance similar to that exhibited by non-inoculated soybeans (Figure 4). These symptoms were suggestive of nitrogen deficiency. Figure 4. Symbiotic phenotype and nodule morphology of soybean plants inoculated with S. fredii SVQ269 (wt), its rirA derivative SVQ780 (rirA), and the complemented strain SVQ780C (rirAC). Non-inoculated plants are denoted as NI. Scale bars next to nodules correspond to 1 mm. A representative example of at least two experiments is shown. Table 1. Symbiotic capacity of S. fredii SVQ269, its rirA (SVQ780) mutant derivative and the complemented version (SVQ780C) with Glycine max cv. Williams. Glycine max cv.Williams 1 Inoculant Number of Nodules Nodules Fresh Weight (mg) Plant-Top Dry Weight (mg) ARA 2 (Nmoles Acetylene/Plant/Hour) Nodules Size (mm) SVQ269 26.38 ± 2.88 510.33 ± 65.71 1.08 ± 0.15 998.66 ± 175.56 4.04 ± 0.11 SVQ780 23.80 ± 3.38 222.71 ± 16.04 ** 0.62 ± 0.07 * 167.99 ± 27.64 ** 2.64 ± 0.10 ** Figure 4. Symbiotic phenotype and nodule morphology of soybean plants inoculated with S. fredii SVQ269 (wt), its rirA derivative SVQ780 (rirA), and the complemented strain SVQ780C (rirAC). Non-inoculated plants are denoted as NI. Scale bars next to nodules correspond to 1 mm. A representative example of at least two experiments is shown. Table 1. Symbiotic capacity of S. fredii SVQ269, its rirA (SVQ780) mutant derivative and the complemented version (SVQ780C) with Glycine max cv. Williams. Glycine max cv.Williams 1 Inoculant Number of Nodules Nodules Fresh Weight (mg) Plant-Top Dry Weight (mg) ARA 2(Nmoles Acetylene/Plant/Hour) Nodules Size (mm) SVQ269 26.38 ±2.88 510.33 ±65.71 1.08 ±0.15 998.66 ±175.56 4.04 ±0.11 SVQ780 23.80 ±3.38 222.71 ±16.04 ** 0.62 ±0.07 * 167.99 ±27.64 ** 2.64 ±0.10 ** SVQ780C 22.44 ±2.50 453.48 ±54.89 1.18 ±0.12 1008.22 ±118.18 4.37 ±0.10 ** NI 30 0 0.245 ±0.02 ** 0 0 1 Numbers are mean ( ± standard error of the mean, SEM) values per plant. Ten plants were tested for each G. max/inoculant combination. Plants were grown for six weeks in a plant growth chamber. All the treatments were individually compared with the values of the parental strain, using the nonparametric test of Mann–Whitney. The presence of one or two asterisks denotes a significant difference with respect to data obtained in plants inoculated with the wild-type (pvalue < 0.05 or < 0.01, respectively). 2 Acetylene Reduction Assays (ARA). 3 Non-Inoculated (NI). Int. J. Mol. Sci. 2019,20, 787 9 of 16 Int. J. Mol. Sci. 2018, 19, x FOR PEER REVIEW 9 of 16 Figure 5. Optical microscopy of nodules induced in G. max cv. Williams by S. fredii wild-type strain SVQ269 (a–b); rirA mutant SVQ780 (c–d); and complemented strain SVQ780C (e–f). Bars correspond to 50 µm. The results revealed that infected cells in nodules induced by the mutant contained fewer bacteria than those in nodules formed by the wild-type strain. The complemented strain SVQ780C regained the wild-type phenotype in symbiosis with soybean, including efficient nitrogen fixation (Figure 4 and Table 1) and degree of bacterial infection in nodule cells (Figure 5). Altogether, these data indicate that RirA is important for efficient symbiosis with soybean plants. 3. Discussion In this work, we investigated the role of RirA, the rhizobial iron regulator, in S. fredii HH103. This study represents the first approach to tackle iron homeostasis in a broad-host range Rhizobium. Despite the pivotal role of iron in the establishment of efficient Rhizobium-legume symbiosis, studies dealing with how the microsymbiont controls the intracellular iron pools have only been performed Figure 5. Optical microscopy of nodules induced in G. max cv. Williams by S. fredii wild-type strain SVQ269 ( a – b ); rirA mutant SVQ780 ( c – d ); and complemented strain SVQ780C ( e – f ). Bars correspond to 50 µm. The results revealed that infected cells in nodules induced by the mutant contained fewer bacteria than those in nodules formed by the wild-type strain. The complemented strain SVQ780C regained the wild-type phenotype in symbiosis with soybean, including efficient nitrogen fixation (Figure 4and Table 1) and degree of bacterial infection in nodule cells (Figure 5). Altogether, these data indicate that RirA is important for efficient symbiosis with soybean plants. 3. Discussion In this work, we investigated the role of RirA, the rhizobial iron regulator, in S. fredii HH103. This study represents the first approach to tackle iron homeostasis in a broad-host range Rhizobium. Despite the pivotal role of iron in the establishment of efficient Rhizobium-legume symbiosis, studies dealing with how the microsymbiont controls the intracellular iron pools have only been performed in three rhizobia: B. japonicum,R. leguminosarum and S. meliloti. Whereas important insights have been Int. J. Mol. Sci. 2019,20, 787 16 of 16 39. Acosta-Jurado, S.; Alias-Villegas, C.; Navarro-Gómez, P.; Zehner, S.; Murdoch, P.D.; Rodríguez-Carvajal, M.A.; Soto, M.J.; Ollero, F.J.; Ruiz-Sainz, J.E.; Gottfert, M.; et al. The Sinorhizobium fredii HH103 MucR1 global regulator is connected with the nod regulon and is required for efficient symbiosis with Lotus burttii and Glycine max cv. Williams. Mol. Plant Microbe Interact. 2016,29, 700–712. [CrossRef] 40. Santos, R.; Herouart, D.; Sigaud, S.; Touati, D.; Puppo, A. Oxidative burst in alfalfa-Sinorhizobium meliloti symbiotic interaction. Mol. Plant Microbe Interact. 2001,14, 86–89. [CrossRef] 41. Bueno, P.; Soto, M.J.; Rodríguez-Rosales, M.P.; Sanjuan, J.; Olivares, J.; Donaire, J.P. Time-course of lipoxygenase, antioxidant enzyme activities and H 2 O 2 accumulation during the early stages of Rhizobium-legume symbiosis. New Phytol. 2001,152, 91–96. [CrossRef] 42. Ramu, S.K.; Peng, H.M.; Cook, D.R. Nod factor induction of reactive oxygen species production is correlated with expression of the early nodulin gene rip1 in Medicago truncatula.Mol. Plant Microbe Interact. 2002 ,15, 522–528. [CrossRef] [PubMed] 43. Shaw, S.L.; Long, S.R. Nod factor inhibition of reactive oxygen efflux in a host legume. Plant Physiol. 2003 , 132, 2196–2204. [CrossRef] [PubMed] 44. Chang, C.; Damiani, I.; Puppo, A.; Frendo, P. Redox changes during the legume-rhizobium symbiosis. Mol. Plant 2009,2, 370–377. [CrossRef] [PubMed] 45. Damiani, I.; Pauly, N.; Puppo, A.; Brouquisse, R.; Boscari, A. Reactive oxygen species and nitric oxide control early steps of the legume-Rhizobium symbiotic interaction. Front. Plant Sci. 2016 ,7, 454. [CrossRef] [PubMed] 46. Sambrook, J.; Fritsch, E.F.; Maniatis, T. Molecular Cloning: A Laboratory Manual; Cold Spring Harbor Laboratory Press: Cold Spring Harbor, NY, USA, 1989. 47. Beringer, J.E. R factor transfer in Rhizobium leguminosarum.J. Gen. Microbiol. 1974 ,84, 188–198. [CrossRef] [PubMed] 48. Nogales, J.; Domínguez-Ferreras, A.; Amaya-Gómez, C.V.; van Dillewijn, P.; Cuéllar, V.; Sanjuán, J.; Olivares, J.; Soto, M.J. Transcriptome profiling of a Sinorhizobium meliloti fadD mutant reveals the role of rhizobactin 1021 biosynthesis and regulation genes in the control of swarming. BMC Genom. 2010 ,11, 157. [CrossRef] 49. Becker, A.; Schmidt, M.; Jäger, W.; Pühler, A. New gentamicin-resistance and lacZ promoter-probe cassettes suitable for insertion mutagenesis and generation of transcriptional fusions. Gene 1995 ,162, 37–39. [CrossRef] 50. Schäfer, A.; Tauch, A.; Jager, W.; Kalinowski, J.; Thierbach, G.; Pühler, A. Small mobilizable multi-purpose cloning vectors derived from the Escherichia coli plasmids pK18 and pK19: Selection of defined deletions in the chromosome of Corynebacterium glutamicum.Gene 1994,145, 69–73. [CrossRef] 51. Schwyn, B.; Neilands, J.B. Universal chemical assay for the detection and determination of siderophores. Anal. Biochem. 1987,160, 47–56. [CrossRef] 52. Pfaffl, M.W. A new mathematical model for relative quantification in real-time RT-PCR. Nucleic Acids Res. 2001,29, e45. [CrossRef] [PubMed] 53. Hidalgo, A.; Margaret, I.; Crespo-Rivas, J.C.; Parada, M.; Murdoch Pdel, S.; López, A.; Buendía-Clavería, A.M.; Moreno, J.; Albareda, M.; Gil-Serrano, A.M.; et al. The rkpU gene of Sinorhizobium fredii HH103 is required for bacterial K-antigen polysaccharide production and for efficient nodulation with soybean but not with cowpea. Microbiology 2010,156, 3398–3411. [CrossRef] [PubMed] 54. Buendía-Clavería, A.M.; Moussaid, A.; Ollero, F.J.; Vinardell, J.M.; Torres, A.; Moreno, J.; Gil-Serrano, A.M.; Rodríguez-Carvajal, M.A.; Tejero-Mateo, P.; Peart, J.L.; et al. A purL mutant of Sinorhizobium fredii HH103 is symbiotically defective and altered in its lipopolysaccharide. Microbiology 2003 ,149, 1807–1818. [CrossRef] [PubMed] 55. Buendía-Clavería, A.M.; Chamber, M.; Ruiz-Sainz, J.E. A comparative study of the physiological characteristics, plasmid content and symbiotic properties of different Rhizobium fredii strains. Syst. Appl. Microbiol. 1989,12, 203–209. [CrossRef] © 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).