The Role of Fnr Paralogs in Controlling Anaerobic Metabolism in the Diazotroph Paenibacillus polymyxa WLY78

The members of the nitrogen-fixing Paenibacillus spp. have great potential to be used as a bacterial fertilizer in agriculture. However, the functions of the fnr gene(s) in nitrogen fixation and other metabolisms in Paenibacillus spp. are not known. Here, we found that in P. polymyxa WLY78, Fnr1 and Fnr3 were responsible for regulation of numerous genes in response to changes in oxygen levels, but Fnr5 and Fnr7 exhibited little effect. Fnr1 and Fnr3 indirectly or directly regulated many types of important metabolism, such as nitrogen fixation, Fe uptake, respiration, and electron transport. This study not only reveals the function of the fnr genes of P. polymyxa WLY78 in nitrogen fixation and other metabolisms but also will provide insight into the evolution and regulatory mechanisms of fnr in Paenibacillus.

ABSTRACT

Fnr is a transcriptional regulator that controls the expression of a variety of genes in response to oxygen limitation in bacteria. Genome sequencing revealed four genes (fnr1, fnr3, fnr5, and fnr7) coding for Fnr proteins in Paenibacillus polymyxa WLY78. Fnr1 and Fnr3 showed more similarity to each other than to Fnr5 and Fnr7. Also, Fnr1 and Fnr3 exhibited high similarity with Bacillus cereus Fnr and Bacillus subtilis Fnr in sequence and structures. Both the aerobically purified His-tagged Fnr1 and His-tagged Fnr3 in Escherichia coli could bind to the specific DNA promoter. Deletion analysis showed that the four fnr genes, especially fnr1 and fnr3, have significant impacts on growth and nitrogenase activity. Single deletion of fnr1 or fnr3 led to a 50% reduction in nitrogenase activity, and double deletion of fnr1 and fnr3 resulted to a 90% reduction in activity. Genome-wide transcription analysis showed that Fnr1 and Fnr3 indirectly activated expression of nif (nitrogen fixation) genes and Fe transport genes under anaerobic conditions. Fnr1 and Fnr3 inhibited expression of the genes involved in the aerobic respiratory chain and activated expression of genes responsible for anaerobic electron acceptor genes.

IMPORTANCE The members of the nitrogen-fixing Paenibacillus spp. have great potential to be used as a bacterial fertilizer in agriculture. However, the functions of the fnr gene(s) in nitrogen fixation and other metabolisms in Paenibacillus spp. are not known. Here, we found that in P. polymyxa WLY78, Fnr1 and Fnr3 were responsible for regulation of numerous genes in response to changes in oxygen levels, but Fnr5 and Fnr7 exhibited little effect. Fnr1 and Fnr3 indirectly or directly regulated many types of important metabolism, such as nitrogen fixation, Fe uptake, respiration, and electron transport. This study not only reveals the function of the fnr genes of P. polymyxa WLY78 in nitrogen fixation and other metabolisms but also will provide insight into the evolution and regulatory mechanisms of fnr in Paenibacillus.

INTRODUCTION

Most biological nitrogen fixation is catalyzed by molybdenum-dependent nitrogenase, which is distributed within bacteria and archaea. This enzyme is composed of two metalloproteins: MoFe protein and Fe protein (1). Nitrogenase is an oxygen-sensitive enzyme, and both the MoFe and Fe proteins are irreversibly damaged by oxygen (2). O₂ exposure leads to inappropriate oxidation of the metalloclusters, decrease of protein secondary structure, and further degradation (3). Exposure to oxygen irreversibly inactivates the Mo-, V-, and Fe-nitrogenases (3–5). To avoid oxygen inactivation, diazotrophs (nitrogen-fixing organisms) have evolved different strategies. One of the strategies is to tightly control the transcription of nitrogen fixation (nif) genes in response to the external oxygen concentration.

Fnr (fumarate and nitrate reduction) protein is a global regulator that binds a [4Fe-4S] cluster to monitor the oxygen status in the cell and then controls transcription of many genes in response to changes in oxygen levels (6–9). Fnr is widely distributed in Gram-negative bacteria (e.g., Escherichia coli) (10) and Gram-positive bacteria (e.g., Bacillus subtilis) (11). Fnr-related transcriptional regulators of the Crp/Fnr (cyclic AMP-binding protein/fumarate nitrate reduction regulatory protein) family have been reported to be involved in nitrogen fixation of some Gram-negative diazotrophs (12–15). For example, Fnr proteins are indirectly involved in controlling the activity of NifA in Herbaspirillum seropedicae SmR1 by regulating respiratory activity in relation to oxygen availability (16, 17). The Fnr protein of Klebsiella oxytoca is required to relieve inhibition of NifA activity by its partner regulatory protein NifL under anaerobic conditions (14). In symbiotic Bradyrhizobium japonicum and Sinorhizobium meliloti, transcription of nifA and fix genes is predominantly controlled by the oxygen-responsive two-component FixL-FixJ system, together with FixK, which is a member of the Crp/Fnr superfamily, or by the redox-sensing system RegS-RegR (12, 13). In Rhizobium leguminosarum UPM791, FnrN is responsible for the expression of the high-affinity oxidase encoded by fixNOQP which supports growth under microaerobic conditions and is essential for nitrogen fixation (15).

Paenibacillus polymyxa WLY78 can fix nitrogen under anaerobic or microaerobic and nitrogen-limited conditions and has a nif operon composed of 9 genes, nifBHDKENX-hesA-nifV, under the control of a σ⁷⁰-dependent promoter in front of the nifB gene (18). Recently, we have revealed that GlnR of P. polymyxa WLY78 activates nif transcription under anaerobic and nitrogen-limited conditions, but GlnR together with glutamine synthetase (GS, glnA product) represses nif transcription under excess nitrogen and anaerobic conditions (19).

Here, we searched the genome of P. polymyxa WLY78 and found that there are four genes coding for Fnr proteins. A total of 12 fnr deletion mutants, including single, double, triple, and quadruple fnr deletion mutants, were constructed by homologous recombination. The growth rates and nitrogenase activities among these fnr mutants and wild-type (WT) P. polymyxa WLY78 were comparatively analyzed. Each of the Δfnr1, Δfnr3, Δfnr5, and Δfnr7 single deletion mutants was effectively complemented by its corresponding fnr gene and by the B. subtilis fnr gene. His-tagged Fnr1 and His-tagged Fnr3 proteins expressed and purified in E. coli under aerobic conditions were used to verify the target genes by electrophoretic mobility shift assay (EMSA). Genome-wide transcription analysis in P. polymyxa WLY78 and the Δfnr13 double mutant was performed.

RESULTS AND DISCUSSION

Identification of fnr genes in P. polymyxa.

Analysis of the P. polymyxa WLY78 genome showed four fnr-like genes (named fnr1 [S6001676], fnr3 [S6003218], fnr5 [S6004820], and fnr7 [S6005182]) (18). There was 39.98 to 53.63% identity among the four proteins Fnr1, Fnr3, Fnr5, and Fnr7 of P. polymyxa at the amino acid level (see Table S1 in the supplemental material). The highest (53.63%) identity was found between Fnr1 and Fnr3. Fnr1 and Fnr3 are more similar to each other than to Fnr5 and Fnr7. Like P. polymyxa WLY78, the three strains P. polymyxa M1, P. polymyxa E681, and P. polymyxa SC2 have four fnr genes. Each of the four fnr genes shows 99.44 to 100% identity with its corresponding gene from the different P. polymyxa strains (Table S1). However, some Paenibacillus species or strains, such as Paenibacillus polymyxa EBL06, Paenibacillus polymyxa Sb3-1, and Paenibacillus jamilae NS115, have only one Fnr, which has 16.34 to 34.78% identity with the four Fnr proteins of P. polymyxa WLY78. Also, Fnr1, Fnr3, Fnr5, and Fnr7 proteins of P. polymyxa share 50.94%, 52.99%, 45.45%, and 43.19% identities with the B. subtilis Fnr protein, respectively. Also, Fnr1, Fnr3, Fnr5, and Fnr7 proteins of P. polymyxa have 41.57%, 45.08%, 19.01%, and 21.82% identities with Bacillus cereus Fnr.

The four Fnr proteins of P. polymyxa WLY78 contain the predicted N-terminal receiver domain and C-terminal DNA-binding domain, which are characteristic features of the Crp/Fnr protein family (7). The [4Fe-4S]²⁺ cluster of B. subtilis Fnr is coordinated by three C-terminally located cysteine residues at positions 227, 230, and 235 and one aspartate residue at position 141 (7, 20). Similarly to B. subtilis Fnr, Fnr1 and Fnr3 proteins of P. polymyxa WLY78 have these conserved cysteine and aspartate residues, but Fnr5 and Fnr7 proteins of P. polymyxa WLY78 lack these conserved residues (Fig. 1A). The data suggest that Fnr1 and Fnr3 proteins of P. polymyxa WLY78 show high similarity with B. subtilis Fnr and B. cereus Fnr in sequence and structure.

FIG 1.

Homology analysis of Fnr proteins and phylogenetic analysis selected from Crp/Fnr superfamily in Firmicutes. (A) Alignments of Fnr proteins from P. polymyxa, B. cereus, and B. subtilis. Conserved cysteines required for binding of [4Fe-4S]²⁺ are indicated by red arrows. The double-underlined sequence represents the region of the N-terminal DNA-binding domain. The red underlined sequence represents the region of the sensory domain. Bs, B. subtilis; Bc, B. cereus; Pp, P. polymyxa WLY78. (B) The phylogenetic tree was constructed using the neighbor joining method, with bootstrap analyses of 1,000 cycles.

A phylogenetic analysis showed that the four P. polymyxa Fnr proteins fall into 3 groups (Fig. 1B). Fnr1 and Fnr3 are in the clade with the Fnr group of Bacillaceae. Fnr5 is near the clade with the Fnr group from Listeria and the FixK group of Sporanaerobacter and Clostridiales, while Fnr7 is divergent from the Fnr and FixK group of Bacillaceae. The data are consistent with the protein homology analysis.

Influence of fnr on growth under anaerobic conditions.

To explore the regulatory function of the four Fnr proteins of P. polymyxa WLY78, 12 unmarked fnr deletion (Δfnr) mutants, including single, double, triple, and quadruple deletion mutants, were constructed as described in Materials and Methods. The number in the Δfnr mutant designation indicates which fnr gene is deleted (e.g., Δfnr1 indicates deletion of the fnr1 gene and Δfnr13 indicates deletion of both fnr1 and fnr3 genes).

As Fnr protein is known to sense oxygen and plays a major role in altering gene expression during the switch from aerobic to oxygen-limiting conditions, the influence of fnr on the growth of P. polymyxa WLY78 under anaerobic conditions is here investigated. P. polymyxa WLY78 and multiple fnr deletion mutants were cultivated in nitrogen-deficient medium with Casamino Acids under anaerobic and aerobic conditions (Fig. 2A). Except for the Δfnr57 double fnr deletion mutant, all of the fnr deletion mutants showed lower growth rates than P. polymyxa WLY78 did. Compared to wild-type P. polymyxa WLY78, each single fnr deletion mutant showed a low growth rate. The Δfnr1357 quadruple fnr deletion mutant showed the lowest growth rates among all of the 12 Δfnr mutants, suggesting that the four fnr genes play roles under anaerobic conditions. Notably, the Δfnr1 and Δfnr3 single deletion mutants and the Δfnr13 double fnr deletion mutant showed very low growth rates, suggesting that Fnr1 and Fnr3 proteins play an important role in anaerobic metabolism in response to oxygen.

FIG 2.

The growth curve and nitrogenase activity of the fnr deletion mutants. (A) Influence of the fnr deletion on growth under anaerobic conditions. P. polymyxa WLY78 and the fnr deletion mutants were cultivated in nitrogen-deficient medium with Casamino Acids and no oxygen. (B) Influence of the fnr deletion on nitrogenase activity under anaerobic conditions. The nitrogenase activity of P. polymyxa WLY78 and the fnr deletion mutants was measured by acetylene reduction assay when grown anaerobically in nitrogen-deficient medium.

Effects of fnr on nitrogenase activity.

Since nitrogenase is very sensitive to O₂, nitrogen fixation is performed under anaerobic or microanaerobic conditions. To determine if Fnr proteins are related to nitrogen fixation, the nitrogenase activities of wild-type P. polymyxa WLY78 and multiple fnr deletion mutants grown anaerobically in nitrogen-deficient medium were measured by using the method of the reduction of acetylene to ethylene (21, 22). As shown in Fig. 2B, the nitrogenase activities of Δfnr1 and Δfnr3 mutants were decreased to about 50% of the wild-type level, while the activities of Δfnr5 and Δfnr7 mutants were decreased to about 73 to 79% of the wild-type level, and the nitrogenase activities of Δfnr37 and Δfnr13 mutants were decreased to about 36% and 10% of the wild-type level, respectively. Notably, the nitrogenase activities of Δfnr137 and Δfnr1357 mutants were nearly lost. The data are consistent with the growth rates of these mutants observed as described above. The results imply that the four fnr genes, especially fnr1 and fnr3, play roles in nitrogen fixation.

Furthermore, complementation of Δfnr1, Δfnr3, Δfnr5, and Δfnr7 mutants with the corresponding P. polymyxa fnr gene and the B. subtilis fnr gene under the control of its own promoter was performed. As shown in Fig. S1 in the supplemental material, fnr1, fnr3, fnr5, and fnr7 from P. polymyxa WLY78 in complemented strains (Δfnr1C, Δfnr3C, Δfnr5C, and Δfnr7C) restored the nitrogenase activity of their corresponding mutants to more than 90% of wild-type activity. Complementation with His-tagged Fnr1 and His-tagged Fnr3 in complemented strains (Δfnr1Chis and Δfnr3CHis) also restored the nitrogenase activity of the corresponding mutant to more than 90% of the wild-type activity. Moreover, we found that the B. subtilis fnr gene greatly improved the nitrogenase activity of the four single fnr deletion mutants, especially the activities of Δfnr1 and Δfnr3 mutants. Also, the B. subtilis fnr gene greatly restored the nitrogenase activities of the Δfnr13, Δfnr137, and Δfnr1357 multiple deletion mutants. The data confirm that the four fnr genes of P. polymyxa WLY78, especially fnr1 and fnr3, play an important role in nitrogen fixation. The data suggest that Fnr1 and Fnr3 of P. polymyxa and Fnr of B. subtilis are similar in function.

nifH transcription in the fnr deletion mutants.

The levels of nifH transcript in different mutants were assayed by quantitative real-time PCR (qRT-PCR) (Fig. S2A). The nifH gene in the Δfnr3 mutant was expressed at basal level. The levels of nifH transcript in Δfnr1, Δfnr5, and Δfnr7 mutants were decreased to about 40%, 70%, and 90% of wild-type levels, respectively, while the levels of nifH transcript in Δfnr13 and Δfnr1357 mutants were nearly lost. Furthermore, the effects of fnr on nif expression were determined by measuring the β-galactosidase activity of P. polymyxa WLY78 and fnr deletion mutants carrying a transcriptional lacZ fusion to the nif promoter region (Pnif-lacZ fusion). Compared to the wild type, the Δfnr1, Δfnr3, Δfnr13, and Δfnr1357 mutants nearly lost β-galactosidase activities (Fig. S2B), consistent with the nitrogenase activities and the levels of nifH transcript observed in these fnr mutants.

Overall, deletion mutation of fnr1 or fnr3 led to decreases in growth rate, nitrogenase activity, and nifH transcript, and double deletion of fnr1 and fnr3 resulted in more serious impacts on growth, nitrogenase activity, and nifH transcript than single deletion mutations of fnr1 or fnr3. Especially, the quadruple deletion mutation of fnr1, fnr3, fnr5, and fnr7 resulted in the lowest growth rates among all fnr deletion mutants and near loss of the nitrogenase activity and nifH transcript. It was reported previously that there are three fnr-like genes and that Fnr1 and Fnr3 were indirectly involved in controlling the activity of NifA in H. seropedicae (16). Fnr1 and Fnr3 of H. seropedicae directly regulate discrete groups of promoters (groups I and II, respectively), while Fnr3-Fnr1 heterodimers regulate a third group (group III) promoters (23). The specific interaction between Fnr1 and Fnr3 of H. seropedicae has been determined by using two-hybrid assays (23). Whether a heterodimer is formed between Fnr1 and Fnr3 or between Fnr1 or Fnr3 and Fnr5 or Fnr7 of P. polymyxa WLY78 needs to be studied in the near future.

Prediction and verification of Fnr target genes.

To decipher the Fnr regulon of P. polymyxa WLY78, its target genes were predicted. According to the known Fnr-binding sequence of Bacillus and Paenibacillus in RegPrecise (https://enigma.lbl.gov/regprecise/), the position weight matrix (PWM) of the Fnr-binding site was constructed using MEME (http://meme-suite.org). The Fnr-binding consensus motif, composed of a 16-bp palindromic sequence, 5′-TGTGA-N₆-TCACA-3′, was determined. Then, we used the 16-bp Fnr consensus binding motif to scan the regions from −350 to +50 bp relative to the translational start codon (ATG) of genes in the P. polymyxa WLY78 genome with the MAST application (http://meme-suite.org) (24). A total of 143 putative Fnr target genes with an E value of ≤10 (the smaller the E value, the greater the probability) from the P. polymyxa WLY78 genome were identified (Table S2). As annotated by the COG (Clusters of Orthologous Groups), the 143 putative target genes were allocated to 12 groups by biological function. Of the 143 putative target genes, 19 belong to regulatory genes, 23 genes are related to energy metabolism, 11 genes are related to carbon metabolism, 54 genes are related to other metabolisms, and 36 are the genes whose functions are unknown or unclassified (Table S2). As shown in Table S2, there is 1 Fnr-binding site in most of the 143 putative target genes, and there are 2 to 3 Fnr-binding sites in some putative target genes.

To investigate the accuracy of the predicted Fnr-binding site in the putative target genes by electrophoretic mobility shift assays (EMSAs), Fnr1 with His₆ tags at its N terminus and Fnr3 with His₆ tags at its C terminus were expressed and purified in E. coli under aerobic conditions. Both of the purified recombinant protein solutions did not exhibit the characteristic brown color, suggesting that the [4Fe-4S]²⁺ cluster was oxidized by O₂. EMSA was done with the purified Fnr1 and Fnr3 proteins under aerobic conditions. As shown in Fig. 3, both Fnr1 and Fnr3 proteins could bind to the promoter regions with the predicted Fnr-binding sites of the 8 operons: qoxABCD (encoding cytochrome aa₃ quinol oxidase), narGHJI (encoding nitrate reductase), ndh (encoding NADH dehydrogenase), hemN3 (encoding oxygen-independent coproporphyrinogen III oxidase), hydEG (encoding [FeFe] hydrogenase), nrdDG (encoding anaerobic ribonucleoside triphosphate reductase), pflBA (encoding formate acetyltransferase), and resDE (encoding two-component regulatory proteins). However, the promoter regions of the cydABCD operon (encoding cytochrome bd ubiquinol oxidase) and narK (encoding nitrate/nitrite transporter) were bound only by Fnr3. Moreover, EMSA showed that neither Fnr1 nor Fnr3 could bind to the promoter regions of nif and feoAB (ferrous-iron transporter FeoAB), consistent with the fact that there was no Fnr-binding site in the promoter regions of these genes. However, EMSA showed no binding of Fnr1 and Fnr3 to the promoter region of glnRA with an Fnr-binding site.

FIG 3.

Fnr-binding sites predicted by software and verification by electrophoretic mobility shift assays (EMSAs). (A) Consensus sequence of the predicted Fnr-binding sites. (B) In vitro binding of NHis₆-Fnr1 and Fnr3-CHis₆ to the promoter region of some Fnr target genes. The DNA fragment (179 to 442 bp) with a concentration of 0.03 pmol was used. The minus sign in lane 1 indicates EMSAs without NHis₆-Fnr1 or Fnr3-CHis₆. Lanes 2 to 5 contained increasing concentrations (0.05, 0.2, 2, and 6 μM) of NHis₆-Fnr1 or Fnr3-CHis₆. S and N indicate competition assays with a 100-fold excess of unlabeled specific probe and nonspecific competitor DNA, respectively. Arrowheads, free probes; Brackets, DNA-protein complexes.

The above-described data demonstrated that His-tagged Fnr1 and His-tagged Fnr3 of P. polymyxa WLY78 could bind to the promoter regions with the Fnr-binding site (5′-TGTGA-N₆-TCACA-3′), suggesting that the aerobically purified Fnr1 and Fnr3 proteins were active forms. A similar report found that both B. cereus Fnr tagged with His at its C terminus (Fnr_His) and Fnr tagged with streptomycin (Strep) at its N terminus (_StrepFnr) were active when expressed and purified in E. coli under oxic conditions. In vitro, the aerobically purified B. cereus Fnr as a monomer bound to the promoter regions of fnr itself, resDE, plcR, and the structural enterotoxin genes hbl and nhe (25). Unlike B. cereus Fnr, B. subtilis Fnr existed in an inactive state under aerobiosis, due to the [4Fe-4S]²⁺ cluster of Fnr being converted by O₂ to [2Fe-2S]²⁺. The B. subtilis Fnr formed a stable dimer under aerobic and anaerobic conditions independently of Fe-S cluster formation, but DNA binding of Fnr was dependent on the presence of an intact [4Fe-4S]²⁺ cluster (11). As a member of the Crp/Fnr family of transcription factors, Fnr should function as a dimer in vivo. It is known that many transcription factors as dimers bind to their specific DNA sites, and there are two pathways to form dimeric protein-DNA complexes. Dimer pathway implies that the protein can dimerize first and then associate with DNA, and monomer pathway means that two protein monomers bind DNA sequentially and form their dimerization interface while bound to DNA (26). It was proposed previously that B. cereus Fnr takes a sequential monomer-binding pathway to form a dimer, but B. subtilis Fnr as a homodimer binds to its specific DNA-binding site and activates transcription (11). Our results suggest that Fnr1 and Fnr3 of P. polymyxa WLY78 behaved as B. cereus Fnr did. Thus, we deduce that in vivo Fnr1 and Fnr3 proteins of P. polymyxa WLY78 may bind to the specific promoter region by a sequential monomer-binding pathway just as B. cereus Fnr did.

Transcriptome sequencing (RNA-Seq) analysis of wild-type and Δfnr13 strains.

To assess the effects of Fnr1 and Fnr3 proteins on global gene expression under anaerobic conditions, genome-wide transcription analysis of P. polymyxa WLY78 and Δfnr13 mutant cultured under N₂-fixing conditions (without O₂ and NH₄⁺) was performed. Transcripts showing statistically significant differences with a q value (P adjusted) of ≤0.05 and a |log₂ fold change (FC)| of ≥1 were accepted as candidate differential expression genes (DEGs). Of the 5,661 genes contained in the genome of P. polymyxa WLY78, 301 genes, including 202 genes and operons, were differentially expressed in the Δfnr13 mutant compared to the wild type (Table S3). Of the 301 genes, 116 were markedly upregulated, indicating that they were directly or indirectly repressed by Fnr, and 185 were significantly decreased, suggesting that they were directly or indirectly activated by Fnr.

Influence of fnr genes on transcription of the nif and glnRA genes.

Our previous study showed that the 9 genes nifBHDKENX-hesA-nifV are organized as a nif operon in P. polymyxa WLY78 (18). In this study, we found that the expression levels of the 9 genes within the nif operon in the Δfnr13 mutant were significantly downregulated by 6.51 to 7.47 log₂FC (Fig. 4A). The data are consistent with the decreased nitrogenase activity and nifH transcript of the Δfnr13 mutant. However, there was no predicted Fnr-binding site in the promoter region of the nif gene operon and EMSA also showed that Fnr1 or Fnr3 did not bind to the promoter region of the nif gene operon. These results indicate that Fnr1 and Fnr3 indirectly activated the expression of the nif gene operon under anaerobic conditions. It is known that GlnR, a global regulator of nitrogen metabolism, is required for the nif gene transcription under anaerobic and nitrogen-limited conditions (19). However, expression of the glnRA operon was upregulated in the Δfnr13 mutant, suggesting that glnR expression was not in coordination with the nif gene expression. Although there is one Fnr-binding site in the promoter region of the glnRA operon, EMSA showed that there was no binding of Fnr1 or Fnr3 to its promoter. We do not know whether Fnr5 or Fnr7 could bind to the promoter region of the glnRA operon. In contrast to our results, the combined deletions in both the fnr1 and fnr3 genes in H. seropedicae led to higher expression of nifA, nifB, and nifH, which was probably a consequence of their influence on respiratory activity in relation to oxygen availability (16). It was shown previously that Fnr was required for relief of NifL inhibition in K. oxytoca under anaerobic conditions (14).

FIG 4.

Differential expression of the nif genes and iron transport genes. (A) Differential expression of the 9 nif genes. (B) Differential expression of the genes involved in iron transport. FC in log₂FC indicates fold change (the read count ratio of Δfnr13 and wild-type strains).

Influence of the fnr1 and fnr3 genes on transcription of the Fe transporter genes.

Fe is an essential element for nitrogenase. Fe is the soluble Fe²⁺ form (ferrous iron) under anaerobic conditions or at acidic pH, and the major route for bacterial ferrous iron uptake is via the Feo (ferrous iron transport) system, composed of FeoA, FeoB, and FeoC (27, 28). At neutral pH, iron is the poorly soluble Fe³⁺ form (ferric iron), which is often biologically unavailable (29). Many bacteria excrete ferric chelators, called siderophores, to take up Fe³⁺. Usually, bacteria take up ferric complexes, including ferric hydroxamate (FhuCDBA), ferric citrate (YfmCDEF), ferric-heme, and the ferric-bacillibactin uptake system (FeuABC) (30).

Our study showed that 36 Fe transporter genes in the Δfnr13 mutant were downregulated 1.59 to 7.65 log₂FC (Fig. 4B and Table S3). Of the 36 Fe transporter genes, only the feoA-feoB operon is involved in Fe²⁺ uptake and the other 34 genes belong to Fe³⁺ transport systems. The highest differentially expressed genes, feoA and feoB, were downregulated 7.18 to 7.65 log₂FC. The 13 fhu genes belonging to the ferric hydroxamate system were downregulated from 5.32 to 1.59 log₂FC. Especially, transcriptions of yfmCDE, involved in the ferric citrate transport system, and isdHCBAE, involved in the ferric-heme transport system, were also downregulated in the Δfnr13 mutant. The data are consistent with our recent reports that all of the Fe transporter genes were upregulated under N₂-fixing conditions (without O₂ and NH₄⁺) (31). As described above, there are no Fnr-binding sites in the promoter regions of the 36 Fe transporter genes and EMSA also showed that Fnr1 or Fnr3 did not bind to the promoter region of the feoAB operon. These results suggest that Fnr1 and Fnr3 indirectly activated the expression of Fe transporter genes under anaerobic conditions. As shown in Table S2, there are two Fnr-binding sites in the promoter region of the fur3 gene whose predicted product is a regulatory protein. There are predicted Fur3-binding sites in the promoter region of some Fe transporter genes, such as feoAB, efeO, and fhuD. Thus, we deduce that Fnr1 and Fnr3 of P. polymyxa WLY78 indirectly activated the expression of Fe transporter genes by regulating expression of Fur3, which directly controls transcription of Fe transporter genes under anaerobic conditions.

Influence of fnr genes on transcription of respiration and energy metabolism genes.

Based on the genome sequence, the respiratory chain of P. polymyxa WLY78 is shown in Fig. 5A. It is composed of several dehydrogenases that transfer electrons to an intramembrane pool of menaquinone and some terminal oxidases responsible for reoxidation of menaquinol. The terminal oxidases include at least two types: one consisting of a cytochrome bd-type quinol oxidase and the second one consisting of cytochrome aa₃ oxidase.

FIG 5.

Differential expression of the genes related to anaerobic respiration and energy metabolism. (A) Schematic representation of the probable components of anaerobic respiration and energy metabolism in P. polymyxa WLY78 based on the genome annotation. The gray area represents the components whose transcripts are downregulated in the Δfnr13 mutant. (B) Differential expression of the genes represented in the schema in panel A. FC in log₂FC indicates fold change. The horizontal line above or below genes indicates that these genes are in the same transcription unit. The asterisk indicates that the promoter region of the gene contains a predictive Fnr-binding site.

The dehydrogenases that play an important role in respiration in Gram-positive Corynebacterium glutamicum include a non-proton-pumping NADH dehydrogenase, encoded by the ndh gene; malate:quinone oxidoreductase, encoded by the mqo gene; and succinate dehydrogenase, encoded by the sdhCAB genes (32). Here, we found that there were 13 genes encoding dehydrogenases that were differentially expressed (Table S3). Of these genes, ndh and sdhABC, the major dehydrogenase genes in the respiratory chain, were upregulated 1.45 to 2.57 log₂FC. Other dehydrogenase genes, such as yutJ (NADH dehydrogenase), yugK (probable NADH-dependent butanol dehydrogenase), hcaD [NAD(FAD)-dependent dehydrogenases], and ldh (l-lactate dehydrogenase), were upregulated 1.76 to 8.44 log₂FC, while glpD, alkH (aldehyde dehydrogenase), adhE, fdhD (formate dehydrogenase), and adhP genes were obviously downregulated 2.48 to 5.87 log₂FC (Table S3). EMSA showed that NHis₆-Fnr1 and Fnr3-CHis₆ could bind to the promoter regions of ndh with the predicted Fnr-binding site (Fig. 3B). qoxABCD, encoding cytochrome aa₃-type oxidase, and cydABCD, encoding cytochrome bd-type oxidase, were upregulated 1.8 to 4.5 log₂FC.

Many bacteria are able to grow anaerobically using alternative electron acceptors, including nitrate or fumarate (33). We found that anaerobic electron acceptor genes narGHJI (nitrate reductase), nasABCD (nitrite reductase), and narK (nitrate/nitrite transporter) were downregulated from 0.6- to 4.6-fold in the Δfnr13 mutant. As described above, the Fnr-binding sites in the upstream region of narGHJI and narK were predicted and confirmed by EMSA. Thus, the results suggest that Fnr1 and Fnr3 directly activated expression of narGHJI and narK in anaerobiosis and indirectly activated expression of nasABCD. The results are consistent with previous studies showing that the expression of narGHJI was intensely induced by anaerobic conditions and the induction was dependent on Fnr in B. subtilis (34). It was reported previously that ResD-ResE (two-component regulatory proteins) and Fnr were indispensable for nitrate respiration in B. subtilis (35, 36). Here, we show also that expression of resDE, whose promoter has one Fnr-binding site, was downregulated 1.4 to 1.6 log₂FC in the Δfnr13 mutant, consistent with the report that B. cereus Fnr regulated expression of resDE (25). The data indicate that Fnr1 and Fnr3 inhibited expression of the genes involved in the aerobic respiratory chain and activated expression of genes responsible for anaerobic electron acceptor genes.

Transcriptional analysis of the potential electron transporters for nitrogenase.

Nitrogen fixation is a process in which electrons originating from low-potential electron carriers, such as flavodoxin or ferredoxin molecules, were transferred to molecular N₂. In K. oxytoca, the electron was produced by pyruvate:flavodoxin oxidoreductase (encoded by nifJ) during the tricarboxylic acid cycle (TCA), and then a flavodoxin (encoded by nifF) mediated electron transfer to the Fe protein of nitrogenase (37).

At present, we do not know the specific electron transfer system for nitrogen fixation in P. polymyxa WLY78. As shown in Table S3 and Fig. 6A, two transcripts, hydEG, located on the plus strand, and COG0196 fdhF hycB hydAN aspA hydFG, located on the minus strand, were significantly downregulated. hydA encodes Fe-Fe hydrogenase, whose synthesis relies on maturation factors HydF (GTPase), HydE, and HydG (38), while hydB encodes ferredoxin and fdhF encodes formate dehydrogenase. These genes were downregulated in the Δfnr13 mutant. It has been reported for Clostridium that electrons produced by the oxidation of pyruvate are transferred to the acceptor ferredoxin and then the ferredoxin can act as an electron donor to reduce Fe-Fe hydrogenase HydA to produce hydrogen (39). In addition, the expression of fldA (flavodoxin), hemN1 and hemN3 (hemN encodes oxygen-independent coproporphyrinogen III oxidase), and COG1249 (encoding FAD-dependent oxidoreductase) was also downregulated. In contrast, hmp (flavohemoprotein), wrbA (multimeric flavodoxin), ywnB (NADH-flavin reductase), ribE (riboflavin synthase), and groS-groL (chaperonin) were upregulated (Fig. 6B). These differently expressed genes may play important roles in transferring electron to nitrogenase. However, fldB (flavodoxin), flr (flavoredoxin), ydfE (flavoprotein oxygenases), porG-porA (pyruvate:ferredoxin oxidoreductase), ywcH3 (flavin-dependent oxidoreductases), and ywcH1 (flavin-dependent oxidoreductases) were not differentially expressed (Table S3).

FIG 6.

Differential expression of the genes related to electron transport and carbon metabolism. (A) Differential expression of the genes in electron transport. (B) Differential expression of the genes in carbon metabolism. (C) Schematic representation of the probable components of carbon metabolism (glycolysis, TCA cycle, and fermentation) in P. polymyxa WLY78 based on the genome annotation. Blue and red indicate the components whose transcripts are downregulated and upregulated in the Δfnr13 mutant, respectively. P, phosphate; DHAP, dihydroxyacetone phosphate; PEP, phosphoenolpyruvate. The horizontal line above genes indicates that these genes are in the same transcription unit. The asterisk indicates that the promoter region of the gene contains the predictive Fnr-binding sites.

Influence of fnr genes on transcription of genes involved in carbon metabolism.

Genes involved in carbon metabolism (such as glycolysis, the Krebs [TCA] cycle, and fermentation) are shown in Fig. 6C and Table S3. The downregulated (1.33 to 3.28 log₂FC) genes in the Δfnr13 mutant were involved in glycolysis. These genes included ptsG (encoding a glucose-specific component in the phosphotransferase system [PTS]), pfkA (encoding ATP-dependent 6-phosphofructokinase), fbaA (encoding fructose-bisphosphate aldolase), pgm (encoding β-phosphoglucomutase), and pykA (encoding pyruvate kinase). However, Fnr-binding sites were not found in the upstream regions of these genes, suggesting that Fnr indirectly affected expression of the genes involved in glycolysis.

Many genes that participate in formate and ethanol metabolism were significantly downregulated in the Δfnr13 mutant, such as pflBA (encoding formate acetyltransferase), fdhD (encoding formate dehydrogenase), adhE (encoding aldehyde-alcohol dehydrogenase), and alkH (encoding aldehyde dehydrogenase). Multiple Fnr-binding sites in the upstream regions of pflBA and adhE were predicted, and EMSA also showed the binding of Fnr1 and Fnr3 to the promoter of pflBA (Fig. 3B and Table S2). This implies that Fnr1 and Fnr3 may directly regulate expression of these genes under anaerobic conditions. In contrast, ldh, encoding l-lactate dehydrogenase, was significantly upregulated by 6.1-fold, but there was no predicted Fnr-binding site in the promoter region of this gene.

Many genes in the Krebs cycle were significantly upregulated, from 1.21- to 4.21-fold, and they included citZ (encoding citrate synthase), citB (encoding aconitate hydratase), icd (encoding isocitrate dehydrogenase), odhAB (encoding 2-oxoglutarate dehydrogenase), sucCD (encoding succinyl coenzyme A [CoA] synthase), and sdhABC (encoding succinate dehydrogenase). However, there were no predicted Fnr-binding sites in the upstream region of these genes. These data suggest that Fnr1 and Fnr3 indirectly activated expression of genes involved in glycolysis and indirectly inhibited expression of genes involved in the Krebs cycle in P. polymyxa WLY78.

Taken together, Fnr1 and Fnr3 of P. polymyxa WLY78 indirectly activated expression of nif genes and Fe transport genes. However, Fnr1 and Fnr3 inhibited expression of the genes involved in the aerobic respiratory chain and activated expression of genes involved in carbon metabolism and responsible for anaerobic electron acceptor genes, which might provide O₂ protection and energy for nitrogenase. EMSA showed that the aerobically purified Fnr1 and Fnr3 could bind to the specific target DNA in vitro as B. cereus Fnr did.

MATERIALS AND METHODS

Strains and media.

P. polymyxa WLY78 used here was isolated from the rhizosphere of bamboo by our laboratory (40). P. polymyxa and Δfnr mutants were routinely grown at 30°C in LB or LD medium (per liter, containing 5 g NaCl, 5 g yeast, and 10 g tryptone) with shaking. When appropriate, antibiotics were added in the following concentrations: 12.5 mg/ml tetracycline, 5 mg/ml erythromycin, and 100 mg/ml ampicillin for maintenance of plasmids.

Nitrogen-deficient medium was used for assay of nitrogenase activity. Nitrogen-deficient medium contained (per liter) 10.4 g Na₂HPO₄, 3.4 g KH₂PO₄, 26 mg CaCl₂·2H₂O, 30 mg MgSO₄, 0.3 mg MnSO₄, 36 mg ferric citrate, 7.6 mg Na₂MoO₄·2H₂O, 10 μg p-aminobenzoic acid, 5 μg biotin, 4 g glucose as carbon source, and 2 mM glutamate as nitrogen source (40).

Nitrogenase activity assays.

For nitrogenase activity assays, P. polymyxa WLY78 and Δfnr mutants were grown in 50 ml LD medium (supplemented with antibiotics when necessary) in 250-ml test tubes shaken at 250 rpm for 16 h at 30°C. The cultures were collected by centrifugation, washed three times with sterilized water, and then resuspended in nitrogen-deficient medium containing 2 mM glutamate to a final optical density at 600 nm (OD₆₀₀) of 0.3 to 0.5. Then, 3 to 5 ml of suspension was transferred to a 26-ml test tube which was sealed with a rubber stopper. The headspace in the tube was then vacuumed and filled with argon gas (41). After C₂H₂ (10% of the headspace volume) was injected into the test tubes, the cultures were incubated at 30°C with shaking at 250 rpm. After incubation for 4 to 8 h, 100 μl of gas was withdrawn through the rubber stopper with a gastight syringe and manually injected into the gas chromatograph (HP6890) to quantify ethylene production. All treatments were in three replicates, and all the experiments were repeated three or more times.

β-Galactosidase assays.

To confirm whether deletion of fnr genes affects nif gene transcription, P. polymyxa WLY78 and 12 fnr mutants were transformed with a recombinant plasmid carrying the nif promoter-lacZ fusion (Pnif-lacZ fusion) (19). β-Galactosidase activity was assayed according to the method described by Wang et al. (19).

Identification and sequence alignment of P. polymyxa Fnr proteins.

The sequences of Fnr1, Fnr3, Fnr5, and Fnr7 from P. polymyxa WLY78 were aligned with that of the Fnr from Bacillus subtilis subsp. subtilis strain 168 (reference sequence NP_391612.1) and Bacillus cereus F4430-73 (reference sequence KMP55664.1) using Clustal W software. The conserved domains in the Fnr proteins were investigated by sequence searching to the Pfam database (http://pfam.sanger.ac.uk/). The secondary structure elements in the Fnr proteins were defined by ESPript 3.0 algorithm (42).

Phylogenetic analysis.

In the nonredundant NCBI database, amino acid sequences were obtained by performing a BLASTP search. Multiple gene alignments were carried out with molecular evolutionary genetics analysis (MEGA) (43). The neighbor-joining trees were constructed, and 1,000 bootstraps were done by using the MEGA 7.0.14 software.

Construction of Δfnr mutants.

Here, 12 Δfnr mutants, including Δfnr1, Δfnr3, Δfnr5, and Δfnr7 single mutants; Δfnr13, Δfnr17, Δfnr35, Δfnr37, and Δfnr57 double mutants; Δfnr137 and Δfnr357 triple mutant; and the Δfnr1357 quadruple mutant, were constructed. The unmarked, single, and multiple fnr deletion mutants were constructed via homologous recombination using the suicide plasmid pRN5101 as described previously (19). The upstream and downstream fragments flanking the coding region of fnr1, fnr3, fnr5, and fnr7 were PCR amplified from the genomic DNA of P. polymyxa WLY78, respectively. The primers used for deletion mutagenesis are listed in Table 1. The upstream and downstream fragments of four fnr genes were then fused with BamHI/HindIII-digested vector pRN5101 in Gibson assembly master mix (New England Biolabs), generating the four recombinant plasmids. Then, each of these recombinant plasmids was transformed into P. polymyxa WLY78 as described previously (19), and the single-crossover transformants were screened for erythromycin resistance (Em^r). Subsequently, marker-free deletion mutants (the double-crossover transformants) were selected from the initial Em^r transformants after several rounds of nonselective growth at 39°C. The marker-free deletion mutants were confirmed by PCR amplification and DNA sequencing analysis. The multiple fnr deletion mutants were constructed via the same method in the single fnr deletion mutant background.

TABLE 1.

Primers used in this study^a

Primer	Sequence (5′→3′)	PCR product
fnr1upF	ACGATGCGTCCGGCGTAGAGGATCCGCTGTGCTTGAGTTGATAG	Upstream region of fnr1
fnr1upR	ACGATGCGTCCGGCGTAGAGGATCCCTATAAGATCATATCATGT
fnr1dnF	GCAGCACACCGTTCCATACAAGTGGC	Downstream region of fnr1
fnr1dnR	CGCAAAAGACATAATCGATAAGCTTTGCCGTCCAATAGGGC
fnr3upF	ACGATGCGTCCGGCGTAGAGGATCCACCGATGCTTGCTATTGAAC	Upstream region of fnr3
fnr3upR	GATACAGCTTGTGAACTATGCCCTCG
fnr3dnF	AGTTCACAAGCTGTATCTCCTTCCCAG	Downstream region of fnr3
fnr3dnR	CGCAAAAGACATAATCGATAAGCTTCACATCAGGGTGGTTCATC
fnr5upF	ACGATGCGTCCGGCGTAGAGGATCCAAGACATACCGCTCCAGC	Upstream region of fnr5
fnr5upR	CGAGCCACTTGAGTGAACTGGCAGGAC
fnr5dnF	TTCACTCAAGTGGCTCGTAGAGTATGG	Downstream region of fnr5
fnr5dnR	CGCAAAAGACATAATCGATAAGCTTGGAGGCAATGGTTTATGG
fnr7upF	ACGATGCGTCCGGCGTAGAGGATCCTCGGATGTGTGGATGCTTT	Upstream region of fnr7
fnr7upR	CACGCAATAATTTGCGAGCCGTCTTCA
fnr7dnF	CTCGCAAATTATTGCGTGCTTCAAGCG	Downstream region of fnr7
fnr7dnR	CGCAAAAGACATAATCGATAAGCTTCCCTATGGTCTGGTATTG
fnr1CF	AAAGAGCTCACACGAGGAGGACGCTAA	fnr1 gene for constructing the complementary strains
fnr1CR	CGCAAGCTTCCCGCCTAATGGTATTCATA
fnr3CF	AAAGAGCTCAACCGATTGCAAGTAATG	fnr3 gene for constructing the complementary strains
fnr3CR	CACAAGCTTCGATTTATTATATCCGGCAA
fnr5CF	AAAGTCGACGATTGAAGGTGAGCTTAG	fnr5 gene for constructing the complementary strains
fnr5CR	CACAAGCTTACTACATGCTCCTATTACA
fnr7CF	AAAGGATCCTATCACCTCAAGAGCGGC	fnr7 gene for constructing the complementary strains
fnr7CR	CAAGTCGACTGCTGGAGAGTGGTTTTC
fnrCBsF	AAAGGATCCTTTCAGAGGTGGCGTTA	fnr of B. subtilis for constructing the complementary strains
fnrCBsR	CCGAAGCTTCAGTCAATATTGCAAATC
fnr1p1F	CGCGAATTCGGGCAGTGAGATCACCAT	Promoter region of fnr1
fnr1p1R	AAAGGATCCGTTAGCGTCCTCCTCGTG
Fnr1F	AACGAGCTCCATGCAAGCCACTTGTA	fnr1 coding region of fnr1 for constructing expression vector
Fnr1R	CCGCTCGAGTTAGAGGCGACAAATTT
Fnr3F	CGGGAGCTCAATGATGAAAGATCGAG	fnr3 coding region of fnr1 for constructing expression vector
Fnr3R	CCGCTCGAGTATCCGGCAAATCTCA
QnifHF	AACAGCCGGAATACGGACC	nifH for qRT-PCR
QnifHR	ACCTGCCAGCTCTTCATACTC
Q1288F	GGTTGCGACACACATACTCG	feoA for qRT-PCR
Q1288R	ACTTCCGATGCCAGTACCAT
Q2029F	GATACACCGGCTAACGAAGC	moaA for qRT-PCR
Q2029R	ATCTTCCACTCCGTCAATGG
Q3365F	TGTCAGATGATTGCCAGCTC	narG for qRT-PCR
Q3365R	CCTGCCGAATAACTCACCAT
Q2206F	CATGAGGATGACGAACTGGA	hemN3 for qRT-PCR
Q2206R	CGCCAATAGGTCATGTTGTG
Q687F	CGGCTTACTTCGGTATTCCA	ndh for qRT-PCR
Q687R	TTGGTTACCGGATTGCTTGT
Q3369F	TGATGGTCACTGATGCCAAG	narK for qRT-PCR
Q3369R	TGGAAGCCGGAGTTGAGTAT
PhemN3F	TTTGGCAATCCGCAGTTG	302-bp promoter region of hemN3 for EMSA
PhemN3R	CAAATCCCAACCTTCCACTT
PndhF	TCTGGCAACTGACCCTGTCT	282-bp promoter region of ndh for EMSA
PndhR	ATGTCCATCCTCCGTTCTCG
PnrdDF	ACCGACGACCCTATTCATTT	274-bp promoter region of nrdD for EMSA
PnrdDR	TTCCTCCAGCACCTTCATC
PpflBF	GAAAGGCAAGGCTAAGGA	248-bp promoter region of pflB for EMSA
PpflBR	CTCTCAATCACCGACATG
PfeoAF	ACGCCGCATACACATTGCT	220-bp promoter region of feoA for EMSA
PfeoAR	ACTGTCCCATTACTCGTCT
PglnRF	TCCGCAGATGCTCCTATTC	217-bp promoter region of glnR for EMSA
PglnRR	TCGTCGCCCATTTGTCA
PresDF	GCCATTTCACACCCAGTCA	300-bp promoter region of resD for EMSA
PresDR	CGGTTGCTTTGTTCTGACA
PnarGF	CTTTAGCAGAGGGAGGCTT	179-bp promoter region of narG for EMSA
PnarGR	CGATGACCAACACTCCTTT
PnarKF	ACAGGCGAAGAACTGCT	305-bp promoter region of narK for EMSA
PnarKR	ACCATGTATAGCTCCCC
PqoxAF	AACCAACGCAATAATCGC	315-bp promoter region of qoxA for EMSA
PqoxAR	AGGATGATAACTGTAGGG
PcydAF	AGAGGGCTGTTGATAGAAT	442-bp promoter region of cydA for EMSA
PcydAR	CCGTGCCAAATCCAATA
PhydEF	TGAGGCGGAAGTTGAAGA	180-bp promoter region of hydE for EMSA
PhydER	TCAGCCAGACGCAGTAA
nifpF	AATTAAAGTTTCTTATCTC	259-bp promoter region of nif operon for EMSA
nifpR	TTATATTCCCTTACATAAAT

^aAll primers are from this work.

Expression and purification of Fnr1 and Fnr3 in E. coli.

The coding regions of fnr1 and fnr3 were PCR amplified from the genomic DNA of P. polymyxa WLY78, respectively. These PCR products were cloned into pET-28b(+) (Novagen) to construct tagged Fnr proteins with His tag at the N terminus of Fnr1 and C terminus of Fnr3, respectively, and then transformed into E. coli BL21(DE3). The recombinant E. coli strains were cultivated at 37°C in LB broth supplemented with 50 μg/ml kanamycin until mid-log phase, when 0.2 mM isopropyl-β-d-thiogalactopyranoside (IPTG) was added, and incubation continued at 16°C for 8 h. Cells were collected and disrupted in the lysis buffer (50 mM NaH₂PO₄, 300 mM NaCl, 10 mM imidazole) by sonication on ice. Recombinant proteins NHis₆-Fnr1 and Fnr3-CHis₆ in the supernatant were purified on Ni₂-nitrilotriacetic acid (NTA) resin (Qiagen, Germany) according to the manufacturer’s protocol. Fractions eluted with 250 mM imidazole were dialyzed into binding buffer [20 mM HEPES, pH 7.6, 1 mM EDTA, 10 mM (NH₄)₂SO₄, 1 mM dithiothreitol (DTT), 0.2% Tween 20, 30 mM KCl) for electrophoretic mobility shift assays (EMSAs). Primers used here are listed in Table 1.

EMSAs.

EMSAs were performed as described previously using a DIG Gel Shift kit (2nd Generation; Roche, USA) (19). The promoter fragments of predicted target genes or operons were PCR amplified from the genomic DNA of P. polymyxa WLY78. The primers used here and DNA fragment sizes are listed in Table 1. The DNA fragments were labeled at the 3′ end with digoxigenin (DIG) using terminal transferase and used as probes in EMSAs. Each binding reaction mixture (20 μl) consisted of 0.3 nM labeled probe, 1 μg poly(dA-dT), and various concentrations (0, 0.05, 0.2, 2, and 6 μM) of purified His-tagged Fnr (apo-Fnr) in the binding buffer. Reaction mixtures were incubated for 30 min at 25°C, analyzed by electrophoresis using a native 5% polyacrylamide gel run with 0.5× Tris-borate-EDTA (TBE) as running buffer at 4°C, and electrophoretically transferred to a positively charged nylon membrane (GE Healthcare, United Kingdom). Labeled DNAs were detected by chemiluminescence according to the manufacturer’s instructions and recorded on X-ray film.

Bacterial RNA extraction and transcriptomic analysis.

P. polymyxa WLY78 wild type (WT) and the Δfnr13 mutant were grown in nitrogen-deficient medium under anaerobic conditions in 250-ml test tubes shaken at 250 rpm for 8 h at 30°C. The cultures were quickly collected by centrifugation at 4°C under anaerobic condition and stored in liquid nitrogen for further use. This experiment was repeated three times.

For bacterial RNA extraction, bacterial cultures at each experimental time point were harvested and rapidly frozen in liquid nitrogen. Total RNAs were extracted with RNAiso Plus (TaKaRa, Japan) according to the manufacturer’s protocol. Removal of genomic DNA and synthesis of cDNA were performed using the PrimeScript RT reagent kit with gDNA Eraser (TaKaRa, Japan). The concentration of purified RNA was quantified on a NanoDrop ND-1000 spectrophotometer (NanoDrop Technologies, Thermo Fisher Scientific, USA).

Illumina HiSeq 4000 sequencing from the total RNA was completed at Novogene Bioinformatics Technology Company (Beijing, China) according to a default Illumina stranded RNA protocol. Differential expression analysis of two groups (two biological replicates per condition) was performed using the DESeq R package (1.18.0) (44). DESeq provides statistical routines for determining differential expression in digital gene expression data using a model based on the negative binomial distribution. The resulting P values were adjusted using the Benjamini and Hochberg approach for controlling the false-discovery rate. The differences of transcript level with an adjusted P value of <0.05 determined by DESeq were considered to be significant, and the genes were assigned as differentially expressed genes (DEGs). The DEGs were annotated using the KEGG (Kyoto Encyclopedia of Genes and Genomes) database (http://www.genome.jp/kegg/). Gene Ontology (GO) enrichment analysis of differentially expressed genes was implemented by the GOseq R package, in which gene length bias was corrected. All of the raw reads are archived in the NCBI Sequence Read Archive (SRA) database. Transcriptional analysis was performed in triplicate, and the reproducibility of the biological repeats was high (mean R² = 0.929).

qRT-PCR analysis.

Transcription levels of genes among P. polymyxa WLY78 and fnr deletion mutants were compared by quantitative real-time RT-PCR (qRT-PCR) analysis. Primers used for qRT-PCR are listed in Table 1. qRT-PCR was performed on an Applied Biosystems 7500 Real-Time System (Life Technologies), and results were detected by the SYBR Green detection system with the following program: 95°C for 15 min, 1 cycle; 95°C for 10 s and 65°C for 30 s, 40 cycles. The relative expression level was calculated using the threshold cycle (ΔΔC_T) method, and 16S rRNA was set as internal control. Triplicate assays using RNAs extracted in three independent experiments were performed for each target gene.

Data availability.

The RNA-Seq data have been deposited in NCBI’s Sequence Read Archive database (accession numbers SRR10737462 to SRR10737463).