The Pseudomonas stutzeri-Specific Regulatory Noncoding RNA NfiS Targets katB mRNA Encoding a Catalase Essential for Optimal Oxidative Resistance and Nitrogenase Activity

Protection against oxygen damage is crucial for survival of nitrogen-fixing bacteria due to the extreme oxygen sensitivity of nitrogenase. This work exemplifies how the small ncRNA NfiS coordinates oxidative stress response and nitrogen fixation via base pairing with katB mRNA and nifK mRNA. Hence, NfiS acts as a molecular link to coordinate the expression of genes involved in oxidative stress response and nitrogen fixation. Our study provides the first insight into the biological functions of NfiS in oxidative stress regulation and adds a new regulation level to the mechanisms that contribute to the oxygen protection of the MoFe nitrogenase.

content using a 15 N dilution technique under two water regimen conditions (27). Similar to most diazotrophs, A1501 can fix nitrogen only under microaerobic conditions and exhibits optimal nitrogenase activity at an oxygen concentration of 1.0% (22). The genes responsible for the oxygen protection mechanisms in A1501 have yet to be determined. Previous studies have demonstrated that a novel P. stutzeri-specific ncRNA, NfiS, whose synthesis was dramatically increased under nitrogen fixation conditions, plays a crucial role in optimizing nitrogen fixation via base pairing with the nitrogenase gene nifK mRNA (28). An nfiS mutant strain was more sensitive to 20 mM H 2 O 2 than the wild-type (WT) strain, whereas the overexpression of NfiS led to enhanced resistance, suggesting that NfiS is essential for optimal resistance to oxidative stress. However, its target genes under oxidative stress have yet to be defined, and the regulatory mechanisms involved in the oxidative stress response remain unknown. In this study, we show that nfiS mutation led to significant downregulation of oxidative stress response genes, especially katB and oxyR. Furthermore, rapid adaptive stress responses and efficient protection of nitrogenase against oxidative injuries in A1501 were experimentally confirmed by mutant strain construction and complementation experiments. We report here that NfiS regulates optimal oxidative stress resistance via direct binding with katB mRNA, which results in enhanced stability of the transcript, or via the indirect transcriptional activation of OxyR. This global regulation might be a conserved and widespread mechanism in P. stutzeri strains.

RESULTS
NfiS deletion caused the decrease in expression of the oxidative stress response genes. We previously reported that an nfiS deletion mutant strain (A1701) displayed increased sensitivity to H 2 O 2 -induced oxidative stress (28), implying that this ncRNA might act as a regulator of the oxidative stress response. The A1501 genome contains four catalase genes (katA, katB, katE, and katG) and a peroxide sensor gene, oxyR, which are highly conserved among different Pseudomonas strains (see Table S1 in the supplemental material), suggesting common regulatory mechanisms (5,29). The effect of H 2 O 2 shock on gene expression level was then determined in LB medium. In the wild-type cells, all selected genes, except katE, showed a significant increase after H 2 O 2 shock (Fig. 1A). This is in agreement with the observation that increasing the concentration of H 2 O 2 from 0 mM to 12 mM was correlated with an increase in the total catalase activity, with optimal activity observed at a concentration of 4.0 mM H 2 O 2 (see Fig. S1 in the supplemental material). In particular, the most dramatic increase (Ͼ9.0fold) was noted for katB (Fig. 1A), suggesting a functional relevance to oxidative stress. In the A1701 ΔnfiS mutant strain, except for katE, all oxidative stress response genes were significantly downregulated, especially katB and oxyR, whose expression was reduced by 15 and 10 times, respectively (Fig. 1B), suggesting that expression of these genes is NfiS dependent via unknown mechanisms. Computational target prediction using the online tool sTarPicker revealed two possible NfiS binding sites to katB mRNA, which were located on the typical stem-loop structures (see Fig. S2A and B in the supplemental material). The predicted site 1 extends from positions 6 to 23 of NfiS and is complementary to the 5= translated region between nucleotides (nt) 62 and 79 of katB mRNA, while the predicted site 2 (from 141 to 153) is complementary to nucleotides 66 to 78 of katB mRNA ( Fig. S2C and D). In contrast, no possible NfiS binding sites were identified in other selected genes' mRNAs. We thus hypothesized that NfiS regulated optimal oxidative stress resistance via direct binding with katB mRNA.
Two base-pairing regions contribute to NfiS regulation of oxidative stress resistance. As mentioned above, predictions of the interaction between NfiS and katB mRNA revealed two base-pairing regions, termed sites 1 and 2 ( Fig. 2A). To determine whether the two complementary regions are essential for NfiS regulatory functions, we constructed three complementation plasmids, namely, pLAnfiS-A1501 expressing the WT nfiS gene, pLAnfiS-tru1 expressing a truncated nfiS gene carrying the first 54 nucleotides with base-pairing site 1, and pLAnfiS-tru2 expressing a truncated nfiS carrying the 200 remaining nucleotides (between nucleotides 55 and 254) with basepairing site 2 (Fig. 2B). We found that expression of the WT or truncated nfiS genes partially restored H 2 O 2 resistance of the nfiS mutant to WT levels (Fig. 2B), indicating that the truncated nfiS genes with either site 1 or 2 functionally substitute for the WT nfiS gene.
Site-directed mutagenesis was performed on the two putative base-pairing sites of NfiS to determine if base-pairing sequences are required for NfiS-mediated regulation. As shown in Fig. 2A, two putative base-pairing sequences, 5=-CAGCAGCAGCAAGGCG GC-3= and 5=-AGCAGCCGCUUGG-3= (unpaired nucleotides are underlined), were converted to GUCGUCGACGUUCCCCCG and UCGUCGCCGUUCC, leading to complete mismatches. The resulting complementation plasmids (pLAnfiS-mS1, expressing a mutated nfiS gene lacking site 1, pLAnfiS-mS2, expressing a mutated nfiS gene lacking site 2, and pLAnfiS-dmS12, expressing a mutated nfiS gene with double mutation of two sites) were used for further complementation of ΔnfiS (A1701). The results showed that either pLAnfiS-mS1 (A1717) or pLAnfiS-mS2 (A1718) partially restored oxidative stress resistance to the WT levels, whereas pLAnfiS-dmS12 (A1719) with a double mutation of the two sites failed to restore oxidative stress resistance, indicating that each of two sites is functional, but not sufficient, for NfiS-mediated regulation of oxidative stress resistance.
NfiS interacts directly with katB mRNA at two base-pairing sites. To obtain experimental evidence for the predicted interactions between NfiS and katB mRNA, microscale thermophoresis (MST) was used to determine the dissociation constant (K d ) in vitro using a set of 254-nt full-length NfiS and 70-nt katB mRNA oligonucleotides containing WT or point-mutated base-pairing regions. N-NfiS-wt containing either WT base-pairing site 1 or 2 of NfiS bound to N-katBR-wt exhibited dissociation constants (K d ) of 10.72 Ϯ 1.34 M (Fig. 3A), suggesting a direct interaction of either site 1 or site 2 with katB mRNA. We also measured the binding affinity using N-NfiS1-com or N-NfiS2-com oligonucleotides containing compensatory mutations resulting in high match levels to N-katBR-wt ( Fig. 3B and C). As expected, both compensatory oligonucleotides bound to katB mRNA tightly, suggesting stronger interactions. In contrast, N-NfiS1-mut6 and N-NfiS2-mut4 containing the point-mutated base-pairing sequences on stem-loops (33% and 31% mismatch, respectively) of NfiS possessed weaker affinity for katB mRNA ( Fig. 3D and E). The effect of katB mRNA mutations on base pairing with NfiS was investigated using 70 nt of the 5= translated region of katB mRNA containing two mismatch mutations (N-katBR1-mut6 and N-katBR2-mut4) and two compensatory mutations (N-katBR1-com and N-katBR2-com). The data indicated that the two mismatch oligonucleotides showed weaker binding affinity for NfiS ( Fig. 3F and G), whereas the two compensatory oligonucleotides displayed a stronger interaction with NfiS than the corresponding N-katBR-wt oligonucleotides ( Fig. 3H and I). The oligonucleotide pairs used for the MST analysis and the resulting dissociation constants observed are reported in Fig. 3J. These data strongly suggest that the mutations in katB mRNA have a stronger influence on the K d for the interaction between NfiS and katB mRNA because mutations in the katB target site disrupt interactions with both NfiS binding sites, whereas mutations in individual NfiS binding sites still leave the other binding site available for the interaction. Moreover, we observed that the half-life of katB mRNA in the nfiS mutant was approximately 2-fold less than in A1501, while the half-life of the nfiS mutant was restored to the WT level by the complementation plasmid (see Fig. S3 in the supplemental material). These data are in agreement with the base pairing of the two sites of NfiS (sites 1 and 2 defined above [ Fig. S2]) with the 5= translated region of katB mRNA, allowing for increased stability of katB mRNA.
NfiS-mediated posttranscriptional regulation of katB may be a widespread mechanism among P. stutzeri strains. Conservation of the structural genes for NfiS in P. stutzeri strains was previously reported (28). Indeed, katB genes were also conserved in P stutzeri strains or other Pseudomonas species, exhibiting a relatively high identity that ranged from 98% to 79% (see Table S1 and Fig. S4 in the supplemental material). These findings are consistent with the conservation of the two base-pairing sites identified in stem-loops of NfiS ncRNAs (Fig. 4A). The match levels between NfiS and the katB mRNA ranged from 67% to 89% for site 1 and from 69% to 77% for site 2, and the complementary sequences from katB mRNAs were also conserved, with identities of A1501 katB ranging from 89% to 100% (Fig. 4B). These results indicate that the two base-pairing sites of NfiS for katB mRNA are structurally conserved in P. stutzeri strains.
P. stutzeri ATCC 17588 is a nondiazotrophic strain. We have shown previously that the nfiS gene of ATCC 17588 (nfiS-ATCC) could restore the oxidative stress-resistant phenotype of the A1501 nfiS mutant, but it could not restore nitrogenase activity (28). The identity between the nucleotide sequences of nfiS-ATCC and nfiS-A1501 is only 79%, while the identities for complementary sequences between NfiS-ATCC and A1501 katB mRNA are 83% for site 1 and 77% for site 2. This finding raised the question as to whether the two sites of NfiS-ATCC are functionally conserved among P. stutzeri strains, particularly between ATCC 17588 and A1501. To this end, we constructed two complementation plasmids: pLAnfiS-ATCC-tru1, expressing a truncated nfiS-ATCC with base-pairing site 1, and pLAnfiS-ATCC-tru2, expressing a truncated nfiS base-pairing site 2. These two plasmids were then introduced into A1701 (the ΔnfiS strain) to yield A1725 and A1724. We observed that both WT and truncated nfiS-ATCC restored partial resistance compared with nfiS-A1501 (Fig. 4C), strongly suggesting that two the basepairing sites of NfiS for katB mRNA are structurally and functionally conserved in P. stutzeri strains.
katB is involved in the oxidative stress response-probably in an oxyRdependent manner. In model systems, such as P. aeruginosa (5) and Pseudomonas putida (29), katB expression is induced in response to increased H 2 O 2 levels by activation of the transcriptional regulator OxyR, and KatB is known as the most pivotal enzyme for detoxifying exogenous H 2 O 2 . In line with this, we predicted that in addition to posttranscriptional regulation by NfiS, the katB expression in A1501 was also induced and activated by OxyR at a transcriptional level in response to H 2 O 2 . To confirm this, we constructed two single-gene mutants, insertion mutants (A1401 and A1921) lacking either katB or oxyR, and their complemented strains (A1406 and A1926) expressing the WT katB and oxyR genes, respectively, and monitored their oxidative stress resistances. As shown in Fig. 5A, both the katB mutant (A1401) and oxyR mutant (A1921) were more sensitive than WT strain A1501 following exposure to 12 mM H 2 O 2 . Consistent with their increased sensitivity to oxidative stress, both mutants displayed decreased catalase activities compared to the WT strain (Fig. 5B), indicating their involvement in protecting against oxidative stress. In addition, further analysis of the A1501 katB promoter region led to the identification of a putative OxyR binding site (see Fig. S5 in the supplemental material); the sequence consisted of four tetranucleotide elements spaced at 7-bp intervals, similar to those experimentally characterized for the OxyRbinding sequences in P. aeruginosa and P. putida (5,29) and to the consensus sequence determined in E. coli (30), strongly suggesting a putative OxyR-dependent activation of katB expression. This possibility was checked by measuring the expression levels of H 2 O 2 -inducible catalase genes under the oxyR mutation background, which show that the oxyR mutation resulted in a significant decrease in the expression of H 2 O 2 -inducible catalase genes, especially katB, whose expression was almost completely suppressed in the oxyR mutant (Fig. 5C). Complementation assays with plasmids carrying either WT oxyR or WT katB genes confirmed the role of oxyR and katB in the oxidative stress response. These findings suggest that katB expression is dependent on OxyR and is essential for optimal oxidative stress resistance. KatB is required for optimal nitrogenase activity. In biological systems, H 2 O 2 is a key endogenous reactive oxygen species (ROS) from metabolic processes or exogenous ROS from the environment. H 2 O 2 can be detoxified by catalases, leaving O 2 as a by-product. O 2 supports sufficient production of ATP for nitrogenase but can also rapidly limit the activity and repress the synthesis of this enzyme. As yet, the combined effect of O 2 and H 2 O 2 on nitrogen fixation has not been analyzed. To this end, the nitrogenase activity of A1501 was examined at different O 2 (initial concentrations from 0.5% up to 2.0%) and H 2 O 2 (initial concentrations from 0.2 mM up to 2.0 mM) concentrations. As shown in Fig. 6A, 1.0% is the optimal concentration of O 2 for optimal nitrogenase activity of A1501, whereas lower or higher concentrations of O 2 resulted in partial or total loss of activity. At this optimal concentration of O 2 (1.0%), increasing concentrations of H 2 O 2 from 0.2 mM to 2.0 mM led to a continued decrease in nitrogenase activity, which was completely inhibited by 2.0 mM H 2 O 2 (Fig. 6B). In contrast, at a suboptimal O 2 concentration of 0.5%, an increase in nitrogenase activity of the wild-type strain was observed in the presence of H 2 O 2 up to 0.6 mM. As was also observed for the wild-type strain A1501, a relatively low level of H 2 O 2 stress can lead to an increase of nitrogenase activity in the katB mutant under O 2 -insufficient conditions (Fig. 6B). We thus presumed that H 2 O 2 at low concentrations was detoxified by KatB, leaving O 2 as a by-product to support nitrogen fixation under O 2 -insufficient conditions.
In addition, it is apparent that the nitrogenase activity of the katB mutant was more susceptible to increasing H 2 O 2 concentrations than the wild-type strain. Especially under O 2 -sufficient conditions, the absence of katB caused a more significant decrease in nitrogenase activity (Fig. 6B). Consistent with observations mentioned above, the H 2 O 2 -inducible catalase gene, katB, and its regulator gene, oxyR, were upregulated by approximately 13-fold and 17-fold, respectively, under nitrogen fixation conditions  (Fig. 6C). We also determined the nitrogenase activities of A1501, the two mutants, and the corresponding complemented strains. The mutation of the oxyR or katB genes led to a decrease in nitrogenase activity by approximately 70 or 33% compared to the WT strain, respectively (Fig. 6D). These results suggest that KatB is required for optimal nitrogenase activity, especially at high H 2 O 2 concentrations and under O 2 -sufficient conditions. We hypothesized that in A1501, nitrogen-fixing conditions led to a significant increase of KatB synthesis (and hence activity), which may be a possible defensive mechanism of nitrogenase activity against oxidative damage.

DISCUSSION
P. stutzeri A1501 is a root-associated nitrogen-fixing bacterium that can colonize the root surface and invade the root tissues of the host plant. During this process, oxygen appears to be a key regulatory signal controlling the nitrogen fixation process through complex regulatory networks. As shown by its importance in other well-studied Pseudomonas strains, ncRNA-mediated regulation is a key component of stress adaptation and gene regulation in nitrogen-fixing P. stutzeri strains. The data presented in this article reveal that NfiS, a P. stutzeri-specific ncRNA, is a multifunctional regulatory RNA that acts by RNA-RNA interactions with two different mRNAs, nifK and katB. We previously showed that NfiS was recruited by nifK mRNA as a novel activator to optimize the nitrogen fixation process in response to specific environmental cues (28). Indeed, the nfiS gene was expressed under nitrogen-fixing conditions from an RpoNdependent promoter, and its expression was strongly impaired in strains carrying mutations in genes controlling the nitrogen fixation process-rpoN, ntrC, and nifA-as well as in an rpoS-deficient background (28). We thus proposed a new regulatory mechanism coupling the oxidative stress response and optimal nitrogen fixation, in which NfiS directly pairs with the mRNAs of both a catalase gene, katB, and a nitrogenase gene, nifK, thereby enhancing the synthesis and activities of both catalase and nitrogenase. The fact that NfiS binding site 1 for katB mRNA overlaps the binding site for nifK mRNA is puzzling. However, binding to either mRNA may depend on physiological conditions. At oxygen concentrations not compatible with nitrogen fixation, NfiS may control katB mRNA using binding sites 1 and 2, while site 1 would be available for binding nifK mRNA when oxygen concentrations are compatible with nitrogen fixation. Furthermore, secondary structure analysis of the katB and nifK mRNAs predicted the formation of a hairpin structure in their coding region downstream of the start codon. Such a hairpin structure is generally believed to affect translation efficiency. These findings suggest that NfiS acts as a pleiotropic riboregulator that integrates adaptation to oxidative stress with other cellular metabolisms and helps to protect nitrogen-fixing cells against oxidative damage.
The genus Pseudomonas is one of the most diverse and ecologically significant groups of bacteria on the planet (31). P. stutzeri is a remarkable member of this genus, with exceptional physiological capacities. The diversity within the species is not limited to physiological traits and is also reflected at the genetic level. Catalases are ubiquitous enzymes that detoxify H 2 O 2 and have been extensively characterized in P. aeruginosa, a species closely related to nitrogen-fixing P. stutzeri (26). P. aeruginosa produces three catalases (KatA, KatB, and KatE) and exhibits a high catalase-specific activity (32,33). The katB gene encoding an H 2 O 2 -inducible catalase is essential for optimal resistance to H 2 O 2 (34). Multiple catalases are also encoded in the P. stutzeri genome, but the role of each catalase in the response to oxidative stress is still unknown. In P. stutzeri A1501, katB inactivation significantly decreased oxidative stress resistance as well as catalase activities. Similar results were observed in P. aeruginosa, suggesting that KatB is the most pivotal enzyme for detoxifying exogenous H 2 O 2 . However, different systems are involved in the regulation of the katB gene in A1501 and other Pseudomonas strains. To our knowledge, NfiS is the first described case of a bacterial small RNA involved in the regulation of the oxidative stress response via a direct base-pairing interaction with katB mRNA. Although NfiS is P. stutzeri specific, the complementary sequences of katB mRNA with NfiS are highly conserved and widely distributed not only in P. stutzeri but also in other Pseudomonas strains. Based on these data, we propose that the involvement of ncRNAs in the posttranscriptional regulation of the katB gene, although not identical to that of NfiS, might be a conserved and widespread mechanism among Pseudomonas spp., independent of their ecological features.
Oxidative stress is one of the major challenges for nitrogen-fixing bacteria in almost all environments. Most nitrogen-fixing bacteria are unable to fix nitrogen at high oxygen tensions due to inactivation of nitrogenase. An unusual case is Azotobacter vinelandii, a soil bacterium that can fix nitrogen under aerobic conditions while simultaneously protecting its nitrogenase from oxygen damage (35). This bacterium is equipped with particular physiological mechanisms, such as the autoprotection of nitrogenase with the involvement of catalase (36), the respiratory protection of terminal oxidases (35), the oxygen barrier of the alginate capsule (37), and conformational protection of nitrogenase due to association with an FeSII protein, named "Shethna protein" (38,39). In addition, the roles of the A. vinelandii RpoS, Kat1, and RgsA ncRNAs in the survival of oxidative stress have been well documented (40,41). Another striking example is rhizobial infection, during which reactive oxygen species such as H 2 O 2 are transiently produced. The symbiotic nitrogen-fixing model bacterium Sinorhizobium meliloti possesses three distinct catalases (KatB, KatA, and KatC) to cope with oxidative stress (42). KatB activity was detected throughout the growth of S. meliloti in minimal medium and represented approximately 90% of the total catalase activity, suggesting an important role of this bifunctional catalase in free-living bacteria (43). In addition, the katB katC double mutant displayed reduced nitrogen fixation and abnormal infection, indicating that these two catalases are essential for the establishment of symbiosis (44).
Oxygen control, which is tightly controlled in response to the external oxygen concentration, is exerted first at the transcriptional level of the nif operons and then at the level of the nitrogenase activity. At the transcriptional level, oxygen repressed nitrogenase synthesis much more rapidly than ammonia did in Azotobacter (45). In most Proteobacteria, including P. stutzeri A1501, NifA is the transcriptional activator of other nif operons (46)(47)(48)(49)(50). In K. pneumoniae strains that can fix nitrogen anaerobically, NifL responds to oxygen and prevents NifA-mediated activation of nif gene expression (46,47). In P. stutzeri, whose optimal nitrogenase activity was observed at an oxygen concentration of 1%, the expression of nifLA was also controlled by oxygen (22,49). Transcription of nif genes from Rhodobacter capsulatus, a bacterium that does not contain an nifL gene, is inhibited by oxygen, probably through direct inactivation of NifA (50). At the next level, oxygen causes rapid and irreversible damage to nitrogenase enzymes (51). In Azotobacter, the FeSII ferredoxin "Shethna protein" plays a protective role against oxygen damage by forming a reversibly inactive complex with nitrogenase. This mechanism is also likely present in Klebsiella pneumoniae and Gluconacetobacter (52) but may be absent in P. stutzeri A1501 because the gene encoding an FeSII protein is absent from the A1501 genome.
Together, these findings suggest that oxygen supports sufficient production of ATP for nitrogenase but can also rapidly limit the activity and repress the synthesis of this enzyme. Therefore, nitrogen-fixing bacteria must develop dynamic and intricate strategies to quickly fine-tune gene expression in response to environmental stimuli. In addition to specific regulatory proteins, ncRNAs have been identified as components of regulatory networks and modulate various physiological processes, especially stress adaptation (53). Furthermore, posttranscriptional regulation by ncRNAs allows a cell to respond to the signal in an extremely flexible and sensitive manner (14,54). Few regulatory RNAs and their targets have been functionally characterized in nitrogenfixing bacteria. Regulatory ncRNA-mediated responses to environmental stimuli provide an adaptive advantage to nitrogen-fixing bacteria, such as observed for A. vinelandii ArrF (55), S. meliloti EcpR1 (56), P. stutzeri NfiS (28), and Methanosarcina mazei sRNA 154 (57). A single sRNA often modulates a particular physiological response via multiple target genes and could be used by cells to integrate multiple signals into gene expression programs (15,58). This study and other preliminary results also suggest that in P. stutzeri A1501, NfiS-mediated processes might work together in a more sophisti-cated manner than previously believed. The working model for NfiS in P. stutzeri (Fig.  7) tentatively illustrates the regulatory roles of NfiS in integrating adaptation to oxidative stress with other cellular metabolisms and will broaden our knowledge of ncRNA-based regulatory mechanisms in environmental microorganisms. At the posttranscriptional level, NfiS directly pairs with the mRNAs of both the catalase gene katB and the nitrogenase gene nifK, thus linking total catalase activity and optimal nitrogen fixation. At the transcriptional level of global regulation, NfiS expression is strongly decreased in rpoN, ntrC, and nifA, which supports the former conclusion that nfiS is a nitrogen fixation-regulated gene (28); a significant decrease is also observed in the oxyR mutant. Thus, NfiS expression depends on OxyR, NifA, NtrC, and RpoN, implying a more complex regulatory circuitry. In addition, NfiS, known to be involved in survival of oxidative stress in other systems, may also play a similar role (28), which remains to be elucidated.
The regulatory mechanisms of the oxidative stress response in P. stutzeri A1501 have not been fully elucidated. We previously showed that NfiS synthesis was significantly induced by sorbitol and that the NfiS mutant strain displayed increased sensitivity to sorbitol, suggesting that NfiS in P. stutzeri is also involved in the regulation of the osmotic stress response (28). Therefore, the possibility that NfiS contains another base-pairing site with target genes involved in osmotic stress response remains to be explored. Future studies should help to understand how P. stutzeri A1501 defends itself against reactive oxygen species generated under various conditions and to determine why the bacterium is highly adaptable to environmental changes.

MATERIALS AND METHODS
Strains, plasmids, oligonucleotides, media, and culture conditions. All the strains and plasmids used and constructed in this study are listed in Table S2 in the supplemental material. P. stutzeri A1501

FIG 7
Model for the NfiS-mediated regulatory network and oxidative stress response in P. stutzeri A1501, indicating data supported by our own experiments in this study and the previous study (40). NfiS acts as a pleiotropic riboregulator to optimize catalase and nitrogenase activities via direct base pairing with katB mRNA and nifK mRNA or via the indirect transcriptional activation of OxyR. This global regulation integrates adaptation to oxidative stress with other cellular metabolic processes and helps to protect nitrogen-fixing cells against oxidative damage. Under O 2 -insufficient conditions, H 2 O 2 at low concentrations is detoxified by H 2 O 2 -inducible catalases, leaving O 2 as a by-product to support nitrogen fixation. The black box represents the putative RpoN-dependent promoters, the red box represents the putative OxyR-dependent promoter, and the white box represents the upstream activator NifA sequences. Solid red or green arrows indicate transcriptional regulation and posttranscriptional regulation, respectively, and dashed lines represent unknown mechanisms. For more details, see the Discussion. and mutant derivatives were grown in LB or in minimal lactate medium (medium K [pH 6.80]) at 30°C (22). Spectinomycin (40 g ml Ϫ1 ), tetracycline (10 g ml Ϫ1 ), or kanamycin (35 g ml Ϫ1 ) was added to the medium when needed. For cloning procedures, E. coli TOP10 competent cells (CWBIO) were grown in LB broth or on LB agar plates.
Construction of katB and oxyR nonpolar insertion mutants and complementation plasmids. The katB and oxyR genes were inactivated by homologous suicide plasmid integration using pK18mob as a vector (59). Oligonucleotide primers were designed to generate internal gene fragments of 145 bp for katB and 290 bp for oxyR by PCR, enabling the creation of mutations in katB and oxyR without preventing the transcription of their downstream genes (PST3567 and PST0135, respectively). The amplicons obtained were doubly digested with EcoRI/BamHI and EcoRI/HindIII and then cloned into pK18mob. The resulting plasmids were introduced into A1501 by triparental mating, generating katB and oxyR nonpolar insertion mutant strains, named A1401 and A1901, respectively (Table S2). Correct recombination was confirmed by PCR, followed by nucleotide sequencing of the amplicons obtained. DNA fragments containing WT genes for katB or oxyR with their promoter and terminator regions were amplified by PCR to construct complementation plasmids. A 1,800-bp fragment containing katB and a 1,453-bp fragment containing oxyR were doubly digested with HindIII/BamHI and EcoRI/SalI, respectively, and then ligated into pLAFR3 to yield the complementation plasmids pLAkatB and pLAoxyR. Both plasmids were then introduced into A1401 or A1921 by triparental mating, generating strains A1406 and A1926.
Construction of complementation plasmids carrying the WT or truncated or mutated nfiS genes. Two plasmids carrying the WT nfiS gene from two different P. stutzeri strains and seven plasmids carrying truncated or mutated nfiS genes were constructed using the broad host plasmid pLAFR3 as a vector (60) to complement the constructed A1701 deletion strain. Nine nfiS DNA fragments with their promoter and terminator regions were synthesized by BGI Co., Ltd. Restriction enzyme sites (HindIII and BamHI) were incorporated into synthesized fragments, which were subsequently cloned into the pLAFR3 vector, resulting in the following plasmids: pLAnfiS-A1501, expressing an A1501 WT nfiS; pLAnfiS-tru1, expressing a truncated nfiS carrying the first 54 nucleotides with base-pairing site 1; pLAnfiS-tru2, expressing a truncated nfiS carrying the 200 remaining nucleotides with base-pairing site 2; pLAnfiS-mS1, expressing a mutated nfiS lacking base-pairing site 1; pLAnfiS-mS2, expressing a mutated nfiS lacking base-pairing site 2; pLAnfiS-dmS12, expressing a mutated nfiS with a double mutation of the two sites; pLAnfiS-ATCC, expressing an ATCC 17588 WT nfiS; pLAnfiS-ATCC-tru1, expressing an ATCC 17588 truncated nfiS carrying the first 44 nucleotides with site 1; and pLAnfiS-ATCC-tru2, expressing an ATCC 17588 truncated nfiS carrying the 200 remaining nucleotides with site 2 (Table S2).
Cell survival under H 2 O 2 stress. The cell susceptibility of P. stutzeri A1501 and its derivatives to H 2 O 2 was assayed as previously described (61). Strains were grown overnight in LB broth at 30°C and were transferred into fresh LB broth up to an optical density at 600 nm (OD 600 ) of 0.6. Then, 12 mM H 2 O 2 was added to the medium, and the cultures were incubated at 30°C and 220 rpm for 10 min. Serial 10-fold dilutions of OD-standardized cultures were spotted on LB plates. Plates were incubated at 30°C for 24 h prior to colony enumeration. The survival rate was expressed as the percentage of the number of colonies in the treated samples compared with that in the untreated A1501 sample used as a control.
qRT-PCR. Total RNA was isolated with an innuPREP RNA minikit (Analytik Jena) and was reversetranscribed into cDNA, which was diluted to 100 ng l Ϫ1 . Quantitative real-time PCR (qRT-PCR) assays were performed using total RNA preparations obtained from three independent cultures (three biological replicates). The gene-specific primers listed in Table S3 in the supplemental material were designed based on the full genome sequence of P. stutzeri A1501, and the 16S rRNA gene was used as the endogenous reference gene to normalize the expression of target genes in each cDNA template. The relative mRNA concentration was calculated by the comparative threshold cycle (2 Ϫ⌬⌬CT ) method. The target gene copy numbers were determined in triplicate using the 7500 real-time PCR system and ChamQ SYBR qPCR master mix. All procedures were carried out according to the manufacturers' recommendations. Data were analyzed using ABI Prism 7500 sequence detection system software (Applied Biosystems).
Computational target prediction. The sTarPicker prediction method (62) was used to predict interactions between NfiS and all annotated open reading frames (ORFs) of A1501. The interaction region on mRNA levels was defined as 100 bp up-and downstream of the start codon.
Microscale thermophoresis measurements. MST experiments were performed according to Zhan et al. (28). A modified overlap extension PCR method was used to amplify the full-length wild-type NfiS and its mutated products using mutagenic primers (Table S3). A set of full-length ncRNAs (N-NfiS probes) for MST was then transcribed by GenePharma using a MAXIscript kit (Thermo Fisher) from PCR-generated templates (see Table S4 in the supplemental material). In addition, another set of Cy3-labeled 70-nt single-stranded RNA (ssRNA) oligonucleotides containing WT or mutated base-pairing regions of katB mRNA (N-katB competitors) were synthesized by GenePharma Company (Table S4). Four microliters of sample containing 200 nM labeled probe and increasing concentrations of a nonlabeled competitor (from 5 nM to 150 M) were loaded on standard treated silicon capillaries (Monolith NT.115 series capillaries; catalog no. MO-K002). The measurements were carried out using a Monolith NT.115 instrument (NanoTemper Technologies GmbH) at 26°C in diethyl pyrocarbonate (DEPC)-treated water with 40% excitation power and medium MST-Power. The dissociation constants (K d ) were calculated as described previously (63). Data analyses were performed using Nanotemper Analysis software (Nano-Temper Technologies).
Nitrogenase activity assays. Nitrogenase activity was determined by the acetylene reduction test (64) using a derepression protocol (22), as follows. Cells from an overnight culture in LB medium were centrifuged and resuspended in a 60-ml flask containing 10 ml of N-free minimal medium K at an OD 600 of 0.1. The suspension was incubated at 30°C, with vigorous shaking, under an argon atmosphere containing 1.0% oxygen and 10% acetylene. Gas samples (0.25 ml) were taken at regular intervals to determine the amount of ethylene produced. Samples were analyzed on a polydivinylbenzene porous bead GDX-502 column using a gas chromatograph SP-2100 fitted with a flame ionization detector (Beijing Beifen-Ruili Analytical Instrument Co., Ltd.). The content of ethylene in the gas samples was determined by reference to an ethylene standard. Each experiment was repeated at least three times. Protein concentrations were determined by the Bio-Rad protein assay. For the nitrogenase activity of A1501 and katB mutant A1401, different concentrations of H 2 O 2 (0.2, 0.4, 0.6, 0.8, 1.0, and 2.0 mM) were added to the samples immediately after inoculation of the bacterial samples, and then the flasks were incubated under an argon atmosphere containing 10% acetylene and 0.5 or 1.0% oxygen.
Catalase activity assays. The total catalase activity of the samples was determined using a catalase assay kit (Beyotime Institute of Biotechnology, Jiangsu, China) based on the protocols provided by the manufacturer. The catalase activity assays were carried out at 37°C in flasks in the presence of H 2 O 2 (concentration of 12 mM). Cells were washed twice with phosphate-buffered saline (PBS) and collected, lysed by cell lysis buffer, and centrifuged at 10,000 ϫ g for 15 min at 4°C. Then the clear supernatants were collected, and the assay was performed immediately. The H 2 O 2 concentration of this kit can be calculated by C (mM) ϭ 22.94 ϫ A 240 according to the protocol, where C is the concentration of H 2 O 2 and A 240 is the absorbance of the reaction solution at 240 nm. Samples were prepared and added to the 96-well plate. Then the mixtures were incubated at 25°C for 15 min at least (but no more than 45 min), and the absorbance was finally determined at 520 nm. The experiments were performed three times independently, and the catalase levels were normalized against total protein levels.
Stability measurements of katB mRNA. For determination of the half-life of katB mRNA under oxidative stress conditions, 12 mM H 2 O 2 was added to LB medium-grown cultures (OD 600 of 0.6) as indicated above. Rifampin (40 mg/ml) was added immediately after H 2 O 2 shock. Then 2-ml samples were collected at different times (0, 1, 3, 5, and 7 min). RNAlater (Sigma) was added to 2 volumes of RNAprotect. After incubation at room temperature for 5 min, samples were centrifuged for 2 min at 12,000 rpm, and pellets were quickly frozen in liquid nitrogen and stored at Ϫ80°C until use. Total RNA was isolated using an innuPREP RNA minikit (Analytik Jena), and cDNA was synthesized from total RNA using a PrimeScript RT reagent kit with gDNA Eraser (Perfect Real Time) and was used to estimate mRNA levels by qRT-PCR. The primers used to measure the half-life of katB transcripts are listed in Table  S3, and qRT-PCR was performed as described above. The relative mRNA concentration was calculated by the comparative threshold cycle (2 Ϫ⌬⌬CT ) method with 16S rRNA as the endogenous reference. Gene expression for each time point was normalized to the endogenous reference. Data are presented as the percentage of katB mRNA levels relative to the amount of these mRNAs at time point zero.

SUPPLEMENTAL MATERIAL
Supplemental material for this article may be found at https://doi.org/10.1128/JB .00334-19.