Pseudomonas aeruginosa Nonphosphorylated AlgR Induces Ribonucleotide Reductase Expression under Oxidative Stress Infectious Conditions

ABSTRACT Ribonucleotide reductases (RNRs) are key enzymes which catalyze the synthesis of deoxyribonucleotides, the monomers needed for DNA replication and repair. RNRs are classified into three classes (I, II, and III) depending on their overall structure and metal cofactors. Pseudomonas aeruginosa is an opportunistic pathogen which harbors all three RNR classes, increasing its metabolic versatility. During an infection, P. aeruginosa can form a biofilm to be protected from host immune defenses, such as the production of reactive oxygen species by macrophages. One of the essential transcription factors needed to regulate biofilm growth and other important metabolic pathways is AlgR. AlgR is part of a two-component system with FimS, a kinase that catalyzes its phosphorylation in response to external signals. Additionally, AlgR is part of the regulatory network of cell RNR regulation. In this study, we investigated the regulation of RNRs through AlgR under oxidative stress conditions. We determined that the nonphosphorylated form of AlgR is responsible for class I and II RNR induction after an H2O2 addition in planktonic culture and during flow biofilm growth. We observed similar RNR induction patterns upon comparing the P. aeruginosa laboratory strain PAO1 with different P. aeruginosa clinical isolates. Finally, we showed that during Galleria mellonella infection, when oxidative stress is high, AlgR is crucial for transcriptional induction of a class II RNR gene (nrdJ). Therefore, we show that the nonphosphorylated form of AlgR, in addition to being crucial for infection chronicity, regulates the RNR network in response to oxidative stress during infection and biofilm formation. IMPORTANCE The emergence of multidrug-resistant bacteria is a serious problem worldwide. Pseudomonas aeruginosa is a pathogen that causes severe infections because it can form a biofilm that protects it from immune system mechanisms such as the production of oxidative stress. Ribonucleotide reductases are essential enzymes which synthesize deoxyribonucleotides used in the replication of DNA. RNRs are classified into three classes (I, II, and III), and P. aeruginosa harbors all three of these classes, increasing its metabolic versatility. Transcription factors, such as AlgR, regulate the expression of RNRs. AlgR is involved in the RNR regulation network and regulates biofilm growth and other metabolic pathways. We determined that AlgR induces class I and II RNRs after an H2O2 addition in planktonic culture and biofilm growth. Additionally, we showed that a class II RNR is essential during Galleria mellonella infection and that AlgR regulates its induction. Class II RNRs could be considered excellent antibacterial targets to be explored to combat P. aeruginosa infections.

Previous studies have revealed the relationship between FimS-AlgR and the RNR. We previously described that the AlgR regulation of class I and II RNRs depends on its phosphorylation state (8). Additionally, AlgR is linked with the oxidative stress response system (17), and it is known that the expression of the Escherichia coli RNR is activated under oxidative stress conditions (18). We have also discovered that AlgR is the transcription factor responsible for the induction of P. aeruginosa class I and II RNRs under oxidative stress (8).
In this study, we delved into the regulation of class I and II RNRs by AlgR in the presence of H 2 O 2 in different biological growth stages, in a planktonic culture, in a biofilm, and during G. mellonella infection. To determine gene regulation variability, we measured RNR expression in other P. aeruginosa strains (laboratory and clinical isolates). To obtain a clinical perspective on RNR expression under oxidative conditions, one of the strains used was a clinical strain isolated from a CF patient (12). In this work, we confirmed that AlgR, in its nonphosphorylated state, is responsible for class I and II RNR transcriptional induction under oxidative stress conditions. We demonstrated that AlgR is essential for nrdA and nrdJ expression in biofilms under stress conditions. Finally, we observed that ROS are produced during G. mellonella infection and that nrdJ is the principal RNR expressed, with AlgR being essential for its induction. The different nrdJ and nrdA induction patterns help to elucidate the essential role of NrdJ in ensuring bacterial survival and suggest a molecular pathway used by P. aeruginosa during bacterial infection to restore the dNTP pool and repair damaged DNA.

RESULTS
AlgR expression is induced under oxidative stress conditions. AlgR is a transcription factor that is part of the fimS-algR operon. FimS-AlgR is a two-component system of P. aeruginosa that can regulate several metabolic pathways in response to external signals. However, the expression of algR can be activated through two different promoter regions, PfimS and PalgR (7). Figure 1A shows a scheme of the two promoters in fimS-algR. The first promoter is located upstream of fimS; when it is transcriptionally activated, the transcript produced encodes the FimS and AlgR proteins. The second promoter is located upstream of algR within the fimS coding region. This second promoter responds to specific sigma factors, such as AlgU and RpoS, to activate its transcription, and the mRNA obtained encodes only the protein AlgR (7).
Previous studies have linked AlgR with the bacterial response system to oxidative stress (8,19). However, the specific mechanism by which AlgR is activated under oxidative stress conditions has not been clarified. In this work, we aimed to examine the direct link between AlgR and oxidative stress. First, we transcriptionally fused the PfimS promoter and the PalgR promoter to gfp, obtaining pETS-PfimS and pETS-PalgR, respectively (see Materials and Methods). Our goal was to determine whether the expression of algR was induced under oxidative stress conditions and, if so, which promoter was responsible for this specific activation. Figure 1B shows that during the exponential phase (OD 600 [optical density at 600 nm] = 0.5), algR expression slightly increased after H 2 O 2 was added. When algR expression was measured in the stationary growth phase (OD 600 . 2.5), PalgR had a higher basal expression level (2.7 times higher) than PfimS. In addition, only pETS-PalgR displayed significantly increased algR expression after the addition of H 2 O 2 . To test our system and as a control, we transcriptionally fused the promoter regions of two P. aeruginosa catalases, katA and katB, to gfp and analyzed whether the H 2 O 2 concentration used was suitable for generating oxidative stress. The catalases katA and katB are part of the oxidative stress response system in P. aeruginosa. katA, the main catalase of Pseudomonas, is constitutively expressed to remove the H 2 O 2 produced in the bacterial respiratory network. When ROS levels are high, other proteins, such as KatB, are expressed to remove excess oxidative stress (15). Figure 1C shows that both catalases, pETS-PkatA and pETS-PkatB, were transcriptionally induced when 1 mM H 2 O 2 was added for 30 min during the exponential and stationary growth phases. This clearly indicated that the conditions used in this study created enough ROS to generate oxidative stress growth conditions. Nonphosphorylated AlgR controls the expression of class II (nrdJ) and Ia (nrdA) RNRs under oxidative stress conditions. Crespo et al. (8) reported that during the exponential growth phase, RNR expression increases under oxidative stress conditions because the master transcriptional regulator AlgR directly binds to an AlgR box in the promoter region of each RNR. We aimed to elucidate how oxidative stress regulates the different RNR genes more deeply via AlgR. Additionally, because regulation by AlgR is linked to the AlgR phosphorylation state, we studied how the phosphorylated form of AlgR modulates nrd expression (8).
The promoter regions of nrdA and nrdJ were transcriptionally fused to the reporter gene gfp to obtain pETS-PA and pETS-PJ, respectively. In the promoter region of nrdA, one AlgR box was found upstream from the coding region. In the nrdJ promoter region, two AlgR boxes were found (8). The RNR promoter regions with their respective mutated AlgR boxes were transcriptionally fused to gfp as well, yielding pETS-PA-Dbox1 and pETS-PJ-Dbox112. These plasmids were transformed into different P. aeruginosa strains: PAO1 and PAO1 DalgR, PA14 and PA14 DalgR, and the clinical P. aeruginosa CF isolate PAET1 and PAET1 DalgR ( Fig. 2 and 3). The DalgR strains were complemented with pUCP-AlgR, which encodes the wild-type (WT) AlgR protein, or with pUCP-D54N, which encodes the AlgR protein with a mutation in amino acid 54 (D54N) to prevent phosphorylation (3).
To study the expression of RNRs under oxidative stress conditions, we treated the bacterial cultures with 1 mM H 2 O 2 for 30 min when the cultures reached the exponential (OD 600 = 0.5) and stationary (OD 600 . 2.5) growth phases. We observed that nrdJ (class II RNR) was transcriptionally induced in both the exponential and stationary growth phases ( Fig. 2A). Moreover, nrdJ induction was eliminated in a P. aeruginosa PAO1 DalgR strain The algR-specific promoters are indicated with lines. (B) Gene expression of the two different promoter regions of algR, PfimS-algR and PalgR, which were transcriptionally fused to gfp (pETS-PfimS and pETS-PalgR, respectively), in the exponential (OD 600 [optical density at 600 nm] = 0.5) and stationary phases (OD 600 . 2.5) after the addition of 1 mM H 2 O 2 or the equivalent volume of water for 30 min. (C) The promoter regions of katA and katB were transcriptionally fused to gfp (pETS-PkatA and pETS-PkatB, respectively). The strains were incubated with 1 mM H 2 O 2 or an equivalent volume of water for 30 min in the exponential (OD 600 = 0.5) and stationary phases (OD 600 . 2.5). Three independent experiments were performed; error bars indicate standard deviation. Statistical analysis to determine significant differences between the H 2 O 2 and their H 2 O counterpart samples was performed using Student's unpaired t test (**, P , 0.01; ***, P , 0.001).
AlgR Induces Ribonucleotide Reductase Gene Expression mSystems and when the plasmid pETS-PJ-Dbox112 was used ( Fig. 2A), indicating the specific dependence on AlgR during transcription. In addition, the complementation of the algR deletion with pETS-D54N restored nrdJ induction under oxidative stress conditions in the exponential growth phase. Both the pETS-AlgR and pETS-D54N plasmids complemented the algR deletion in the stationary phase. However, the induction was significant only when AlgR was not phosphorylated.
To unravel the molecular mechanism involved in the transcriptional gene induction of class I and II RNRs in strains other than the laboratory strain PAO1, we measured the expression of nrdJ and nrdA in the laboratory strain P. aeruginosa PA14 and the clinical strain P. aeruginosa PAET1. We observed that the promoter region of nrdJ was conserved with 99% identity, while that of nrdA was conserved with 100% identity, among the three strains (Fig. S1). Thus, we surmised that measurement of nrdA and nrdJ expression would provide information on how the genetic background of each strain affects the expression of class I and II RNRs under oxidative stress conditions. When using the laboratory strain P. aeruginosa PA14, we observed that the basal expression level of nrdJ was similar to that of nrdJ in PAO1 (Fig. 2B). Figure 2B shows that the oxidative stress generated after incubation with 1 mM H 2 O 2 for 30 min induced nrdJ expression during the exponential and stationary growth phases. AlgR caused this gene induction, as indicated by the finding that the increased nrdJ expression was eliminated in a PA14 DalgR strain and when using the plasmid pETS-PJ-Dbox112. Expression was restored when the gene deletion was complemented with AlgR. In the exponential phase, the induced expression of nrdJ was higher when pETS-AlgR was used than when pETS- D54N was used. In the stationary phase, the induction of nrdJ was reestablished after AlgR-D54N was used.
We also measured RNR expression in the clinical strain P. aeruginosa PAET1, isolated from a chronic CF patient. Figure 2C shows the basal nrdJ expression in the clinical isolate PAET1 and nrdJ induction under incubation with 1 mM H 2 O 2 for 30 min. The graphs show that P. aeruginosa PAET1 nrdJ expression was lower than that in the laboratory strain PAO1. nrdJ expression increased after the addition of H 2 O 2 in the exponential and stationary phases. The induction was abolished in the DalgR strain or when the AlgR boxes of the promoter region were mutated. When the algR deletion was complemented with AlgR and AlgR-D54N, nrdJ induction was restored. When AlgR-D54N was used, the expression of nrdJ was higher than that when AlgR was used. This may imply that nrdJ induction under oxidative stress is due to nonphosphorylated AlgR.
The same type of regulation pattern was found when the general transcription of nrdA (class I RNR) under oxidative stress conditions was analyzed. The strains were incubated for 30 min with 1 mM H 2 O 2 when they reached exponential and stationary growth phases. Figure 3A shows that nrdA expression increased during the exponential and stationary phases in P. aeruginosa PAO1. This induction was abolished in PAO1 DalgR, and when we used the plasmid pETS-PA-DAlgRbox1, which carries a mutation in its AlgR binding box, the transcriptional induction was found to be specific to the transcription factor AlgR. The algR deletion was complemented with AlgR-D54N, restoring nrdA expression. nrdA showed lower basal expression in the P. aeruginosa laboratory strain PA14 than in PAO1 (Fig. 3B), demonstrating its strain variability. nrdA expression was also induced after 1 mM H 2 O 2 was added (by 1.7 times in the exponential growth phase and 1.2 times in the stationary phase). This induction was abolished in the DalgR strain and when using the plasmid pETS-PA-DAlgRbox1, again showing that AlgR was responsible for nrdA induction. The induction was restored when DalgR was complemented with the nonphosphorylated AlgR.
Finally, the expression of nrdA in the clinical isolate PAET1 is shown in Fig. 3C. The overall expression of nrdA was lower in PAET1 than in the PAO1 laboratory strain (by 26 times in the exponential phase and 2.5 times in the stationary phase). Even so, the induction pattern observed in the strains PAO1 and PA14 was also observed in PAET1. When 1 mM H 2 O 2 was added, nrdA was induced in the exponential and stationary phases, and this induction was abolished in both the DalgR strain and when the AlgR binding box was removed in the plasmid pETS-PA-DAlgRbox1. nrdA induction was reestablished when the DalgR strain was complemented with AlgR-D54N in the exponential phase and with AlgR and AlgR-D54N in the stationary phase.
During biofilm growth, class II and I RNR expression under oxidative stress conditions is AlgR-dependent. The different environmental characteristics inside a biofilm also affect the transcriptional activation of RNRs. While nrdA expression is activated in the upper layers of the biofilm, where oxygen is available, the class II (nrdJ) and III (nrdD) RNRs are activated in the deeper biofilm layers. Additionally, we previously demonstrated the essential role of nrdJ during biofilm development in P. aeruginosa PAO1 (8,11,14).
We used flow-cell chambers to grow different P. aeruginosa strain biofilms and measured RNR gene expression under oxidative stress conditions (see Materials and Methods). Luria-Bertani (LB) medium supplemented with 4 mM H 2 O 2 was added to a mature biofilm (96 h) to generate oxidative stress. Negative controls were treated with LB medium. Following an incubation period of 4 h, the biofilm was dyed and imaged under a confocal microscope (see Materials and Methods). The biofilm biomass was stained with SYTO60, which appears gray in the images ( Fig. 4 and 5); the DNA damage due to the oxidative stress present in the biofilm was dyed with CELLRox Orange, shown in magenta; and the RNR-specific expression was measured using green fluorescent protein (GFP) fluorescence (nrdJ and nrdA), shown in green ( Fig. 4 and 5). Figure 4A shows that nrdJ expression (pETS-PJ) was transcriptionally induced under oxidative stress conditions and almost completely eliminated when the AlgR boxes 1 and 2 were mutated (pETS-PJ-Dbox112) (Fig. 4A). This may indicate that during biofilm infection, AlgR is the primary transcription factor that regulates nrdJ expression under oxidative stress conditions. In addition, because biofilms show oxygen gradient heterogenicity, we wanted to study their pleiotropic regulation with the Anr transcriptional regulator. Anr is a transcription factor regulating several pathways related to oxygen tension and NO (20,21). We previously mutated the Anr box in nrdJ (14) and, because it is an important regulator of oxygen-depletion conditions, we wanted to determine its involvement in nrdJ expression in biofilms under oxidative stress conditions. nrdJ expression was similar when the single AlgR boxes were mutated and when a double mutation was present in the AlgR and Anr binding sites (pETS-PJ-Dbox112-DAnrbox), thus demonstrating that Anr does not play an important role during oxidative stress conditions (Fig. 4A). Figure 4B shows that when the P. aeruginosa PAO1 DalgR strain was complemented with AlgR or AlgR-D54N under oxidative stress conditions, nrdJ expression increased to the wild-type level. Furthermore, nrdJ induction was higher than that in the wild-type strain when the nonphosphorylated AlgR protein was used (Fig. 4B). These results indicate that AlgR plays a key role in nrdJ regulation during biofilm infection under oxidative stress conditions. nrdJ expression was quantified using COMSTAT software and plotted on a graph as shown in Fig. 4C.
The same expression pattern was observed when the expression of nrdA (class Ia RNR) was examined during biofilm formation (Fig. 5). Figure 5A shows images of P. aeruginosa PAO1 with pETS-PA and PAO1 with pETS-PA-Dbox1 after the addition of LB medium supplemented with 4 mM H 2 O 2 or pure LB medium. The images show that nrdA expression was induced under oxidative stress conditions during biofilm growth (Fig. 5C). The induction was removed when a promoter with a mutated AlgR box (pETS-PA-Dbox1) was used, demonstrating a direct role of AlgR in its transcriptional activation (Fig. 5A). When P. aeruginosa PAO1 DalgR was complemented with AlgR or AlgR-D54N, it was observed that nrdA expression was restored after addition of H 2 O 2 (Fig. 5B). The nrdA expression was not restored to the wild-type level but was closer when AlgR-D54N was used for complementation (Fig. 5C).
nrdJ plays an important role during Galleria mellonella infection in response to oxidative stress. Galleria mellonella is an outstanding alternative in vivo model to study bacterial infections because its innate response mimics that of mammals (22). G. mellonella has been previously used to study RNR expression during infection (23). Because Galleria larvae have autofluorescence, plasmids which produce GFP cannot be used to monitor gene expression during infection. Thus, the promoter regions used in this study were transcriptionally fused to luxCDABE genes that produce bioluminescence (23). After injecting the Galleria larvae, we made relative luminescence (RL) measurements for each larval group at 8, 14, 16, and 18 h postinfection. We determined the expression induction of each gene by comparing its expression at 14, 16, and 18 h with its expression at 8 h postinfection (the initial stage of infection). Figure 6A shows the induction of katA and katB expression during infection. It was determined that the expression of katA (pLUX-PkatA) was induced at 14, 16, and 18 h (113-, 865-, and 2,449-fold induction, respectively). In addition, katB (pLUX-PkatB) was AlgR Induces Ribonucleotide Reductase Gene Expression mSystems induced at 14, 16, and 18 h, but its expression was lower than that of katA (51-, 85-, and 1,220-fold induction, respectively). These results may indicate that as the amount of ROS increases during infection, the expression of bacterial catalases increases as well to protect the bacteria from damage. The induction of algR from the two promoters (pLUX-PfimS and pLUX-PalgR) was measured (Fig. 6B). We found that the expression of PalgR was higher than that of PfimS throughout the whole infection course, with the highest difference at 18 h postinfection (3,579-and 6,098-fold induction for PfimS and PalgR).
We measured the expression of nrdJ (pLUX-PJ), nrdJ with the AlgR boxes 1 and 2 mutated (pLUX-PJ-Dbox112), nrdJ with the Anr box mutated (pLUX-PJ-DAnrbox), and nrdJ with AlgR boxes 1 and 2 mutated and the Anr box mutated (pLUX-PJ-Dbox112-DAnrbox) (Fig. 6C). The results showed that nrdJ expression was highly induced during infection, with nrdJ showing the highest induction at 18 h postinfection (130,257-fold induction). The expression of nrdJ decreased significantly when the AlgR boxes were mutated and when the Anr box was mutated (3,519-and 2,669-fold induction at 18 h, respectively). However, the most dramatic decrease in nrdJ expression was observed when both the AlgR and Anr boxes were mutated (825-fold induction). The expression observed when the AlgR boxes and the AlgR and Anr boxes were mutated was 158 times lower than that of the wild-type promoter, revealing the importance of both transcription factors during infection.
Finally, we measured the expression of the class I RNR gene nrdA with the WT promoter (pLUX-PA) and the promoter with the AlgR box mutated (pLUX-PA-Dbox1). The graph in Fig. 6D shows that the highest induction in nrdA expression occurred at 18 h Notably, the induction of nrdJ expression was much larger than that of nrdA expression (37 and 13 times, respectively, compared with that of the AlgR box mutant promoter counterpart), which demonstrates the key role of the class II RNR gene nrdJ, compared to nrdA, which was not the primary RNR gene activated during the infection of G. mellonella.
The bioluminescence produced by the different promoters was visualized using the ImageQuant LAS 4000, and images were taken at 18 h postinfection (Fig. 6E). The amount of bioluminescence shown in the images of the larvae depends on the expression of each gene, which is in accordance with the results shown in Fig. 6A to D.

DISCUSSION
RNRs are essential enzymes in the life of any cell. Bacterial genomes commonly encode several RNRs to facilitate adaptation to different environmental conditions; thus, the expression and activation of each RNR class are tightly regulated. One of the main RNR transcriptional regulators is AlgR, which is part of the two-component system FimS-AlgR. AlgR is a global regulator, and FimS catalyzes its phosphorylation (8). The phosphorylation state of AlgR determines which regulatory pathways are activated or inactivated (3). We previously demonstrated that while nrdA is activated by phosphorylated AlgR in planktonic growth, nrdJ needs nonphosphorylated AlgR to be activated during early biofilm formation (8,12).
The transcription of algR is carried out through two different promoters regulated by specific transcription factors (7). Environmental signals activate algR transcription through a specific promoter. In this study, we found that the expression of algR was induced under oxidative stress conditions, especially during the stationary growth phase (Fig. 1). The algR expression levels were not very high, probably due to its role as a global bacterial regulator. Global regulators are tightly regulated because dramatic changes in their expression can modify several metabolic pathways in the cell. However, although the activation was low, it was mainly observed through PalgR, whose transcription is carried out after the binding of sigma factors such as RpoS and AlgU (7). The activation through the PalgR promoter may indicate that AlgR does not depend on FimS when ROS are produced; thus, it may be nonphosphorylated or simply found in small amounts. Thus, new experiments using rpoS and algU mutant strains should be performed in the future to unravel their exact roles under oxidative stress conditions. Other studies have linked AlgR to the oxidative stress defense system in P. aeruginosa, but its role is not yet clear (19). It is possible that one of the main roles that AlgR plays against ROS is activating the algD operon to produce alginate, as it is known that alginate scavenges ROS, protecting bacteria (15). Due to this, studying algD or algC when ROS are present could be another acceptable way to delve into the roles of AlgR and alginate under oxidative stress conditions.
AlgR is a key factor in the regulation of class II (nrdJ) and class I (nrdA) RNRs and is involved in the control of the total dNTP pool in the cell. We determined that AlgR is responsible for inducing nrdJ expression under oxidative stress conditions (Fig. 2). It seems that nrdJ depends directly on AlgR binding, as its induction was removed when the AlgR binding boxes of the nrdJ promoter region were mutated and in a DalgR strain (8). nrdJ induction was restored in a higher expression pattern when the algR mutation was complemented with the protein AlgR in its unphosphorylated state (pUCP-D54N). Using a fimS mutant strain could have been another acceptable way to study the phosphorylation state of AlgR under oxidative stress conditions. In the laboratory P. aeruginosa PA14 strain and the clinical isolate P. aeruginosa PAET1, we observed similar patterns of expression. nrdJ was induced when H 2 O 2 was present, and this induction was abolished in a DalgR strain and when using PnrdJ-Dbox112. AlgR-D54N restored nrdJ induction, indicating that AlgR was not phosphorylated. We observed that the basal expression of nrdJ in P. aeruginosa PA14 and PAET1 was lower than that in the PAO1 laboratory strain. We hypothesize that the reduced expression values observed may be due to differences in the genetic contexts of the three strains. However, the experiments confirmed that nonphosphorylated AlgR was the factor responsible for inducing nrdJ under oxidizing conditions. When we measured the expression of nrdA under oxidative stress conditions, we found that class I RNRs were transcriptionally induced in P. aeruginosa PAO1, PA14, and PAET1 (Fig. 3) (8). This induction was absent in a DalgR strain and when the AlgR binding box in the nrdA promoter was mutated (PnrdA-Dbox1). As shown in Fig. 2, nrdA expression in P. aeruginosa PA14 and in the clinical isolate P. aeruginosa PAET1 was lower than that in the laboratory strain PAO1. However, the experiments were useful enough to confirm that AlgR in its nonphosphorylated state was the factor responsible for inducing nrdA expression when H 2 O 2 was present.
During infection, the production of ROS is one of the main defensive mechanisms against bacteria (15). To evaluate whether AlgR regulates RNR expression under oxidative stress conditions during infection, we measured RNR induction in a continuous biofilm, which can simulate an infection-like situation (24). Other studies have already measured ROS in biofilms and have shown the importance of this growth condition during infection (25). In our experiments, we used a continuous flow biofilm to measure nrdJ and nrdA expression under oxidizing conditions (Fig. 4). The results showed that nrdJ expression was transcriptionally induced when H 2 O 2 was present and that this induction was completely abolished when the AlgR boxes on the promoter region were mutated. When the Anr box found in the nrdJ promoter was mutated together with the AlgR boxes, we showed that nrdJ expression was even lower. These results demonstrate the critical role of AlgR in regulating nrdJ during biofilm formation. Anr is a transcription factor which regulates genes involved in anaerobic conditions (21), and it was shown to have a role in the regulation of nrdJ during biofilm formation (Fig. 4). However, the use of an anr mutant strain or a nrdJ promoter with the Anr box mutated could contribute to a better understanding of the role of Anr on its own. Because the complementation of DalgR with AlgR and AlgR-D54N restored nrdJ expression under oxidative stress conditions, we determined that AlgR in its nonphosphorylated state was responsible for inducing nrdJ during biofilm formation. These results agree with those of previous studies showing that while phosphorylated AlgR regulates the initial steps of infection, nonphosphorylated AlgR controls the later steps, such as the production of alginate (7,8).
In addition, we observed that the expression of nrdA was transcriptionally induced under oxidative stress conditions and that this induction was abolished when the AlgR box was mutated, confirming that AlgR is responsible for nrdA induction in biofilms. When we evaluated whether AlgR bound to the promoter in its phosphorylated or nonphosphorylated state, we observed that nrdA expression did not reach WT levels; however, AlgR-D54N produced slightly higher expression than wild-type AlgR (Fig. 5).
Finally, we used the in vivo model G. mellonella to measure the expression of several genes (Fig. 6). Other authors have linked the hemocytes of G. mellonella with the production of ROS inside the larvae (26). To test whether the bacteria inside the larvae could sense the oxidative stress produced during infection, we measured the expression of the P. aeruginosa catalases katA and katB. We observed that during G. mellonella infection, the expression of the constitutive catalase katA was increased at all the time points measured, and the expression of katB, the catalase whose expression is activated only when large amounts of ROS are detected (15), was activated at 16 and 18 h postinfection. After confirming that P. aeruginosa sensed the oxidative stress produced, we measured the expression of the remaining genes. We observed that algR expression was induced throughout the infection process, with higher levels of induction factors observed with PalgR than with PfimS. When measuring the expression of the RNR, we observed that nrdJ showed the highest induction at 18 h postinfection, but its expression was completely eliminated when the AlgR boxes and the Anr box were mutated, confirming the remarkable roles of these transcription factors during infection (8,21). The high induction of nrdJ at the latest time point suggests the important role of class II RNR in infections. However, when investigating the expression of nrdA, we observed that it decreased when the AlgR box was mutated, but it was not completely eliminated; thus, other factors may regulate the expression of nrdA. These results highlight the essential role of nrdJ in G. mellonella infection as the principal transcriptionally active RNR gene.
It is important to mention that all the experiments were conducted using plasmid constructions, which may not reflect the endogenous expression levels inside the cell and could yield certain modifications of the final levels depending on the system studied. However, our goal was not to investigate the specific amount of RNR production but to understand all actors taking place in their transcriptional regulation.
We have designed a schematic diagram to summarize the steps followed by P. aeruginosa under oxidative stress conditions (Fig. 7). Under nonoxidative conditions and after specific signals are sensed, the transmembrane protein FimS is autophosphorylated and phosphorylates the regulatory protein AlgR (3). Phosphorylated AlgR induces the expression of nrdA during the exponential growth phase and in the initial stages of biofilm colonization (8).
During the stationary growth phase, the sigma factors AlgU and RpoS bind to the promoter region of algR. AlgU is a sigma factor (s 22 ) that is usually sequestered by the anti-sigma factor MucA. AlgU regulates genes important for alginate formation (27). RpoS is a sigma factor (s 38 ) whose expression is cell cycle-dependent. It is regulated by cell density, and it regulates the transition to stationary phase (28). Once AlgU and RpoS bind to PalgR, most of the AlgR protein produced is not phosphorylated (7). The nonphosphorylated AlgR favors the expression of nrdJ in the stationary phase of a planktonic culture and in biofilm.
Under oxidative stress conditions, AlgR expression is generated from PalgR, and class I and class II RNRs are induced by nonphosphorylated AlgR in planktonic, biofilm and infection conditions (Fig. 1-6). These results suggest that AlgU and/or RpoS are the factors responsible for sensing ROS when they are present. AlgU has been linked with the oxidative stress defense system in P. aeruginosa (29). In addition, other proteins related to AlgU and MucA, such as AlgW and MucP, have been related to oxidative stress. We hypothesize that under oxidative conditions, AlgW or MucP cleaves MucA, freeing AlgU (15). The freed AlgU may activate algR expression, and AlgU and AlgR together may activate the transcription of their target genes. In addition, the RNR NrdA and NrdJ would produce the dNTPs needed to repair the damaged DNA. Once the ROS are eliminated by the multifaceted system of P. aeruginosa, AlgU would be sequestered by MucA again. On the other hand, the RpoS counterpart in E. coli, in addition to being involved in the transition to stationary phase, plays a role in the sensing of several stresses, including oxidative stress. However, no conclusive studies have confirmed a notable role of RpoS in oxidative stress-sensing in P. aeruginosa beyond a slight relationship (30). Taking all of this into account, we believe that AlgU is the sigma factor involved in the activation of algR under oxidative stress conditions. However, there are still many gaps that need to be filled. Experiments using DalgU and DrpoS strains would help to shed light on the oxidative stress defense system and its relationship with the RNR regulatory network.
Conclusion. In conclusion, these results indicate that algR expression is induced after the addition of H 2 O 2 in exponential and stationary phases and that algR expression under oxidative stress conditions is mostly obtained through the promoter PalgR. These results may help unravel the role of AlgR in the oxidative stress response system. Additionally, we observed that under planktonic and biofilm conditions with oxidative stress, the expression of nrdJ and nrdA was transcriptionally induced by AlgR in its nonphosphorylated state in all P. aeruginosa strains tested. Finally, we showed that during FIG 7 Schematic representation of AlgR regulation with and without oxidative stress. Hypothetical regulatory model of the molecular pathway used by P. aeruginosa with or without oxidative stress conditions. Question marks indicate the putative pathways that oxidative stress may follow to induce algR expression and consequently class I (nrdA) and class II (nrdJ) ribonucleotide reductase (RNR) expression under specific growth conditions. Sources of information for each event are indicated in the manuscript. Artwork created with www.biorender.com.
AlgR Induces Ribonucleotide Reductase Gene Expression mSystems G. mellonella infection, the ROS produced by the larvae can be sensed by the bacteria and that nrdJ plays a major role in the production of dNTPs during an infection.

MATERIALS AND METHODS
Bacterial strains, plasmids, and growth conditions. The bacterial strains and plasmids used are listed in Table S1. E. coli and P. aeruginosa strains were routinely grown in Luria-Bertani (Scharlab, Spain) medium at 37°C. Liquid cultures were shaken at 200 rpm. When necessary, antibiotics were added at the following concentrations: 50 mg/mL ampicillin and 10 mg/mL gentamicin for E. coli; and 100 mg/mL gentamicin, 40 mg/mL tetracycline, and 300 mg/mL carbenicillin for P. aeruginosa.
DNA manipulation and plasmid construction. Recombinant DNA manipulations were performed using standard protocols (31). The molecular biology kits and enzymes used in this study were purchased from Thermo Fisher Scientific (Spain) except as otherwise stated and were used following the manufacturers' instructions. DNA fragments were amplified using Phusion High-Fidelity DNA polymerase or DreamTaq Green PCR MasterMix with the primers listed in Table S2. DNA fragments were isolated from agarose gels using a GeneJet Gel Extraction kit. Plasmid DNA was extracted using a GeneJET Plasmid Miniprep kit and transferred into P. aeruginosa cells via electroporation using a Gene Pulser XCell electroporator (Bio-Rad) as previously described (32). All the constructs obtained were verified with DNA sequencing by Eurofins Genomics.
The Anr binding box in the nrdJ promoter region was mutated using PCR-based site-directed mutagenesis. The primer pairs 7/8 and 9/10 were used to amplify two fragments of the templates pETS-PJ (pETS180) and pETS-PJ-Dbox112 (pETS211) to generate the DNA fragments PnrdJ-DAnrbox and PnrdJ-DAlgRbox112-DAnrbox, respectively. Each fragment was gel-purified and used as a template for a second round of PCR with the primers 7/10. The resulting amplicons were ligated into the pJET1.2b vector. The resulting plasmids and the pETS130-GFP plasmid were digested with BamHI-SmaI, and ligation was performed using the enzyme T4 ligase to obtain the plasmids pETS-PJ-DAnrbox (pETS232) and pETS-PJ-Dbox112-DAnrbox (pETS233). These plasmids were electroporated into P. aeruginosa PAO1. Each construct was verified by DNA sequencing.
Deletion of the algR gene in P. aeruginosa strain PA14 and the clinical PAET1 strains. The plasmid pEX100Tlink was used to obtain P. aeruginosa PA14 and PAET1 algR mutant strains. First, we performed PCR to amplify the upstream and downstream regions of algR using the primer pairs 11/12 and 13/14 and the chromosomal DNA of PA14 and PAET1 as the templates. The amplified fragments were gel-purified and ligated into pJET1.2b using the enzyme T4 ligase. The plasmids obtained and the vector pEX100Tlink were digested using the restriction enzymes HindIII-BamHI and BamHI-SacI. The gel-purified fractions were ligated with the digested pEX100Tlink plasmid to obtain pEX100Tlink::algR9-9algR (pETS242) and pEX100Tlink::algR9-9algR (pETS243) for PA14 and PAET1, respectively. Afterward, the plasmids pETS242, pETS243, and pUCGmlox were digested with the restriction enzyme BamHI. The digested fractions were gel-purified and ligated with the enzyme T4 ligase to obtain the plasmids pEX100Tlink::algR9-Gmlox-9algR (PA14; pETS244) and pEX100Tlink::algR9-Gmlox-9algR (PAET1; pETS245). These final constructs were transformed into the E. coli S17.1 helper strain.
The PA14 DalgRGmlox mutant (pETS131) and PAET1 DalgRGmlox mutant (pETS133) of P. aeruginosa were generated by introducing pETS244 and pETS245, respectively, from E. coli S17.1 by conjugation. LB medium supplemented with 5% sucrose was used to counterselect the gentamicin-resistant transconjugants PA14 DalgRGmlox and PAET1 DalgRGmlox. Next, the plasmid pCM157 was electroporated into PA14 DalgRGmlox and PAET1 DalgRGmlox. The mutant strains were grown on LB broth supplemented with tetracycline to remove the gentamicin resistance cassette via the expression of the cre recombinase (33). The pCM157 was then removed from the mutant strains by three successive growth cycles in LB broth without tetracycline. The selected PA14 DalgRlox mutant (pETS132) and PAET1 DalgRlox mutant (pETS134) of P. aeruginosa were sensitive to gentamicin and tetracycline.
Green fluorescent protein gene reporter assay. P. aeruginosa bacterial cultures were grown on LB without antibiotics at 37°C and 200 rpm to OD 550 = 0.5 (exponential phase) and OD 550 . 2 (stationary phase). Upon reaching the desired OD 550 , three independent 1-mL samples of each strain were collected. The samples were centrifuged for 10 min at 5,000 rpm, and the cell pellets were fixed with 1 mL of freshly prepared phosphate-buffered saline (PBS) solution containing 2% formaldehyde and stored in

SUPPLEMENTAL MATERIAL
Supplemental material is available online only. FIG S1, PDF file, 1.8 MB.