Pseudomonas aeruginosa ttcA encoding tRNA-thiolating protein requires an iron-sulfur cluster to participate in hydrogen peroxide-mediated stress protection and pathogenicity

During the translation process, transfer RNA (tRNA) carries amino acids to ribosomes for protein synthesis. Each codon of mRNA is recognized by a specific tRNA, and enzyme-catalysed modifications to tRNA regulate translation. TtcA is a unique tRNA-thiolating enzyme that requires an iron-sulfur ([Fe-S]) cluster to catalyse thiolation of tRNA. In this study, the physiological functions of a putative ttcA in Pseudomonas aeruginosa, an opportunistic human pathogen that causes serious problems in hospitals, were characterized. A P. aeruginosa ttcA-deleted mutant was constructed, and mutant cells were rendered hypersensitive to oxidative stress, such as hydrogen peroxide (H2O2) treatment. Catalase activity was lower in the ttcA mutant, suggesting that this gene plays a role in protecting against oxidative stress. Moreover, the ttcA mutant demonstrated attenuated virulence in a Drosophila melanogaster host model. Site-directed mutagenesis analysis revealed that the conserved cysteine motifs involved in [Fe-S] cluster ligation were required for TtcA function. Furthermore, ttcA expression increased upon H2O2 exposure, implying that enzyme levels are induced under stress conditions. Overall, the data suggest that P. aeruginosa ttcA plays a critical role in protecting against oxidative stress via catalase activity and is required for successful bacterial infection of the host.

Purified TtcA binds an oxidant-sensitive iron-sulfur cluster. To detect iron-sulfur cluster-TtcA ligation, P. aeruginosa TtcA expression in Escherichia coli and TtcA protein purification were performed as described in the Methods. The purified TtcA was then subjected to UV-visible spectroscopy scanning analysis to determine the presence of iron-sulfur clusters. The results in Fig. 2a show significant absorption at 420 nm in the UV-visible spectrum of the purified TtcA, suggesting the presence of a [Fe-S] cluster ligated with the protein, similar to the results of previous studies investigating the characteristics of iron-sulfur cluster proteins 23,24 . This finding was supported by an in silico analysis of the iron-sulfur cluster binding domains in the P. aeruginosa TtcA sequence, which contained two separate Cys-X-X-Cys motifs indicative of [Fe-S] cluster ligation.
Oxidative damage occurs when ROS oxidize an exposed Fe 2+ atom in the [4Fe-4S] cluster through a metalbased oxidation mechanism, resulting in the ejection of an iron atom from the cluster and subsequent reduction of the cluster to the inactive [3Fe-4S] + oxidation state 25,26 . To determine the effects of H 2 O 2 on [Fe-S] cluster integrity, purified TtcA was incubated with various concentrations of H 2 O 2 prior to performing UV-visible spectroscopy. The results showed decreases in TtcA absorbance at 420 nm that were H 2 O 2 concentration-dependent (Fig. 2a), suggesting that ligation of the [Fe-S] cluster to TtcA provided targets for H 2 O 2 -mediated oxidation (5-50 mM), resulting in the destabilization of iron-sulfur clusters bound to the protein. Treatment of the protein with a high concentration (0.5 M) of H 2 O 2 led to the total loss of [Fe-S] clusters bound to TtcA, as shown in Fig. 2b. Together with the previously described results, we found that P. aeruginosa TtcA contains the ROS-sensitive [Fe-S] cluster as its cofactor, similar to TtcA in E. coli, which contributes to the thiolation of cytosine 32 in tRNA 19 ; however, the importance of this cofactor for extended physiological function, particularly under oxidative stress conditions, still needs to be further investigated.
The ΔttcA mutant shows increased susceptibility to H 2 O 2 and sodium hypochlorite. To evaluate the physiological function of the ttcA in P. aeruginosa PAO1 against various stresses, a gene deletion mutant SCientifiC REpORTS | (2018) 8:11882 | DOI: 10.1038/s41598-018-30368-y (∆ttcA) was constructed in PAO1, as described in the Methods. Resistance levels against stresses including: osmotic (high salt, 5 mM NaCl), heat (high temperature, 50 °C), acidic (pH 4), basic (alkaline, pH 9) and oxidative stresses, were determined using a plate sensitivity assay and were compared to that of wild-type PAO1. The oxidants used in this study included: H 2 O 2 , sodium hypochlorite (NaOCl), organic hydroperoxides (cumene hydroperoxide [CHP]), superoxide generators (paraquat [PQ]), a thiol-depleting agent (N-ethylmaleimide [NEM]) and an intracellular iron chelating agent 2,2′-dipyridyl (DIPY). There were no significant differences in the resistance levels of the ∆ttcA mutant and wild-type PAO1 against high salt, high temperature, acidic pH, basic pH, organic hydroperoxides, superoxide generators, the thiol-depleting agent and the iron chelator (Fig. 3a). However, the ∆ttcA mutant exhibited 50-fold lower resistance to H 2 O 2 and an 8-fold reduction in the percent survival against NaOCl compared to PAO1 (Fig. 3a). The sensitive phenotype of the ∆ttcA mutant against both H 2 O 2 and NaOCl was complemented by the expression of a single copy of ttcA in Tn7 site (Fig. 3a), indicating that TtcA plays an important role in the H 2 O 2 -mediated and NaOCl-derived stress response. In PAO1, the cellular detoxification of H 2 O 2 primarily depended on catalase activity levels; however, other mechanisms, such as thiol-peroxidase activity (Tpx) and supporting systems, including haem biosynthesis, were also required to achieve fully responsive functionality against H 2 O 2 in P. aeruginosa. NaOCl is a bleaching agent that acts as a strong oxidizing agent and can disturb several enzymatic mechanisms, both redox and non-redox, including reactions in tRNA modification processes 27,28 . Moreover, NaOCl has been shown to generate intracellular ROS, which may increase H 2 O 2 levels and lead to mutant susceptibility 29 . Although our results showed that P. aeruginosa TtcA particularly affects the oxidative stress response, it may also alter the translational process of other biological pathways since it could function in the translational control.
Furthermore, PAO1 containing an extra copy of functional ttcA did not elevate the levels of resistance against these tested oxidants (Fig. 3a), suggesting that other components in the tRNA modification process are required or another detoxification system compensates for oxidant sensitivities. In addition to TtcA in P. aeruginosa PAO1, TrmJ, another tRNA-modifying enzyme, has also been shown to function in the oxidative stress response of P. aeruginosa PA14 18 .
[Fe-S] cluster-ligated cysteine residues are required for the physiological function of TtcA. To assess the important role of [Fe-S] clusters in TtcA-mediated protection against stress conditions, the site-directed mutagenesis of ttcA and a complementation assay were performed. Amino acids were changed from cysteine The alignments were performed using the CLUSTALW algorithm. Underlined and bold letters indicate the amino acids responsible for the PP-loop motif and conserved cysteines in TtcA, respectively. The asterisk, colon, and period symbols indicate identical residues, conserved substitutions, and semi-conserved substitutions, respectively. The numbers on top of the alignments indicate the positions of the amino acids. (b) Gene organization of ttcA (PA1192) in the P. aeruginosa PAO1 genome. (c) Northern analysis of ttcA transcript. Total RNA samples were prepared form exponential phase cultures of PAO1 induced with indicated concentrations of H 2 O 2 and hybridised with 32 P-labeled ttcA specific probe. The gene is transcribed mostly as a monocistronic transcript of approximately 825 bp compared to the size range of ribosomal RNAs. The lower panel shows 16S and 23S ribosomal RNAs at size of 1,536 bp and 2,891 bp, respectively, as a loading control and a size marker.
SCientifiC REpORTS | (2018) 8:11882 | DOI:10.1038/s41598-018-30368-y (C) to serine (S) at different positions in the TtcA, including a cysteine next to the PP-loop motif C38; putative cysteines for [Fe-S] cluster ligation at C115, C118 and C206; and other conserved cysteines, C184 and C203, using pUC18-mini-Tn7T-Gm-ttcA; then, mutated genes were transformed and integrated into the chromosome of the ∆ttcA mutant. A plate sensitivity assay using lethal concentrations of H 2 O 2 and NaOCl was performed to compare the susceptibility of bacterial growth between the transformed ∆ttcA mutants. The results in Fig. 3b show that increased susceptibility to H 2 O 2 in the ∆ttcA mutant was completely restored to wild-type PAO1 levels in ∆ttcA mutants containing either the native ttcA cassette (WT), the site-directed ttcA cassette with C38S, or C184S. However, H 2 O 2 susceptibility in the ∆ttcA mutant containing the site-directed ttcA cassette with either C115S, C118S, C203S or C206S demonstrated similar levels as the ∆ttcA mutant (Fig. 3b), indicating no phenotypic restoration among these site-directed mutant strains. Therefore, the four cysteines (C115, C118, C203 and C206) were required for fully functional TtcA to play role in the H 2 O 2 -mediated stress response. Moreover, a similar pattern was obtained with the NaOCl complementation assay, as shown in Fig. 3c, indicating that the site-directed ttcA cassette containing the cysteine residues (either C115S, C118S, C203S or C206S) was unable to restore NaOCl susceptibility of the ∆ttcA mutant to wild-type PAO1 levels, resulting in a sensitivity level similar to the ∆ttcA mutant. This suggested that these four cysteines were also required for the TtcA functionality in the NaOCl-mediated stress response. Similar observations regarding the importance of this conserved Cys-X-X-Cys motif in the TtcA protein have been reported for the thiolation of the cytidine in position 32 of tRNA in S. Typhimurium 21 and in E. coli 19 . Mutation of C219 to alanine in E. coli TtcA, mapped as C203 of P. aeruginosa TtcA, exhibited a 50% reduction in tRNA thiolation activity 19 , which was supported by the observed importance of this cysteine residue for the full function of TtcA. However, the exact function of these four cysteine residues in the TtcA, i.e. their possible involvement in the [Fe-S] cluster ligation, is under investigation.
The ΔttcA mutant exhibits decreased total catalase activity via KatA function. In several pathogenic bacteria, the cellular detoxification of H 2 O 2 mainly depends on catalase activity levels. The two major catalases KatA and KatB are responsible for cellular H 2 O 2 detoxification in P. aeruginosa PAO1 11,30 . To investigate the involvement of TtcA in the H 2 O 2 -mediated stress response through catalase activity, a total intracellular catalase activity assay was performed in wild-type PAO1 and the ∆ttcA mutants. The results showed that total catalase activity in the ∆ttcA mutant was 39% and 41% lower than that in wild-type PAO1 under the exponential and stationary phases, respectively, while the ∆ttcA mutant harbouring a functional ttcA cassette at the Tn7 site showed catalase activity levels similar to that of wild type (Fig. 4a). This result suggested that TtcA is required for full catalase activity in P. aeruginosa PAO1 under both the exponential and stationary phases.
To determine whether TtcA has roles in KatA or KatB activity, a catalase gel activity assay was performed. The results in Fig. 4b show that KatA activity in the ∆ttcA mutant was decreased compared to that in wild-type PAO1, and activity was restored by the expression of the functional ttcA cassette at the Tn7 site, as shown in the complemented strain. However, KatB activity levels were slightly increased in the ∆ttcA mutant compared to those in wild-type PAO1 (Fig. 4b). These data support the hypothesis that TtcA might have a direct role in KatA activity and may disrupted the H 2 O 2 -responsive pathways. To confirm the contribution of TtcA to KatA activity, double ∆katA∆ttcA and katB − ∆ttcA mutants were constructed and used to determine H 2 O 2 susceptibility levels in a plate sensitivity assay. If TtcA plays a role in KatA function, the sensitivity of the double mutant should be the same as that of the katA mutant. By contrast, if TtcA plays no role in the maintenance of KatA activity, the sensitivity in the double mutant should be additive compared to the single katA mutant. The results in Fig. 4c show that the ∆katA mutant was more than 1,000-fold more sensitive to H 2 O 2 , while the ∆ttcA mutant was approximately 50-fold less sensitive to H 2 O 2 compared to the sensitivity of wild-type PAO1. Additionally, the double ∆katA∆ttcA mutant exhibited H 2 O 2 susceptibility levels similar to the ∆katA mutant under a range of H 2 O 2 concentrations with differing lethality (Fig. 4c), suggesting that TtcA contributes to KatA activity against H 2 O 2 toxicity. Moreover, the results in Fig. 4d show that the katB − mutant was approximately 100-fold more sensitive to H 2 O 2 , while the double katB − ∆ttcA mutant exhibited increased H 2 O 2 susceptibility (20-fold) relative to the katB − mutant (Fig. 4d), suggesting that TtcA contributes to H 2 O 2 resistance primarily through regulation of KatA activity.

Deletion of ttcA causes a change in the expression of genes involved in the oxidative stress response.
To test whether the deletion of ttcA contributed to decreased KatA activity either at the transcriptional level or at the post-transcriptional level, expression analysis of katA in the ∆ttcA mutant compared to wild-type PAO1 was performed using real time RT-PCR analysis. The results in Fig. 5a show that katA expression in the ∆ttcA mutant was approximately three-fold higher than that in PAO1 under conditions lacking an oxidant, and increased katA expression in the ∆ttcA mutant was fully restored to wild-type levels by the extra copy of functional ttcA inserted at the Tn7 site. This suggests that decreased KatA activity in the ∆ttcA mutant does not result from altered katA expression at the transcriptional level; however, it may arise from post-transcriptional control, as it has previously been shown that TtcA has roles in translational control 19 , and decreased KatA activity was observed in this study. To extend our gene expression analysis, the expression profile of genes involved in the oxidative stress response, such as katB, oxyR, and tpx, was determined by performing real time RT-PCR analysis. The results in Fig. 5a show a partial increase in katB, oxyR and tpx expression in the ∆ttcA mutant compared to the expression in wild-type PAO1. Furthermore, the expression of katB and tpx was slightly increased in the ∆ttcA mutant under H 2 O 2 exposure, but there was no significant difference among these strains under  Fig. 4b that KatB activity was slightly increased in the ∆ttcA mutant compared to the activity in wild-type PAO1. However, the real time RT-PCR analysis could reflect an effect on both transcriptional alteration and the stability of mRNAs.
Furthermore, to observe the translational efficiency of the katA transcript, Western blot analysis was performed using an ectopic 6His-tagged katA expression vector on the ∆katA mutant background to compare the native ttcA (∆katA) and ttcA deletion (∆katA∆ttcA) strains. The results shown in Fig. 5b indicate that the relative amounts of 6His-KatA in the ttcA deletion mutant (∆katA∆ttcA/pkatA-6His, 36%) were dramatically lower than those in the native ttcA mutant (∆katA/pkatA-6His, 100%) and were partially restored by the extra copy of ttcA under Tn7-mediated expression (∆katA∆ttcA/pkatA-6His, 67%). These results suggested a defect in the translational efficiency of 6His-tagged katA expression in the absence of functional ttcA and indicated that ttcA plays roles in the oxidative stress response at the post-transcriptional level via KatA activity, and the disruption of functional ttcA alters the global expression profile of genes involved in oxidative stress management, including induction of katB and expression of tpx. Western analysis showed only the steady state level of KatA in the cell and therefore could not exclude an effect on KatA degradation.
The ΔttcA mutant shows attenuated virulence in a Drosophila host model. The full function of KatA is required for bacterial virulence in several model host systems, as shown in previous studies 31 , and TtcA has been shown to respond to oxidative stress via KatA activity; therefore, the contribution of ttcA to the bacterial pathogenicity of P. aeruginosa was evaluated using Drosophila melanogaster as a pathogen-host model. As shown in Fig. 6a, feeding the flies with cultured PAO1 resulted in 50.8 ± 12.5% (after incubation for 15 hours) and 36.0 ± 6.7% (after incubation for 24 hours) fly survival compared with 100 ± 0% (at both time points) fly survival when LB medium was fed to the flies as a negative control. Feeding the flies with ΔttcA mutants resulted in 1.6-fold and 2.1-fold increases in fly survival (81.7 ± 7.4% and 76.7 ± 7.2% after incubation for 15 and 24 hours, respectively) compared with feeding with PAO1 (Fig. 6a). Thus, ttcA deletion attenuated the virulence of P. aeruginosa PAO1 in the tested model (p < 0.01). The attenuated virulence phenotype of the ΔttcA mutant was restored in a ΔttcA mutant expressing a functional copy of ttcA (57.2 ± 8.2% and 40.5 ± 6.3% fly survival after incubation for 15 and 24 hours, respectively). The attenuated virulence phenotype was consistent with H 2 O 2 sensitivity levels in the ΔttcA mutant (Fig. 3a). In several plant and animal pathogenic bacteria, defects in peroxide detoxification or repair systems, such as knockout of catalase, methionine sulfoxide reductase or iron-sulfur cluster regulator Bacterial cultures were grown to the exponential cell phase prior to RNA extraction. Relative expression was analysed using the 16 S rRNA gene as the normalizing gene and expressed as the fold expression relative to wild-type PAO1 levels. The data shown are the mean and SD from three biologically independent experiments. The asterisks indicate statistically significant differences (p < 0.01) compared with PAO1 levels. (b) Western blot analysis of 6His-KatA levels in P. aeruginosa strains was determined using a mouse anti-6His antibody. Crude proteins were prepared from an equal amount of P. aeruginosa culture, and electrophoresis was carried out using 12.5% SDS-PAGE with protein markers. The fulllength blot was shown in the Supplementary Fig. S2. genes, render the mutant strains attenuated for virulence in model hosts 7,[31][32][33] . Hydrogen peroxide is one of the key components of innate immunity generated by host cells to eradicate invading microbes. In human hosts, H 2 O 2 is produced within the phagolysosomes of phagocytic cells to kill engulfed pathogens 2 . Thus, defective protection against H 2 O 2 toxicity in bacteria would reduce survival within the host. Hence, the attenuated phenotype may result from the reduced ability of the ΔttcA mutant to mitigate exposure to H 2 O 2 during host interactions.
Moreover, the four cysteine residues in TtcA were required for fully functionality in the oxidative stress response via catalase activity. To investigate the requirement of these TtcA cysteine residues in bacterial virulence, complementation with the site-directed ttcA mutants was evaluated in a Drosophila feeding assay. The results shown in Fig. 6b are similarly to those in Fig. 6a, indicating that feeding the flies with cultured PAO1 either with or without Tn7-mediated insertion of a ttcA expression cassette and incubation for 18 hours resulted in approximately 50% fly survival; however, feeding with ΔttcA mutants resulted in an approximately two-fold increase in fly survival. Feeding with the ΔttcA::Tn-TtcA mutant resulted in a fly survival level similar to that of the PAO1 strain (Fig. 6b). Substitution of these particular cysteine residues, either C115, C118, C203, or C206, with serine in the functional ttcA expression cassette and insertion into the ΔttcA mutant chromosome did not restore fly survival levels, while replacing one of the other conserved cysteines (either C38 or C184) in TtcA caused the phenotypic restoration of fly survival to wild-type PAO1 levels (Fig. 6b). This indicated that these four particular cysteine residues, including putative amino acids for iron-sulfur cluster ligation (C115, C118 and C206), were required for the complete functionality of TtcA in bacterial pathogenicity. Several P. aeruginosa genes involved in iron-sulfur cluster biogenesis, including IscR, have been shown to play a role in bacterial virulence, which may correlate with TtcA function in the H 2 O 2 -mediated oxidative stress response through catalase activity. ttcA expression is increased in response to H 2 O 2 and NaOCl exposure. Adaptive gene expression is a key component of bacterial defence against environmental stresses. The expression of many genes involved in oxidative stress protection and repair processes is frequently induced by exposure to oxidants 7,34-36 . The expression patterns of ttcA in PAO1 cultivated under inducing concentrations of various oxidants were determined using real time RT-PCR. The results illustrated that exposure of PAO1 to organic hydroperoxides, superoxide anion-generating agents, a thiol-chelating agent and an iron-chelating agent did not induce ttcA expression (Fig. 7a). By contrast, H 2 O 2 and NaOCl treatment of PAO1 highly induced ttcA expression by 13.4 ± 1.5-fold and 2.9 ± 1.1-fold, respectively (Fig. 7a). The induction of ttcA expression by H 2 O 2 and NaOCl treatment correlated with physiological analysis indicating that TtcA contributes to protection against H 2 O 2 and NaOCl. To extend the range of the oxidant-induced gene expression profile, several concentrations of oxidants were applied to bacterial cultures and analysed by real time RT-PCR. The results in Fig. 7b showed that PAO1 cultures were induced with H 2 O 2 at concentrations ranging from 0.2 mM to 1 mM and in a dose-dependent manner, which was similar to the gene expression profile obtained for genes in the OxyR regulon, including katA, katB, ahpB and ahpCF 11,30 . This hinted at the possibility that ttcA expression is regulated by OxyR, the global transcriptional regulator responding to H 2 O 2 . Moreover, extending the concentration range for NaOCl treatment from 0.001% to 0.02% showed that ttcA expression was not significantly altered under 0.005% NaOCl exposure compared to that in untreated PAO1 (Fig. 7c). This suggested that, unlike H 2 O 2 induction, induction in response to high concentrations of NaOCl (including a sublethal dose of 0.02%), which slightly induced ttcA expression (2.5-fold), may arise from NaOCl reactions generating oxidative stress and probably did not arise via direct NaOCl reactions with the regulator. The NaOCl induction mechanism is under further investigation.
OxyR modulates the expression of ttcA to control catalase activity under stress exposure. ttcA promoter analysis was performed and physically mapped in silico, and the results are presented in Fig. 7d. ttcA is located next to PA1191 with a 102-bp intergenic region. To characterize the ttcA promoter, putative +1 sites were investigated using 5′ RACE. The +1 site of ttcA was mapped to a cytosine located 28 bp upstream of its translational ATG start codon (Fig. 7d). Two sequences (GGGCTG and GCGTAAAAT, separated by 18 bp) that resembled the E. coli ơ 70 -35 and -10 promoter motifs were identified. Given the limited intergenic space and a putative promoter sequence analysis, the ttcA and PA1191 promoter motifs might overlap with each other. The canonical OxyR promoter recognition sequence was previously proposed to be ATAG-N7-CTAT-N7-ATAG-N7-CTAT 11 . We mapped the P. aeruginosa ttcA promoter region and found an upstream sequence (TCGGcgtcgctTTGTgcgtaaaATAGccagcttTTCT) that matched 56% (9 of 16 bases) of the OxyR promoter recognition sequence; therefore, we considered this sequence a putative OxyR binding domain of ttcA (Fig. 7d). This putative binding domain overlapped the -10 promoter region, implying an OxyR derepression mechanism for ttcA expression in response to oxidative stress. OxyR is a member of the LysR family of transcription regulators, which often use extended palindromic DNA sequences as binding boxes to modulate target gene expression, and diverse consensus sequences for OxyR binding boxes in target gene promoters have been proposed 30 . Direct binding of OxyR to the ttcA promoter was determined by electrophoresis mobility shift assay (EMSA). The results in the Fig. 8a demonstrated that purified OxyR binds specifically to the putative ttcA promoter in the presence of the reducing agent, dithiothreitol (DTT), suggesting that reduced OxyR is required for the ttcA promoter binding. These results support the hypothesis that OxyR directly binds to the ttcA promoter and may regulate ttcA expression.
To assess whether OxyR regulated the induction of ttcA expression upon exposure to oxidative stress, ttcA expression levels were examined in an ΔoxyR mutant (ΔoxyR/pBBR) and a complemented ΔoxyR/pBBR-OxyR strain using real time RT-PCR. oxyR mutant strains were constructed in PAO1 as described in the Methods. The results showed that under uninduced conditions, the expression of ttcA in the ΔoxyR mutant was approximately 15-fold higher than ttcA levels in PAO1, with p < 0.05 (Fig. 8b). H 2 O 2 and NaOCl treatments did further not enhance the expression of ttcA in the ΔoxyR mutant. The expression of oxyR from the pBBR1MCS-4 vector in the mutant led to the repression of ttcA expression to levels similar to those observed in PAO1 (Fig. 8b). Furthermore, the oxidant-induced expression of ttcA expression in the complemented strain was restored to wild-type levels  (Fig. 8b). These data strongly suggest that OxyR is a transcriptional repressor of ttcA expression. Thus, reduced OxyR likely functions as a transcriptional repressor of ttcA expression in the absence of the inducers H 2 O 2 and NaOCl. However, due to the presence of oxidants, oxidized OxyR either activates or derepresses ttcA expression, leading to upregulated ttcA expression, increased catalase activity, and increased resistance to H 2 O 2 and NaOCl. OxyR controls a core regulon of oxidative stress defensive genes and other genes involved in the regulation of iron homeostasis, quorum-sensing, protein synthesis and tRNA modification 37,38 . Our results indicate that OxyR is involved in oxidative stress defence through diverse paths of control against H 2 O 2 as well as NaOCl.

Conclusion
In this study, the physiological role of tRNA modification through Fe-S cluster-ligated TtcA in the pathogenic bacterium P. aeruginosa is presented. Under either H 2 O 2 -or NaOCl-mediated stress, direct transcriptional regulation through kat gene expression may not be sufficient to control cellular catalase activity, and translational control through tRNA modification is required. Herein, we proposed an additional model (Fig. 8c) involving OxyR regulation to control catalase activity via both direct transcription and indirect translation of TtcA under oxidative stress conditions. During bacterial infection, P. aeruginosa OxyR upregulates katA 39 and derepresses ttcA expression to increase catalase activity in response to H 2 O 2 generated via host defence mechanisms. It is possible that the consequence of ttcA inactivation is a malfunction in tRNA thiolation, which may affect the efficiency of translation, leading to low cellular levels of KatA and the impairment of the oxidative stress response as well as the attenuation of virulence. Although, extended exposure to H 2 O 2 may cause the disruption of Fe-S cluster-ligating enzymes including TtcA, the low cellular level of active Fe-S clusters would allow the IscR to turn on the ISC system for rebuilding the Fe-S cluster 6 and to increase the tpx expression in order to detoxify the H 2 O 2 36 . Together with four essential cysteine residues, including putative [Fe-S] cluster coordinators, TtcA has been shown to play an important role in the oxidative stress response and to facilitate bacterial survival during infection of the host, which emphasizes the critical role of the intracellular function of iron-sulfur cluster biogenesis and tRNA modification via IscR and OxyR regulation to mitigate oxidative stress and promote bacterial pathogenicity.

Methods
Bacterial strains, plasmids and growth conditions. Both E. coli and P. aeruginosa (PAO1, ATCC15692) strains were aerobically cultivated in Lysogeny broth (LB from BD Difco, USA) at 37 °C unless otherwise stated. Exponential phase cells (OD 600 of 0.5) were used in all experiments. All plasmids used in this study are listed in Supplementary Table S1. Molecular techniques. General molecular techniques were performed according to standard protocols 40 .
Transformation into P. aeruginosa strains was carried out using electroporation as previously described 41 . The oligonucleotide primers used in this study are listed in Supplementary Table S2.
Northern blot analysis. Total RNA isolation, gel electrophoresis, blotting, and hybridization were performed as previously described 35 . For analysis of ttcA expression, 20 µg of purified total RNA was loaded into the gel. Radioactively labeled probes were prepared using [α-32 P]dCTP and a DNA labeling bead (Amersham, GE Healthcare). A 247-bp fragment of the ttcA coding region used as a gene-specific probe was amplified from pBBR-TtcA using primers BT4675 and BT4676.
Construction of P. aeruginosa ΔttcA mutants. The ttcA deletion mutant was constructed using homologous recombination with an unmarked Cre-loxP system as previously described 42 . A 1,271-bp right-flank (RF) containing the C-terminal of the ttcA coding region and a 1,093-bp left-flank (LF) containing the N-terminal was separately amplified from PAO1 genomic DNA using primers EBI1009 and EBI1010 and primers EBI1007 and EBI1008, respectively. The RF fragment was digested with PstI and the 1,010-bp RF fragment was isolated and cloned into pUC18::Gm r6 digested with HindIII/blunted and PstI yielding pUCttcAR::Gm r . The LF fragment was digested with NcoI and the 931-bp was isolated and cloned into pUCttcAR::Gm r digested with MunI/blunted and NcoI yielding pUCΔttcA::Gm r . The constructed plasmid resulted in the deletion of 721 bp of the ttcA coding region. pUCΔttcA::Gm r was transferred into PAO1, and the ΔttcA::Gm r mutants were selected for the Gm r and Cb s phenotypes. An unmarked ΔttcA mutant was created using the Cre-loxP system to excise the Gm r gene as previously described 42 . To construct the ΔkatAΔttcA mutant, the pUCΔttcA::Gm r was transferred into ΔkatA mutant 36 and followed by similar methods.
Construction of plasmid and mini-Tn7 harbouring ttcA-coding regions. A pBBR-TtcA for ectopic expression of ttcA was constructed by amplifying the full-length ttcA with primers BT4673 and BT4674. The 868-bp PCR products were cloned into the medium-copy-number expression vector pBBR1MCS-4 43 cut with SmaI, yielding pBBR-TtcA. Single-copy complementation was performed using a mini-Tn7 system 41 . The full-length ttcA were cut from pBBR-TtcA and cloned into pUC18-mini-Tn7T-Gm-LAC 41 prior to transposing into either PAO1 or mutant strains, generating overexpressed (PAO1::Tn-ttcA) or complemented (ΔttcA::Tn-ttcA) strains. Construction of katB − and ∆ttcAkatB − mutants. The katB knockout mutant was constructed by an insertional inactivation method using pKNOCK vector 44 as previously described 35 . The katB fragment amplified from PAO1 genomic DNA with primers, BT5639 and BT5640, was cloned into pKNOCK Gm digested with SmaI, generating pKNOCK Gm katB, which was introduced into PAO1 by conjugation. The trans-conjugants were selected by the Gm r phenotype. To construct a double ∆ttcAkatB − mutant, the pKNOCK Gm katB was introduced into the genome of ∆ttcA mutant. The mutants were confirmed by PCR and DNA sequencing.
Expression and purification of P. aeruginosa TtcA. 6His-tagged TtcA from P. aeruginosa was expressed using the pQE-30Xa expression system (Qiagen, Germany) under oxygen-limited conditions as previously described 34 . The full-length ttcA gene was amplified from PAO1 genomic DNA with the primers EBI1035 and EBI1036. An 835-bp PCR product was digested with HindIII before ligation into pQE-30Xa digested with StuI/ blunted and HindIII to generate pQE-30Xa-ttcA for the high-level expression of ttcA containing an N-terminal 6His-tag. An E. coli M15 strain harbouring pQE-30Xa-ttcA was grown to an OD 600 of 1.0 before being induced with 100 μM IPTG for 60 min with 40 rpm shaking. Purification of 6His-tagged TtcA was carried out using a nickel-nitrilotriacetic acid (Ni-NTA) agarose column as previously described 34 . The purity of the TtcA protein was more than 95%, as judged by a major band corresponding to the 32.3-kDa protein observed on SDS-PAGE. UV-Visible spectrophotometry was used to analyse the relative amount of [Fe-S] cluster using the ratio of absorbance 420 nm and 280 nm.
Plate sensitivity assay. A plate sensitivity assay was performed to determine the oxidant resistance level as previously described 7 . Briefly, exponential phase cells were adjusted to OD 600 of 0.1 before making 10-fold serial dilutions. 10 μl of each dilution was then spotted onto LB agar plate containing testing reagents. The plates were incubated overnight at 37 °C before the colony forming units (CFU) were scored. Percent survival was defined as the percentage of the CFU on plates containing oxidant divided by the CFU on plates without oxidant.
Hydrogen peroxide, NaOCl, heat and pH susceptibility test. A susceptibility assay was performed to determine the stress resistance level as previously described 35 . In short, exponential-phase cultures were normalized to an OD 600 of 0.1 before treating with lethal concentrations of either H 2 O 2 , NaOCl, heat, acidic pH, or basic pH for 30 min at 37 °C. After treatment, cells were immediately washed twice with fresh LB broth. Cells that survived the treatment were scored using a viable cell count. The resistance levels against these stresses were expressed as the % survival, defined as the percentage of the CFU with treatment divided by the CFU without treatment.
Catalase activity assays. Total catalase activity in P. aeruginosa cells was measured by spectrophotometrically monitoring the decomposition of hydrogen peroxide 6 . Briefly, the reaction was performed by mixing bacterial lysate with 30 mM H 2 O 2 in 50 mM phosphate buffer pH 7.0. The absorbance changes at A 240 were recorded at time intervals and calculated as the specific activity of catalase (U mg −1 protein). One unit of catalase was defined as the amount of enzyme required to hydrolyse 1 µmol of H 2 O 2 per min at 25 °C, pH 7.0, and the molar extinction e 240 was equal to 0.041 cm 2 µmol −1 .
The gel activity of Kat was intensely measured from native PAGE of P. aeruginosa cell extracts, which were stained for Kat activity as previously described 45 . Thirty milligrams (unheated) of protein were loaded, and protein concentrations were estimated using Bradford assay (Bio-Rad, USA). The stained gel was renatured, and catalase activity was visualized following a previously described method 46 with some modifications by washing twice before soaking with horseradish peroxidase (Sigma, USA) and then removing this enzyme. The gel was immediately soaked in 5 mM H 2 O 2 and stained with 3,3′diaminobenzidine. Catalase activity was visualized as colourless bands against a brownish background.
Western blot analysis. Western blot analysis was performed as previously described 47 . In brief, crude protein was extracted and isolated before mixing with 6X protein loading dye and boiling for 10 min. The stained protein was run under 12.5% SDS-PAGE and transferred to a Hybond PVDF membrane (GE Healthcare) in a semi-dry transfer cell (Bio-Rad). The transferred membrane was blocked and hybridized with anti-6His-peroxidase primary antibody (Roche, Switzerland) and developed with Ultra TMB-Blotting Solution (Thermo Scientific) according to the manufacturer's recommendation.
Drosophila virulence test. The virulence of P. aeruginosa was investigated using the Drosophila melanogaster feeding assay as previously described 7 . Shortly, 800 µL of P. aeruginosa cultures were overlaid to completely cover the surface of the corn flour Drosophila medium in a glass fly culture vial. One-week-old adult flies were starved for 3 hours prior to the feeding assay. Twenty flies were added to each vial and incubated at 25 °C before the number of the viable flies was observed at different time points. The experiments were performed in a double-blind fashion and were analyzed from nine experiments using three different batches of flies.
Real time RT-PCR was conducted using a SYBR ® FAST qPCR kit (KAPA Biosystems, USA). The reaction was run on an Applied Biosystems StepOnePlus thermal cycler under the recommended fast protocol condition. The specific primer pairs used for ttcA, katA, katB, oxyR, and tpx were BT4675-BT4676, BT5637-BT5638, BT5639-BT5640, EBI163-EBI164 and BT3186-BT3787, respectively 35 . The primer pair for the 16 S rRNA gene was BT2781-BT2782, which was used as the normalizing gene. Relative expression analysis was calculated using StepOne software and is presented as expression fold-change relative to the level of uninduced conditions. Data shown are the means with standard deviations (SD) from three biologically independent experiments. 5′ rapid amplification of cDNA ends (RACE). 5′ RACE was performed using a 5′/3′ RACE kit (Roche, Switzerland) as previously described 34 . Essentially, DNase I-treated total RNA was reverse transcribed using specific primers EBI341 as SP1 primers. The first-strand DNA (cDNA) was purified, and poly(A) was added to the 5′-terminus of the cDNA using terminal transferase. Next, poly(A)-tailed cDNA was PCR-amplified using the specific SP2 primer BT4991 and an anchored oligo(dT) primer. The purified PCR product was cloned into the pGEM-T Easy vector, and the +1 site was identified from the DNA sequences.
OxyR purification and electrophoresis mobility shift assay. 6His-tagged OxyR from P. aeruginosa was purified using the pQE-30Xa expression system in a manner similar to that used for TtcA purification. The purity of OxyR was more than 90% as a 35 kDa major band observed in SDS-PAGE. EMSA was performed as previously described 34 using a 32 P-labeled probe containing the putative ttcA promoter. The probe was amplified with 32 P-labeled BT4990 and BT4991 primers. Binding reactions consisting of labeled probe in 20 µl of reaction buffer containing 20 mM Tris-HCl (pH 7.0), 50 mM KCl, 1 mM EDTA, 5% glycerol, 50 μg ml −1 BSA, 5 μg ml −1 calf thymus DNA, 0.5 μg ml −1 poly[dI-dC], 50 μg ml −1 salmon sperm DNA, 100 mM DTT, and purified OxyR were incubated at 25 °C for 30 min. Gel electrophoresis and visualization was done as previously described 34 . Statistics. Group data are presented as means ± standard deviation (SD). The Student's t-test was used to determine differences between means using the function of Excel (Microsoft, Washington) and the SPSS (version 17.0; SPSS Inc.) statistical package. Unless otherwise is stated, p values of <0.05 were considered significant.
Ethics statement. All P. aeruginosa and D. melanogaster were raised, maintained and all experiments were conducted following procedures, MUSC2016-002 and MUSC60-039-389, approved by the Committee of Biosafety, Faculty of Science, Mahidol University (MUSC) and the MUSC-Institutional Animal Care and Use Committee (IACUC), respectively.