The strains, plasmids and primers used in this study are shown in Supplementary Table S1 .

DNA-free RNA was prepared from S. maltophilia KJΔChrR cells and reverse transcribed into cDNA using the ChrA-C primer ( Supplementary Table S1 ). cDNA was used as the template for PCR using the primer sets of RpoEcQ-F/R, ChrRQ-F/R, and ChrAQ-F/R ( Supplementary Table S1 ). The ChrAQ-F/R primer sets were used as a control for DNA contamination check.

In-frame deletion mutants were constructed using allelic replacement strategy as described previously (Yang et al., 2009). Two DNA fragments flanking the deleted region were amplified by PCR and subsequently cloned into pEX18Tc to generate pΔRpoEc, pΔChrR, and pΔChrA. The primer sets used were RpoEcN-F/R and RpoEcC-F/R for pΔRpoEc, RpoEcC-F/R and ChrRC-F/R for pΔChrR, as well as ChrRC-F/R and ChrAC-F/R for pΔChrA ( Supplementary Table S1 ). The resulting pEX18Tc-derived constructs were transported to S. maltophilia strain by conjugation. Transconjugants selection and mutants confirmation were performed as described previously (Yang et al., 2009). The double and triple mutants were constructed from the single mutant sequentially using the same protocol.

Tert-Butyl hydroperoxide (tBOOH) and Rose Bengal (RB) were used for the generation of single oxygen. The tBOOH reacts with peroxynitrite to generate singlet oxygen (Di Mascio et al., 1997). Rose Bengal is a UV/VIS absorbing molecule capable of absorbing and using light energy to excite oxygen to singlet oxygen (DeRosa and Crutchley, 2002). For cell viability, the logarithmic-phase bacterial cells tested of 2 × 10⁵ CFU/μL were 10-fold serially diluted. Five microliters of the bacterial suspension were spotted onto the LB agar with and without tBOOH or RB as indicated. For the RB test, two plates were prepared. One plate was covered with foil seal to create “dark” conditions in which little or no singlet oxygen is produced. The other plate was kept in the light, representing the singlet oxygen-stressed condition. After 18-h incubation at 37°C, the bacterial viability was imaged. For growth curve, overnight culture was inoculated into LB broth with and without tBOOH or RB at an initial OD₄₅₀ of 0.15. Bacterial growth was monitored for 24 h.

For cell viability, the logarithmic-phase bacterial cells of 2×10⁵ CFU/μL were serially 10-fold diluted. Five microliter bacterial aliquot was spotted onto LB agar with and without MD or antibiotic as indicated. After a 24-h incubation at 37˚C, the cell viabilities were recorded. For growth curve, overnight culture was inoculated into LB broth with and without MD or antibiotic at an initial OD₄₅₀ of 0.15. Bacterial growth was monitored for 24 h.

For the construction of a transcriptional fusion of the rpoEc promoter with a promoterless xylE gene, the 356-bp DNA segment upstream of rpoEc was PCR-amplified using rpoEcN-F/R primer pair ( Supplementary Table S1 ) and cloned into pXylE (Chen et al., 2011) to generate pRpoEc_xylE.

C23O, encoded by xylE, can convert catechol to an intensely yellow products, which can be spectrophotometrically quantified. One unit of enzyme activity (Uc) was defined as the amount of enzyme that converts 1 nmol of catechol per minute. The specific activity (Uc/OD₄₅₀) of the enzyme was defined as units per OD450 unit of cells. All data were reported from experiments performed in triplicate.

Overnight cultures of KJ, KJΔChrR, and KJΔChrRΔRpoEc were inoculated into fresh LB broth at an initial OD₄₅₀ of 0.15. After 5-h culture at 37°C, total RNA was prepared for RNAseq transcriptome analysis. RNA isolation, rRNA depletion, adapter-ligated cDNA library, and cDNA sequencing were carried out as described previously (Wu et al., 2022). The sequencing reads were mapped to the genome of K279a (Crossman et al., 2008). The total number of reads per gene was normalized by transcripts per kilobase million (TPM). The RNA-seq data have been deposited in GenBank under BioProject accession numbers SAMN41918898 for KJ, SAMN41918899 for KJΔChrR, and SAMN41918900 for KJΔChrRΔRpoEc.

RNA was isolated from logarithmical phase bacterial cells and converted to cDNA using the High Capacity cDNA Reverse Transcription Kit (Applied Biosystems) following the manufacturer’s instructions. Real-time PCR was performed using the TaqMan Universal PCR Master Mix (Applied Biosystems) and an ABI Prism 7000 Sequence Detection System (Applied Biosystems). The primer sets used are listed in Supplementary Table S1 . The 16S rRNA was used to normalize the gene expressions. Fold change was calculated using the ΔΔC_T method (Livak and Schmittgen, 2001). Three biological replicates were performed.

The σ^E-ChrR pair homologs are distributed among α-proteobacteria and γ-proteobacteria, such as R. sphaeroides, Vibrionaceae, and Pseudomonas (Glaeser et al., 2011). To identify the σ^E-ChrR homolog in S. maltophilia, we performed BLASTP analysis using P. aeruginosa ChrR (accession No. CRR21278) as the query. The search results revealed a single candidate, Smlt2378, showing 94% protein identity with P. aeruginosa ChrR ( Figure 1A ). Next, genomic organization surrounding smlt2378 was surveyed. smlt2377 encoded an ECF σ factor, demonstrating that the σ^E-ChrR pair is conserved in S. maltophilia. Eleven base pairs downstream of smlt2378 and in the same orientation, a third open reading frame of 693 bp (Smlt2379) was identified ( Figure 1B ). smlt2379 encoded a 23.6-kDa cytoplasmic protein with an NAD(P)-binding domain that highly functioned as an oxidoreductase. Genomic organization strongly suggested the presence of the smlt2377-smlt2378-smlt2379 operon, which was verified via reverse transcription-polymerase chain reaction (RT-PCR) ( Figure 1C ). Based on the results presented in this study, we designated smlt2377-smlt2378-smlt2379 as rpoEc-chrR-chrA. Next, we examined the conservation of rpoEc-chrR-chrA operon in S. maltophilia. Among the 12 strains surveyed, all harbored this operon, indicating high conservation of the rpoEc-chrR-chrA operon in S. maltophilia. In addition, we also noticed that there are nine genes (smlt2375-smlt2367) located upstream of rpoEc-chrR-chrA operon and divergently transcribed ( Figure 1D ). This genomic organization highly suggested that the nine genes may form an operon and regulated by σ^Ec, which will be further expounded later.

To characterize rpoEc-chrR-chrA operon, its genes were deleted from S. maltophilia KJ, either alone or in combination, to yield KJΔRpoEc, KJΔChrR, KJΔChrA, KJΔChrRΔRpoEc, KJΔChrRΔChrA, KJΔRpoEcΔChrA, and KJΔRpoEcΔChrRΔChrA ( Supplementary Table S1 ). These mutants exhibited no observable growth defects on LB agar ( Figure 2A ) or broth (data not shown).

A study on σ^E-ChrR pair of R. sphaeroides indicated its contribution to the singlet oxygen stress response (Anthony et al., 2004). We were interested in understanding whether the rpoEc-chrR-chrA operon played a role in the alleviation of singlet oxygen stress. Rose Bengal (RB) and tert-butyl hydroperoxide (tBOOH) methods were used to evaluate singlet oxygen tolerance (Di Mascio et al., 1997; Brahmachari and Karmakar, 2020). All the mutants tested displayed almost comparable viability to wild-type KJ in RB- and tBOOH-containing LB agar or broth ( Figures 2A, B ), indicating that rpoEc-chrR-chrA hardly contributes to singlet oxygen alleviation in our assay.

Involvement of the rpoE_C-chrR-chrA operon in superoxide tolerance was assessed using an MD tolerance assay. Inactivation of chrR increased bacterial tolerance to MD, and complementation of KJΔChrR with a plasmid containing chrR reversed MD tolerance to wild-type levels ( Figure 3A ). Furthermore, further inactivation of rpoEc or chrA in KJΔChrR partially restored MD tolerance ( Figure 3A ). Thus, σ^Ec activation increases MD tolerance, and ChrA is involved in this regulatory circuit.

Next, we investigated the role of rpoE_C-chrR-chrA operon in H₂O₂ tolerance. All tested strains exhibited comparable H₂O₂ susceptibilities ( Figure 3B ), tentatively ruling out the contribution of rpoE_C-chrR-chrA operon to H₂O₂ tolerance.

S. maltophilia is intrinsically resistant to various antibiotics (Crossman et al., 2008). Clinically, CAZ and fluoroquinolones are the choices for the treatment of S. maltophilia infection. Stress alleviation systems can cross-protect bacteria from antibiotics (Shin et al., 2020). Therefore, we assessed the involvement of rpoEc-chrR-chrA operon in susceptibility to CAZ and levofloxacin (LEV) susceptibility. The viabilities of wild-type KJ and its derived mutants in MH agar or broth containing CAZ and LEV were assessed. Of the three single-deletion mutants, KJΔChrR showed compromised viability in a medium containing CAZ, and viability was restored by ChrR complementation ( Figure 4 ). Furthermore, KJΔChrRΔRpoEc and KJΔChrRΔChrA displayed viability almost comparable to wild-type KJ ( Figure 4 ). Collectively, ΔchrR-mediated rpoEc and chrA upregulation contributed to an increase in CAZ susceptibility. With respect to LEV susceptibility, all rpoEc-chrR-chrA operon-associated mutants displayed viability comparable to wild-type KJ ( Figure 4 ), tentatively ruling out the involvement of rpoEc-chrR-chrA operon in LEV susceptibility.

The mechanisms responsible for CAZ susceptibility in S. maltophilia can be attributed to β-lactamase-mediated and non-β-lactamase-mediated mechanisms (Huang et al., 2021). To assess whether β-lactamase is involved in the ΔchrR-mediated CAZ susceptibility increase, CAZ-induced β-lactamase activities of KJ, KJΔChrR, KJΔChrRΔRpoEc, and KJΔChrRΔChrA were determined. All mutants tested displayed comparable CAZ-induced β-lactamase activities with wild-type KJ ( Supplementary Figure S1 ), indicating that σ^Ec activation-mediated increase of β-lactam susceptibility is irrelated to β-lactamase activity.

We constructed a P_rpoEc and xylE transcriptional fusion construct, pRpoEc_xylE ( Supplementary Table S1 ), to assess the expression of rpoEc-chrR-chrA operon by C23O determination. Plasmid pRpoEc_xylE was transformed into the wild-type and its derived rpoEc-chrR-chrA operon-associated mutants ( Supplementary Table S1 ) to assess the autoregulation circuit. KJ(pRpoEc_xylE) exhibited moderate C23O activity, supporting the intrinsic expression of rpoEc-chrR-chrA operon. Compared to that of wild-type KJ, the promoter activity of rpoEc-chrR-chrA operon was significantly increased in KJΔChrR and reverted to wild-type levels in KJΔChrRΔRpoEc and KJΔChrRΔChrA ( Figure 5 ). These results are consistent with the previous understanding of σ^E-ChrR system that ChrR functions as an anti-σ^E and σ^E imposes autoregulation on its own expression (Donohue, 2019). A more interesting finding is that ChrA played a positive role in ΔchrR-mediated rpoEc upregulation ( Figure 5 ). To further clarify this, the promoter activity of rpoEc-chrR-chrA operon in KJΔChrA was assessed. KJΔChrA(pRpoEc_xylE) displayed lower C23O activity than KJ(pRpoEc_xylE) ( Figure 5 ), supporting that ChrA plays a positive role in regulating the transcriptional expression of rpoEc-chrR-chrA operon.

Given the contribution of rpoEc-chrR-chrA operon to MD tolerance ( Figure 3A ) and CAZ susceptibility ( Figure 4 ), we wondered whether MD and CAZ were the stimuli that induced the expression of rpoEc-chrR-chrA operon. The C23O activity expressed by KJ(pRpoEc_xylE) under MD and CAZ challenges was determined. MD- and CAZ-treated KJ(pRpoEc_xylE) exhibited C23O activity comparable to the untreated counterpart ( Supplementary Figure S2 ), indicating that MD and CAZ are not the stimuli that activate σ^Ec.

To elucidate the mechanism underlying the σ^Ec activation-mediated increase in MD tolerance and CAZ susceptibility, RNA-seq transcriptome analysis of wild-type KJ, KJΔChrR, and KJΔChrRΔRpoEc was performed once. Differentially expressed genes (DEGs) were considered significant if the change in transcripts per kilobase million (TPM) between KJ and KJΔChrR was greater than three-fold. Twenty-four DEGs were revealed ( Supplementary Table S2 ). Of them, 23 genes were upregulated and one was downregulated in KJΔChrR ( Supplementary Table S2 ). Furthermore, the TPM values of the 24 DEGs in KJΔChrRΔRpoEc were significantly reverted ( Supplementary Table S2 ), indicating that these 24 genes were members of the ChrR-σ^E regulatory circuit. To validate the transcriptome data, we performed qRT-PCR to probe the genes rpoEc, smlt2373, smlt2382, cytB, L1, and L2. The results supported the reliability of the transcriptome results ( Supplementary Figure S3 ). By inputting the 500-base region upstream of the 24 DEGs into the motif search program MEME (Bailey et al., 2009), we generated a putative consensus DNA-binding motif of σ^Ec (σ^Ec box) ( Figure 6A ). Figure 6B shows the putative σ^Ec box located upstream of rpoEc and smlt2375.

As β-lactam resistance of S. maltophilia is strongly linked to PG homeostasis and β-lactamase induction (Huang et al., 2021), we checked the TPM values of L1, L2, and 37 known PG homeostasis-associated genes. No significant DEGs were identified among the 39 genes ( Supplementary Table S3 ). β-lactamase activity and transcriptome analyses ( Supplementary Figure S1 , Supplementary Table S3 ) suggested the involvement of a non-β-lactamase mechanism in the σ^Ec activation-mediated increase in β-lactam susceptibility.

Among the identified DEGs ( Supplementary Table S2 ), a nine-gene cluster, smlt2375-smlt2367, was the most significantly upregulated ( Table 1 ). The nine genes were located upstream of rpoEc-chrR-chrA operon and divergently transcribed ( Figure 1 ). Furthermore, these genes were simultaneously upregulated in KJΔChrR but significantly reverted in KJΔChrRΔRpoEc ( Table 1 ). These observations strongly suggest that these nine genes form an operon regulated by σ^Ec. The annotations and locations of the proteins encoded by smlt2375-smlt2367 cluster are summarized in Table 1 . The nine-gene cluster appeared to be involved in membrane lipid modification as it encoded a fatty acid desaturase (Smlt2374), an oxidoreductase (Smlt2373), two cyclopropane-fatty-acyl phospholipid synthase (CFA synthases) (Smlt2371 and Smlt2368), and five hypothetical proteins ( Table 1 ). To investigate the role of the nine-gene upregulation in ΔchrR-mediated increase in MD tolerance and CAZ susceptibility, a nine-gene deletion mutant was constructed in wild-type KJ and KJΔChrR, yielding KJΔCfa9 and KJΔChrRΔCfa9, respectively ( Supplementary Table S1 ). KJΔCfa9 and wild-type KJ showed comparable MD tolerance and CAZ susceptibility; however, compared to those in KJΔChrR, MD tolerance and CAZ susceptibility were almost reverted to wild-type levels in KJΔChrRΔCfa9 ( Figures 3A , 4 ), indicating that smlt2375-smlt2367 genes upregulation contributes to ΔchrR-mediated increase in MD tolerance and CAZ susceptibility.