Implication of the σE Regulon Members OmpO and σN in the ΔompA299–356-Mediated Decrease of Oxidative Stress Tolerance in Stenotrophomonas maltophilia

ABSTRACT Outer membrane protein A (OmpA) is the most abundant porin in bacterial outer membranes. KJΔOmpA299–356, an ompA C-terminal in-frame deletion mutant of Stenotrophomonas maltophilia KJ, exhibits pleiotropic defects, including decreased tolerance to menadione (MD)-mediated oxidative stress. Here, we elucidated the underlying mechanism of the decreased MD tolerance mediated by ΔompA299–356. The transcriptomes of wild-type S. maltophilia and the KJΔOmpA299–356 mutant strain were compared, focusing on 27 genes known to be associated with oxidative stress alleviation; however, no significant differences were identified. OmpO was the most downregulated gene in KJΔOmpA299–356. KJΔOmpA299–356 complementation with the chromosomally integrated ompO gene restored MD tolerance to the wild-type level, indicating the role of OmpO in MD tolerance. To further clarify the possible regulatory circuit involved in ompA defects and ompO downregulation, σ factor expression levels were examined based on the transcriptome results. The expression levels of three σ factors were significantly different (downregulated levels of rpoN and upregulated levels of rpoP and rpoE) in KJΔOmpA299–356. Next, the involvement of the three σ factors in the ΔompA299–356-mediated decrease in MD tolerance was evaluated using mutant strains and complementation assays. rpoN downregulation and rpoE upregulation contributed to the ΔompA299–356-mediated decrease in MD tolerance. OmpA C-terminal domain loss induced an envelope stress response. Activated σE decreased rpoN and ompO expression levels, in turn decreasing swimming motility and oxidative stress tolerance. Finally, we revealed both the ΔompA299–356-rpoE-ompO regulatory circuit and rpoE-rpoN cross regulation. IMPORTANCE The cell envelope is a morphological hallmark of Gram-negative bacteria. It consists of an inner membrane, a peptidoglycan layer, and an outer membrane. OmpA, an outer membrane protein, is characterized by an N-terminal β-barrel domain that is embedded in the outer membrane and a C-terminal globular domain that is suspended in the periplasmic space and connected to the peptidoglycan layer. OmpA is crucial for the maintenance of envelope integrity. Stress resulting from the destruction of envelope integrity is sensed by extracytoplasmic function (ECF) σ factors, which induce responses to various stressors. In this study, we revealed that loss of the OmpA-peptidoglycan (PG) interaction causes peptidoglycan and envelope stress while simultaneously upregulating σP and σE expression levels. The outcomes of σP and σE activation are different and are linked to β-lactam and oxidative stress tolerance, respectively. These findings establish that outer membrane proteins (OMPs) play a critical role in envelope integrity and stress tolerance.

Recently, we reported that ompA is highly expressed in logarithmic-phase S. maltophilia KJ cells (38). We characterized an in-frame deletion ompA mutant of S. maltophilia KJ, originally termed KJDOmpA (39) and later renamed KJDOmpA 299-356 (40). We also reported that the truncated OmpA protein can be stably embedded in the outer membrane but loses contact with peptidoglycan (PG) (40). KJDOmpA 299-356 exhibits decreased conjugation ability and swimming motility (39), as well as increased susceptibility to b-lactams (40). Transcriptome analysis revealed that the expression levels of the three s factors, rpoN, rpoP, and rpoE, are significantly altered in KJDOmpA 299-356 (39,40). Notably, rpoN downregulation is the key factor contributing to the swimming defect in KJDOmpA 299-356 , and upregulated s P is involved in the DompA 299-356 -mediated increase in b-lactam susceptibility via the s P -NagA-L1/L2 regulatory circuit (39,40). Here, we aimed to further investigate the effect of DompA 299-356 on oxidative stress tolerance and elucidate the underlying mechanism.
We previously characterized several oxidative stress alleviation systems in S. maltophilia, including enzymatic and nonenzymatic systems (32)(33)(34)(35)(36)(37). To understand the FIG 1 Menadione tolerance of wild-type KJ and its derived mutants. Logarithmic-phase bacterial cells (2 Â 10 5 CFU/mL) were 10-fold serially diluted. Then, 5-mL aliquots of the cells were spotted onto Luria-Bertani (LB) agar plates with and without 40 mg/mL menadione (MD). Bacterial growth was observed after 24 h of incubation at 37°C. All experiments were performed at least thrice, and one was selected as a representative experiment.
Role of OmpO and s N in Oxidative Stress Tolerance Microbiology Spectrum mechanism involved in the DompA 299-356 -mediated decrease in MD tolerance, we analyzed the transcriptome results of KJ and KJDOmpA 299-356 (39), focusing on genes involved in oxidative stress alleviation (see Table S1 in the supplemental material). A change in KJ and KJDOmpA 299-356 gene expression greater than 3-fold was considered statistically significant. Of the 27 genes analyzed, none exhibited significant alterations in transcript levels (Table S2), suggesting that an unidentified determinant is responsible for the DompA 299-356 -mediated decrease in MD tolerance. OmpO (Smlt0387) expression is downregulated in KJDOmpA 299-356 . To further identify the putative candidates responsible for the DompA 299-356 -mediated decrease in MD tolerance, we rechecked the transcriptome results (39), focusing on the top five upregulated and downregulated genes. Smlt0387 was highly expressed in wild-type KJ and was the most downregulated gene (approximately a 203-fold decrease in expression levels) in KJDOmpA 299-356 (Table 1). Smlt0387 is annotated as a hypothetical protein in several sequenced S. maltophilia genomes. Based on the findings of this study, we designated Smlt0387 as OmpO. OmpO encodes a 190-amino acid (aa) protein. Subcellular location prediction (https://www.psort.org/psortb/) and signal peptide prediction (https://services.healthtech.dtu.dk/service.php?SignalP) indicated that OmpO was an OMP with a 19-aa signal peptide.
To verify whether OmpO is indeed an OMP, the OMP profiles of wild-type KJ and KJDOmpO, an ompO isogenic mutant, were analyzed via sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). KJDOmpO lacked a protein band that was present in KJ (band A in Fig. 2). This band was excised from the gel and characterized using liquid chromatography-tandem mass spectrometry (LC-MS/MS). The LC-MS/MS results correlated band A with Smlt0387 (OmpO). The expected mature OmpO protein had an expected molecular weight of 19.0 kDa, which matches the location of band A in the gel (Fig. 2). Furthermore, the fragmentation patterns of band A showed the absence of the predicted signal peptide (1 to 19 aa residues) (Table S3), confirming that OmpO is an OMP with a 19aa signal peptide. Interestingly, another protein with a molecular weight smaller than OmpO was upregulated in KJDOmpO (band B in Fig. 2). This band was characterized by LC-MS/MS, and the results correlated band B with Smlt0184 (Table S4).
To further clarify the biological significance of OmpO, phylogenetic analysis of OmpO and other OMPs, including TolC family and b-barrel proteins (OmpA, OmpX, phospholipase A, general porins, substrate-specific porins, and TonB-dependent transporters), was conducted. OmpO was phylogenetically closely related to OmpX of E. coli (Fig. 3).  showed an MD tolerance level reverted to near the wild-type level, whereas KJDOmpODOmpA 299-356 exhibited low MD tolerance ( Fig. 1), indicating that the extent of MD tolerance is proportional to the expression level of ompO. Therefore, OmpO downregulation in KJDOmpA 299-356 is involved in the DompA 299-356 -mediated decrease in MD tolerance. As OmpO is related to oxidative stress tolerance, we designated Smlt0387 as OmpO in this study.
The involvement of OmpO levels in MD tolerance was revealed in the DompA 299-356 genetic background, and we wondered whether a similar effect could be observed in the wild-type KJ genetic background. Thus, we prepared an ompA deletion construct of KJ cells for an MD tolerance assay. KJDOmpO displayed MD tolerance comparable to that of its parent strain, KJ (Fig. 1).
Upregulated RpoE (Smlt3555) and downregulated rpoN (Smlt2297) levels in KJDOmpA 299-356 are involved in the DompA 299-356 -mediated decrease in MD tolerance. We recently demonstrated that the expression levels of the three s factors are significantly altered (upregulated s P and s E levels and downregulated s N levels) in KJDOmpA 299-356 (40). To assess whether the alteration of these s factor expression levels is associated with the DompA 299-356 -mediated decrease in MD tolerance, KJDRpoEDOmpA 299-356 , KJDRpoPDOmpA 299-356 , and KJL2-RpoNDOmpA 299-356 , a strain of KJDOmpA 299-356 complemented with rpoN expression, were subjected to an MD tolerance assay. The MD tolerance assay revealed that KJDRpoEDOmpA 299-356 and KJL2-RpoNDOmpA 299-356 had an MD tolerance level that was nearly reverted to the wildtype level, whereas KJDRpoPDOmpA 299-356 exhibited MD tolerance comparable to that of KJDOmpA 299-356 (Fig. 1). These results indicated that rpoE upregulation and rpoN downregulation in KJDOmpA 299-356 are involved in the DompA 299-356 -mediated decrease in MD tolerance.
The linkage between rpoE upregulation and the DompA 299-356 -mediated decrease in MD tolerance highly suggested the occurrence of ESR in KJDOmpA 299-356 . The involvement of the rpoE-rseA system in ESR has been characterized in several bacteria (18), including the rpoE-rseA-mucD operon in S. maltophilia (38). RseA functions as an antis E factor; thus, the loss of RseA function induces s E activation, mimicking ESR (18). To assess the relationship between ESR and MD tolerance, the MD tolerance assay was performed with KJDRseA and KJDRseADRpoE. Compared to wild-type KJ, KJDRseA exhibited a decreased MD tolerance and KJDRseADRpoE showed an MD tolerance level that had reverted to the wild-type level (Fig. 1), indicating that an ESR system reduces the resistance of bacteria to MD-induced oxidative stress.
r E activation downregulates ompO expression. The next question was whether s E upregulation in KJDOmpA 299-356 imposes a negative effect on ompO expression. Thus, the ompO transcript levels in KJ, KJDOmpA 299-356 , and KJDRpoEDOmpA 299-356 were compared by quantitative reverse transcription PCR (qRT-PCR). The ompO transcript was downregulated in KJDOmpA 299-356 and reverted to near wild-type levels in KJDRpoEDOmpA 299-356 (Fig. 4A). We further determined the impact of s E activation on ompO expression in the wild-type KJ. Plasmid pOmpO xylE , containing a P ompO -xylE transcriptional fusion construct, was introduced into KJ, KJDRseA, and KJDRseADRpoE to generate KJ(pOmpO xylE ), KJDRseA(pOmpO xylE ), and KJDRseADRpoE(pOmpO xylE ), respectively. KJ(pOmpO xylE ) exhibited significantly higher catechol 2,3-dioxygenase (C23O) activity than the vector-only control strain, KJ(pRKXylE) (Fig. 4), indicating that the ompO gene is highly expressed in logarithmic-phase KJ cells, consistent with the transcriptome results (Table 1). KJDRseA(pOmpO xylE ) showed lower C23O activity than KJ Role of OmpO and s N in Oxidative Stress Tolerance Microbiology Spectrum (pOmpO xylE ). However, C23O activity was comparable in KJDRseADRpoE(pOmpO xylE ) and KJ(pOmpO xylE ) (Fig. 4B), indicating that s E activation has a negative effect on ompO expression. A similar strategy was used to investigate the regulatory effect of s N on ompO expression. The rpoN expression levels of KJ cells under our test condition were first verified by RT-PCR (data not shown). C23O activity was then determined in KJDRpoN (pOmpO xylE ). The results demonstrated that rpoN barely affected the expression level of the ompO gene (Fig. 4B).
r E activation negatively regulates r N expression. The involvement of s E and s N in the DompA 299-356 -mediated decrease in MD tolerance was previously established. Next, we   (Fig. 5A). We also observed that the rpoN transcript level in KJDRpoEDOmpA 299-356 had reverted to a level higher than the level in wild-type KJ (Fig. 5A). Next, we used a P rpoN -xylE transcriptional fusion construct (pRpoN xylE ) to investigate the role of s E activation in rpoN expression in wild-type KJ. Compared to wild-type KJ, KJDRpoE(pRpoN xylE ) displayed increased C23O activity (Fig. 5B), in turn indicating that, under a s E -nonactivated condition, free-form s E exists and exerts a negative impact on rpoN expression. We also observed that C23O activity was lower in KJDRseA(pRpoN xylE ) than wild-type KJ. Furthermore, the C23O activity of KJDRseADRpoE(pRpoN xylE ) had reverted to a level even higher than that of wild-type KJ (Fig. 5B), indicating that s E activation attenuates rpoN expression. Collectively, the rpoN transcript level was inversely proportional to the free-form s E level.
To further investigate the correlation between the rpoN transcript level and Role of OmpO and s N in Oxidative Stress Tolerance Microbiology Spectrum swimming motility, the swimming motilities of KJ, KJDRpoE, KJDRseA, and KJDRseADRpoE were evaluated. The swimming motility of KJDRpoE was slightly, but not significantly, higher than that of wild-type KJ (Fig. 5C). However, KJDRseA exhibited decreased swimming motility, and rpoE deletion from the chromosome of KJDRseA reverted the swimming motility to the wild-type level (Fig. 5C). Based on these results, we concluded that s E activation-mediated rpoN downregulation results in compromised swimming motility. Next, we investigated whether rpoN regulates rpoE expression. The plasmid pRpoE xylE , a P rpoE -xylE transcriptional fusion construct, was transfected into KJ and KJDRpoN to evaluate C23O activity. KJDRpoN(pRpoE xylE ) and KJ(pRpoE xylE ) exhibited comparable C23O activities (Fig. 5D).

DISCUSSION
ECF s factors are crucial in bacterial signaling networks, as they allow bacteria to recognize external signals. ECF s factors remain in an inactive state under normal conditions, via either nonexpression or functional restriction by the anti-s factor. In response to stress signals, ECF s factors are activated by upregulating their expression or sequestering them from the anti-s factor. Free ECF s factors recruit the associated RNAP core enzyme, drive the RNAP holoenzyme to bind to specific promoters, and induce gene expression. In the best-known ECF-mediated regulatory circuits, the genes responsible for stress alleviation are generally not expressed under normal conditions, but are upregulated under stressed conditions via ECF s factor-mediated transcription. In this study, we revealed a particular regulatory circuit mediated by s E . Transcriptome and promoter assays (Table 1; Fig. 4B) revealed that ompO is intrinsically expressed. Intrinsic expression of ompO should be driven by the housekeeping s factor, s D . DompA 299-356 -mediated stress induced the upregulation of s E , which drove ompO expression. s E -mediated ompO expression levels may be lower than those mediated by s D . Therefore, ompO downregulation in KJDOmpA 299-356 may be due to the s factor switch from s D to s E .
Based on their structures, OMPs can be mainly classified into two types, namely, the TolC family and b-barrel proteins. The TolC family comprises proteins with an a-helical trans-periplasmic tunnel embedded in the outer membrane via a contiguous b-barrel channel. TolC-like proteins, which are inner membrane-associated periplasmic proteins, usually assemble with integral inner membrane proteins to form a tripartite efflux pump, which is involved in chemical export (41). b-Barrel OMPs, which include OmpA, OmpX, phospholipase A, general porins, substrate-specific porins, and TonB-dependent transporters (42), generally function as channels for the influx or efflux of hydrophilic molecules. Of these b-barrel proteins, OmpA is particularly important, as it tightly attaches the outer membrane to the PG layer via its periplasmic domain (43) and, thus, plays a critical role in envelope stability. Unlike the OmpA protein, most b-barrel OMPs, such as OmpO investigated in this study, are devoid of the periplasmic domain; therefore, most b-barrel OMPs participate in molecule transportation.
A correlation between OMP deletion and a decrease in oxidative stress tolerance has been reported in some bacteria, but most such instances have involved TolC-like OMPs, such as those in Salmonella enterica, Acinetobacter baumannii, Cronobacter sakazakii, and Pseudomonas syringae (44)(45)(46)(47). This is likely due to the deleted TolC-like OMPs being members of the tripartite efflux pumps, which extrude the toxic compounds generated during oxidative stress. Therefore, TolC-like OMP deletion compromises oxidative stress tolerance due to the accumulation of toxic compounds. In this study, we found that KJDOmpA 299-356 is more susceptible to oxidative stress than its parental strain, KJ. These results seem to indicate that OmpA is an outlet for oxidative-stress-mediating toxic compounds. However, the exact determinant leading to the decrease in oxidative stress tolerance in KJDOmpA 299-356 was the downregulation of the expression of another OMP, OmpO, supporting that the biological significance of OmpA is envelope stabilization, rather than molecular transport. OmpA defects cause ESR and s E activation. OmpO, an intrinsically highly expressed OMP, is a member of the s E regulon, and its expression is downregulated upon s E activation. We further established a novel regulatory circuit of DompA 299-356 -s E -ompO involved in the DompA 299-356 -mediated decrease in oxidative stress tolerance.
The association between ECF and oxidative stress adaptation has been widely reported in several microorganisms, such as Porphyromonas gingivalis (s E ), Bacteroides fragilis (EcfO), Bradyrhizobium japonicum (CarQ), and Shewanella oneidensis (s E2 ) (13,(48)(49)(50). ECF acts as a positive regulator that protects bacteria from oxidative stress. Under conditions of oxidative stress, ECF is activated and induces ECF regulon expression to address the oxidative stress. Therefore, the loss of ECF function is generally linked to a decrease in oxidative stress tolerance. Here, we demonstrated that DompA 299-356 -mediated upregulation of rpoE expression exerted a negative effect on oxidative stress tolerance mediated by the downregulation of rpoN and ompO. OmpO is a novel OMP that is associated with oxidative stress tolerance. RpoN is a s 54 family s factor involved in nitrogen assimilation, flagellar motility, type III and VI secretion systems, biofilm formation, and environmental adaptation (51)(52)(53)(54). The role of rpoN in oxidative stress adaptation has rarely been reported, except in Labrenzia aggregata and Edwardsiella tarda (54,55). Here, we found that activated s E had a negative impact on the expression of rpoN. Interplay among s factors is important and widely studied. For example, s E regulates rpoN, rpoH, and rpoD expression levels in E. coli (24,56,57), s T controls rpoU and rpoR expression levels in Caulobacter crescentus (15), and rpoH II expression in Rhodobacter sphaeroides is s E -dependent (58).
The truncated OmpA protein expressed by KJDOmpA 299-356 can be stably embedded in the outer membrane but fails to contact with PG (40), which endows KJDOmpA 299-356 with pleiotropic defects, including compromised swimming motility (39), increased b-lactam susceptibility (40), and decreased oxidative stress tolerance. Based on our previous study (40) and the novel findings of this study, we conclude that KJDOmpA 299-356 experiences PG and envelope stress, leading to upregulation of the s factors, rpoP and rpoE, respectively. s P upregulates the expression of nagA, which decreases L1/L2 expression levels and increases bacterial susceptibility to b-lactams (40). In contrast, s E -mediated ESR downregulates rpoN and ompO expression. OmpO downregulation results in a decrease in oxidative stress tolerance. Furthermore, downregulation of s N decreases swimming motility (39) and oxidative stress tolerance. Multidrug resistance is a challenging issue for the treatment of S. maltophilia infections, and OmpA represents the most abundant OMP in S. maltophilia. Blocking the interaction between OmpA and PG may present an alternative strategy for S. maltophilia infection control.

MATERIALS AND METHODS
Bacterial strains, plasmids, and primers. The bacterial strains, plasmids, and primers used in this study are listed in Table S5.
Outer membrane protein preparation and SDS-PAGE. The purification of outer membrane proteins was carried out as described previously (40). The OMPs were separated by discontinuous SDS-PAGE with a 5% stacking gel and a 15% separating gel. Bands were visualized by staining with 0.1% Coomassie brilliant blue R250 (Bio-Rad) and de-staining with 40% methanol/10% glacial acetic acid.
Construction of deletion mutant KJDOmpO. The deletion mutants were obtained using the double homologous recombination method as described previously (59). The upstream and downstream DNA fragments of ompO were gotten by PCR using the primer pairs OmpON-F/OmpON-R and OmpOC-F/OmpOC-R (Table S5). Next, the two PCR amplicons were subsequently cloned into pEX18Tc to yield plasmid pDOmpO (Table S5). Plasmid pDOmpO was transferred into S. maltophilia KJ by conjugation. The plasmid's conjugation, the transconjugants' selection, and the mutant's confirmation were carried out as described previously (59).
Construction of the P ompO -xylE transcriptional fusion plasmid, pOmpO xylE . The DNA fragment containing the promoter region of the ompO gene was obtained by PCR using primer pair OmpON-F/ OmpON-R (Table S5). The 365-bp PCR amplicon was cloned into pRKxylE, a xylE reporter plasmid, yielding pOmpO xylE . Catechol 2,3-dioxygenase activity determination. Catechol 2,3-dioxygenase (C23O) is encoded by the xylE gene. The C23O activity was measured using 100 mM catechol as the substrate, as described previously (60). The hydrolysis rate of catechol was calculated using 44,000 M 21 cm 21 as the extinction coefficient. One unit of C23O activity (U) was defined as the enzyme amount that converts 1 nmol of substrate per minute. The specific activity was expressed as U/optical density at 450 nm (OD 450 ).
Statistical analysis. Student's t test was used for comparison of means between the groups. The Bonferroni correction method was applied to adjust the P values.
Data availability. The RNA-seq data have been deposited in GenBank under BioProject accession number PRJNA876818.

SUPPLEMENTAL MATERIAL
Supplemental material is available online only. SUPPLEMENTAL FILE 1, DOCX file, 5.9 MB.