LsrR, the effector of AI-2 quorum sensing, is vital for the H2O2 stress response in mammary pathogenic Escherichia coli

Mammary pathogenic Escherichia coli (MPEC) is an important causative agent of mastitis in dairy cows that results in reduced milk quality and production, and is responsible for severe economic losses in the dairy industry worldwide. Oxidative stress, as an imbalance between reactive oxygen species (ROS) and antioxidants, is a stress factor that is common in most bacterial habitats. The presence of ROS can damage cellular sites, including iron-sulfur clusters, cysteine and methionine protein residues, and DNA, and may cause bacterial cell death. Previous studies have reported that Autoinducer 2 (AI-2) can regulate E. coli antibiotic resistance and pathogenicity by mediating the intracellular receptor protein LsrR. This study explored the regulatory mechanism of LsrR on the H2O2 stress response in MPEC, showing that the transcript levels of lsrR significantly decreased under H2O2 stress conditions. The survival cell count of lsrR mutant XW10/pSTV28 was increased about 3080-fold when compared with that of the wild-type WT/pSTV28 in the presence of H2O2 and overexpression of lsrR (XW10/pUClsrR) resulted in a decrease in bacterial survival rates under these conditions. The β-galactosidase reporter assays showed that mutation of lsrR led to a remarkable increase in expression of the promoters of ahpCF, katG and oxyR, while lsrR-overexpressing significantly reduced the expression of ahpCF and katG. The electrophoretic mobility shift assays confirmed that LsrR could directly bind to the promoter regions of ahpCF and katG. These results revealed the important role played by LsrR in the oxidative stress response of MPEC.


Introduction
Escherichia coli is one of the main pathogenic bacteria causing clinical mastitis [1,2]. Mammary pathogenic E. coli (MPEC) is a class of extraintestinal pathogenic E. coli that usually adheres to and infects dairy cow mammary epithelial cells together with other pathogens, forming biofilms. The biofilm can help pathogens to evade the host immune system and continue to multiply in the mammary gland, leading to persistent intramammary infections [3][4][5]. As such, research on the role of biofilms in the pathogenesis of mastitis has become important.
Previous studies showed that bacteria express a series of stress proteins to protect themselves from environmental stimuli and to resist adverse environmental pressures such as antibiotics, oxidative stress, acid-base, and osmotic pressure [6][7][8][9][10]. It is generally understood that bacterial infection can trigger innate immune response [11,12]. Once the pathogenic bacteria infect the host, they are phagocytosed by the host immune cells and macrophages and heterophile cells may produce oxidative bursts in response to pathogens [13,14]. The rapid production of reactive oxygen species (ROS) including superoxide anions (O 2 − ), hydrogen peroxide (H 2 O 2 ) and the highly reactive hydroxyl radical (·OH) can damage many cellular sites, including iron-sulfur clusters, cysteine and methionine protein residues and DNA, and may eventually cause cell senescence and death [15][16][17]. H 2 O 2 is a widely used bactericide for inactivating foodborne pathogens. Over time, a variety of E. coli regulatory pathways has evolved to adapt to H 2 O 2 stress including synthesizing catalases (encoded by katG and katE), alkyl hydroperoxide reductase (Ahp, encoded by ahpCF), and stress protein YciF (encoded by yciF). A range of transcriptional regulators including the transcriptional regulator OxyR and the sigma factor RpoS could regulate the synthesis of these enzymes [6,8,[18][19][20]. Previous studies showed that E. coli resistance to oxidative stress affects both survival and pathogenic potential in the host [21,22]. However, the detailed molecular mechanism of bacterial adapting to H 2 O 2 stress needs to be further explored. Autoinducer 2 (AI-2) is produced by many Gram-negative and Gram-positive bacteria and is considered to be a quorum sensing (QS) signaling molecule involved in interspecies communication [23]. Previous studies have reported that the AI-2 QS system is involved in the bacterial response to oxidative stress [24][25][26]. The absence of luxS (involved in the synthesis of AI-2) in Streptococcus suis significantly enhanced its tolerance to H 2 O 2 , and the mutation of luxS in Deinococcus radiodurans resulted in a significantly higher susceptibility to H 2 O 2 than wildtype strains [25,26]. In E. coli, the process of AI-2 uptake involves an ATP-binding transporter complex encoding by the lsrACDB operon. The expression of the lsr operon was regulated by LsrR, a DNA-binding repressor and LsrK, a cognate signal kinase, whose genes (LsrRK) are located immediately upstream of the lsr operon and divergently transcribed. The LsrR represses the expression of the lsr operon and its own lsrRK operon by binding to the promoters, while phospho-AI-2 can release LsrR repression and then activating lsr operon expression [27][28][29]. The role of LsrR as a transcriptional regulator mediating QS signal AI-2-related cellular functions has been identified in previous studies and its significant influence on gene expression control has been revealed. According to a previous study, 146 genes were significantly affected by LsrR deletion [30], but whether there is a relationship between LsrR and the regulation of bacterial oxidative stress response has not yet been reported.
In the present study, we demonstrated that the survival ability of lsrR mutant under H 2 O 2 stress was significantly increased compared to that of the wild type. The β-galactosidase reporter assays indicated that LsrR had an obvious inhibitory effect on the expression of ahpCF, katG, and oxyR. Electrophoretic mobility shift assays (EMSA) confirmed that LsrR inhibited the expression of ahpcF and katG by direct binding to their promoter regions. This study is therefore helpful to understand the bacterial response to host-derived ROS in MPEC, and might provide potential drug targets for the treatment and prevention of E. coli infection.

Bacterial strains, plasmids, and culture conditions
All the strains and plasmids used in this study are listed in Table 1. The E. coli strain DCM5 was isolated from milk samples in dairy cows with mastitis and was identified by 16S rDNA sequencing. The E. coli cultures were routinely grown at 37 °C in Luria-Bertani (LB) broth (Oxoid, Basingstoke, UK) or on LB agar plates containing 1.5% agar (Oxoid). When needed, antibiotics (Sangon Biotech, Shanghai, China) were used at the following final concentrations: ampicillin (100 μg/mL), chloramphenicol (15 μg/mL), and kanamycin (50 μg/mL).

General DNA manipulation
Genomic DNA from E. coli DCM5 was prepared according to the instructions of TIANamp Bacteria DNA Kit (TianGen Biotech, Bei Jing, China). Plasmid DNA was extracted using a plasmid extraction kit (Sangon Biotech, Shanghai, China). PCR testing was carried out using Taq or PrimeSTAR ® Max DNA Polymerase (Takara Bio Inc., Dalian, China). The PCR products and DNA restriction fragments were purified using a gel purification kit (Sangon Biotech, Shanghai, China). DNA restriction endonuclease (Thermo Fisher Scientific, Waltham, MA, USA) digestion and DNA ligation (Thermo Fisher Scientific) were performed by standard methods. Primer premier 5.0 software was used for sequence analysis and primer design. The primer sequences used are listed in Table 2.

Construction of the lsrR-deficient mutant
Homologous recombination was used for deletion of lsrR from the chromosome of DCM5 based on λ red recombinase system [31]. Briefly, using pKD3 plasmid as a template, the fragment containing the chloramphenicolresistance cassette (cat) inserted within the lsrR gene was amplified with primers DCM5-lsrR-f and DCM5-lsrR-r and then transformed into DCM5 competent cells containing lambda Red recombinase expressed by pKD46 plasmid using electroporation. After electroporation, the cells were added immediately to 900 µL super optimal broth with catabolite repression broth (SOC) without antibiotics, incubated at 37 °C for 60 min, and then mutants were selected by plating cultures on LB plates with 15 μg/mL chloramphenicol. Following incubation for 24 h, the lsrR mutant was selected by PCR amplification using the primers check-lsrR-f and check-lsrR-r. Plasmid pCP20 was subsequently transformed into the   mutant strain to delete the Cm cassette. The Cm-sensitive mutant strain was confirmed by PCR amplification using the primers check-lsrR-f and check-lsrR-r, and the PCR products were further confirmed by DNA sequencing, with the mutant strain was designated as DCM5ΔlsrR (XW10).

Complementation of the lsrR mutant
To construct the complemented plasmid pSTV28-lsrR, the lsrR gene and its promoter region were amplified from the DCM5 strain using the primers lsrR-KpnI-f and lsrR-BamHI-r, and then the PCR products were gelpurified and ligated into the BamHI and KpnI restriction sites of the plasmid pSTV28. The recombinant plasmid pSTV28-lsrR was extracted and further confirmed by PCR using primers M13-f and M13-r, and DNA sequencing. To obtain the complemented strain XW10/pClsrR, the plasmid pSTV28-lsrR was electroporated into the mutant strain XW10. The wild-type strain (WT) and the mutant strain XW10 were transformed with the empty vector pSTV28 to obtain WT/pSTV28 and XW10/ pSTV28, respectively.

Construction of the lsrR overexpressing strain
The strain overexpressing lsrR was constructed as previously described [32]. The lsrR gene was amplified from the DCM5 strain using the primers lsrR-KpnI-f and lsrR-EcoRI-r, and then the PCR products were gel-purified and ligated into the KpnI and EcoRI restriction sites of the plasmid pUC19 to construct the overexpressing plasmid pUC19-lsrR. Subsequently, the recombinant plasmid pUC19-lsrR was extracted and further confirmed by PCR using primers M13-f and M13-r, and DNA sequencing. The purified recombinant plasmid pUC19-lsrR and the control empty vector pUC19 were electroporated into the wild-type strain WT to obtain strains WT/pUClsrR and WT/pUC19, respectively.

Bacterial growth curves
All strains were cultured overnight in fresh LB broth. Then each sample was diluted to an OD 600 of approximately 0.03 in 100 mL of fresh LB broth with 15 μg/mL chloramphenicol or 100 μg/mL ampicillin and grown at 37 °C for 28 h with shaking. The OD 600 was detected at 2 h intervals using a UV/Vis spectrophotometer (Thermo Scientific, Pittsburgh, PA, USA), and the growth curves of these strains were established. The experiment was repeated three times and all samples were measured in triplicate.

H 2 O 2 stress assays
The overnight cultures were diluted to an OD 600 of approximately 0.03 in 3 mL of fresh LB broth with 15 μg/mL chloramphenicol or 100 μg/mL ampicillin, and incubated in culture tubes at 37 °C for 2 h with shaking. Then, 6 μL of 30% H 2 O 2 was added in each experimental group cultures, and the cells continued to be incubated at 37 °C for 1 h with shaking. All cultures from the experimental group and control group (without addition of H 2 O 2 ) were then diluted by serial ten-fold dilution with LB medium, and 3 μL of each dilution was dropped onto the LB agar plates and cultured at 37 °C overnight. At the same time, to perform the viable colony-forming units (CFUs) assays, 100 μL of each dilution was spread onto the LB agar plates and cultured at 37 °C overnight. The colonies on each plate were counted. The results were compared between the test groups and the control groups and all experiments were repeated three times.

β-galactosidase assays
The promoters of oxyR, katG, katE, rpoS, yciF, and ahpCF were amplified and cloned into the promoterless lacZ plasmid pRCL by restriction enzyme digestion and DNA ligation to construct the transcriptional lacZ fusion reporter plasmid. To avoid the lacZ gene on the chromosome of the strain from affecting the experimental results, the lacZ genes from the WT strain and the lsrR-deficient strain were mutated using homologous recombination methods based on λ red recombinase system, and then the constructed lacZ fusion reporter plasmids were introduced into strain WT/ΔlacZ and XW10/ΔlacZ, respectively. The E. coli cells containing the lacZ fusion reporter plasmid were cultured in 100 mL LB broth containing a final concentration of 15 μg/mL chloramphenicol for the specified time. After cultivation, cells were collected and diluted in Z-buffer containing 16.1 g/L Na 2 HPO 4 , 5.50 g/L NaH 2 PO4, 0.75 g/L KCl, and 0.246 g/L MgSO4 to 1 mL and assayed for β-galactosidase activity using ortho-Nitrophenylβ-galactoside (ONPG) as a substrate. The following equation was used to calculate units of enzyme activity: Miller Units: [OD 420 × 1000)/(OD 600 × Volume (mL) × Time (min)]. The experiment was repeated three times independently.

Total RNA isolation, cDNA generation, and real-time reverse transcription-quantitative PCR processing
Real-time reverse transcription-quantitative PCR (RT-qPCR) assays were performed according to a previous study [32]. cDNA abundance was used to normalize the quantity of the target genes. All the real-time RT-qPCR assays were repeated at least three times.

Expression and purification of the LsrR protein
The C-terminal His6-tag LsrR protein was cloned and purified according to a previous study [27]. The open reading frame of the lsrR gene of E. coli DCM5 was amplified with primers lsrR-NcoI-f and lsrR-XhoI-r from DCM5 genomic DNA, and then cloned into the pET28a (+) vector. The recombinant plasmid pET-lsrR was confirmed by PCR using primers T7-f and T7-r, and further confirmed by DNA sequencing. The pET-lsrR vector was then transferred into E. coli BL21(DE3) by a chemical transformation method. The LsrR protein was induced to express in E. coli BL21(DE3) cells by adding isopropyl β-D-thiogalactopyranoside (IPTG) at a final concentration of 0.5 mM. The purification of His6-tagged LsrR fusion protein was performed with a HisTrap high-performance column as previously described [32]. The purified LsrR protein was stored in 10% glycerol at −80 °C until use. The LsrR protein purity was determined by SDS-PAGE, and the protein concentration was measured by an enhanced BCA protein assay kit (Beyotime, Shanghai, China).

Mutation of ahpCF and katG promoter sites and enzyme activity detection
Mutation of binding sequences of the ahpCF and katG promoter was performed according to the previous method [27]. There were 6-bp nucleotide bases (AAC AAT ) in the katG promoter region, and 6-bp nucleotide bases (AAA ACT ) in the ahpCF promoter region, which may be necessary for LsrR binding. High ratios of A and T within these motifs also suggest that LsrR might preferentially bind to AT-rich sequences [27] so these 6-bp nucleotide bases in plasmids pRCL-p ahpCF and pRCL-p katG were replaced by the GCG CGC sequence to obtain plasmids pRCL-p ah-pCFM6 and pRCL-p katGM6 ( Table 1). The mutations were performed by using Mut Express II Fast Mutagenesis Kit V2 (Vazyme Biotech, Nanjing, China). A DNA fragment about 6.0 kb, containing the full length of the plasmid except the 6-bp nucleotide bases, was amplified by PCR using the plasmid pRCL-p ahpCF or pRCL-p katG as the template and M6-p-ahpCF-f/M6-p-ahpCF-r or M6-p-katG-f/M6-p-katG-r as the primers. The remaining steps were performed according to the Mut Express II Fast Mutagenesis Kit V2 instructions. The constructed mutant plasmid was identified by DNA sequencing then the mutant reporter plasmids were introduced into strain WT/ΔlacZ. The strains WTΔlacZ/pRCL-p ahpCFM6 and WTΔlacZ/pRCLp katGM6 were obtained by screening and the expression activity of the mutant promoter was detected by β-galactosidase assays.

Statistical analyses
Statistical analyses were conducted using the Graph-Pad Primer 8.0 (GraphPad Software Inc., GraphPad Prism 8.0.1.244, San Diego, CA, USA, 2018) using a one-way ANOVA method. The test results are shown as the mean ± SD. A paired t test was used for statistical comparisons between groups. The level of statistical significance was set at a P value of ≤ 0.05.

Identification of lsrR mutant and complementary strains of DCM5
The lsrR gene was knocked out using homologous recombination to obtain the lsrR deletion mutant. The schematic diagram of the strategy for deleting the lsrR gene in DCM5 and the primers used for confirmation of the lsrR deletion are shown in Figure 1A. The mutant and complementary strains of lsrR were confirmed by PCR (Figure 1B). A 1660 bp product was amplified from the wild type strain WT/pSTV28, while 1330 bp products were respectively amplified from the mutant strain XW10/ pSTV28 and the complementary strain XW10/pClsrR by using primers check-lsrR-f and check-lsrR-r (Lane 1-3). The plasmid pSTV28 and the complementary plasmid pClsrR were also confirmed by PCR ( Figure 1B, Lane 4-6). The PCR products of 100 bp were amplified from strains WT/pSTV28 and XW10/pSTV28 using primers M13-f/M13-r, and a 1094 bp product was amplified from strain XW10/pClsrR.

lsrR deletion and overexpression did not affect strains growth
To ensure that the growth conditions of the mutant XW10 strain and its parent strain were consistent with that of the complementary strain, the empty vector pSTV28 was transformed into the WT and XW10 strains. Growth curve assays were performed and results showed that the growth curves of XW10/pSTV28 and XW10/pClsrR were like that of WT/pSTV28 when cells were cultured in LB broth with 15 μg/mL chloramphenicol ( Figure 1C). Likewise, to ensure that the growth condition of the WT strain was consistent with that of the lsrR overexpression strain, the empty vector pUC19 was transformed into the WT strain. The growth curve of strain WT/pUClsrR in LB broth with 100 μg/mL ampicillin was like that of strain WT/pUC19 as shown in Figure 1D.

Deletion of lsrR increased bacterial survival ability under the H 2 O 2 stress
With the tolerance of pathogenic bacteria to H 2 O 2 stress being increased, the infecting ability of pathogenic bacteria to the host is also increased. To evaluate the effect of lsrR on the survival of strain DCM5 in the presence of H 2 O 2 , the survival ability of the lsrRdeficient mutant strain XW10/pSTV28 was compared with that of the wild-type strain WT/pSTV28 and the complementary strain XW10/pClsrR. The results showed that there was no difference in the number of colonies grown on the plates of each group in absence of H 2 O 2 (Figure 2A). However, the tolerance of strain XW10/pSTV28 to H 2 O 2 stress was increased when compared with that of strain WT/pSTV28 after treating with H 2 O 2 for 1 h, and the tolerance was partially restored in strain XW10/pClsrR ( Figure 2B). The CFU assays were performed to further analyze the survival rates of strains WT/pSTV28, XW10/pSTV28, and XW10/pClsrR, and the colonies of hydrogen peroxide treatment group and control group on the plates were counted and compared after cultivating in LB broth with 15 μg/mL chloramphenicol overnight. In the absence of H 2 O 2 , there was no difference in the survival rate between the groups as shown in Figure 2C, D. However, with the addition of H 2 O 2 , the survival rate of strain XW10/PSTV28 was increased 3080-fold when compared with that of strain WT/PSTV28, and the survival rate was partially recovered in strain XW10/ PClsrR. These results suggested that LsrR played a vital role in the bacterial oxidative stress response.

Figure 2
Analysis of the H 2 O 2 stress response of strains WT/pSTV28, XW10/pSTV28, and XW10/pClsrR. The strains WT/pSTV28, XW10/ pSTV28 and XW10/pClsrR with an initial OD 600 of 0.03 were inoculated into LB broth containing 15 μg/mL chloramphenicol and incubated for 2 h. Then the cultures of the control group (A) and the experimental group (treated with 6 μL 30% hydrogen peroxide for 1 h) (B) were ten-fold serially diluted, and 3 μL of each dilution was dropped onto LB agar plates and cultured overnight. The above dilutions of the control group (C) and the experimental group (D) were spread onto the LB agar plates. After incubation overnight, the colony forming units were analyzed. ( * * * * P < 0.0001).

Overexpression of lsrR decreased bacterial survival ability under the H 2 O 2 stress
To further evaluate the effect of lsrR on the survival of strain DCM5 in the presence of H 2 O 2 , the survival ability of the lsrR overexpression strain WT/pUClsrR and its parental strain WT/pUC19 were compared under the H 2 O 2 stress condition. The results showed that there was no difference in the number of colonies grown on the plates of each group in absence of H 2 O 2 ( Figure 3A), but strain WT/pUClsrR showed reduced tolerance to H 2 O 2 stress compared with WT/pUC19 after being treated with H 2 O 2 for 1 h ( Figure 3B). To further analyze the survival rates of strains WT/pUC19 and WT/pUClsrR, CFU assays were performed. There was no difference in the survival rate between the groups in the absence of H 2 O 2 as shown in Figures 3C, D. In the presence of H 2 O 2 , the survival rate of WT/pUClsrR was decreased approximately 21-fold when compared with that of WT/pUC19. This data further supported the important role of LsrR on the oxidative stress response of MPEC.

Deletion of lsrR increased the expression of oxyR, ahpCF, and katG
To assess whether the regulatory effects of LsrR on oxyR, rpoS, ahpCF, katE, yciF, and katG were transcriptionally mediated, a system of β-galactosidase report plasmids was constructed to measure expression levels of the promoters of target genes. The data showed that the deletion of lsrR had no effect on the transcriptional activities of the katE, yciF, and rpoS promoters (Figures 4A-C). Compared with the WT strain, the deletion of lsrR significantly increased the transcription activity of the promoters of oxyR (1.95-, 1.71-, and 1.88-fold at 4, 8 and  16 h, respectively), katG (1.98-, 1.73-, and 1.58-fold at 4,  8, and 12 h, respectively), and ahpCF (7.0-, 3.94-, 2.66and 2.76-fold at 4, 8, 12 and 16 h, respectively) seen in Figures 4D-F. Taken together, these results suggested that LsrR decreases the bacterial ability to respond to hydrogen peroxide stress by downregulating the expression of hydrogen peroxide stress-associated genes oxyR, katG and ahpCF.

H 2 O 2 stress decreased the expression of lsrR, and overexpression of lsrR decreased the expression of ahpCF, and katG
To investigate whether the expression level of lsrR in ECM5 was affected by H 2 O 2 stress, 0.5 μL 30% H 2 O 2 was added and the transcript level of lsrR decreased by 7.5-fold as shown in Figure 5A. The RT-qPCR experiments were performed to determine whether Figure 3 Analysis of the H 2 O 2 stress response of strains WT/pUC19 and WT/pUClsrR. A The strains WT/pUC19 and WT/pUClsrR with an initial OD 600 of 0.03 were inoculated into LB broth containing 100 μg/mL ampicillin and incubated for 2 h. Then the cultures of the control group (A) and the experimental group (treated with 6 μL 30% hydrogen peroxide for 1 h) (B) were ten-fold serially diluted, and 3 μL of each dilution was dropped onto LB agar plates and cultured overnight. The above dilutions of the control group (C) and the experimental group (D) were spread onto the LB agar plates. After incubation overnight, the colony forming units were analyzed. ( * * * P < 0.001).
overexpression of lsrR affected the transcript levels of the catalase encoding gene katG and alkyl hydroperoxide reductase encoding gene ahpCF. The data showed that the transcript level of ahpCF and katG of the lsrRoverexpressing strain WT/pUClsrR was decreased by 2.7-and 2.9-fold when compared with its parental strain WT/pUC19 as seen in Figures 5B, C. The expression level of lsrR was inhibited by H 2 O 2 stress and lsrR reduced the expression of ahpCF and katG, so it can be inferred that MPEC can adapt to the H 2 O 2 stress conditions by mediating LsrR expression.

LsrR binding to ahpCF and katG promoters
Our previous study showed that lsrR and lsrA promoter regions have two LsrR binding boxes, AAA ACT and AAA ACT GAA in the p-lsrA-box and AAC AAT and AAG ATT TAA in the p-lsrR-box, and the motifs contain high ratios of A and T [27]. On this basis, it was speculated that a putative binding site of LsrR existed in the promoter regions of ahpCF and katG genes ( Figure 6A) and the results suggested that LsrR might repress the two genes by direct binding to their promoter regions. According to sequence alignment, the putative binding site of LsrR in the oxyR promoter region was not found. To verify the above hypothesis, electrophoretic mobility shift assays were performed to confirm whether LsrR binds to the oxyR, ahpCF and katG promoter regions. A positive control assay confirmed that LsrR could bind to its own promoter region ( Figure 6B). As shown in Figure 6C, D, clearly shifted bands of protein-DNA complex were detected at LsrR concentrations of 4, 8 and 12 mM, and the intensity of the shifted band was enhanced as the amount of LsrR was increased. The results confirmed that LsrR could specifically bind to the promoter regions of ahpCF and katG, indicating that LsrR could directly regulate the transcription of ahpCF and katG. However, the shifted band of the protein-DNA complex was not detected when the promoter region of oxyR was used as a probe ( Figure 6E), suggesting that LsrR cannot bind to the oxyR promoter regions. This suggested that LsrR might regulate the expression of oxyR through several indirect pathways and indicated that it had a complicated regulatory pattern in the regulation of bacterial oxidative stress response.

The β-galactosidase assays of the mutated promoters of ahpCF and katG
Based on sequence alignment, it was speculated that a putative binding site of LsrR existed in both promoter regions of ahpCF and katG genes. The EMSA results showed that LsrR could directly bind to the promoter regions of ahpCF and katG, which led to the hypothesis that these two putative sequences may be crucial for LsrR binding. To further verify this, mutant β-galactosidase reporter plasmids pRCL-p ahpCFM6 and pRCL-p katGM6 were constructed and the data in Figure 7A showed that compared with pRCL-p ahpCF , pRCL-p ahpCFM6 was associated with an apparent increase in the expression of lacZ throughout the growth cycle. The expression of lacZ in pRCL-p katGM6 increased only before 4 h compared with pRCL-p katG as shown in Figure 7B, so these results suggested that the sequence AAA ACT in the promoter region of ahpCF was essential for the binding of LsrR in DCM5 strain. However, mutation of AAC AAT in the promoter region of katG did not reduce the binding ability of LsrR throughout the growth cycle.

Discussion
Bacteria can precisely regulate gene expression to promote their survival in various environments or hosts. Bacterial AI-2 quorum sensing is common and is involved in a variety of biological processes, including signal transduction, drug resistance, pathogenicity, motility, and interbacterial competition [23,[33][34][35][36]. In E. coli, AI-2 can regulate some physiological processes through the AI-2 receptor protein LsrR and the AI-2 metabolic pathway. The previous study showed that AI-2 relies on the intracellular receptor LsrR to increase the β-lactam antibiotic resistance of a broad-spectrum β-lactamase-positive E. coli isolated from cows with mastitis [37]. Another subsequent study indicated that exogenous AI-2 can down-regulate the transcription of folate biosynthesis-related genes folA, folC, luxS, metE, and metH through non-LsrR-dependent pathways, increasing avian pathogenic Escherichia coli (APEC) sensitivity to trimethoprime-sulfamethazole [38]. These studies supported the regulatory role of AI-2 or LsrR in bacterial adaption to environmental stimulus. However, whether there is a relationship between LsrR and the regulation of bacterial oxidative stress response has not been reported yet. This study explored the relationship between LsrR and bacterial oxidative stress regulation in MPEC and clarified the regulatory details of LsrR on the oxidative stress-related genes.
Oxidative stress can trigger multiple responses that have evolved as complex regulatory systems in bacterial pathogens, aiding bacteria to cause persistent infections [39][40][41]. A previous study indicated that the LsrR controls invasiveness of Salmonella Typhimurium by regulating Salmonella pathogenicity island-1(SPI-1) and flagella genes [42]. A subsequent study demonstrated that the lsr operon is involved in mediating the virulence of APEC but the regulatory mechanism remains unclear [43]. In Xanthomonas oryzae pv. oryzae, OxyR degrades hydrogen peroxide in rice by regulating the expression of catalase CatB, and then promotes bacterial infectivity [44]. The data from this study showed that the expression of LsrR was affected by H 2 O 2 stress in the environment and that LsrR can directly regulate the expression of oxidative stress response associated genes ahpCF and katG suggesting that MPEC might respond to the H 2 O 2 stress in an LsrR-dependent pathway. Because the ability to respond to oxidative stress of bacteria is very important for its infection process in the host, it was speculated that LsrR may also be associated with the regulation of bacterial pathogenicity in MPEC, so it was important to explore the regulatory role of LsrR on bacterial pathogenicity or other important physiological processes in future work.
The EMSA results showed that LsrR directly binds to the promoter regions of ahpCF and katG, which led to the hypothesis that p ahpCF : AAA ACT and p katG : AAC AAT may be crucial for LsrR binding. To verify this, mutant β-galactosidase enzyme activity reporter plasmids pRCLp ahpCFM6 and pRCL-p katGM6 were constructed and their β-galactosidase activities were measured. The results suggested that the AAA ACT sequence in the promoter region of ahpCF was essential for the binding of LsrR in DCM5 strain, which was consistent with our previous results [27], but the mutation of the AAC AAT sequence in the promoter region of katG did not promote the transcriptional activity of katG promoter throughout the growth cycle. The previous study also has shown that LsrR binding sequences contained high ratios of A and T, but do not need to share a high level of sequence homology [27]. A subsequent study indicated that mutation of nucleotides in the vicinity of p-lsrR-box sequence AAC AAT -AAG ATT TAA also affected the transcriptional activity of the lsrR promoter, which might be associated with binding of CytR repressor [45], therefore it was speculated that the mutation of the binding site in the katG promoter region may cause the recognition and binding by other transcriptional repressors.
According to present and previous studies, a schematic diagram was made to illustrate the regulatory mode of LsrR on the H 2 O 2 stress response of MPEC ( Figure 8). These data showed that LsrR decreased H 2 O 2 stress tolerance ability of MPEC by directly or indirectly inhibiting the transcriptions of H 2 O 2 stress response associated genes including ahpCF, katG, and oxyR. The transcription of lsrR exhibited a significant decrease in response to the H 2 O 2 stress. These results led to the conclusion that MPEC can adapt to H 2 O 2 stress conditions by reducing the transcription of lsrR, which contributed to increased expression of H 2 O 2 stress response associated genes. But how this bacterium regulates the lsrR expression under the H 2 O 2 stress conditions remains obscure and needs to be further explored.