Autoinducer2 affects trimethoprim‐sulfamethoxazole susceptibility in avian pathogenic Escherichia coli dependent on the folate synthesis‐associate pathway

Abstract Avian pathogenic Escherichia coli (APEC) causes airsacculitis, polyserositis, septicemia, and other mainly extraintestinal diseases in chickens, ducks, geese, pigeons, and other avian species, and is responsible for great economic losses in the avian industry. The autoinducer 2 (AI‐2) quorum sensing system is widely present in many species of gram‐negative and gram‐positive bacteria and has been proposed to be involved in interspecies communication. In clinical APEC strains, whether or not AI‐2 affects the expression of antibiotic‐related genes has not been reported. In this study, we have reported that exogenous AI‐2 increase the susceptibility of APEC strains to trimethoprim‐sulfamethoxazole (SXT) in a folate synthesis‐dependent pathway but not in the LsrR‐dependent manner. Our results further explained that exogenous AI‐2 can down regulate the transcription of the folate synthetase encoding genes folA and folC, and the folate synthesis‐related genes luxS, metE, and metH. Gel shift assays confirmed that LsrR, the AI‐2 receptor, did not bind to the promoters of folA and folC, suggesting that exogenous AI‐2 might influence folate metabolism by a feedback inhibition effect but not in the LsrR‐dependent pathway. This study might provide further information in the search for potential drug targets for prophylaxis of avian colibacillosis and for auxiliary antibiotics in the treatment of avian colibacillosis.

A fixed-dose combination of trimethoprim-sulfamethoxazole (SXT) has been widely used as a broad-spectrum antimicrobial agent since the 1960s (Church, Fitzgerald, Walker, Gibb, & Prendergast, 2015;Mcguinness, 1969). The rationale for the combination of trimethoprim with sulfamethoxazole is that each component blocks a different step in the folate biosynthetic pathway (Wormser, Keusch, & Heel, 1982). Sulfamethoxazole, a sulfonamide drug, is a structural analog of para-aminobenzoic acid and competitively inhibits the synthesis of the intermediary dihydrofolic acid from its precursors, and trimethoprim is a structural analog of the pteridine portion of dihydrofolic acid that competitively inhibits dihydrofolate reductase and, consequently, decreases the production of the physiologically active tetrahydrofolic acid from dihydrofolic acid (Action, 2003;Church et al., 2015;Epstein, Amodio-Groton, & Sadick, 1997). Although each agent alone is bacteriostatic, their blockade of two sequential enzymes results in bactericidal activity when combined (Action, 2003;Church et al., 2015). The synergy between trimethoprim and sulfamethoxazole together provides effective inhibition of enzymes involved in the bacterial synthesis of tetrahydrofolic acid, which is a necessary cofactor in bacterial nucleic acid synthesis (Grim, Rapp, Martin, & Evans, 2005;Sköld, 2011;Wormser et al., 1982). It has been reported that continuous intravenous SXT is a potential therapy to treat E. coli meningitis in rabbits and that the pharmacokinetic characteristics of sulfamethoxazole and trimethoprim given in combination to poultry produces a synergistic level for both antimicrobials and thus is considered to be a useful combination in the management of various avian diseases (Mylotte, Bates, Sergeant, Matson, & Jr, 1981;Queralt & Castells, 1985). The impact of SXT on E. coli in broiler intestinal or digestive tracts has already been reported (Dheilly et al., 2012).
Furthermore, SXT has also been used in the treatment of bovine respiratory tract infections and feline and canine urinary tract infections caused by E. coli (Boothe, Smaha, Carpenter, Shaheen, & Hatchcock, 2012;Morrissey et al., 2016).
Quorum sensing (QS) is a process in which bacteria use chemical molecules as a signaling language to change behavior to adapt to specific environments (Bassler, 1999(Bassler, , 2002Miller & Bassler, 2001).
It has been reported that QS is a regulator of cellular processes such as bioluminescence production, virulence gene expression, cell division, biofilm formation, motility, metabolism, recombinant protein production, and the responsiveness to antibiotics (Ahmed, Petersen, & Scheie, 2007. Autoinducer 2 (AI-2), a signal molecule, is produced to mediate both intra and interspecies communication and has the potential to influence both gene regulation and bacterial fitness. AI-2 signals derive from spontaneous rearrangement of (4S)-4, 5-dihydroxy-2, 3-pentanedione (DPD) and are synthesized by a highly conserved AI synthase, LuxS, which is present in a variety of gramnegative and gram-positive bacteria species. In addition, AI-2 supports cell population-dependent behavior by interacting with central metabolism through the intracellular activated methyl cycle, which is related to the synthesis of tetrahydrofolic acid (Action, 2003;Bushby, 1973;Fuqua, Winans, & Greenberg, 1994;Vendeville, Winzer, Heurlier, Tang, & Hardie, 2005). It has been reported that AI-2 increases the susceptibility in Streptococcus anginosus to ampicillin and erythromycin and decreases the susceptibility in Staphylococcus aureus to vancomycin, and AI-2 is also involved in increased biofilm formation in Streptococcus intermedius (Ahmed et al., 2009;Xue, Zhao, & Sun, 2013). Our previous study confirmed that exogenous AI-2 increased the antibiotic (such as ampicillin, oxacillin, and penicillin-G) resistance of a clinical E. coli strain isolated from a dairy cow with mastitis by upregulating the expression of TEM-type enzyme in an LsrR (the AI-2 receptor)dependent manner (Xue et al., 2016). However, whether or not AI-2 affects the expression of SXT susceptibility genes in APEC strains has not been reported.
In this study, we performed real-time reverse transcription (RT)-PCR experiments and digoxigenin (DIG) gel shift assays to evaluate how AI-2 affects the SXT susceptibility of APEC. In this study, we reported that exogenous AI-2 increses the susceptibility of APEC strains to trimethoprim-sulfamethoxazole (SXT) in a folate synthesisdependent pathway but not in the LsrR-dependent manner. Our results showed that exogenous AI-2 can downregulate the transcription of the folate synthetase encoding genes folA and folC, and the folate synthesis-related genes luxS, metE, and metH. Gel shift assays confirmed that LsrR, the AI-2 receptor, did not bind to the promoters of folA and folC, suggesting that AI-2 might influence folate metabolism by a feedback inhibition effect but not in the LsrR-dependent pathway.

Strains
The host source

| Bacterial strains and culture conditions
The four E. coli strains used in this study were isolated from poultry, such as chickens, ducks, and pigeons (

| Antimicrobial activity assay
The overnight cultures were inoculated into fresh MH broth and diluted to a final concentration of 0.03 by optical density at 600 nm, and 200 μl dilutions were dispensed into 96-well plates (Costar, Corning, Steuben, NY, USA) containing SXT or the same volume of fresh MH medium with antibiotic-free for control group, whereas the test group was with adding a final concentration of 39 μmol/L AI-2.
Every group was four technical replicates. The plates were incubated at 37°C for 16 hr, and then 10-fold serial dilutions of cultures were T A B L E 2 Oligonucleotide primers used in this study method, using the 16s rDNA gene as the housekeeping gene, normalized by subtracting the C t value of 16s cDNA from target cDNA.
All of the real-time RT-PCR assays were repeated at least three times with similar results. The primers used in this study are listed in Table 2, and the PCR amplification efficiency was controlled between 1.93 and 2.09.

| Gel shift assay
The transcriptional regulator LsrR protein was overexpressed and purified according to the methods described previously (Xue, Zhao, Sun, Zhou, & Sun, 2009). Two DNA fragments, p-folA

| Statistical analysis
The data were analyzed using the statistical software SPSS (ver. 19.0, IBM Corp., Armonk, NY) by a one-way ANOVA method; the test results were shown as mean ± SD. The paired t test was used for statistical comparisons between groups. The level of statistical significance was set at a p-value of ≤.05.

| Exogenous AI-2 increased the SXT susceptibility in APEC
To examine the susceptibility of 31 E. coli isolates (21, 3, 1, and 6 from chickens, ducks, geese, and pigeons with avian colibacillosis, respectively) to SXT, antimicrobial susceptibility testing was performed using the broth dilution technique. The minimal inhibitory concentrations (MIC) ( Table 1)  Antibacterial assays were performed to examine the changes in SXT susceptibility of APEC 17, APEC 19, APEC 29, and APEC 40 with the addition of exogenous AI-2. This result showed that the survival rates of the test groups with AI-2 were similar to that of the control groups ( Figure 1a); however, decreased survival rates were observed in the test groups with exposure to SXT and AI-2 compared to the control groups with SXT alone. The survival rates of APEC 17, APEC 19, APEC 29, and APEC 40 with SXT and AI-2 were almost two times, 344 times, 19 times, and 30 times, respectively, lower than that of the control group with SXT alone (Figure 1b). These results indicated that with the addition of exogenous AI-2, susceptibility of APEC 17, APEC 19, APEC 29, and APEC 40 to SXT was increased, suggesting that, AI-2 might play an important role in the regulation of SXT susceptibility in these APEC isolates.

| Exogenous AI-2 increased SXT susceptibility by downregulating the folate synthetase encoding genes folA and folC
To investigate the mechanism of how AI-2 affects SXT susceptibility in APEC, the transcript levels of genes associated with SXT susceptibility and AI-2 QS were assessed by performing real-time RT-PCR experiments. It has already been reported that SXT could lead to bacterial death by affecting the folate metabolic pathway and folA and folC are two important encoding genes in folate metabolism. The transcript levels of folA, folC, and lsrR, which encode dihydrofolate reductase (DHFR), dihydrofolate synthetase (DHFS), and the receptor of AI-2, respectively, were measured in the four APEC strains under different conditions. As shown in Figure 2, the transcript levels of lsrR were increased 2.0-fold, 2.2-fold, 3.1-fold, and 2.2-fold by the addition of exogenous AI-2 in the absence of SXT in APEC 17, APEC 19, APEC 29, and APEC 40, respectively. Accordingly, the transcript levels of folA were decreased 2.9-fold, 2.4-fold, 2.3-fold, and 2.2fold, and the transcript levels of folC were decreased 1.9-fold, 1.9fold, 2.1-fold, and 3.2-fold in APEC 17, APEC 19, APEC 29, and APEC 40, respectively. Furthermore, in the presence of SXT, as shown in Figure 3, the addition of AI-2 resulted in 1.8-fold, 1.6-fold, 2.1-fold, and 1.9-fold increases in the transcription of lsrR, 3.7-fold, 2.6-fold, 3.3-fold, and 2.1-fold decreases in the transcription of folA, and 1.9fold, 1.8-fold, 3.2-fold, and 2.0-fold decreases in the transcription of folC in APEC 17, APEC 19, APEC 29, and APEC 40, respectively.

| LsrR did not bind to the promoters of folA and folC
It has been reported that LsrR contains a helix-turn-helix (HTH) DNA binding domain and is a QS regulator that can regulate the transcription of a variety of genes by binding to their promoters. On the basis of these findings, we hypothesized that exogenous AI-2 might activate LsrR, which then inhibits the transcription of folA and folC by directly binding to their promoter regions. To determine whether or not LsrR can bind to the promoters of folA and folC, we performed DIG-gel shift assays. The results showed that LsrR can not bind to the promoters of folA and folC (Figure 4), indicating that AI-2 might influence folate metabolism in an LsrR-independent pathway.

| Exogenous AI-2 influenced folate metabolism by downregulating the transcription of the folate synthesis-related genes luxS, metE, and metH in APEC
It has been reported that luxS (encodes S-ribosylhomocysteinase), Furthermore, in the presence of SXT, as shown in Figure 7, the addition of AI-2 resulted in 2.9-fold, 5.5-fold, 7.1-fold, and 1.7-fold decreases in transcription of luxS, 2.6-fold, 7.2-fold, 6.6-fold, and

| DISCUSSION
In recent years, the QS system has emerged as a topic of great interest due to its involvement in various biochemical processes. Previous studies reported that AI-2 is an important QS signal molecule in gramnegative and -positive bacteria, including E. coli, suggesting that AI-2 could play an important role in the physiological activity or regulation of E. coli (Ahmed et al., 2009;Vendeville et al., 2005;Xue et al., 2009).
In E. coli model strains, the functional mechanisms of AI-2 have been well studied (Wang, Li, March, Valdes, & Bentley, 2005;Xue et al., 2009 gene TEM, and the regulation was directly related to LsrR which is the receptor of AI-2 (Xue et al., 2016). However, in avian pathogenic E. coli, whether or not AI-2 affects antibiotic-related genes and pathogenic genes has not been reported. In this study, we found that, in the presence of SXT, exogenous AI-2 decreased the physiological activity of APEC, and increased the susceptibility of APEC to SXT by an LsrRindependent pathway. Therefore, we have confirmed that exogenous AI-2 affects SXT susceptibility in APEC strains by a folate synthesisrelated pathway, indicating that the combined use of AI-2 and SXT can minimize the dose of SXT in the treatment.
Previous studies on AI-2 did not explain its metabolic effect on APEC. This study indicated that AI-2 increases the SXT susceptibility of APEC in a folate-dependent manner. The addition of exogenous pre-