Deletion of phoP or phoQ Restores TMP Sensitivity to ΔmgrB

It is known that mgrB encodes MgrB, which feeds back to inhibit the kinase activity of PhoQ, so the MgrB-dependent negative feedback loop plays a crucial role in E. coli PhoQ-PhoP signaling (Lippa and Goulian, 2009; Salazar et al., 2016). We thus hypothesized that the TMP-resistant phenotype resulting from mgrB deletion might be mediated by the PhoP/Q pathway. To prove this hypothesis, we constructed two double-knockout mutants (E. coli W3110 ΔmgrB ΔphoQ and E. coli W3110 ΔmgrB ΔphoP) and tested their susceptibilities to TMP. The results showed that the TMP MICs of both double knockout mutants were identical to that of the wild-type strain, two times that of ΔphoP or ΔphoQ, demonstrating that disruption of the PhoP/Q pathway could partially reversed the effects of mgrB deletion (Table 2).

Deletion of pmrK Does Not Affect Susceptibility of E. coli to TMP

Previous studies have shown that the deletion of mgrB led to the upregulation of the PhoQ/PhoP pathway and the activation of pmrK (part of pmrHFIJKLM operon), eventually causing resistance to polymyxins via modifications of the lipopolysaccharide target in K. pneumoniae (Helander et al., 1996; Cannatelli et al., 2013). To investigate whether pmrK was also involved in the TMP resistance of E. coli ΔmgrB, we both knocked out and overexpressed pmrK in E. coli W3110. Our results showed that pmrK did not affect TMP susceptibility in E. coli (Table 2). In addition, the ΔmgrB mutant was not more resistant to polymyxin B or colistin than was the wild-type strain (Table 3).

Overexpression of tolC Does Not Affect Susceptibility of E. coli to TMP

TolC, an outer membrane protein, is part of the AcrAB-TolC drug efflux system (Tikhonova and Zgurskaya, 2004). This system plays an important role in both intrinsic and acquired resistance to a wide variety of currently available antimicrobial agents (Du et al., 2014). Previous studies have identified a putative PhoP binding site (PhoP box) in the tolC promoter region (Eguchi et al., 2003). As our results indicated that PhoP/Q was involved in the TMP resistance of the ΔmgrB mutant, we speculated that the PhoP/Q pathway might affect TMP sensitivity by modulating the expression of tolC. To test this theory, we overexpressed tolC in E. coli W3110. We found that the overexpression of tolC did not affect TMP susceptibility (Table 2). In addition, our qRT-PCR results suggested that the deletion of mgrB did not affect tolC expression (Figure 1).

Open in a separate window

FIGURE 1

Average transcription levels of the genes phoP, phoQ, folA, and tolC in E. coli W3110 (wild type) and E. coli W3110 ΔmgrB, as measured with qRT-PCR. Expression levels of gapA were normalized as an endogenous control. Data represent mean ± SD from three independent experiments. P values were calculated using t tests (*P < 0.05; **P < 0.01; ***P < 0.001).

Transcriptome Analysis of the ΔmgrB Mutant

To further understand how mgrB deletion led to TMP resistance in E. coli, we compared the global transcription profiles of the E. coli K12 W3110 ΔmgrB mutant and the wild-type E. coli strain using RNA-seq (PRJNA523715). We identified 518 significantly upregulated genes and 433 significantly downregulated genes (951 in total; Supplementary Table 2). As expected, the deletion of mgrB was followed by a significant upregulation of phoQ and phoP. Interestingly, folA also was upregulated. The folA gene encodes the DHFR of E. coli, the target of TMP (Table 4). We verified these results with qRT-PCR (Figure 1).

Co-overexpression of phoP and phoQ Causes TMP Resistance

To further test the involvement of PhoP/Q in the TMP resistant phenotype of ΔmgrB, we constructed three strains of E. coli W3110, overexpressing phoP, phoQ, or phoPQ, and tested their sensitivity to TMP. Our results showed that while overexpressing phoP or phoQ did not affect TMP sensitivity, co-overexpression of phoP and phoQ caused TMP resistance (Table 2).

PhoP Recognizes the folA Promoter in vitro

As the deletion of mgrB was followed by the increased expression of folA, phoP, and phoQ, we hypothesized that MgrB might modulate the expression of folA via the PhoP/Q pathway. To verify this hypothesis, we used electrophoretic mobility shift assay (EMSA) assay to test whether PhoP recognized and bound the promoter regions of folA directly, by using the promoter region of mgrB, a known PhoP substrate (Kato et al., 1999), as the positive control. Phosphorylated PhoP with a poly-histidine tag was prepared, and the biotin-labeled DNA probe containing the promoter region of folA (−231/+26) were amplified by PCR using 3′end biotin-labeled primers. After combined incubation and subsequent electrophoresis, it was clearly that PhoP hindered the mobility of the probe, and the hindering effect increased correlating to the quantities of PhoP, suggesting that PhoP specifically recognize and bind the promoter region of folA (Figure 2).

Open in a separate window

FIGURE 2

Electrophoretic mobility shift assays (EMSAs) of PhoP to the folA promoter region. A varying concentration of the PhoP (0, 0.25, 0.5, 1, 2, and 4 μg) was incubated with PCR-amplified biotin-labeled probes of the promoter regions of folA, and the binding of PhoP (2 μg) to the promoter regions of mgrB was used as the positive control, showing that PhoP recognized and bound to the promoter of folA. The experiment was repeated at least three times and a representative result was presented.

MgrB and PhoP Regulate folA Gene Expression in vivo

To confirm that MgrB and PhoP regulated the expression of folA, the β-galactosidase activities of strains (the wild type, the ΔphoP mutant, the ΔmgrB mutant, and the ΔmgrB ΔphoP mutant) harboring a reporter plasmid pZT102 (Li et al., 2014) (Supplementary Figure 1) carrying a lacZ gene fusion with the folA promoter region (−231/+26) were determined during the exponential growth phase. Our results showed that deletion of mgrB caused a ~2-fold increase in the β-galactosidase activity, and deletion of phoP led to about 40% decrease of the β-galactosidase activity (Figure 3). Further deletion of phoP gene in the ΔmgrB mutant decreased the β-galactosidase activity back to the level of the wild type strain, but not to that of the ΔphoP mutant.

Open in a separate window

FIGURE 3

A comparison of promoter activity of folA in various strains of E. coli using lacZ-fusion analysis. β-Galactosidase activities were measured from exponential phase aerobic cultures in LB of several E. coli W3110 strains: wild-type (WT), null mutants (ΔphoP and ΔmgrB), and a double mutant (ΔmgrB ΔphoP), all of them carrying a reporter plasmid pZT102:folAp-257 containing the promoter region of folA (–231/+26). CK is empty pZT102 plasmid control in E. coli W3110. Means and standard deviations are represented. Data represent mean ± SD from three independent experiments. P values were calculated using t tests (*P < 0.05; **P < 0.01; ***P < 0.001).

Construction of Gene Deletion Mutants, Complemented Strains, and Gene Overexpression Strains

We constructed the E. coli W3110 gene knockout mutants with the λ Red recombination system and verified mutants with junction PCR as previously described (Datsenko and Wanner, 2000). Double-knockout mutants were generated with the same procedure. We amplified the genes mgrB, phoP, phoQ, phoPQ, pmrK, and tolC genes from E. coli W3110 genomic DNA using gene specific primers (see Supplementary Table 6), and then cloned the genes into plasmids pCA24N. The recombinant plasmids were then transduced into either the deletion mutants of E. coli W3110 (to form the complemented strain) or the E. coli wild types (to form the gene overexpression strain). Plasmid pCA24N:phoP was transduced into E. coli BL21(DE3) for the overexpression and purification of the protein PhoP. The details for relevent strains and plasmids are listed in Supplementary Table 5.

Screening for Genes Associated With TMP Resistance or Susceptibility

We cultured the E. coli Keio Knockout Collection (Baba et al., 2006) in 96 Deepwell Multiwell plates in Luria-Bertani (LB) medium (Difco) at 37°C until the OD₆₀₀ was 0.6. We then diluted each well to 10⁷ colony-forming units (cfu) ml^–1 (about OD₆₀₀ of 0.01) with fresh LB containing either 0.1 μg ml^–1 or 1 μg ml^–1 TMP. Plates were statically incubated overnight (12 h) at 37°C. We measured the absorbance of each well at 600 nm with a Microplate Reader (BioTek, Winooski, VT, United States) to monitor the growth of all mutants.

Drug Susceptibility Tests

Bacterial cells were grown to the mid-log phase (OD₆₀₀ of ~0.6), then diluted to 10⁵ cfu ml^–1 in fresh LB. Next, ten-fold serial dilutions were plated onto LB agar solid plates containing different concentrations of TMP (0, 0.01, 0.02, 0.04, 0.08, 0.16, 0.32, 0.64, 1.28, and 2.56 μg ml^–1). 10 μM IPTG was contained when the complementation and/or over-expression strains were tested. Cultures were incubated overnight at 37°C. The MIC (minimum inhibitory concentration) was defined as the lowest concentration of the TMP required to inhibit 99% of the colony-forming bacterial units. Each strain tested two times with at least three different clones.

RNA-Seq

Total RNA was isolated with a RNeasy mini kit (Qiagen, Hilden, Germany). Library constructions were prepared using a TruSeq Stranded Total RNA Sample Preparation kit (Illumina, San Diego, CA, United States), and RNA was sequenced with an Illumina HiSeq 2500 at the Shanghai Biotechnology Corporation (Shanghai, China). The insert size of the purified libraries was confirmed with an Agilent 2100 bioanalyzer (Agilent Technologies, Santa Clara, CA, United States). Bowtie2 v2-2.0.5 was used to map the cleaned reads to the genome of E. coli str. K-12 substr. W3110 (downloaded from the National Center for Biotechnology Information)¹. To estimate fold changes, HTSeq v2.1.1 was used with a reference annotation to generate values for fragments per kilobase of exon model per million mapped reads. Three biological replicates were used for RNA-seq, and p- and q-values were calculated. Differentially expressed genes were selected only if the false discovery rate q-value ≤ 0.05 and fold-change ≥ 2.

Quantitative Real-Time PCR Assays

Total RNA was extracted and cDNA was synthesized with a ReverTra Ace quantitative polymerase chain reaction (qPCR) kit (TOYOBO, Osaka, Japan), following the manufacturer’s instructions. We used real-time PCR (RT-PCR) to quantify gene expression levels using a 7900 HT Sequence Detection System with ABI Power SYBR Green PCR Master Mix (ABI, Hudson, NH, United States). Primers specific to phoQ, phoP, and tolC were designed and synthesized by Sangon Biotech Co., Ltd. (Shanghai, China). Expression levels of gapA mRNA were normalized as an endogenous control. All gene-specific primers are listed in Supplementary Table 6.

Protein Purification and Phosphorylation

Cultures of E. coli BL21 (DE3)/pCA24N:phoP were incubated at 37°C in LB broth medium supplied with corresponding antibiotics until the exponential phase. IPTG (isopropyl β-D-thiogalactoside) was added to produce a final concentration of 0.1 mM, and cultures were then incubated overnight with shaking at 16°C. Bacterial cells were harvested and then disrupted by sonication. The cell lysate was centrifuged at 10,000 × g for 30 min and the protein was isolated from the supernatant by nickel affinity chromatography. The purified His₆-PhoP protein was dialyzed against 50 mM Tris–HCl buffer (pH 7.5). For EMSA assays, the His₆-PhoP was phosphorylated as previously described (Tang et al., 2012) with modifications. Briefly, His₆-PhoP was incubated in phosphorylation buffer (50 mM Tris–HCl [pH 7.5], 50 mM KCl, 10 mM MgCl₂) containing 10 mM acetyl phosphate (Sigma-Aldrich) for 1 h at 37°C.

Electrophoretic Mobility Shift Assay (EMSA)

The biotin-labeled probes containing the promoter regions of the gene folA were amplified using biotin-labeled primers (Supplementary Table 6). EMSA/Gel-Shift Kit (Beyotime, Shanghai, China) was used for the binding of probes and phosphorylated PhoP proteins. We resolved band shifts on 6% non-denaturing TBE (Tris-Boric acid-EDTA)-polyacrylamide gels and subjected the gels to electrophoresis at 80 V for 1 h at 4°C. DNA was transferred to a nylon membrane at 380 mA for 70 min, and ultraviolet cross-linked at 302 nm for 15 min. We subsequently developed membranes using the Chemiluminescent Biotin-labeled Nucleic Acid Detection Kit (Beyotime, Shanghai, China), following the manufacturer’s protocol. We detected signals using a Typhoon 9410 imager (GE Healthcare, Maple Grove, MN, United States).

lacZ-Fusion Construction and β-Galactosidase Assays

The DNA fragments containing promoter regions (including original sequence and sequence with mutation sites) of the folA operon were amplified with BamHI- and HindIII-included primers (see Supplementary Table 6). Amplified fragments were ligated with pZT102 reporter plasmids containing the lacZ operon (obtained from Prof. Shiyun Chen, Wuhan Institute of Virology, Chinese Academy of Sciences, China) to generate a series of lacZ-fused plasmids (Supplementary Table 5). The plasmids were then electroporated into E. coli strains (E. coli W3110, E. coli W3110 ΔphoP, E. coli W3110 ΔmgrB, and E. coli W3110 ΔmgrB ΔphoP). We cultured the recombinant strains at 37°C overnight, then diluted the cultures to 1% with fresh LB broth, grew them again at 37°C, and finally harvested them at an OD₆₀₀ of ~0.7. We washed cell pellets with PBS and assayed β-galactosidase activity as described by Miller (1992).

DNase I Footprinting Assay

The promoter region of folA (−216 to −12) was amplified using FAM-labeled reverse primer (Supplementary Table 6). The purified DNA fragment (800 ng) was added to the reaction mixture (containing phosphorylated PhoP proteins) at 37°C for 30 min as described in EMSA. All of the mixtures were treated with DNase (1 unit, Fermentas China Co., Ltd., Shenzhen, China) at 37°C for 45 s. The results were analyzed with Applied Biosystems 37030XL DNA analyzer (manufactured by Tsingke Company, Wuhan, China).

We thank LetPub (www.letpub.com) for its linguistic assistance during the preparation of this manuscript.