Screening of Anti-Adhesion Agents for Pathogenic Escherichia coli O157:H7 by Targeting the GrlA Activator

Generation of the grlA* Mutant Strain

The primers used in this study are listed in Table 2. It is known that basic residues on the C-terminus of the GrlA protein (Lys 111 and Arg 114) are involved in mechano-sensing and, thus, are critical for the GrlA-mediated activation of LEE1 upon mechanical stimulation [25]. To examine whether the hit compounds from the initial screening exhibited GrlA-inhibitory activity regardless of the mechano-sensing status of GrlA, we constructed the mechano-sensing defective mutant grlA*. Briefly, point mutations on the basic residues of grlA (K111A, R114A) were introduced into the plasmid-cloned wild-type (WT) grlA gene by site-directed mutagenesis, and then the truncated grlA locus in the ΔgrlA mutant strain was substituted with the grlA* using the CRISPR/Cas9 system, as described below.

First, a custom single guide RNA (sgRNA)-expressing plasmid, pTargetF_grlA, was constructed (Table 1). For this purpose, near the 5' end of the truncated grlA gene sequence in ΔgrlA was inspected for the N20 region containing a PAM site (nGG) with high %GC content. NEBaseChanger (New England Biolabs, USA; https://nebasechanger.neb.com/) was then used to design the primers required for the introduction of the detected N20 region as a sgRNA sequence into the pTargetF plasmid.

Next, a repair template consisting of the gene to be inserted (grlA*) and 300 bp up- and down-stream homology arms were prepared. The WT grlA gene and the flanking regions were cloned into the pBluescript SK II (+) plasmid, and the point mutations (K111A and R114A) were introduced via site-directed mutagenesis. Due to the presence of the sgRNA-matching sequence in the resulting construct, other silence mutations (ATC to ATT for Ile 130 and AGG to CGG for Arg 132) were additionally introduced to the grlA* gene. This manipulation prevents the sgRNA-guided Cas9 protein from recognizing and cutting the matching part after knock-in of the grlA*. The resulting mutated sequence (grlA with K111A, R114A, and the silence mutations at Ile 130 and Arg 132) was amplified with the flanking homology arms by PCR and used as a repair template.

Finally, the repair template and pTargetF_grlA were co-electroporated into the EDL933 ΔgrlA strain, which expresses Cas9 from the pCas plasmid. After recovery, the cells were spread on LB agar plates containing kanamycin and streptomycin and incubated overnight at 30°C. The correctly mutated construct was then identified by PCR and sequencing, and the remaining plasmids (pTargetF_grlA and pCas) were removed as described previously [28].

High-throughput Screening (HTS) of Small Molecules

To construct the HTS system, reporter strains carrying the GrlA-expressing plasmid and reporter plasmid were created. The grlA, grlA*, and promoter region of ler (384 bp) were amplified by PCR and then treated with restriction enzymes. The resulting products and linearized plasmids (pJK1113 for grlA and grlA*; pBBR_lux for the promoter region of ler) were ligated to create an arabinose-inducible GrlA-expressing plasmid (pSY2101), an arabinose-inducible GrlA*-expressing plasmid (pSY2201), and a reporter plasmid (pSY2104) containing the GrlA-activated ler promoter fused to the bioluminescence (lux) operon. The heterologous host reporter strains, SY2102 and SY2201, were constructed by co-transforming E. coli DH5α with the reporter plasmid pSY2104 and either pSY2101 or pSY2201, respectively. The homologous host reporter strains, SY2110 and SY2204, were constructed by transforming E. coli EDL933 having either the WT grlA or grlA*, respectively, with the reporter plasmid pSY2104. As a positive control for the homologous host reporter system, SY2111 was constructed by transforming the EDL933 ΔgrlA strain with the reporter plasmid pSY2104.

The HTS was performed as described previously with a slight modification [29]. The heterologous host reporter strains SY2102 or SY2201 were cultured to an absorbance at 600 nm (A₆₀₀) of 0.2 in fresh LB containing 0.0001% or 0.0002% L-(+) arabinose, respectively, with the appropriate antibiotics. One hundred microliters of culture were transferred to each well of a 96-well microtiter plate containing a 20 μM concentration of each molecule or 2% DMSO. The plates were incubated at 37°C with shaking, and the relative luminescence unit (RLU) was measured three times at hourly intervals using a Spark microplate reader (Tecan, Switzerland). For these screenings, the DMSO-treated reporter E. coli without L-(+) arabinose was set as a positive control, while the DMSO-treated reporter E. coli with L-(+) arabinose was set as a negative control. The homologous host reporter strains and their positive control strain SY2111 were cultured to an A₆₀₀ of 0.2 (for SY2110 and SY2111) or 0.5 (for SY2204 and SY2111) in fresh LB with the appropriate antibiotics, and the RLU value was calculated hourly. The percentage of inhibition was calculated using the following equation:

Cell Culture and Bacterial Adherence Assay

HeLa cells (ATCC CCL-2) were kindly provided by Dr. S.H. Choi at Seoul National University and cultured at 37°C with 5% CO₂ in DMEM medium supplemented with 10% heat-inactivated fetal bovine serum (FBS; Thermo Fisher Scientific) and 1% antibiotic/antimycotic (Thermo Fisher Scientific). Cell adhesion assays were performed using the previously described method [30], with a slight modification. Briefly, the HeLa cells were seeded and grown in 24-well cell culture plates until confluent. Before bacterial treatment, monolayers of the HeLa cells were washed three times with pre-warmed Dulbeccós Phosphate Buffered Saline (DPBS, Thermo Fisher Scientific), and the fresh DMEM without antibiotic/antimycotics and FBS were added to each well. The HeLa cells were then treated with the indicated bacterial strains at a multiplicity of infection (MOI) of 100 (2 × 10⁷ CFUs/well) and further incubated for 2 h in the presence (20 μM) or absence of hit molecules at 37°C with 5% CO₂. After washing out the non-adherent bacterial cells with DPBS, the HeLa cells were lysed with 1% Triton X-100. Finally, serial dilutions of the lysates were plated on LB agar and incubated overnight at 37°C. The colonies were counted the next day to determine the total number of adherent bacteria per well.

Total RNA Extraction and Quantitative Real-Time PCR (qRT-PCR)

The preparation process for total RNA extraction from the HeLa cell-attached EDL933 was described elsewhere [31]. Briefly, the HeLa cells were seeded in a 24-well cell culture plate at a density of 2 × 10⁵ cells per well and treated with EDL933 at an MOI of 100, as described above. After 2 h of co-incubation in the presence or absence of hit molecules, the HeLa cells were washed three times with DPBS to remove the non-adherent bacteria. To collect the total RNA from the attached bacteria, 400 μl of DPBS and 800 μl of RNAProtect Bacteria Reagent (Qiagen, Germany) were added to each well, followed by scraping and centrifugation of the cells. The pellet was kept at -20°C until total RNA extraction. The total RNA was extracted from the pellets using an RNeasy kit (Qiagen), and all the samples were treated with DNaseI to remove any contaminated genomic DNA. The extracted RNA was kept at -80°C. The quantity and quality of the RNA in the samples were determined using a NanoQuant plate and Spark microplate reader (Tecan).

For qRT-PCR analysis, complementary DNA (cDNA) was synthesized using the iScript Reverse Transcription Supermix (Bio-Rad Laboratories, USA). For each well, 10 μl of 2X iQ SYBR supermix (Bio-Rad), 0.5 μl of forward primer, 0.5 μl of reverse primer, 7 μl of Nuclease-free water, and 2 μl of cDNA template were added, and amplification was performed using QuantStudio Real-Time PCR System (Applied Biosystems, USA). The genes that were closest to the promoter of each LEE operon were examined. The relative expression levels of the transcripts were calculated using the expression level of the rpoA housekeeping gene as the internal reference for normalization. The specific primers used for the amplification of the cDNAs are listed in Table 3.

Construction of the Reporter Strain for HTS

To screen the chemical libraries for GrlA inhibitors, an HTS system was constructed, as described in the Materials and Methods section. Since the E. coli DH5α strain lacks the Ler, GrlA, and LEE operons, this heterologous host would enable us to investigate the GrlA-dependent response of LEE gene expression with few potential crosstalks or signal inputs mediated by other regulatory proteins and pathways. Therefore, the heterologous host (E. coli DH5α) reporter strain was constructed by co-transforming an arabinose-inducible GrlA-expressing plasmid (pSY2101) and a GrlA-dependent lux reporter plasmid (pSY2104) (Fig. 2A). Since GrlA binds directly to the LEE1 promoter to activate the transcription of ler, the expression of the lux operon would be diminished if potential hit molecules interfered with the activity of GrlA. The feasibility of the construct as a reporter strain was examined via treatment with various concentrations of arabinose, and varying levels of bioluminescence emission were confirmed depending on the arabinose concentration (Fig. 2B).

Open in a separate window

Fig. 2

Reporter strain for the HTS of the GrlA inhibitor.

(A) A heterologous host E. coli DH5α reporter strain contains pSY2101 expressing GrlA under the P_BAD promoter and pSY2104 carrying luxCDABE genes under the GrlA-activated promoter P_LEE1. (B) The constructed reporter strain was verified by L-(+) arabinose induction, with various concentrations (w/ v) indicated. The one-way ANOVA with multiple comparison was used to determine the statistical significance between the control (no arabinose) and other samples at 4h (ns, non-significant; ****, p < 0.0001).

HTS Screening Using WT grlA Reporter Strains

To identify the anti-virulence molecule that prevents EHEC adherence to host cells via GrlA inhibition, a total of 8,201 compounds from the Korea Chemical Bank were screened at a concentration of 20 μM using the heterologous host reporter strain SY2102. After the initial screening, 80 compounds with a 50 or higher %inhibition were classified as “possible hits” and examined further in triplicate. It should be noted that the compounds that showed reproducible results in the triplicate experiments were selected as “initial hits.” The compounds that showed growth inhibition were excluded from the hits because the purpose of this HTS was to identify the molecules that would not result in selective pressure on E. coli. Colored chemicals were also excluded because the read-out of the screening system was based on bioluminescence. Consequently, a total of 16 compounds were selected as initial hit molecules (Fig. 3A).

Open in a separate window

Fig. 3

Validation of the initial hit molecules from the HTS.

(A) Validation using the heterologous host reporter strain SY2102. The strain was grown in LB medium supplemented with 0.0001% L-(+) arabinose and 2% DMSO or each chemical (final concentration of 20 μM). As a positive control, SY2102 was grown without L-(+) arabinose. (B) Validation using the homologous host reporter strain SY2110. The strain was grown in LB medium supplemented with 2% DMSO or each chemical (20 μM). As a positive control, SY2111 was grown in LB medium supplemented with 2% DMSO. After 5 h of incubation, the luminescence and absorbance were measured to calculate the % inhibition. The data were represented as the means ± SD from three independent experiments. The red arrows indicate the selected six compounds at this validation step with the homologous host reporter strain SY2110.

The molecules were further validated using the homologous host reporter strain SY2110, and 6 of 16 compounds were chosen at this step (Fig. 3B). The selection criteria were: compounds that exhibited greater than 80% inhibition and were not toxic to the homologous host reporter strain. The molecular characteristics, such as the chemical structure and molecular weight, of these six compounds are listed in Table 4.

Table 4

Molecular characteristics of the hit compounds.

Open in a separate window

Open in a separate window

Validation Using grlA* Reporter Strains

GrlA is a membrane-bound, mechano-sensing transcriptional regulator, the mechanism of which is liberation from the bacterial inner membrane [25]. Therefore, a possible scenario for the selected hits is that the inhibitor traps the GrlA in the inner membrane. To examine this scenario, we introduced point mutations (K111A and R114A) in the grlA that resulted in the expression of a membrane binding-defective mutant, GrlA*. All six compounds were then evaluated using the heterologous and homologous host reporter strains harboring grlA*. As shown in Figs. 4A and ​and4B,4B, the compounds 199F7 and 182H11 did not exhibit inhibition activity either for the heterologous or homologous host reporter strains with grlA*, respectively, suggesting that they may affect the liberation of GrlA from the membrane, not the binding of GrlA to the LEE1 promoter. The compounds 188C9 and 51H8 were also excluded from the further investigation since both excessively reduced the RLU beyond the level of the positive control, suggesting an adverse effect on the reporter plasmid maintenance/replication and not GrlA (Fig. 4B). From the screening and validation, two hit molecules (187F9 and 185D7) were selected as putative GrlA inhibitors.

Open in a separate window

Fig. 4

Validation of the hit molecules using strains expressing grlA*.

An additional set of heterologous (SY2201) and homologous (SY2204) host reporter strains expressing grlA* were prepared and used for the hit validation. (A) Strain SY2201 was grown in LB medium supplemented with 0.0002% L-(+) arabinose and the indicated hit molecules (20 μM). The negative and positive controls were the same strain grown with or without L-(+) arabinose, respectively. (B) Strain SY2204 was grown in LB with the indicated hit molecules (20 μM). Strains SY2204 and SY2111 treated with 2% DMSO were used as a negative and positive control, respectively. The RLU of each sample was calculated after 4 h of incubation. The data were represented as the means ± SD from three independent experiments. The red arrows indicate the selected two compounds at these validation steps. The Student’s t-test was used to determine the statistical significance between the negative control and other samples (ns, non-significant; ****, p < 0.0001; ***, p < 0.001; **, p < 0.01).

This work was supported by the National Research Foundation of Korea funded by the Ministry of Science and ICT (NRF-2020R1F1A1070168) and the Ministry of Food and Drug Safety in 2022 (20162MFDS142) to B.S.K. The chemical library used in this study was kindly provided by the Korea Chemical Bank (www.chembank.org) of Korea Research Institute of Chemical Technology. We thank Dr. B.T. Kim at Korea Research Institute of Chemical Technology for depositing the Grlactin into the chemical library. We also thank Dr. S.H. Choi at Seoul National University for sharing the E. coli EDL933 ΔgrlA strain and HeLa cells. pCas and pTargetF were gifts from Sheng Yang (Addgene plasmid #62225 and #62226).