The vector pNMD46772 containing the ColM-coding cma gene under the control of a 35S promoter and selective bar gene under the control of a nos promoter was transformed to Agrobacterium tumefaciens strain GV3101. The transformed bacteria were grown on selective LB (Bertani, 1951) agar medium containing rifampicin 50 mg/L, gentamicin 25 mg/L and kanamycin 100 mg/L. Colonies were confirmed to be transformed with vector pNMD46772 using PCR analysis with primers specific for the cma gene.

Agrobacterium-mediated transformation of lettuce ‘Odeskyj curling’; with the vector pNMD46772 containing the cma gene was carried out as described in Shcherbak et al. (2013). After 48 h of co-cultivation with diluted Agrobacterium suspension supplemented with 100 mM acetosyringone, explants were placed on modified B5 (Gamborg et al., 1968) regeneration medium (B5 medium salts with 2.5% sucrose, 3 mg/L kinetin, 0.5 mg/L NAA) supplemented with 5 mg/L phosfinothricin (PPT) for selection. As shoot appeared, each one was removed and placed on the MS (Murashige and Skoog, 1962) medium supplemented with 0.5 mg/L of indole-3-butyric acid for rooting. Well established plants were transferred into the soil and grown in a greenhouse.

For genetic transformation of mizuna, hypocotyls of 5-7-day-old seedlings were used as explants. After co-cultivation with Agrobacterium, the explants were cultivated on callus induction MS medium containing 2,4-D (2 mg/L) and BAP (0.5 mg/L) for one week. Further regeneration of shoots took place on a nutrient medium containing zeatin (2 mg/L), BAP (1 mg/L) and NAA (0.5 mg/L). PPT at concentrations of 5 mg/L was used for the first step of mizuna transformants selection and after two weeks of cultivation on the regeneration medium the PPT concentration was increased to 10 mg/L. Regenerated plants were also rooted, transferred to the soil and grown in a greenhouse.

Genomic DNA was isolated from plant leaves using the CTAB protocol for isolating DNA from plant tissues. PCR was carried out in 20 μl reaction volume containing 50 ng DNA, 200 μM each of forward and reverse primers, 200 μM dNTPs and 1 U Taq DNA polymerase (Thermo Scientific, USA). Amplification was performed on a Mastercycler® personal (Eppendorf, USA). Sequences of the primers used were as follows:

for bar gene amplification.

for cma gene amplification.

The products of the amplification (411 bp when using primers for bar gene and 1003 bp for cma gene) were separated on 1% (w/v) agarose gels.

RT-PCR was performed for previously selected transgenic plants to detect cma gene transcripts. Total RNA was prepared from 100 mg of transgenic plant leaf tissue using an GeneJET RNA Purification Kit (Thermo Fischer Scientific, Waltham, USA) and reverse-transcribed using a RevertAid First Strand cDNA Synthesis Kit (Thermo Scientific, USA) with oligo dT primers. Genomic DNA from RNA preparation was removed by DNase I (RNase-free) treatment for 30 min at 37°C. A negative control for each sample consisted of the same reaction mixture without reverse transcriptase. The obtained cDNA was used as a template for PCR with gene-specific primers.

Leaf material of transgenic and non-transgenic plants was ground in liquid nitrogen and total soluble protein (TSP) extracts were prepared with 5 volumes of extraction buffer (50 mM HEPES pH7.0, 10 mM K acetate, 5 mM Mg acetate, 10% glycerol, 0.05% Tween-20, 300 mM NaCl). After incubation for 20 min on ice and centrifugation for 10 min at 13000 rpm at 4°C, supernatant was used for immunoblot analysis and determination of antimicrobial activity. Protein concentration in the extracts was determined by the Bradford method using Bio-Rad Protein Assay (Bio-Rad Laboratories, Hercules, USA) and bovine serum albumin (BSA) (Sigma-Aldrich, Saint Louis, USA) as a standard.

Plant leaf material was harvested, frozen in liquid nitrogen and stored at -80°C. After 262 days, part of the material was thawed and incubated for 1 day at 40°C in an oven for drying. For final drying, material was incubated at room temperature for an additional 10 days. Dried material was further stored at room temperature. Fresh and frozen plant material was extracted with 5 volumes of extraction buffer, while dried material was extracted with 20 volumes of extraction buffer.

To allow direct comparison of activities for fresh and dried plant material, the weight factor for normalization of different extraction volumes and the weight loss during drying process were calculated as follows: Weight factor = fresh weight/dry weight * 5 (vol fresh extraction)/20 (vol dry extraction).

ColM was transiently expressed in N. benthamiana and purified according to Stephan et al. (2017) with minor modifications. For purification plant material was extracted with the buffer containing 20 mM citric acid (pH4.0), 20mM NaH₂PO₄ and 30 mM NaCl in a 5:1 (v/w) buffer:biomass ratio. Ground leaf material admixed with pre-chilled extraction buffer was incubated on ice for 10 min followed by centrifugation at 10,000 x g for 15 min at 22°C. The supernatant was filtered using Miracloth (Merck KGaA, Darmstadt, Germany) and mixed with 10 mg/ml diatomaceous earth (Sigma-Aldrich, Saint Louis, USA) followed by incubation of the filtrate for 20 min at room temperature and centrifugation for 15 min at 10,000 x g at 22°C. The resulting supernatant was filtrated using filter discs with a pore size of 8–12 µm and the final filtrate was used for further purification by column chromatography. The filtrate was loaded on a HiTrap™Capto™MMC (GE Healthcare, Munich, Germany) column equilibrated with extraction buffer. The column was washed in two steps. The first wash step was performed with extraction buffer. The second wash step was done using diluted (30%) elution buffer consisting of 50 mM Na₂HPO₄ (pH 7.84), 10 mM citric acid and 1 M NaCl. The elution of ColM was carried out in a linear gradient from 30 to 100% of elution buffer concentration. The eluted fractions were analyzed by SDS-PAGE and Coomassie staining using Instant Blue™ staining solution (Expedeon, San Diego, CA, USA). The colicin containing fractions were pooled and dialyzed overnight at 4°C against 20 mM Na₂HPO₄ (pH 7.5), 10 mM citric acid and 50 mM NaCl. After dialysis, purified protein was frozen in liquid nitrogen and finally freeze-dried by lyophilization using freeze dryer Alpha 1–2 LDplus (Martin Christ Gefriertrocknungsanlagen, Osterode am Harz, Germany). Purified ColM batch 12 was used as a reference standard for antimicrobial activity and in immunoblot analysis.

Protein-containing plant extracts were mixed 1:1 with 2xLaemmli buffer and denatured at 95°C for 10 min. The total proteins were separated by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) using a 15% polyacrylamide gel under reducing conditions and then transferred to a PVDF membrane. The membrane was incubated with a rabbit polyclonal ColM antibody 30365 (diluted 1:5,000) or monoclonal ColM antibody 3F9 (diluted 1:2,000) overnight. Polyclonal ColM antibody was generated against purified full length ColM by BioGenes GmbH (Berlin, Germany). Monoclonal ColM antibody 3F9 was generated against purified full-length ColM by Fraunhofer Institute for Cell Therapy and Immunology (IZI, Halle (Saale), Germany). Secondary antibodies were IgG (whole molecule) peroxidase affinity isolated antibody (Sigma-Aldrich, Saint Louis, USA), anti-mouse (diluted 1:5,000) or anti-rabbit (diluted 1:10,000). Determination of ColM concentration in extracts was done semi-quantitatively by comparison of different amounts of extracts with known amounts of purified ColM batch 12 reference protein.

Microbiological studies were carried out on two laboratory strains of E. coli (DH10B and XL1-Blue) and a set of reference strains (Table 1). The bacteria were incubated at 37°C in LB medium (Bertani, 1951).

Semi-quantitative determinations of antimicrobial bacteriocin activity by a spot-on-lawn soft-agar overlay assay were performed as described in Schulz et al. (2015). Specific activity was calculated by normalization of arbitrary colicin activity units to TSP concentrations determined via Bradford assay to exclude effects of different extraction efficiencies.

Each respective bacterial test culture was grown in LB medium until OD600 of 0.2 and diluted to approximately 2x10³ CFU/µl. 5 µl of diluted culture were spotted on LB medium supplemented with 0.75% (w/v) agar, 0.1 mg/ml BSA (AppliChem, Darmstadt, Germany) and a dilution series of purified ColM (10 µg/ml to 1 ng/ml). Plates were incubated for 16-20 h at 37°C before MIC readout. The MIC was defined as the lowest concentration of ColM that inhibited visible growth of the test culture.

We generated transgenic plants of lettuce (Lactuca sativa L.) ‘Odeskyj curling’ and mizuna (Brassica rapa subsp. nipposinica var. laciniata) using Agrobacterium-mediated transformation with a vector carrying the ColM-coding gene (cma). The vector pNMD46772 contains two expression cassettes: a cma gene under the control of a 35S promoter and the selective bar gene under the control of a nos promoter. Expression of the bar gene in transgenic plants provides for their resistance to the herbicide phosphinothricin (PPT). After five- or six-weeks incubation of lettuce cotyledons on the regeneration medium supplemented with the selective agent PPT (5 mg/L), green shoots were formed on the explants (Figure 1A). Regenerated plants developed roots after transfer to the medium supplemented with 0.5 mg/L of indole-3-butyric acid (Figure 1B). The resulting rooted lettuce plants were transferred to soil and grown in a greenhouse (Figure 1C). Mizuna plants were regenerated on the zeatin-containing media after 3-5 weeks of cultivation (Figure 1D). The selected plant lines resistant to 5 mg/L of PPT were cultivated under aseptic conditions (Figure 1E). Two lines were transferred to soil and grown in a greenhouse (Figures 1F, G). Greenhouse-maintained lettuce and mizuna plants bloomed and formed seeds in 3-4 months.

All putative transgenic plants were initially screened for the integration of bar and cma genes using primers which should produce fragments of 411 bp and 1003 bp respectively. Amplification of the DNA of all plants revealed the expected fragments whose position is indicated by the arrows (Figures 2A, B). This result confirms the presence of cma and bar genes in the genome of the plants generated on the selective medium. A total of 5 independent transgenic lettuce plants and 2 mizuna plants were confirmed as PCR-positive for the transgenes integration. The purified RNA of transgenic plants along with wild type plants were used for cDNAs synthesis by reverse transcription. PCR product (1003 bp) obtained from the cDNAs by PCR with primers complementary to the cma gene confirmed ColM expression at the transcription level for transgenic plants of lettuce and mizuna (Figure 2C). A negative control for each sample consisted of the same reaction mixture without reverse transcriptase and no PCR products have been detected (lanes (1- to 6-), Figure 2C).

In an initial screen for antibacterial activity, the transgenic plants were tested by placing explants on agar medium covered with diluted bacterial suspension. Clear bacterial growth inhibition zones were observed around the explants of transgenic lettuce plants grown in a greenhouse (Figure 3А) and two transgenic mizuna plant lines maintained in vitro (Figure 3B).

Antimicrobial activity was also evaluated semi-quantitatively using a soft-agar overlay assay with plant extracts based on the highest dilution of each that caused a clearing effect. The results of these experiments are presented in Table 2. The data showed very comparable levels of antibacterial activity in different transgenic lines and even between the two analyzed species. However, a decrease or complete absence of antibacterial activity in some plants of T1 generation was also noted. The antibacterial activity of two T1 generation plants of the transgenic lettuce line Ls#21 (Ls#21/T1#3 and Ls#21/T1#5) differed by almost 10 times (Table 2). This result was also confirmed by analysis of recombinant ColM expression detected using immunoblot: the lower antibacterial activity values correlated with less pronounced immunoblot signals (Figure 4A, these lines are marked in blue).

The same correlation between the antibacterial activity of the extracts, determined by soft-agar overlay assay, and the intensity of the immunoblot signals was observed when analyzing plants of the T2 generation of transgenic lettuce. Table 3 summarizes results of analysis of the transgenic trait associated with ColM gene expression in the T2 generation of lettuce plants.

These results show that lettuce transgenic line Ls#6.2 retained significant antibacterial activity in almost all plants of the next generations T1 and T2. Only three plants of T2 generation of Ls#6.2/T1#2 line showed very low or no antibacterial activity (Table 3). One of the plants of T1 generation of lettuce line Ls#21 (namely Ls#21/T1#3) had high antibacterial activity as did all of its T2 descendants (Table 3; Figure 4B). Another T1 plant of line Ls#21 (Ls#21/T1#5) had low antibacterial activity and so did most of its offspring, except line Ls#21/T1#5/T2#4 (Table 3; Figure 4B). This line showed high antibacterial activity despite its parental line showing low activity. This observation could be the result of silencing of the ColM gene in transgenic line Ls#21/T1#5 (T1 generation) and the reversion of silencing in one of the descendants. In addition, the offspring of Ls#21/T1#3/T2#3 was analyzed for ColM expression by immunoblot. All 20 analyzed plants of the T3 generation, which were selected on PPT-containing medium, showed ColM expression (Figure 4C).

For lettuce and mizuna transgenic plants, we selected lines that had high antibacterial activity and calculated the concentration of recombinant ColM in total soluble protein (TSP) extracts generated by addition of 5 volumes of extraction buffer. The content of ColM in transgenic plant extracts (mg/ml) and fresh weight (FW) of plant biomass (mg/g) was calculated via antibacterial activity (Table 4) and by semiquantitative immunoblot analysis (Figure 5).

After finding high antibacterial activity in ColM-containing transgenic plant extracts against E. coli DH10B, we subsequently evaluated the activity of the extracts against a set of reference strains of E. coli, including two EHEC pathogenic serotypes (Figure 6). Аntibacterial activity against the emerging pathogenic serotype O104:H4 was higher than against the other EHEC serotype, O157:H7. In addition, ColM-containing extracts of edible plants, as well as purified ColM, also inhibited the growth of MDR E. coli strains. All three strains, including NCTC13352 that produces beta-lactamase and is therefore resistant to beta-lactam antibiotics, NCTC13919 that produces carbapenemases conferring resistance to carbapenems, and ATCC BAA-2452 that is resistant to both classes of these antibiotics, were sensitive to ColM-containing plant extracts.

Plant-derived purified recombinant ColM was used for determined Minimal Inhibitory Concentrations (MIC) for analyzed E. coli strains. The results showed that ColM was active against most of the E. coli strains used in the study with MIC values ranging from 4.0 to 78.0 ng/ml (Table 5). Only for E. coli strain ATCC 35150 O157:H7 the MIC of recombinant ColM was significantly higher (5 µg/ml). The MIC value serves as the basis for assessing the sensitivity or resistance of the pathogen to an antimicrobial agent. The data presented in Table 4 show that the amount of ColM in transgenic lettuce plants extracts and mizuna line Br#1/T1#12 ranged from 1.06 to 3.36 µg/ml and these values are significantly higher than the MIC of purified recombinant ColM for the studied pathogenic bacterial strains (except for ATCC 35150 O157:H7). These results suggest the feasibility of using the obtained edible transgenic plants expressing ColM to help control foodborne disease.

For edible plants producing pharmaceuticals, the issue of long-term storage conditions without loss of recombinant protein activity is highly important. Therefore, in our studies we also evaluated the antibacterial activity of dried samples of transgenic lettuce and mizuna plants expressing ColM. In these studies, leaf material stored at -80°C for 9 months was subsequently dried at 40° C and antibacterial activity was evaluated after 1 week, 4 weeks and 3 months of storage at room temperature. The results are shown in Figure 7.

Except for an initial drop in ColM antibacterial activity after 9 months of storage at -80°C, no additional activity decrease was observed that could be correlated to either the drying procedure or the length of time of additional storage. These results underscore the advantages of edible transgenic plants over other methods of recombinant ColM production. The levels of specific antimicrobial activity observed in plant extracts (Figure 6), and the stability of ColM under various storage regimes (Figure 7), suggest a new and inexpensive method for producing recombinant pharmaceutical products that does not require costly protein purification steps and is suitable for transportation and long-term storage without loss of activity.