Outbreak strains of L. monocytogenes (FSL N1-227, FSL R2-499, FSL N3-013, FSL J1-177, FSL C1-056) developed for use in challenge studies were obtained from Cornell University (Fugett et al., 2006). C. maltaromaticum ATCC-PTA 5313, carnocyclin A producer, and Leuconostoc gelidum UAL187, leucocin A producer, were obtained from the University of Alberta Lactic Acid Bacteria culture collection. All strains were stored in 30% glycerol in media at -80°C until needed.

Prior to use in experiments, cultures were streaked onto All Purpose Tween agar (APT; Becton, Dickinson and Company, Franklin Lakes, NJ, United States), and a single colony was inoculated into broth media. C. maltaromaticum ATCC-PTA 5313 and Leuconostoc gelidum UAL187 were grown anaerobically at 25°C in APT broth and semi-defined Casamino Acids (CAA) broth without Tween^® 80 (Hastings et al., 1991; Balay et al., 2017), respectively. Strains of L. monocytogenes were grown at 32°C in APT broth.

Leucocin A was purified using a protocol described in Balay et al. (2017). Carnocyclin A was purified using a modified protocol of Martin-Visscher et al. (2008). C. maltaromaticum ATCC PTA-5313 was grown in 1 L APT broth at 25°C for 24 h. For purification, the culture, including the cells, was loaded onto an Amberlite^® XAD-16N column (100 g/L; 2.5 cm × 50 cm; 5 ml/min; Sigma-Aldrich^®, St. Louis, MO, United States) at 6°C. Resin was pre-conditioned with 100% isopropanol (IPA) and rinsed with sterile distilled water (dH₂O). The column was washed sequentially at a flow rate of 10 ml/min with 500 ml of dH₂O, 500 ml of 30% (v/v) ethanol (EtOH), 500 ml of 40% (v/v) IPA, 500 ml 70% (v/v) IPA, and carnocyclin A was eluted with 500 ml 70% (v/v) IPA, pH 2. The fraction was concentrated using a Buchi^® rotary evaporator (Brinkman Instruments, Westbury, NY, United States) at 30°C under vacuum to 10 ml. The concentrated fraction was loaded onto a Sep-Pak^® Vac 12cc (2 g) C18 cartridge (Waters, Milford, MA, United States), which had been pre-washed with 50 ml methanol followed by 50 ml dH₂O. The column was washed consecutively with 50 ml of dH₂O, 50 ml of 30% (v/v) EtOH, 50 ml of 30% (v/v) acetonitrile (ACN), 50 ml 40% (v/v) IPA, 50 ml 70% (v/v) IPA, and carnocyclin A was eluted with 100 ml 70% (v/v) IPA, pH 2. This fraction was concentrated to 2 ml using a rotary evaporator and lyophilized (Labconco^® FreezeZone, Kansas City, MO, United States). The lyophilized sample was suspended in 5 ml of dH₂O for reverse-phase high performance liquid chromatography (RP-HPLC; Beckman System Gold, Beckman Coulter, Inc., Mississauga, ON, Canada). Samples of 1 ml were injected onto a C8 column (5 μm particle size, 10 mm × 250 mm, Vydac 208TP510, Hichrom, Berkshire, United Kingdom). The RP-HPLC mobile phase consisted of (A) IPA and (B) H₂O, each containing 0.1% trifluoroacetic acid. A gradient RP-HPLC method was used, consisting of a 8 min hold at 20% A; 30 min increase of A from 20 to 86%; 5 min hold at 86% A; and a 3 min decrease from 86 to 20% A, with a flow rate of 1 ml/min and ultraviolet detection set at 220 nm. Carnocyclin A eluted as a sharp peak (retention time 23 to 24 min). After purification, all carnocyclin A fractions were combined, lyophilized and stored at -20°C until needed. For experimental use, carnocyclin A was suspended in dH₂O to the desired concentration.

The purified samples were subjected to matrix-assisted laser desorption ionization-time of flight mass spectrometry (MALDI-TOF) to confirm the presence of leucocin A or carnocyclin A.

For growth experiments, basal media supplemented with individual carbohydrates was used. For 1 L of basal medium the following formulation was used: 10 g Proteose Peptone No. 3 (Becton, Dickinson and Company), 10 g Beef Extract (Becton, Dickinson and Company), 5 g Yeast Extract (Becton, Dickinson and Company), 1 g Tween^® 80, 2 g disodium phosphate, 1 ml sterile basal media salt solution [10 g magnesium sulfate, 5 g manganese sulfate dissolved in 100 ml dH₂O], and for each liter of basal medium, 20 g of a carbohydrate (glucose, sucrose, fructose, mannose, or cellobiose) was added.

To determine growth kinetics, five series of four wells in separate 96 well microtiter plates(s) for each strain were filled with 50 μl of each of the five basal media supplemented with individual carbohydrates. An additional 100 μl of each basal media was added to the first well of each series as a negative control. For each microtiter plate, an overnight culture of one strain of L. monocytogenes was centrifuged, suspended in sterile 0.1% (w/v) peptone water (Becton, Dickinson and Company) and serially diluted with peptone water with the final dilution made in basal media. To the second well of each series, 50 μl of basal media and 50 μl of one strain of L. monocytogenes was added. The third series of wells contained basal media, L. monocytogenes and leucocin A; and the fourth series included L. monocytogenes and carnocyclin A. The final concentration of bacteriocins was 3.3 mM and cell density was ∼10⁴ CFU per well, confirmed by plate count on APT agar. The concentration of bacteriocin was chosen as it is equal or higher than the MIC against L. monocytogenes determined on a critical dilution assay on agar after incubation for 24 h (data not shown; Balay et al., 2017) but did not prevent growth of resistant derivatives upon prolonged incubation. Preliminary experiments indicated that the response of the five strains of L. monocytogenes to challenge with bacteriocins was similar when pre-cultures were prepared with APT (glucose) or supplemented basal media containing carbohydrates other than glucose.

Each microtiter plates was covered with a clear optical adhesive film (MicroAmp, Applied Biosystems, Thermo Fischer Scientific, Burlington, ON, Canada) to prevent evaporation, loss of culture or contamination and placed in a spectrophotometer (Multiskan Ascent, Thermo Electron Corporation, Shanghai, China) and incubated at 25°C for 72 h. Each plate was shaken every 2 h for 10 s at 700 RPM, and the optical density (630 nm) of each well was recorded. Growth curves were generated using Mulitskan Ascent Software (Version 2.6, Thermo Fischer Scientific). The experiment was replicated independently three times.

To characterize phenotypical differences of the subpopulation that grew in the presence of either bacteriocin from the growth experiments (see section “Growth Kinetics of L. monocytogenes Grown in Different Carbohydrates Treated with Leucocin A or Carnocyclin A”), two strains were selected for further testing; FSL R2-499 and FSL C1-056. Both strains of L. monocytogenes were grown in basal media supplemented with either mannose or cellobiose and treated with 3.3 mM leucocin A or carnocyclin A.

To compare growth kinetics of the original cultures compared to that each subpopulation first a growth experiment under similar conditions as outlined in Section “Growth Kinetics of L. monocytogenes Grown in Different Carbohydrates Treated with Leucocin A or Carnocyclin A,” followed by isolation of individual colonies from the subpopulation that grew, which then were used in a second growth experiment with the same conditions as the first growth experiment.

To determine the growth kinetics for the original cultures FSL R2-499 and FSL C1-05, two series (one for each carbohydrate) of four wells was used in separate 96 well microtiter plates(s) for each strain. The total well volume used was 200 μl. The first well of each series was filled with 200 μl of each basal media supplemented with individual carbohydrates (mannose or cellobiose) as a negative control. For each microtiter plate, an overnight culture of one strain of L. monocytogenes was centrifuged, suspended in sterile 0.1% (w/v) peptone water (Becton, Dickinson and Company) and serially diluted with peptone water with the final dilution made in basal media. To the second well of each series, 150 μl of basal media and 50 μl of one strain of L. monocytogenes was added. The third series of wells contained basal media, L. monocytogenes and leucocin A; and the fourth series included L. monocytogenes and carnocyclin A. The final concentration of bacteriocins was 3.3 mM and cell density was ∼10⁶ CFU per well, confirmed by plate count on APT agar. Each microtiter plates was covered with a clear optical adhesive film (MicroAmp, Applied Biosystems, Thermo Fischer Scientific, Burlington, ON, Canada) to prevent evaporation, loss of culture or contamination and placed in a spectrophotometer (Multiskan Ascent, Thermo Electron Corporation, Shanghai, China) and incubated at 25°C for 40 h. Each plate was shaken every 2 h for 10 s at 700 RPM, and the optical density (630 nm) of each well was recorded. Growth curves were generated using Mulitskan Ascent Software (Version 2.6, Thermo Fischer Scientific).

Following the first growth experiment seals were removed from microtiter plates and a loopful of culture was obtained from each well of the both series. The culture was then streaked on APT agar to obtain single colonies. Plates were incubated at 32°C for 24 h before individual colonies were picked and cultured into APT broth at 32°C for 24 h. The same growth experiment then followed for the isolated subpopulations as outlined above with the original cultures. The experiment was replicated independently four times.

Growth curves were fitted to a modified logistic model (Zwietering et al., 1990) to assess how maximal growth rates (µ) and lag phase (λ) were affected by the presence of leucocin A or carnocyclin A when L. monocytogenes was grown in different carbohydrates. SigmaPlot 12.5 was used for all curve fit procedures.

Data were analyzed using the PROC ANOVA of SAS Studio (Release 3.5 University Edition, SAS Institute, Inc., Cary, NC, United States) with strain, carbohydrate and bacteriocins as categorical variables and lag phase and growth rate as dependent variables. To determine significant differences among means, Tukey test with a 95% confidence interval was used to determine significant interactions. To determine the specific impact of the main effects of strain, carbohydrate or bacteriocin, one-way ANOVA was done with differences among means determined with a Tukey test.

The novel protocol for purification of leucocin A and carnocyclin A recovered 2.3 and 4.8 mg/L of leucocin A and carnocyclin A, respectively. Purification yields of leucocin A and carnocyclin A were 5-and 2.5-fold higher, respectively, than the yields reported by Hastings et al. (1991) and Martin-Visscher et al. (2008). The molecular mass of the purified leucocin A and carnocyclin A was 3,930.03 and 5,863.50 Da, respectively, matching prior reports (Hastings et al., 1991; Martin-Visscher et al., 2008).

Individual growth curves were fitted to a modified logistic model with an R² ≥ 0.99 with an overall fit of R²= 0.995 as shown in the unity plot (Figure 1). Each of the experimental growth curves with the modeled curve for each growth condition is shown in Supplementary Figures S1–S7. L. monocytogenes FSL N1-227 grown in fructose with carnocyclin A and FSL J1-177 grown in glucose with carnocyclin A gave a growth curve that did not precisely fit the sigmoidal shape of the logistic model (Supplementary Figures S1, S4), but the R² was > 0.99. The modeled growth curves for each strain of L. monocytogenes grown in the presence of each carbohydrate and leucocin A or carnocyclin A are shown in Figure 2. Leucocin A at a concentration of 3.3 mM did not inhibit the growth of any of the strains of L. monocytogenes in the presence of any of the tested carbohydrates (Figure 2A), demonstrating that all strains developed resistance against leucocin A. Only two of the five strains shows an extended lag phase in presence of leucocin A in presence of some but not all carbohydrates (Table 1), indicating that the development of resistance is depended on the strain and the carbohydrate source. Carnocyclin A extended the lag phase and reduced the growth rate of all strains; the extent of inhibition again depended on individual strains and/or carbohydrates (Figure 2B).

Growth rate was plotted as a function of lag phase duration for each strain grown in each carbohydrate based on bacteriocin treatment to assess trends in the data (Figure 3). In the absence of either bacteriocin, the lag phases of all strains of L. monocytogenes were of similar duration, although there was greater variation among their subsequent growth rates (Figure 3). In the presence of carnocyclin A this situation was reversed, in that the lag phase of L. monocytogenes varied considerably, but their subsequent growth rates were broadly similar. In the presence of leucocin A there was less variation among lag phases, and among subsequent growth rates, than in the absence of bacteriocin, or in the presence of carnocyclin A (Figure 3).

The two-way interaction means for the growth rate and lag phase of cultures grown in media with individual carbohydrates, across all bacteriocin treatments are shown in Figure 4. There were significant interactions between strain and carbohydrate that resulted in differences in growth rate and lag phase for individual carbohydrates (Figures 4A,B). For FSL N1-227 and FSL N3-013, the maximal growth rate was significantly lower when strains were grown in sucrose, whereas growth in cellobiose resulted in a significantly higher growth rate for all strains (Figure 4A). The lag phases of L. monocytogenes FSL C1-056 and FSL J1-177 were significantly longer when grown in mannose, glucose, and fructose (Figure 4B).

There were significant two-way interactions between strain and bacteriocin treatment, and carbohydrate and bacteriocin treatment (Figures 5, 6, respectively).

For each strain, regardless of carbohydrate, leucocin A resulted in a significantly lower growth rate than the untreated control, while carnocyclin A resulted in a significantly lower growth rate than strains treated with leucocin A or the untreated control (Figure 5A). When treated with leucocin A, FSL R2-499 had the fastest growth rate (P < 0.05) and FSL N3-013 had the slowest growth rate (P < 0.05; Figure 5A). All strains had the same lag phase when not treated with bacteriocin, but treatment of strains FSL C1-056 and FSL J1-177 with leucocin A resulted in a significantly longer lag phase than that observed for the untreated control (Figure 5B). Treatment with carnocyclin A significantly increased the lag phase for all cultures compared to the untreated control (Figure 5B). The lag phase of strains FSL N1-227, FSL C1-056, and FSL J1-177 were significantly longer when treated with carnocyclin A compared to those treated with leucocin A.

Regardless of strain, growth in cellobiose resulted in the highest growth rate (P < 0.05) within each bacteriocin treatment (Figure 6A). Growth in media with carnocyclin A and any of the carbohydrates significantly reduced growth rates (Figure 6A). Sucrose resulted in the slowest growth rate for the untreated control and cells treated with leucocin A, whereas fructose resulted in the slowest growth rate for all strains treated with carnocyclin A (Figure 6A). Compared to the untreated control, there was no difference (P > 0.05) in lag phase of cultures treated with leucocin A regardless the carbohydrate, but the lag phase increased (P < 0.05) when cultures were treated with carnocyclin A (Figure 6B). Growth in glucose, fructose, or mannose increased (P < 0.05) the lag phase in media with carnocyclin A compared to cultures grown in the presence of leucocin A (Figure 6B).

The impact of strain, bacteriocin or carbohydrate was determined by one-way ANOVA (Table 1). Strain did not affect growth rate when cultures were grown in glucose or cellobiose without bacteriocin (Table 1). Growth rate was significantly reduced for strain FSL N1-227 when grown in sucrose, fructose, and mannose compared to the untreated control group (Table 1). Strain had no effect on the lag phase for the untreated control group regardless of the carbohydrate (Table 1).

Leucocin A significantly decreased the growth rates of all strains when grown in glucose, fructose, or cellobiose, except for FSL R2-499 (Table 1). Strain FSL R2-499 was not affected by leucocin A except when grown in mannose, which significantly decreased the growth rate compared to the untreated control (Table 1). When FSL N1-227 was grown in mannose it was resistant to leucocin A as there was no effect on growth rate compared to the untreated control, which was the exact opposite for FSL R2-499 (Table 1). Strain FSL C1-056 had the slowest growth rate when grown in fructose, mannose, or cellobiose, and treated with leucocin A (Table 1). However, when FSL C1-056 was grown in glucose or sucrose it was one of the fastest growing strains despite having a significantly increased lag phase compared to the untreated control in sucrose and mannose (Table 1). In all carbohydrates, when FSL C1-056 was treated with leucocin A there was a significant increase in lag phase (Table 1).

Carnocyclin A significantly decreased the growth rates of all strains compared to the untreated control, but fructose and cellobiose caused variation in growth rates among strains (Table 1). Carnocyclin A significantly increased the lag phase for all strains when grown in glucose, fructose, or cellobiose compared to the untreated control. There was no difference in lag phase across strains when grown in glucose and treated with carnocyclin A. When grown in sucrose FSL R2-499 had increased resistance to carnocyclin A as the lag phase was not affected by bacteriocin treatment. Strain FSL C1-056 had an increase in lag phase when grown in sucrose in the presence of carnocyclin A compared to strains FSL N1-227 and FSL R2-499. Similarly, FSL C1-056 had the longest lag phase compared to all other strains when grown in fructose and treated with carnocyclin A. Strain FSL C1-056 had significantly longer lag phase when treated with carnocyclin A and grown in mannose when compared to strains FSL R2-499and FSL N3-013 (Table 1).

The growth kinetics of cultures was analyzed after growth in the presence of both bacteriocins to compare the maximal growth rate and lag phase to that of original cultures.

The original cultures prepared in absence of bacteriocin grew in both sugars (Figures 7A,B and Supplementary Table S1); growth parameters were slightly modified, which reflects the different culture conditions for the inoculum of the respective cultures. L. monocytogenes FSL C1-056 had longer lag phases in both carbohydrates when treated with either bacteriocin compared to FSL R2-499 (Figure 7 and Supplementary Table S1). Carnocyclin A reduced the maximal growth rate for both strains grown in both carbohydrates compared to that of the control (Figure 7 and Supplementary Table S1).

After pre-adaptation of L. monocytogenes to bacteriocins, neither carnocyclin A nor leucocin A increased the lag phase of the cultures (Figure 7 and Supplementary Table S1). This indicates that adaptation to the bacteriocins eliminates or reduces the time for development of resistance upon sub-culture in presence of bacteriocins. One noticeable exception was L. monocytogenes FSL C1-056. The lag phase of this strain in mannose or mannose–carnocyclin A did not differ after pre-adaptation to carnocyclin A; however, the strain still showed an extended lag phase to carnocyclin A after adaptation in presence of cellobiose (Supplementary Table S1).