Antimicrobial Efficacy of Pelargonic Acid Micelles against Salmonella varies by Surfactant, Serotype and Stress Response

The antimicrobial properties of Pelargonic acid (PA), a component of tomatoes, makes it an attractive candidate as a food additive and sanitizer. The antimicrobial efficacy of PA emulsions generated using surfactants: Tween 80, Triton X100, Sodium Dodecyl Sulfate (SDS) and Quillaja Saponin was evaluated against Salmonella serotypes Newport, Oranienburg and Typhimurium. Micelle/dropletsize, and minimal inhibitory concentrations (MIC) were determined. Surfactant type and concentration significantly influenced the antimicrobial efficacy of PA (p < 0.05). Overall, Salmonella Newport was the most (p < 0.05) susceptible serotype to PA emulsions. PA emulsions generated with 1.00% SDS had the highest (p < 0.05) antimicrobial activity, with MIC of 7.82 mM against S. Newport and 15.62 mM against S. Oranienburg/S. Typhimurium, respectively. Addition of PA to Trypticase Soy Broth resulted in a decreased growth rate and an increased lag phase duration. Cells exposed to PA formed elongated filaments (>5 µm). Additionally, Salmonella serotypes Typhimurium and Newport also formed floccular biofilms. PA emulsions at a concentration of 31.25 mM generated using 1% SDS and 1% Quillaja saponin resulted in >6 log CFU/ml reduction in Salmonella population. Althought all PA emulsions evalauted inhibited Salmonella, morphological changes to this antimicrobial varied substantially among the Salmonella serotypes tested.

www.nature.com/scientificreports www.nature.com/scientificreports/ Safe (GRAS) status and is a component of the tomato exometabolome. Recently the antimicrobial activity of PA against Salmonella was demonstrated 11 indicating that the fatty acid had better antimicrobial activity when dispersed as a micelle using Quillaja saponin than its unmodified form, hence highlighting its potential as a GRAS antimicrobial food additive or as a sanitizer.
Though a variety of fatty acids demonstrate antimicrobial activity, their efficacy is often diminished by their poor solubility in water and phase separation in an aqueous medium 11 . Emulsification of fatty acids can increase their miscibility and dispersion in water 12 , resulting in improved antimicrobial activity 11,13 . Surfactants such as Tween 80, SDS and Quillaja saponin, a triterpene glycoside from the bark of the Quillaja saponaria Molina can be used to produce emulsions of fatty acids and essential oils that have antimicrobial properties. Surfactants can influence the antimicrobial efficacy of an emulsion due to their charges and their placement on the colloidal particle, thereby influencing the particle size and the amount of antimicrobial compound that is in contact with the target bacterial cell [14][15][16][17] .
The surfactant choice and concentration can either, synergistically strengthen or impede the antimicrobial activity of an emulsion 16,18 . Previous research indicated that exposure to sublethal concentration (20 mM) of PA induced a transitional morphological change (filament formation) in Salmonella 11,19,20 . Formation of diverse phenotypes by bacteria in challenging environments is a strategy to overcome antimicrobial hurdles 21,22 . Determining morphological changes in Salmonella during the exposure to PA emulsions formed using different surfactants could provide a better understanding of the antimicrobial's mechanistic action of and means of adaptation by the pathogen.
The minimal inhibitory concentrations (MIC) of aqueous emulsions of PA generated using Tween 80, Triton X100, SDS and Quillaja saponin were determined against Salmonella serotypes Newport isolated from cattle, Oranienburg isolated from nut and a reference Typhimurium strain for antimicrobial testing. Bactericidal efficacy of the most potent emulsions was evaluated against a cocktail of the Salmonella serotypes after treatment durations of 30 s and 5 min. Growth kinetics and cell physiology of the Salmonella serotypes were tested in presence of inhibitory PA emulsions to study adaptative mechanisms to the antimicrobial. The results of this study are intended to bridge the current knowledge gap on the role of surfactant type in antimicrobial efficacy of PA to aid in the development of fatty acid-based washes and rinses for the food industry.

Materials and Methods
Micelle and droplet size measurements. Droplet size distribution and polydispersity (intensity-based size distribution) measurements of emulsions were conducted at 20 °C by Dynamic Light Scattering that examines the fluctuations in light scattering due to brownian motion of the emulsion droplets, using a Zetasizer Nano ZS device (Malvern, Southbough, MA). Briefly, 5 ml of emulsion samples were diluted in 50 ml of deionized water. Then, 10 ml of the diluted samples were filtered through a 0.40 µm nylon filter (Syringe-filters, Cole Parmer, UK) and vortexed. Mean hydrodynamic diameter of emulsion droplets was calculated via cumulative analysis (z-average) and the size distributions were defined by the polydispersity index (PdI) and the size distribution graph.
Bacterial Culture. Salmonella enterica subsp. enterica serotype Newport (11590 K, beef isolate, GFP labelled), S. Oranienburg (1839, pecan isolate) and S. Typhimurium (CDC 6516-60/ATCC 14028) were obtained from the Center for Food Safety, University of Georgia, Culture Collection. Cultures were revived by two successive 24 h transfers in 10 ml Tryptic Soy Broth (TSB, Acumedia Neogen Corporation. Lansing, MI) with antibiotics (100 µg/ ml Ampicillin for S. Newport and Typhimurium and 50 µg/ml Nalidixic acid for S. Oranienburg), at 35 °C. Then, the cultures were streaked on Tryptic Soy Agar (TSA, Acumedia, Neogen Corporation. Lansing, MI) with antibiotics and incubated at 35 °C for 24 h. Individual colonies were confirmed serologically using a latex agglutination (Oxoid, Ogdensberg, NY). Isolated colonies were then streaked on XLT4 (Xylose Lysine Deoxycholate agar with tergitol 4, Acumedia, Neogen Corporation, Lansing, MI) at 35 °C for 24 h and observed for typical black colonies. Stock cultures were prepared by streaking a colony of each Salmonella serotype on TSA with antibiotics and incubating the plate at 35 °C for 24 h. The colonies were then scraped with a sterile loop and suspended in Phosphate Buffered Saline (PBS) and adjusted to a population of ca. 6 Log CFU/ml. Emulsion preparation. A 1 M stock emulsion of n-Pelargonic acid, 97% (Pelargonic acid, Acros Organics, New Jersey, USA) in water was prepared using TritonX-100, (ICN Biomedicals, Inc., Aurora, OH), Tween 80 (Fisher Chemical, Fair Lawn, NJ), Sodium Dodecyl Sulfide (SDS, Fisher Chemical), Sodium lauroyl sarcosinate (Sarkosyl, Fisher Chemical) and saponin from quillaja bark (Sigma-Aldrich, St. Louis, MO) at concentrations of 1%, 0.1% and 0.01% (w/v). Stock emulsions (1 M) were prepared by adding 1.58 g of PA to 10 ml of water with 1%, 0.1%, and 0.01% (w/v) of surfactant. The mixture was agitated on a vortexer (Vortex-2 Genie, Scientific Industries, Bohemia, NY, USA) at maximum speed setting of 10 (3200 rpm) for 1 min to form the emulsion.

MIC determination.
Minimal inhibitory concentration of the emulsions was determined by the modified 96 well plate Resazurin assay 23 . Briefly, stock solutions of the PA emulsions were serially diluted into Iso-sensitest broth (Oxoid Ltd., Hants, UK) containing Resazurin (Biotium, Hayward, CA) (0.001 g/ml) and inoculated with ca. 5 Log CFU/ml of Salmonella. Serial dilutions were performed to obtain PA concentrations ranging from 500 mM to 2.73 mM of each emulsion. The 96 well plates were incubated for 24 h at 37 °C and the wells containing the serial dilutions of the emulsions were observed for a change in color from blue to pink indicating bacterial growth. The lowest dilution at which no bacterial growth was determined (blue well) was considered the minimum inhibitory concentration. The MICs of the individual surfactants at concentrations of 1%, 0.1% and 0.01% (w/v) and PA without surfactant that was suspended in water by vortexing for 1 min were determined as comparative controls using the modified 96 well plate Resazurin assay. Each test was performed three times and an average of the three MIC values were used. www.nature.com/scientificreports www.nature.com/scientificreports/ Bactericidal activity. Concentration of PA emulsions that were inhibitory were evaluated for their bactericidal activity against Salmonella serotypes. After determining the MIC as described in the previous section, wells containing the lowest inhibitory concetration of the pelargonic acid emulsion were neutralized using Dey-Engley neutralizing broth (Neogen Corp., Lansing, MI) and enumerated on TSA by spread plating with appropriate dilutions. Emulsions that resulted in a reduction of bacterial populations ≥3 log CFU/ml were considered bactericidal 24 . Each experiment was performed three times and an average of the three MIC values were used.

Antimicrobial efficacy against Salmonella cocktail. A PBS suspension consisting of a cocktail of
Salmonella serotypes Newport, Typhimurium and Oranienburg at a concentration of 9 Log CFU/ml was exposed to 31.25 mM and 15.62 mM of PA-Quillaja Saponin (1%) and PA -SDS (1%). The concentrations of emulsions selected were inhibitory to all three serotypes of Salmonella. Exposure durations tested were for 30 s and 5 min. Following exposure, the fatty acid was neutralized using Dey-Engley neutralizing broth and enumerated by spread plating after appropriate diltuons. Each experiment was performed three times.
Growth Kinetics of Salmonella in presence of PA emulsion. Salmonella serotypes Newport, Typhimurium and Oranienburg were grown in presence of TSB at 37 °C with 15.6 mM of PA emulsion generated using 1% Quillaja saponin and TSB with 7.80 mM of PA emulsion generated using 1% SDS. The concentration of the emulsion used was inhibitory to S. Newport (p < 0.05) and was sub-inhibitory to S. Typhimurium and S. Oranienburg. Initial population of 5 log CFU/ml of Salmonella serotypes were grown in 50 ml centrifuge tubes (VWR International, Radnor, PA) of TSB at 37 °C. The cells were neutralized using Dey-Engley neutralizing broth and enumeratd at 0, 2, 4, 8, 12, 16, 20 and 24 h by droplet plate method on TSA with appropriate serial dilutions. Controls for the experiment were grown without the presence of PA emulsions with identical culture conditions. Growth curves were plotted and analyzed using MATLAB (The Math Works, Natick, MA, Version 2017). Baranyi model 25 was used as the primary model to predict the bacterial population as a function of time.

Assessment of morphology. Cell morphology of Salmonella serotypes grown in TSB amended with
15.62 mM of PA emulsion generated using 1% Quillaja saponin and TSB with 7.80 mM of PA emulsion generated using 1% SDS at, 37 °C after 24 h of growth were evaluated through phase contrast microscopy (Nikon Eclipse Ci-L; Nikon Corp., Japan). Wet mount preparations of 20 µl and studied using a Nikon 100X Plan Apo objective (Nikon Eclipse 50i; Nikon Corp., Japan) with the Nikon Elements software through with a scale of 5 µm to differentiate and enumerate cells that were larger than 5 µm (filaments, F) and cells that were smaller than 5 µm (regular, R). and aggregates of cells (>5 cells) were considered as floccules. Direct counts of filamentous cells and regular-sized cells were determined from 3 experimental replicates (3 fields per slide) for each treatment and time point. The average counts of filamentous cells (F) and regular cells (R) were used to determine the F:R ratio. Each field had between 50 to 200 cells.
The structure of the floccular biofilms was studied using Confocal Laser Scanning Miroscopy (CLSM). Floccular biofilm was further confirmed by staining with FilmTracer calcein red-orange (Molecular Probes, Eugene, OR) biofilm stain. Speciment staining with FilmTracer calcein red-orange was conducted according to manufacturer's instructions. After an hour of incubation at room temperature in the dark, the sample was observed using a Zeiss 880 laser scanning confocal microscope with a 63×/1.4 NA oil immersion Plan Apochromatic objective at 576/590 nm.

Statistical analysis.
The experiments for this study were designed using a randomized blocked design factorial treatment arrangement, repeated measures with sampling, blocked on replication ( Table 1). The statistical analysis of the data was performed with analysis of variance (ANOVA) using JMP Pro 14 (SAS Institute, Cary, NC). Tukey-Kramer honestly significant difference test was used to determine differences between mean values of the MIC and droplet size of different tests at a 95% confidence level.
To evaluate the correlation between antimicrobial efficacy of PA and average micellar/droplet diameter of PA emulsions, the ordinary least squares regression analysis was applied to the MIC values and the corresponding average micellar/droplet diameter.

Results
Micelle characteristics. The micelle size of PA emulsions varied depending on the type and concentration of surfactant. In PA emulsions generated using SDS, micelle diameter was positively associated with surfactant concentration, specifically 1.00% SDS emulsion had the largest (3220.50 ± 133.50 nm) micelles, significantly (p < 0.05) smaller micelles (1499.67 ± 112.90 nm) in emulsions produced with 0.10% SDS, and those generated with 0.01% SDS had the smallest micelles diameters (289.70 ± 27.44 nm) ( Table 1). In contrast, in all other surfactants tested, no positive association between micelle diameter and surfactant concentration was observed. Nevertheless, the diameter of micelles varied significantly (p < 0.05) with surfactant concentration, except for Quillaja saponin and TritonX 100 at 1.00 and 0.10% (Table 1). The antimicrobial activity of PA emulsions generated using Quillaja saponin against all three Salmonella strains indicated that a positive correlation existed between higher micellar size and inhibitory activity. This correlation was highest for S. Newport as compared to other two serotypes. For Tween treatment, there was very weak correlation for S. Newport and S. Oranienburg, whereas the 'r' value was 0.38 for S. Typhimurium. For Triton, the correlation coefficient ranged between −0.25 and 0.10, whereas for SDS, the correlation coefficient was strongly negative for S. Newport (−0.86) (Fig. 2). The correlation coefficient could not be calculated for the other two strains as the MIC values were constant for all the experimental micellar sizes.
The average pH of 1 M stock solutions of the pelargonic acid emulsions generated using the various surfactants was 3.55 ± 0.18. Polydispersity index varied with both, the type and concentration of the surfactant. The positive www.nature.com/scientificreports www.nature.com/scientificreports/ association between the value of the polydispersity index and surfactant concentration, was observed only for SDS (Table 1).

Inhibitory activity of PA emulsions against Salmonella. The MIC of PA against all three Salmonella
serotypes was 125 ± 0.00 mM. Quillaja Saponin did not inhibit Salmonella growth at 1%, 0.1% and 0.01%. When emulsions of PA were formed using Quillaja saponin, the antimicrobial activity, as denoted by a decrease in MIC, increased significantly (p < 0.05). Of the three serotypes of Salmonella tested, S. Newport was most susceptible to PA (p < 0.05). The lowest concentration of PA emulsions that was inhibitory against S. Newport was 7.82 ± 0.00 mM and was produced using 1% SDS (p < 0.05). The MIC of PA emulsion produced using 1% SDS for S. Oranienburg and S. Typhimurium, was 15.62 ± 0.00 mM (Fig. 1). No significant difference in the MIC of PA emulsions produced using 1%, 0.1% and 0.01% SDS was observed against S. Oranienburg and S. Typhimurium (Fig. 1).
No significant differences in susceptibility was observed among the Salmonella serotypes when treated with 0.01% Quillaja saponin generated PA emulsions (MIC of 27.77 ± 6.88 mM) (Fig. 1).
Non-ionic surfactants, Tween 80 or Triton X 100 were not as effective as SDS and Quillaja saponin in improving the antimicrobial activity of PA. Emulsions of PA generated using 1.00% Tween 80 or 1.00% Triton X 100 were inhibitory to the Salmonella serotypes tested at 31.25 ± 9.02 mM (Fig. 1). PA emulsions generated using 0.1% Tween 80 had an average MIC of 48.61 ± 16.47 mM against the Salmonella serotypes tested. MIC of PA emulsions  www.nature.com/scientificreports www.nature.com/scientificreports/ generated using 0.1% SDS and Quillaja saponin were significantly (p < 0.05) more efficacious against Salmonella than those of Tween 80 and Triton X100 (39.62 ± 15.37 mM) (Fig. 1). S. Newport was significantly more susceptible to PA emulsions generated using 0.01% Tween 80 (39.79 ± 15.51 mM) and TritonX 100 (39.62 ± 15.37 mM) than other serotypes (p<0.05). The average MIC of PA emulsions generated using 0.01% Tween 80 and Triton X100 against S. Oranienburg and Typhimurium was 57.29 ± 12.16 mM.
Growth kinetics of Salmonella. Salmonella serotypes Newport, Oranienburg and Typhimurium were grown in TSB containing PA emulsions generated using Quillaja saponin and SDS. The emulsions were used at a concentration that was inhibitory to S. Newport, 5.62 mM, PA with 1% Quillaja saponin and 7.82 mM PA with 1% SDS. Growth rates and physiological adapatations to PA were compared among the three serotypes of Salmonella to understand the effects of PA on bacterial growth and the physiological adaptations undergone by the organisms to overcome PA associated challenge. Figure 3 shows the fitted Baranyi models ( Baranyi and Roberts, 1994) for Salmonella serotypes Newport, Oranienburg, Typhimurium, grown in TSB (control) and TSB amended with 15.62 mM PA with 1% Quillaja saponin and 7.82 mM PA with 1% SDS. While the control samples demonstrated a full growth curve with the maximum population density of 10.57 log CFU/ml, none of the PA treated samples could reach the maximum population density after the initial growth. As such, a full Baranyi model was fitted to the control sample growth data, whereas a reduced Baranyi model was fitted to the growth data showing a growth pattern in both the treated www.nature.com/scientificreports www.nature.com/scientificreports/ samples. As it is evident from the growth curves, the lag phase duration was the minimum for control samples for all three strains (2.90, 2.72, and 2.54 h for S. Newport, S. Typhimurium, and S. Oranienburg, respectively). For the same strains and for samples treated with SDS and pelargonic acid, the lag phase durations were calculated as 3.36, 2.53, and 5.84 h, respectively. The maximum growth rates (log CFU/h) for these treatments were estimated as 0.99, 0.97, and 0.58, as compared with the control sample growth rates of 1.26, 1.62, and 0.87 log CFU/h for S. Newport, S. Typhimurium, and S. Oranienburg, respectively. These values suggest that although lag phase duration was not affected much by SDS generated PA emulsion, the maximum growth rate was reduced considerably. The samples treated with Quillaja Saponin generated PA emulsions were further impacted, as reflected by the lag phase durations of 4.84, 1.09, and 5.84 h, and maximum growth rates of 0.42, 0.30, and 0.43 log CFU/h for S. Newport, S. Typhimurium, and S. Oranienburg, respectively. All three strains had a considerable drop of population within 12-hours of treatment with either PA emulsion. www.nature.com/scientificreports www.nature.com/scientificreports/ Morphological diversity. Filament formation was significantly influenced by the type of Salmonella serotype (p < 0.05) exposed to PA emulsions with S. Oranienburg forming the significantly higher number of filaments (p < 0.05) ( Table 2). Exposure of S. Oranienburg cells to 7.80 mM of PA emulsion generated using 1% SDS for 24 h resulted in the highest number of filament formation with an F:R of 2.18 ± 1.53 (p < 0.05). S. Typhimurium formed the lowest number of filaments after exposure to the PA emulsions generated using both surfactants (p < 0.05) with F:R of 0.24 ± 0.18 and 0.07 ± 0.03 for Quillaja saponin and SDS generated PA emulsions respectively ( Table 2).
Floccule formation was significantly influenced by both the serotype (p < 0.0001) of Salmonella and the PA emulsions (p < 0.05) the cells were exposed to. Exposure of S. Typhimurium to 7.80 mM of PA emulsion generated using 1% SDS resulted in significantly higher floccules than all other treatments (p < 0.05) that resulted in floccule formation. S. Oranienburg did not form floccules upon exposure to both SDS and Quillaja saponin generated PA emulsions (Table 2, Fig. 7).
Microscopy analysis of PA-challenged S. Typhimurium cells after 24 h growth in TSB indicated that the cells formed aggregates and were encased in a film of exopolymeric substance (Fig. 4). Observation using confocal laser scanning microscopy using optical Z sectioning demonstrated that cellular aggregates (herein reffered as floccules) had a thickness of 10-20 µm and consisted of Salmonella cells and EPS (Figs. 4-panel E and F, 5, 6).    www.nature.com/scientificreports www.nature.com/scientificreports/ CLSM micrographs of the stained floccules further confirmed that both filaments and regular size cells were constitutents of these floccular biofilms (Figs. 5, 7).
Salmonella Newport was the most susceptible serotype to the concentration of PA used in TSB. In response to both Quillaja saponin-and SDS-generated emulsions, S. Newport formed filaments (> 5 µm) and regular sized cells (Figs. 4, 7). The filaments formed by S. Newport were thicker than filaments formed by other serotypes with certain filaments displaying cellular dysplasia indicating structural damage to certain filaments (Fig. 4, panel b, c).
Antimicrobial efficacy of PA emulsions against a Salmonella cocktail. Emulsions containing 31.25 mM PA generated using 1% SDS was most effective against the Salmonella cocktail after an exposure duration of 30 s, resulting in a 4.7 log CFU/ml reduction (p < 0.05). Quillaja saponin (1%) based PA emulsions (31.2 mM) had significantly lower antimicrobial efficacy and reduced Salmonella by 2.15 log CFU/ml (p < 0.05)  www.nature.com/scientificreports www.nature.com/scientificreports/ (Fig. 8). No significant differences were observed in antimicrobial activity between PA emulsions generated using 1% SDS (1.62 log CFU/ml) or 1% Quillaja saponin (0.77 log CFU/ml) when used at a concentration of 15.62 mM (Fig. 8).
After 5 min of exposure no significant differences were observed in antimicrobial activity between PA emulsions generated using 1% SDS (6.85 log CFU/ml) or 1% Quillaja saponin (6.21 log CFU/ml) when used at a concentration of 31.25 mM (Fig. 9). The antimicrobial efficacy of PA emulsions tested at 31.25 mM was significantly higher than PA emulsions tested at 15.26 mM (p < 0.05) (Fig. 9). PA (15.26 mM) emulsions generated using 1% SDS or Quillaja saponin resulted in 4.61 and 3.87 log CFU/ml reduction of Salmonella population in the cocktail. www.nature.com/scientificreports www.nature.com/scientificreports/ PA concentration and duration of exposure significantly influenced antimicrobial activity of the emulsions (p < 0.05) (Fig. 9).

Discussion
PA is a fatty acid commonly found in plants such as tomatoes and is an FDA approved food additive 26 . PA was found to have inhibitory activity against Salmonella at a concentration of 125 mM when dispersed in water by physical agitation 11 . When emulsified using Quillaja saponin, the inhibitory activity of PA significantly increased 11 . Hence in the current study, emulsions of PA were formed using a variety of surfactants at different concentrations and evaluated for their inhibitory activity against Salmonella Serotypes Newport, Typhimurium and Oranienburg. Of the surfactants tested in the study Quillaja saponin, SDS and Tween 80 are approved by the FDA as GRAS and TritonX 100 is approved for the dispersal of pesticides and applications associated with food packaging 27 .
All the surfactants used in the current study improved the inhibitory activity of PA against Salmonella in comparison to the control (125 mM) (Fig. 1). The antimicrobial properties of a micelle depends on its availability in the aqueous phase 28 . Emulsions of essential oils generated using surfactants such as Tween 80 or Triton X100, which are non-ionic in nature, had lower antimicrobial efficacy against Salmonella Enteritidis, Escherichia coli O157:H7, and Listeria monocytogenes than the unmodified essential oils [28][29][30] . The current study determined that PA emulsions formed using Tween 80 and Triton X100 had lower (p < 0.05) inhibitory activity against all three Salmonella serotypes, in comparison to Quillaja saponin and SDS (Fig. 1). It was observed that surfactant type and concentration significantly affected both, inhibitory and bacteridical activities of PA (p < 0.05). The concentration of surfactant at which its molecules aggregate to form micelles is known as the critical micelle point (CMC) 31,32 . The CMC of SDS is 2.1 g/l, Tween 80 is 0.013 g/l, Triton X100 is 0.15 g/l and Quillaja saponin is 0.8 g/ l 29,31,33,34 . Surfactants used at 1% were all above the CMC and could have contributed to the antimicrobial activity of micelles. Tween 80, Triton X100 and Quillaja saponin exceeded the CMC at 0.1% and Tween 80 was concordant to the CMC at 0.01%.
Among the surfactants tested, SDS based emulsions had the largest PA micelles, as well as a high level of variation in the size of the colloidal particles generated (as denoted by the polydispersity index) compared to emulsions generated using other surfactants.
S. Newport was generally more sensitive to PA emulsions than serotypes Typhimurium and Oranienburg (p < 0.05). Differences in the sensitivity of Salmonella serotypes to antimicrobial compounds can occur and can result in higher incidences of specific serotypes in certain food commodities 35 . For instance, S. Oranienburg has been isolated from dry spices and has a higher tolerance to natual antimicrobial compounds in spices 36 .  www.nature.com/scientificreports www.nature.com/scientificreports/ Pelargonic acid exposure to bacterial cells at lethal concentrations results in cell membrane damage and lysis 11 , similar to bacterial cells exposed to carvacrol, the active component of Oregano oil 11,37 . Analysis of growth curves of the Salmonella serotypes exposed to MIC of PA emulsions generated using 1% SDS and Quillaja saponin respectively, resulted in a reduction in growth rate and an increase in lag phase duration (Fig. 3). The formation of bacterial filaments, which are elongated structures containing many cells that remain attached due to a block in the divisome, could contribute to the reduction in growth rate as filaments are detected as a single colony during plating on bacteriological media 19,38 .
Microscopy indicated that while cells of all three serotypes formed filaments when exposed to PA emulsions, the ratios of filaments to regular cells were different (Table 2). Further, S. Typhimurium and S. Newport cells formed free-floating aggregates of floccular biofilms, whereas S. Oranienburg did not, indicating that stress response to PA varied among Salmonella serotypes. Changes in bacterial morphology among Salmonella serotypes occurred both during the use of Quillaja saponin and SDS as surfactants for the formation of PA emulsions. S. Oranienburg cells formed filaments that were longer than 5 µm and several filaments exceeded 100 µm in length (Fig. 4). Filaments are transient phenotypes that contain chains of cells connected one to another due to incomplete cell division 20 . Filaments disintegrate when cell division is completed and result in spontaneous release of many individual cells 19,20 . S. Typhimurium and S. Newport also formed filaments, but they occurred less frequently and were usually shorter (30≥µm) than S. Oranienburg filaments (100≥µm), ( Table 2, Figs. 4, 7). Apart from filaments, S. Typhimurium and less frequently S. Newport formed aggregates that were encased in EPS (Figs. 5, 6). Autoaggregation of cells is an important precursor to biofilm formation 39,40 . The free-floating aggregates of S. Typhimurium and S. Newport were encased in a dense polymeric matrix that was observed through fluorescent staining and the use of LSCM (Figs. 5, 6). The formation of floccular biofilms as well as filamentation suggest another, a less explored bacterial strategy to mitigate antimicrobial exposure (Fig. 6) 40 . Filaments formed by S. Newport, the serotype most susceptible to PA, often appeared to be disfigured indicating possible structural damage to the cells (Fig. 3). All three serotypes displayed phenotypic diversity by forming filaments (Figs. 4, 5, 7), floccular biofilms and regular size cells when exposed to PA indicating that the formation of diverse phenotypes might be a strategy employed by Salmonella to overcome stressors such as sanitizer exposure 38 .
Emulsions have many applications in the food industry and can be used in antimicrobial formulations and coatings [41][42][43][44] . Antimicrobial emulsions can improve the safety of foods, feeds and food contact surfaces. PA emulsions (31.2 mM) generated using 1% SDS resulted in a 4.7 log CFU/ml reduction of the Salmonella cockatail containing all three serotypes in 30 s and 6.8 log CFU/ml in 5 min (Figs. 8,9). A similar reduction of Salmonella serotypes was observed using a PA emulsion generated using Quillaja saponin after 5 min but not at 30 s, indicating that the type of surfactant selected might also effect the time taken for the interaction between the fatty acid and the target pathogen.

Conclusions
The current study demonstates the antimicrobial activity of PA against Salmonella and for the first time reveals that surfactant type plays an important role in influencing the antimicrobial efficacy of PA emulsions. Further, the observation that Salmonella employs multiple stress responses among a clonal population of cells when challenged with an antimicrobial, indicates that phenotypic diversity may play an important role in antimicrobial resistance. This new information will help industry better evaluate their choices of antimicrobials, as inhibitory concentrations of certain sanitizers could vary greatly from bactericidal concentrations, depending on the type of surfactant and surfactant concentration chosen. The finding that the use of surfactants can reduce the amount of PA required for inhibitory activity against Salmonella by 50% or greater, could help in reducing the environmental footprint, the cost associated with sanitizer production and the development of alternatives to biocides currently used by the food industry.
www.nature.com/scientificreports www.nature.com/scientificreports/ Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.