Effect of interaction with food constituents on plant extracts antibacterial activity

The Gaillac red wine powder and Cinnamon cassia essential oil were selected for their in vitro antibacterial activity against Staphylococcus aureus CNRZ3 and Listeria innocua LRGIA 01, respectively. In order to assess the potential application of Gaillac wine powder to the preservation of raw meat, its antibacterial activity was assayed in Mueller Hinton broth supplemented with up to 20% (w/w) bovine meat proteins (bovine meat protein content): Gaillac wine powder as well as resveratrol, a stilbene polyphenol present in red wine, lost their antibacterial activity, likely as a result of interactions of Gaillac wine antibacterial molecules with bovine meat proteins at the expense of their interactions with Staphylococcus aureus CNRZ3 cells. Cinnamon cassia essential oil antibacterial activity assays in Tryptone Soya broth, skimmed, semi-skimmed, and whole milk showed that its antibacterial activity was significantly reduced by milk fat globules but not by milk proteins: it could thus be used for the preservation of skimmed milk. The developed methodology based on the use of microbiological media mimicking the composition of perishable foods or of liquid foods such as sterilized milk with various milk fat contents could be used for the rapid screening of antibacterial plant extracts of interest for perishable foods preservation.


Introduction
There is a growing interest in antimicrobial plant extracts as alternatives to some synthetic food preservatives for sanitary, environmental, regulatory, and marketing reasons (Oulahal et al. 2017).Plant extracts for food preservation containing non-volatile (e.g.polyphenols) and volatile (e.g.essential oils) antimicrobial molecules can be distinguished.In previous studies, Gaillac red wine powder (Bouarab-Chibane et al. 2017) and Cinnamon cassia essential oil (Trinh et al. 2015) were selected based on their in vitro antibacterial activity against several pathogenic or food-spoiling bacterial strains.Gaillac red wine powder is rich in phenolics, while trans-cinnamaldehyde is the main component of C. cassia essential oil (EO).The fact that these plant extracts had major compounds containing either a phenyl structure and/or aldehyde groups is consistent with Dorman and Deans (2000) observation that antibacterial activity of plant extracts is namely due to the presence of hydrophilic functional groups, such as hydroxyl groups of phenolic components and aldehyde groups.Due to their respective organoleptic properties (odour, taste, colour), Gaillac red wine extract and C. cassia EO could be added in meat and milk, respectively.However, in vitro screening of antibacterial activity of plant extracts is usually performed in microbiological media (Mueller-Hinton (MHB) or Tryptone Soya (TSB) broth in the present case), while perishable foods such as raw beef patties or cow's milk have a far more complex composition and structure.This difference of composition and structure may result in interactions of antibacterial plant molecules with food components such as proteins or fat.Moreover, the structure of most of perishable foods is hetero-geneous: this could also result in a heterogeneous distribution of antibacterial molecules in such foods.As a consequence, target bacterial cells might be exposed to a reduced concentration of plant antimicrobial molecules due to their binding to other food components or to the distribution of target bacterial cells in another phase of heterogeneous foods than antimicrobial molecules.This has been suggested by Weiss et al. (2015) as the main cause of the loss of antibacterial activity of most of food preservatives added to foods in comparison to their activity in model media.This observation led us to propose that in vitro screening of antimicrobial plant extracts intended for food preservation should not only be performed in microbiological media but also in more complex media supplemented with food components and/or mimicking the heteregoneous structure of many foods.Therefore, increasing concentrations of bovine meat extract were added in MHB in order to mimick the protein content of bovine meat (20% (w/w)).This allowed to study the effect of proteins present in meat on the antibacterial activity of Gaillac red wine and of resveratrol.Resveratrol is a stilbene polyphenol present in red wines known for its antioxidant and antibacterial activities.In order to check the effect of the presence of casein micelles and dispersed fat in cow's milk on C. cassia EO antibacterial activity, its minimal inhibitory (MIC) and bactericidal (MBC) concentrations in TSB, skimmed milk, semiskimmed milk, and whole milk were compared.

Materials
Chemicals and reagents.Resveratrol, transcinnamadehyde, sodium azide, Nisaplin® (a 2.5% (w/w) nisin commercial preparation), beef extract powder prepared from beef muscle (75% (w/w) protein) and all reagents unless otherwise stated were purchased from Sigma (St Quentin Fallavier, France).Skimmed, semi-skimmed and whole milk powder were purchased from Régilait (St Martin Belleroche, France).Their respective compositions once resuspended in distilled water are stated in Table 2.

Plant extracts preparation and characterization
Gaillac red wine powder.Dry extract from Gaillac red wine was directly obtained by vacuum evaporation of ethanol and water.Total phenolic content and reducing sugars content of Gaillac red wine powder were 81.2±4.2mggallic acid equivalent per g and 80.2±3.3mgglucose equivalent per g, respectively (Bouarab-Chibane et al. 2017).C. cassia essential oil.C. cassia EO was extracted by hydrodistillation from leaves harvested at Yen Bai (Vietnam) by the laboratory of Chemical Analysis, Institute of Natural Products Chemistry (Vietnamese Academy of Science and Technology).Its composition was determined by gas chromatography-mass spectrometry analysis performed in the same laboratory.Transcinnamaldehyde was the major component: it represented approximately 90% (w/w) of the EO (Trinh et al. 2015).
Bacterial strains.The bacterial strains used as test microorganisms were Staphylococcus aureus CNRZ3 and Listeria innocua LRGIA01 (isolated in a dairy, collection of BioDyMIA laboratory (Université Claude Bernard Lyon 1, Bourg en Bresse, France)) for Gaillac red wine powder or resveratrol and C. cassia EO antibacterial activity assays, respectively.S. aureus and L. innocua strains were stored at -40°C in MHB or TSB (Biokar, Beauvais, France) with 15 % (v/v) of glycerol, respectively.One milliliter of the stock culture was transferred to 9 mL of MHB or TSB and incubated for 8h at 37°C for S. aureus or 30°C for L. innocua.Then, 1mL of this first culture was inoculated again with 9mL of fresh MHB or TSB and incubated overnight at 37°C or 30°C.Finally, a third culture was made from the second one (same conditions) and used for antibacterial susceptibility testing.

Methods
Antibacterial activity assays of Gaillac red wine powder and of resveratrol.In vitro antibacterial activity was assessed by monitoring the growth of S. aureus in the presence or absence (control) of Gaillac red wine powder or resveratrol at a final 1g.L -1 concentration in MHB supplemen-ted or not with beef extract at 37°C for 24h.Gaillac red wine powder and resveratrol were pre-diluted at a 10g.L 1 concentration in a 10% (w/w) DMSO solution.Beef extract was added to MHB to reach a 5, 10, 15, or 20% (w/w) final protein concentration.Briefly, 270μL of MHB with or without beef extract alone (control) or supplemented with Gaillac red wine powder or resveratrol (final concen-trations: 1g.L -1 ) were mixed either with 30 µL of bacterial inocula (5.0 x 105 cfu.mL -1 (Colony Forming Units) ) or 30µL of sterile MHB with or without beef extract (control) in each well of the microplate and incubated at 37°C for 24h in a Bioscreen C® plate reader (Oy Growth Curves AB Ltd., Helsinki, Finland).The optical density of the culture was monitored every 15min, in the 420nm -580nm wavelength range (OD420-580).Negative control wells containing 2000 IUmL -1 nisin were also monitored.All measurements were performed in triplicate.After screening, S. aureus cells were enumerated, 10µL of the suspension from each inoculated well were spotted on plates containing Mueller Hinton Agar, and then incubated at 37°C for 24h.Data were expressed as log cfu per mL.

Determination of resveratrol partitioning coefficient between a phase with and a phase without beef extract proteins.
A 20% (w/w) bovine meat proteins suspension was prepared by resuspending 26.66% (w/w) of beef extract powder (75% (w/w) protein) in distilled water at pH 7.0.A part of this suspension was ultrafiltrated through 3 kDa cut-off ultrafiltration membranes (3kDa Microsep TM Advance Centrifugal Device, Pall Life Sciences, USA) in order to remove bovine meat proteins with a molecular weight exceeding 3kDa.In order to estimate the "affinity" of resveratrol for bovine meat proteins with a molecular weight exceeding 3kDa, 1 mL of the ultrafiltrate of the 20% (w/w) bovine meat proteins suspension was placed in a 3.5-5kDa cut-off dialysis tube (Spectra/ Por Float-A-Lyzer G2, G235029, cut-off threshold 3.5-5 kDa).This 1 mL tube was dialyzed against 20 mL of the 20% (w/w) bovine meat proteins suspension supplemented with 1g.L -1 of resveratrol for 2 weeks at 6°C in order to reach the equilibrium of distribution of resveratrol between the phase with bovine meat proteins and the phase without bovine meat proteins.The system was continuously stirred in dark.After 2 weeks, the resveratrol concentration in dialysis tubes (initial concentration: 0g.L -1 ) was assayed by Reversed-Phase High Performance Liquid Chromatography A partitioning coefficient of resveratrol between the phase with bovine meat proteins (with a molecular weight exceeding 3kDa) and the phase without these proteins was thus calculated as follows: log P bovine meat proteins > 3 kDa / ultrafiltrate (< 3 kDa) = [resveratrol] in 20% (w/w) bovine meat proteins suspension / [resveratrol] in ultrafiltrate (<3 kDa) of 20%(w/w) bovine meat proteins suspension

Determination of Minimal Inhibitory (MIC) and Bactericidal (MBC) concentrations of
Cinnamon cassia EO.Skimmed, semi-skimmed or whole milk powder were suspended in distilled water before being sterilized for 15 min at 110°C.They were then supplemented with serial dilutions of C. cassia EO and inoculated with L. innocua LRGIA 01 strain to determine corresponding MIC and MBC values.First, an overnight L. innocua LRGIA01 culture was diluted in TSB to a final concentration of 105cfu.mL - .One hundred microliters of this bacterial suspension were used to inoculate 10 mL of either TSB (1% (w/w)), skimmed, semi-skimmed, or whole milk supplemented or not (control) with serial dilutions of C. cassia EO.After 24h incubation at 30°C, the MICs of C. cassia EO were determined using micro-dilution broth method.The lowest concentration of the EO resulting in no growth of the strain was taken as the MIC.10µL from each medium were streaked on solid agar medium and incubated for 24h at 30°C.The lowest concentration at which no visible growth on solid agar medium was obtained was considered as the MBC.

Results and Discussion
Effect of bovine meat proteins addition on the antibacterial activity of Gaillac red wine powder and resveratrol.S. aureus growth in MHB supplemented with up to 20% (w/w) bovine meat proteins was monitored in the presence or absence of Gaillac red wine powder or of resveratrol (Table 1).In MHB, both Gaillac red wine powder and resveratrol at 1g.L -1 were bacteriostatic but not bactericidal: this means that no growth occurred over 24h incubation at 30°C, while regrowth of S. aureus cells was observed after 24h incubation in fresh MHB without these antibacterial agents.The antibacterial activity of phenolics present in red wines has already been reported by several authors (Rodriguez- Vaquero et al. 2007Vaquero et al. , 2013;;Friedman 2014).The addition of 20% (w/w) bovine meat proteins in MHB annihilated the bacteriostatic effect against S. aureus of 1g.L -1 of Gaillac red wine powder or resveratrol.However, while S. aureus growth was no more inhibited by Gaillac red wine powder, a 38% reduction of the maximal optical density (OD420-580nm) over 24h incubation was still observed in the presence of 1g.L -1 of resveratrol.This means that in the presence of the same concentration of proteins than in bovine meat (20% (w/w)), a 1g.L -1 resveratrol concentration still significantly inhibited S. aureus growth unlike Gaillac red wine powder at the same concentration.This observation is consistent with the fact that in the presence of 5% (w/w) bovine meat proteins in MHB, the bacteriostatic effect of resveratrol is preserved unlike that of Gaillac wine powder.The higher the bovine meat protein concentration, the lower the S. aureus growth inhibitory activity of Gaillac wine powder and to a lesser extent of resveratrol.The antimicrobial activity of Gaillac wine powder is likely due to phenolic compounds.Vaquero et al. (2007) reported that the antibacterial activity of red Food Science and Applied Biotechnology, 2018, 1(1), 77-85 Bouarab-Chibane et al., 2018 Effect of interaction with food constituents...

Page 81
wines against pathogenic bacteria such as S. aureus was mainly due to phenolics.Gaillac red wine has a high total phenolics content (2800 mg gallic acid equivalent.L -1 ).Since it contained 29 g dry matter.L -1 , this corresponds to a 81.2±4.2mggallic acid equivalent per g of Gaillac red wine powder.The mean concentration of resveratrol in Gaillac red wine is about 5 mg.L -1 corresponding thus to about 0.17 mg.g -1 of powder.Resveratrol is thus likely not a significant antibacterial compound since it represents only a 0.17mg.L -1 concentration in MHB supplemented with 1g.L -1 of Gaillac wine powder.Resveratrol is a hydrophobic molecule with a log P octanol-water (o/w) value of 3.4.The decrease of antibacterial activity of resveratrol when increasing beef extract concentration might thus result from hydrophobic interactions with bovine meat proteins.Other phenolic antibacterial compounds present in red wine such as gallic acid (log Po/w = 3.72), catechin (log Po/w = 1.5), quercetin-3-β-glucoside (log Po/w = -0.34),and rutin (log Po/w = -1.25)have different physico-chemical properties which might modulate their interactions with bovine meat proteins and thus the susceptibility of their antibacterial activity to beef extract proteins.In the present case, both Gaillac red wine powder and resveratrol antimicrobial activity decreased in the presence of bovine meat proteins.This is likely due to electrostatic and hydrophobic interactions of antimicrobial plant phenolics with bovine meat proteins as proposed by Weiss et al. (2015).Interactions between proteins and polyphenols have been extensively reviewed (Papadopoulou and Frazier 2004;Ozdan et al. 2013).
In order to investigate the interactions between bovine meat proteins and resveratrol, we determined the partitioning coefficient at 6°C of resveratrol between the phase without proteins with a molecular weight exceeding 3kDa (ultrafiltrate of a 20% (w/w) bovine meat proteins suspension) and a 20% (w/w) bovine meat proteins suspension.At equilibrium, the resveratrol concentration in the 20% (w/w) bovine meat proteins suspension was 7.95-fold higher than in the phase without bovine meat proteins with a molecular weight exceeding 3 kDa.This corresponds to a log P bovine meat proteins > 3 kDa/ultrafiltrate (< 3 kDa) value of 0.9.This clearly indicates the affinity of resveratrol for bovine meat proteins with a molecular weight exceeding 3kDa.The interaction of resveratrol (a relatively hydrophobic stilbene polyphenol with a log Po/w value of 3.4) with bovine meat proteins is likely both due to electrostatic (ionic, hydrogen bonding) and hydrophobic interactions.The affinity of resveratrol for meat proteins substantiates the hypothesis that resveratrol antibacterial activity reduction in their presence likely results from a lower quantity of "free" resveratrol which could interact with target bacterial cells and inhibit their growth.However, the partition coefficient determined to estimate the affinity of resveratrol for bovine meat proteins was determined at 6°C while antibacterial activity assays in the presence of bovine meat proteins were performed at 37°C.At 37°C, electrostatic interactions are weaker, while hydrophobic interactions are stronger than at 6°C.Therefore, the affinity of resveratrol for proteins should also be determined at 37°C.

Effect of milk proteins and milk fat on the anti-L. innocua activity of C. cassia essential oil.
In order to investigate the effect of milk proteins and milk fat on the anti-L.innocua activity of C. cassia EO, a 103cfu.mL - initial L. innocua LRGIA 01 population was incubated for 24 h at 30°C in TSB (1% w/w), skimmed, semi-skimmed or whole milk with or without (control) serial dilutions of C. cassia EO.Although TSB 1% (w/w), skimmed, semi-skimmed, and whole milk had very different compositions (  Cava et al. (2007).
Table 1.Effect of proteins (bovine meat extract) addition to Mueller-Hinton broth on the antibacterial activity of Gaillac red wine powder and of resveratrol against S. aureus CNRZ3 strain (initial population: 5.0 x 10 5 cfu.mL -1 ).
Gaillac red wine powder (1g.L -1 ) Resveratrol (1 g.L -1 )  Taken together, the results indicate that while MIC and MBC values are quite similar in TSB 1% (w/w) and skimmed milk, respective 2.5-fold and 5-fold increases of MIC and MBC were observed between skimmed and semi-skimmed milk.This suggests that C. cassia EO antibacterial activity is more affected by the presence of milk fat globules than by the increase of proteins and carbohydrate concentrations between TSB 1% (w/w) and skimmed milk.The reduction of C. cassia EO anti-Listeria activity in the presence of milk fat globules is substantiated by the fact that a further 2-fold decrease of its MIC was observed in whole milk.The increase of MIC observed in whole milk containing 3.6% (w/w) milk fat in comparison to TSB 1% (w/w) is similar to the increase of MIC in emulsified TSB 1% (w/w) containing 2.5% (w/w) sunflower oil and 1.5% (w/w) soya lecithin reported by Trinh et al. (2013).Since C. cassia EO anti-Listeria activity is mainly due to trans-cinnamaldehyde which represents 90% (w/w) of its components (Trinh et al. 2015), physico-chemical characteristics of transcinnamaldehyde such as its log Po/w value (2.12) might explain why its antibacterial activity is far more affected by the presence of milk fat globules than by the presence of proteins like casein micelles and β-lactoglobulin which represent 80% (w/w) and 10% (w/w) of cow's milk proteins, respectively.Compounds with the highest log Po/w values are likely more susceptible to bind to hydrophobic particles like milk fat globules.As illustrated by these results concerning C. cassia EO, several authors have recently studied or reviewed the effect of the interaction with food components of EOs on their antimicrobial activity (Gutierrez et al. 2008(Gutierrez et al. , 2009;;Perricone et al. 2015).

Conclusions
Gaillac red wine powder antibacterial activity was annihilated when bovine meat proteins were added.This antibacterial activity probably results from interactions of bovine meat proteins with antibacterial phenolics present in red wines such as Gaillac (as examplified with resveratrol).C. cassia EO antibacterial activity only slightly decreased in skimmed milk while it strongly decreased in whole milk and to a lesser extent in semi-skimmed milk.This is likely due to interactions between transcinnamaldehyde, the main component of C. cassia EO, and milk fat globules.A putative explanatory scheme presenting interactions between resveratrol (as an example of phenolics) and meat proteins or between transcinnamaldehyde and milk fat globules at the expense of interactions with S. aureus or L. innocua cells, respectively is presented in Figure 1.In vitro screening of plant extracts antibacterial activity in complexified media should both allow not to consider plant extracts which lose their activity in real foods and provide a better understanding of the interactions between antimicrobial plant molecules such as phenolics and aldehydes and dispersed fat or proteins.

Figure 1 .
Figure 1.Scheme presenting interactions between resveratrol (as an example of polyphenol) and meat proteins (A) or between trans-cinnamaldehyde and milk fat globules (B) at the expense of interactions with S. aureus (A) or L. innocua (B) cells, respectively.