Efficacy of Nisin-Coated Polymer Films To Inactivate Salmonella Typhimurium on Fresh Broiler Skin

Transcription

1 1189 Journal of Food Protection, Vol. 63, No. 9, 2000, Pages Copyright, International Association for Food Protection Efficacy of Nisin-Coated Polymer Films To Inactivate Salmonella Typhimurium on Fresh Broiler Skin NANDINI NATRAJAN AND BRIAN W. SHELDON* Department of Food Science, Box 7624, North Carolina State University, Raleigh, North Carolina , USA MS : Received 10 November 1999/Accepted 7 April 2000 ABSTRACT Nisin is an antimicrobial peptide produced by the food-grade microorganism Lactococcus lactis subsp. lactis. This peptide inhibits many gram-positive bacteria, and when combined with chelating agents it inhibits gram-negative bacteria such as Salmonella sp. The efficacy of packaging films treated with nisin-containing formulations to reduce Salmonella contamination of fresh broiler drumstick skin and increase the refrigerated shelf life was investigated. Three films (5.1 cm 2 ) of varying hydrophobicities (polyvinyl chloride [PVC], linear low density polyethylene, nylon) were coated with one of three liquid formulations (ph 3.5 to 3.8) composed of 100 g/ml nisin and varying concentrations of citric acid, EDTA, and Tween 80. The treated films were applied either wet or dry to 5.1-cm 2 broiler drumstick skin samples inoculated with a nalidixic acidresistant (NAr) strain of Salmonella Typhimurium. After incubation at 4 C for 24 h the populations of surviving Salmonella Typhimurium NAr organisms were recovered from the skin and film samples using a rinse procedure and enumerated on brain heart infusion agar containing 800 ppm NA. Log reductions (untreated versus treated skin) in Salmonella Typhimurium NAr populations ranged from 0.4 to 2.1. Treatment formulation compositions and wet versus dry treatment application also influenced the extent of kill. The shelf life of refrigerated broiler drumsticks was extended by 0.6 to 2.2 days following a 3-min immersion in a nisin-containing treatment solution and subsequent storage in a foam tray pack containing a nisin-treated PVC overwrap and a nisin-treated absorbent tray pad. These findings demonstrated that Salmonella Typhimurium and spoilage microorganism populations on the surface of fresh broiler skin and drumsticks can be significantly reduced using immersion treatments, absorbent tray pads, and packaging films treated with nisin-containing formulations. Food safety has become one of the most visible issues in recent times. Outbreaks of foodborne disease occur worldwide, including in the United States, where the food supply is considered to be one of the safest in the world. These outbreaks have led to consumer concerns and erosion in public confidence regarding the safety of our food supply. The greatest concerns center on the presence of pathogenic microorganisms. The annual number of cases and the associated economics involving foodborne disease outbreaks are of considerable magnitude. It is estimated that foodborne sources cause one-third of all microbially related diseases in the United States with an annual estimated cost of $9.4 billion (4). A partial list of the more common microorganisms responsible for these foodborne diseases includes Salmonella sp., Campylobacter sp., Listeria monocytogenes, Staphylococcus aureus, Escherichia coli O157: H7, and Clostridium perfringens (16). Outbreaks of foodborne disease are most often attributed to inadequate cooking, temperature abuse, use of contaminated raw ingredients, and cross contamination (5). Among other foods, * Author for correspondence. Present address: Department of Poultry Science, Box 7608, North Carolina State University, Raleigh, NC , USA. Tel: ; Fax: ; brian sheldon@ncsu.edu. The use of trade names in this publication does not imply endorsement by the North Carolina Agricultural Research Service nor criticism of similar ones not mentioned. Present address: Molecular Circuitry Inc., 3400 Horizon Drive, King of Prussia, PA 19406, USA. the consumption of contaminated meat and poultry products is often associated with these disease outbreaks. Numerous approaches to reduce pathogens and extend the shelf life of meat and poultry products have been widely investigated with varying success. Additional research is needed to identify new approaches for controlling, reducing, or eliminating pathogens from muscle food products. The purpose of this study was to examine an antimicrobial (bacteriocin) that might achieve similar or better microbial reductions on poultry carcasses than currently approved poultry product disinfectants (i.e., chlorine, trisodium phosphate, sodium chlorite). This nontraditional control might provide a safe alternative for decreasing foodborne disease agents leading to fewer outbreaks. Bacteriocins are commonly found in foods due to their production by the lactic acid bacteria used to ferment dairy, vegetable, and meat products (8). The bacteriocin nisin is a ribosomally synthesized peptide having a relatively broad spectrum of antibacterial activity against microbial foodborne pathogens and spoilage organisms (10). It is produced by the microorganism Lactococcus lactis subsp. lactis during its exponential phase of growth (3, 10). Nisin is a member of the class of antimicrobial substances known as lantibiotics because it contains the unusual amino acid lanthionine. Lantibiotics in general are considered to be potentially useful as food preservatives although only nisin has been recognized by the Food and Drug Administration (1988) as generally recognized as safe. It may be used to

2 1190 NATRAJAN AND SHELDON J. Food Prot., Vol. 63, No. 9 inhibit the outgrowth and toxin production by Clostridium botulinum spores in processed pasteurized cheese spreads and has been approved for use in salad dressings and pasteurized liquid whole egg (6, 8). Nisin has been approved worldwide in over 40 countries for use in food products ranging from canned foods to baby foods (7, 8). It is currently produced by Aplin and Barrett, Dorset, England and marketed under the trade name of Nisaplin. Nisin s spectrum of inhibitory activity is limited to gram-positive microorganisms such as L. monocytogenes, C. botulinum, S. aureus, and some lactic acid-producing bacteria (15). Previous studies conducted in our laboratory and elsewhere showed that, when combined with foodgrade chelating agents such as EDTA, citrate, or phosphate, nisin inhibited gram-negative bacteria including Salmonella serotypes and E. coli O157:H7 (21). The mechanism of this inhibitory action was speculated as the ability of the chelating agent to bind magnesium and other divalent cations located in the outer bacterial membrane leading to a destabilization of the lipopolysaccharide region of the membrane, increased outer membrane permeability to nisin, and eventually lysis of the cell (9, 12, 14, 20). Nisin-mediated cell lysis is associated with its ability to change the surface free energy on the cytoplasmic membrane that leads to a change in cell permeability and ultimately cell lysis (13). Related studies (19) conducted in our laboratory have shown significant reductions in the populations of Salmonella Typhimurium (3.12 to 4.86 log cycles per ml) recovered in the rinse solution from artificially inoculated broiler drumstick skin following a 30-min dip in a nisin-chelator combination in comparison to a 20-ppm chlorine dip treatment (1.9 log cycle reduction). A 1.5- to 3-day extension of broiler drumstick refrigerated shelf life was also achieved using the nisin-chelator combination dip. These experiments demonstrate the potential usefulness of applying nisin-containing treatments to control pathogens and spoilage organisms associated with muscle foods. Primary packaging plays an integral role in preventing further contamination of poultry products throughout marketing and distribution channels. This study was designed to evaluate the potential use of primary packaging materials of varying compositions and physical and chemical properties to act as delivery vehicles for carrying and transferring nisin-containing formulations onto the surfaces of fresh poultry products. The application of antimicrobials at this final processing stage is anticipated to enhance further the safety and shelf life of poultry products throughout the wholesale and retail marketing chain. MATERIALS AND METHODS Bacterial strain. A stock culture of a nalidixic acid-resistant (NAr) strain of Salmonella Typhimurium NAr was obtained from Frank T. Jones, North Carolina State University, Raleigh. This organism is resistant to greater than 1,000 ppm NA. In comparison to 18 other Salmonella species representing both environmental and type collections, this test strain had moderate resistance to the nisin plus EDTA treatment (20). The bacterial culture was maintained at 4 C on sterile brain heart infusion (BHI) agar (Difco Laboratories, Detroit, Mich.) supplemented with 200 ppm NA (Sigma Chemical Company, St. Louis, Mo.) and transferred to fresh BHI broth 24 h prior to experimentation. Reagents. Purified nisin (Applied Microbiology, Inc., New York, N.Y.) was stored at 20 C in a desiccator. Nisin solutions (ph 2.0) were prepared fresh for each experiment in 0.02 N hydrochloric acid (Fisher Scientific Co., Dallas, Tex.). The inhibitory activity of nisin was verified using L. lactis subsp. cremoris ATCC and L. monocytogenes Scott A (North Carolina State University Culture Collection, Raleigh, N.C.) as indicator organisms in an agar well-diffusion assay based on that of Barefoot and Klaenhammer (2). The chelating agents and surfactants used in this study were EDTA (Fisher), citric acid anhydrous (citrate) (Sigma), and Tween 80 (polyoxyethylenesorbitan monooleate) (Sigma). Treatment formulations. Three optimized nisin-containing formulations (designated A, B, and C) were chosen for testing in this study based on their demonstrated inhibitory activity against the Salmonella Typhimurium NAr test strain inoculated on fresh broiler drumstick skin (19). The treatment formulations were prepared fresh for each experiment in distilled water and were composed of: treatment A: 100 g/ml nisin, 7.45 mm EDTA, 3.1% citric acid, ph 3.8; treatment B: 100 g/ml nisin, 3.0% citric acid, 0.61% Tween 80, ph 3.5; and treatment C: 100 g/ml nisin, 5.0 mm EDTA, 3.0% citric acid, 0.5% Tween 80, and ph 3.5. The ph of the solutions was adjusted prior to the addition of surfactant using NaOH pellets. Treatment formulations were sterilized by autoclaving at 121 C/15 psi pressure for 15 min. Standard inactivation protocol. An overnight culture of Salmonella Typhimurium NAr was grown in BHI broth at 37 C to a population density of approximately 10 8 CFU/ml (optical density at 600 nm, 0.4, mid-log phase). A 0.1-ml volume of the diluted culture (in 0.1% peptone water) was spread evenly across the surface of a 5.1-cm 2 skin sample aseptically excised from fresh nonfrozen broiler drumsticks obtained from a local supermarket using a sterile L-shaped glass rod. Following inoculation, the skin samples were held for 5 min to allow for skin adsorption and attachment of the test strain followed by direct application of the skin to the bacteriocin-coated packaging films. Following a specified exposure time, the skin and film samples were vortexed for 2 min in sterile 50-ml polystyrene tubes containing 20 ml of 0.1% peptone water. Serial dilutions in 0.1% peptone water were prepared and samples spread plated in duplicate on BHI agar supplemented with 800 ppm NA. Colonies were enumerated after incubation at 37 C for 48 h. Unless otherwise stated, each experiment included controls of inoculated skin pieces in contact with untreated package film (positive control used for calculating log reductions), treated package film in contact with uninoculated skin (negative control used to confirm the absence of NAr contaminants), and inoculated skin with no package film (for calculating microbial recoveries from the skin). Log 10 reductions in the population of Salmonella Typhimurium NAr organisms were calculated as the difference in the number of organisms recovered from treated (positive control) and untreated (negative control) skin and film samples. All experiments were replicated three times using two samples per treatment (n 2), and statistical differences between means were determined using the Student s t test at P 0.05 (least-squares difference) (18). This protocol was followed for all the objectives under consideration unless otherwise stated. Three separate studies were conducted to examine the effect of nisin-based treatment formulations on Salmonella Typhimurium NAr - contaminated broiler drumstick skin.

3 J. Food Prot., Vol. 63, No. 9 NISIN-COATED POLYMER FILMS 1191 FIGURE 1. Bactericidal effect of packaging films of varying hydrophobicities coated with one of three nisin-containing formulations against Salmonella Typhimurium NAr -contaminated broiler drumstick skin. Treated packaging films were applied either wet or dry to the inoculated skin and the surviving population of Salmonella Typhimurium NAr recovered from the skin and film after 24 h exposure at 4 C. Initial inoculum load was log cells per skin sample. Mean log survivors with different letters are significantly different (n 6, P 0.05). Treatment A, 100 g/ml nisin, 7.45 mm EDTA, 3.1% citric acid, ph 3.8; treatment B, 100 g/ml nisin, 3.0% citric acid, 0.61% Tween 80, ph 3.5; treatment C, 100 g/ml nisin, 5.0 mm EDTA, 3.0% citric acid, 0.5% Tween 80, ph 3.5. Effect of film type and nisin formulation on inhibiting Salmonella Typhimurium NAr. Packaging films of varying hydrophobicities including flexible polyvinyl chloride (PVC) ( 35 dynes/cm 2 ), linear low density polyethylene (LLDPE) ( 30 dynes/cm 2 ), and nylon ( 42 dynes/cm 2 ) were obtained from Borden (PVC, LLDPE; Andover, Md.) and KNF Corp. (nylon; Tamaqua, Pa.) and aseptically cut into 5.1-cm 2 pieces. The treatment formulations were sprayed onto each film type using a liquid atomizer (2 ml volume) and applied either wet or air-dried (15 min air-drying period under a laminar flow hood) to the Salmonella Typhimurium NAr -inoculated (10 6 to 10 7 CFU/skin sample) broiler skin samples (n 2). After incubation at 4 C for 24 h the population of surviving Salmonella Typhimurium NAr organisms were recovered using the previously described rinse procedures. Effect of varying contact time and nisin concentration. Two separate experiments were conducted to determine the influence of film to skin contact time and varying nisin concentrations under refrigeration conditions on the inhibitory activity of a nisincoated PVC film against Salmonella Typhimurium NAr -contaminated broiler drumstick skin (log CFU/skin sample). In the first experiment the lethality of treatment C, applied both wet or air-dried to 5.1 cm 2 monolayered PVC film (n 2), was determined at 4 C under varying contact times (24 h intervals up to 96 h). Twenty-four-hour sampling intervals were chosen to simulate a 4-day time span from the processing plant to display at a retail outlet. The standard inactivation protocol was followed as previously described. In the second experiment, flexible monolayered PVC films (5.1 cm 2 ) were treated with 2 ml of treatment C containing either 0, 100, 150, 175, 200, 300, or 500 g/ml nisin. The treated films were applied wet to Salmonella Typhimurium NAr -inoculated (log CFU/skin sample) broiler drumstick skin and incubated at 4 C. Surviving organisms were recovered at 24-h intervals up to 96 h and enumerated as previously described. Shelf-life study. The objective of this final study was to determine the efficacy of a nisin-based treatment formulation to extend the shelf life of refrigerated whole broiler drumsticks. The experimental design departed significantly from the previous trials in that broiler drumsticks were first immersed in an aqueous nisin formulation and then packaged in a foam traypack containing a nisin-treated absorbent tray pad and a nisin-coated overwrap film. Ninety (trial 1) or 120 (trial 2) fresh nonchilled broiler drumsticks were obtained from a local processor prior to chilling, bagged, and transported on ice (no physical contact with ice) to the laboratory ( 45 min). Forty-five (trial 1) and 60 (trial 2) chilled drumsticks per treatment were submerged for 3 min in 7 liters (trial 1) or 9 liters (trial 2) (1.3 liters and 1.36 liters/kg of tissue, respectively) of one of the following treatment solutions (25 C): (i) treatment C at 50 g/ml nisin or (ii) sterile distilled, deionized water (ph 6.7 to 7.0). Following this immersion treatment and a

4 1192 NATRAJAN AND SHELDON J. Food Prot., Vol. 63, No. 9 1-min drip time, the nisin-treated drumsticks were packaged in commercial foam traypacks (two drumsticks per traypack for trial 1 or three drumsticks per traypack for trial 2) containing a commercial absorbent tray pad (Dri-chick, model , 50 g absorbency, National Converting Co. Inc., Cordova, N.C.) treated with 2 ml of treatment C containing 50 g/ml nisin. The tray was overwrapped with PVC film sprayed with 2 ml of treatment C containing 150 g/ml nisin and refrigerated at 4 C. Control drumsticks were packaged similarly with the exception that the pad and overwrap were not treated with the nisin formulation. At specified sampling times (0, 2, 4, 5, 6, 7, 8, 9, 10 days), two traypacks per treatment (n 4 drumsticks for trial 1, n 6 for trial 2) were individually sampled and the population of mesophilic and psychrotrophic organisms per drumstick enumerated using a drumstick rinse procedure in 20 ml of 0.1% sterile peptone water. The drumstick and peptone water were shaken in a Whirlpak bag (NASCO, Fort Atkinson, Wis.) for 1 min to facilitate the removal of surviving organisms. Pour plates of the rinse solution were prepared in duplicate on BHI agar plates and incubated at 37 C for 48 h to enumerate total mesophiles and 4 C for 10 days to estimate psychrotrophic populations. Data have been reported as the total mesophilic and psychrotrophic populations on a CFU per cm 2 of skin basis. To estimate the total number of microorganisms per cm 2, first the number of CFU per ml of drumstick rinse was multiplied by 20, the drumstick volume rinse, and then converted to log 10 values (19, 22). The CFU per cm 2 counts were then derived by dividing the total number of microorganisms per drumstick by the average surface area of a 114-g drumstick (191.0 cm 2 ) and converted to log 10 values (19, 22). This experiment was replicated twice and the data analyzed and reported separately due to differences in the initial microbial populations detected between trials. RESULTS AND DISCUSSION Effect of film type and nisin formulation on inhibiting Salmonella Typhimurium NAr -contaminated broiler drumstick skin. Results from the three nisin-based treatment formulations (A, B, and C) applied (either wet or dry) to three packaging films of varying hydrophobicities against Salmonella Typhimurium NAr -contaminated broiler drumstick skin are presented in Figure 1. The average inoculum load per drumstick skin sample was 10 7 organisms that yielded an average population recovered from the skin and untreated film of log CFU/ml of rinse. Following a 24-h exposure to treatment A, the average population of Salmonella Typhimurium NAr survivors recovered from the skin PVC film overlay was log 5.1 and 4.3 for the dry and wet film applications, respectively, log 4.7 and 4.2 for the nylon film treatment, and log 4.3 and 4.0 for the LLDPE film treatment (Fig. 1). The formulation without EDTA (treatment B) yielded log survivor populations of 5.36 and 5.36 for PVC, 4.86 and 5.16 for LLDPE, and 4.76 and 4.96 for the nylon dry and wet film applications, respectively. Treatment C that contained all formulation components, yielded log survivor populations of 4.56 and 3.66 for the PVC film, 4.32 and 4.56 for the LLDPE film, and 4.86 and 4.66 for the nylon dry and wet film applications, respectively. The absence of EDTA in treatment B significantly reduced the level of kill (0.63 to 0.67 logs, see main effect means) in comparison to treatments A and C (Table 1). This TABLE 1. Main log reductions in Salmonella Typhimurium populations on inoculated broiler skin as influenced by nisin formulation composition, application method (wet versus dry), film type, and their interactions a Wet Dry A B C Wet Dry Main effects b (treatment, application, film) A c B C Prob F 1.35 A 0.68 B 1.31 A 0.01 Dry d Wet LLDPE e Nylon PVC Treatment interactions (treatment application, film treatment, film application) A B C 1.65 A 1.06 AB 0.60 B 0.76 B 1.45 A 1.16 AB LLDPE Nylon PVC 1.60 A 0.75 BC 1.30 AB 1.35 AB 0.90 ABC 1.0 ABC LLDPE Nylon PVC ABC 0.40 C 1.65 A Prob F a Population reductions were calculated as the difference in Salmonella Typhimurium populations recovered from control samples (skin overlayed with untreated film) stored for 24 h at 4 C versus skin samples overlayed with nisin-coated films. b Means in the same row (main effects) or across rows and columns (treatment interactions) without common letters are significantly different (P 0.05) (n 6). Means without letters are not significantly different. c A, B, C: treatment A, 100 g/ml nisin, 7.45 mm EDTA, 3.1% citric acid, ph 3.8; treatment B, 100 g/ml nisin, 3.0% citric acid, 0.61% Tween 80, ph 3.5; treatment C, 100 g/ml nisin, 5.0 mm EDTA, 3.0% citric acid, 0.5% Tween 80, ph 3.5. Two milliliters of each treatment formulation was sprayed on separate 5.1-cm 2 film pieces. d Dry, wet: Polymer films (LLDPE, nylon, PVC) were sprayed with 2 ml of either treatment A, B, or C and then applied either wet or dry (air-dried for 15 min) to Salmonella Typhimuriuminoculated broiler drumstick skin. e LLDPE, linear low density polyethylene; PVC, polyvinylchloride. reduction in kill was most evident on the LLDPE and PVC films. No significant differences in mean log reductions were detected between the wet and dry treatment applications (Table 1), although wet and dry application differences were noted in the mean log survivor populations (Fig. 1). In general, wet applications involving treatments A and C and PVC films yielded lower populations than dry applications. Log reductions in Salmonella Typhimurium populations produced by the treatment fomulations were not

5 J. Food Prot., Vol. 63, No. 9 NISIN-COATED POLYMER FILMS 1193 FIGURE 2. Effect of film contact time at 4 C on the inhibitory activity of a nisin-based treatment formulation (treatment C applied to PVC films) against Salmonella Typhimurium NAr -contaminated broiler drumstick skin. Initial inoculum load was log cells per skin sample. Mean log survivors with different letters are significantly different (n 6, P 0.05). Treatment C, 100 g/ml nisin, 5.0 mm EDTA, 3.0% citric acid, 0.5% Tween 80, ph 3.5. influenced by the type of polymeric film (main effect), although a significant formulation by film type interaction was observed (Table 1, P 0.07). Significant differences in mean log reductions were observed between treatments B and C on PVC films and between treatments A and B on LLDPE films. The apparent requirement of EDTA to enhance the nisin-mediated kill of Salmonella Typhimurium (treatments A and C) is related to the ability of this chelating agent to bind magnesium and other divalent cations (20). The loss of divalent cations from the outer membrane of gram-negative bacteria such as Salmonella sp. acts to destabilize the lipopolysaccharide layer and produce cells with increased outer membrane permeability. This increase in outer membrane permeability to nisin and other components was proposed to facilitate the inactivation of the cell via bactericidal action at the cytoplasmic membrane (15, 20). In addition to EDTA, other chelators including citric acid monohydrate, sodium phosphate dibasic, and ethylenebis (oxyethylene-nitrilo) tetraacetic acid were also reported as effective in inhibiting Salmonella Typhimurium NAr and other gram-negative bacteria when combined with nisin. In the present study both EDTA and citric acid were introduced in the treatment formulations to maximize chelation as well as utilize the acidulant properties of citric acid. Tween 80 was added to the formulation to decrease the surface tension and interfacial energies between the treatment components (i.e., nisin) and Salmonella Typhimurium cells, thereby facilitating better distribution and interaction of nisin and chelators on the bacterial cell wall and membrane (17). In related studies, the nonionic emulsifier Tween 80 was shown to enhance the biocidal activity of nisin against L. monocytogenes in a high fat whole milk system (11). The authors proposed that Tween 80 displaced nisin from the milk fat globules that led to the restoration or retention of nisin s inhibitory activity. Stevens et al. (21) suggested that the presence of fat in foods (i.e., subcutaneous fat associated with broiler drumstick skin) might decrease nisin s inhibitory activity by partitioning the molecule into hydrophobic regions where they may be unavailable for action against bacteria. To some extent, the population of Salmonella Typhimurium survivors was influenced by film type, although the data are difficult to interpret due to the confounding factors of formulation composition, application method (wet versus dry), and possible unknown porosity variations between the films. For example, in a comparison of treatment C survivor data within the wet film application, the PVC film yielded the lowest surviving population of the three films (around

6 1194 NATRAJAN AND SHELDON J. Food Prot., Vol. 63, No. 9 FIGURE 3. Bactericidal effect of PVC films coated with a nisin-based treatment (treatment C) composed of varying nisin concentrations on Salmonella Typhimurium NAr -contaminated broiler drumstick skin. Initial inoculum load was log cells per skin sample. Mean log survivors (n 6) with different letters are significantly different (P 0.05); (ND nondetectable). Treatment C, 100 g/ml nisin, 5.0 mm EDTA, 3.0% citric acid, 0.5% Tween 80, ph log difference; Fig. 1). PVC was the thinnest film composed of a single monolayer, whereas the other two films were multilayered composites with inner layers of nylon and LLDPE, respectively. The impact of structural differences between the films is yet to be determined. Other significant Salmonella Typhimurium population differences between films were observed for the A-Dry, B-Dry, and C- Dry treatments. As anticipated, the more hydrophobic film (LLDPE) repelled the aqueous nisin formulations to a greater degree than the other films resulting in a coalescence of the treatment solution droplets. Moreover, there appeared to be a greater coalescence of treatment A droplets, the treatment without surfactant. This apparent repulsion between the LLDPE film and treatment solution may have influenced the overall inhibitory activity of the formulations by causing more localized inactivation of the target Salmonella skin population. Further testing under more defined conditions will be required for a full understanding of these experimental factors. Effect of varying contact time and nisin concentration. The average inoculum load spread on each skin sample was log Treatment C contact time and wet versus dry treatment application on PVC films significantly influenced the surviving populations of Salmonella Typhimurium NAr (Fig. 2). In general, longer contact times yielded greater cell death. For the wet treatment application, contact times of 72 h and 96 h yielded significantly lower survivors (1.5 and 1.6 logs, respectively) than at the 24- and 48-h contact times. No significant differences in populations were observed between the 24- and 48-h contact times or between the 72- and 96-h contact times. For the dry treatment application, only the 96-h contact time produced significantly lower Salmonella Typhimurium populations compared to the 24- and 48-h exposure times. These findings agree with our previous trials that suggested that higher film moisture levels (wet applications), and longer contact times result in greater inhibition. Presumably, increased contact time between the treated films and infected skin surfaces led to increased migration of the formulation components onto the skin surface. Previous studies (20) demonstrated that when a 30-min EDTA treatment was sequentially followed by a 30-min nisin treatment, only 1 of 20 Salmonella sp. was inhibited (i.e., greater than 1-log reduction in population). Cells treated with either nisin or EDTA alone were not inhibited. Conversely, cells treated for 30 min with a combination of nisin and EDTA yielded a population reduction ranging from 2.5 to 4.7 log cycles across the 20 Salmonella sp. Because microbial inhibition on skin surfaces by nisin-

7 J. Food Prot., Vol. 63, No. 9 NISIN-COATED POLYMER FILMS 1195 FIGURE 4. Effect of a nisin-based formulation (treatment C) on the refrigerated (4 C) shelf life of broiler drumsticks expressed as CFU/cm 2 (trial 1; n 4). Treatment C, 100 g/ml nisin, 5.0 mm EDTA, 3.0% citric acid, 0.5% Tween 80, ph 3.5. containing formulations is a time-dependent phenomenon, incorporating higher concentrations of nisin and other treatment components may offset the need for longer contact times to achieve the desired population reductions. Thus, the lethality of treatment C containing varying nisin concentrations ranging from 100 to 500 g/ml was determined against Salmonella Typhimurium NAr populations on broiler drumstick skin. The results are expressed as the surviving population of Salmonella Typhimurium NAr recovered from the skin and film rinse solution (Fig. 3). The surviving populations at 24 h ranged from log 3.1 in the 100- g/ml nisin treatment to log 1.5 in the 500- g/ml treatment. At 48 h, the surviving populations ranged between 2.5 and 0.6 logs for the 100- and 500- g/ml nisin treatments, respectively. At 72 h the surviving populations ranged from no detectable organisms (300 and 500 g/ml nisin treatments) to 1.5 logs (100 g/ml nisin). Similarly, at 96 h the surviving populations ranged from nondetectable (175 to 500 g/ml nisin) to 0.3 logs (100 g/ml nisin). Generally, higher concentrations of nisin and longer contact times yielded greater inhibition. Films coated with 300 and 175 g/ml nisin or greater yielded no detectable organisms after 72 and 96 h of incubation, respectively. Besides contact time and concentration, nisin activity is also influenced by incubation temperature. In a study by Stevens et al. (21), higher population reductions of six Salmonella sp. were obtained with increasing incubation temperature ranging from 4 C to42 C. The greatest lethality occurred between 30 and 42 C. Further study is warranted to determine which of these films would be more effective at elevated exposure temperatures such as found under temperature abuse situations. Shelf-life studies. Summaries of the two broiler drumstick shelf-life studies are presented in Figure 4 for trial 1 and Figure 5 for trial 2. Drumsticks submerged in treatment C for 3 min and then packaged in traypacks containing FIGURE 5. Effect of a nisin-based formulation (treatment C) on the refrigerated (4 C) shelf life of broiler drumsticks expressed as CFU/cm 2 (trial 2; n 6). Treatment C, 100 g/ml nisin, 5.0 mm EDTA, 3.0% citric acid, 0.5% Tween 80, ph 3.5. nisin-treated overwrap films and tray pads had lower initial (day 0) mesophilic (0.70 log reduction) and psychrotrophic (1.07 log reduction) populations than drumsticks submerged in distilled, deionized water (Fig. 4). The initial log mesophilic bacterial population averaged 1.52 CFU per cm 2 for the control drumsticks and 0.82 CFU for the treated drumsticks. In both treatments, the mesophilic and psychrotrophic populations increased gradually throughout refrigerated storage with the nisin-treated samples generally having lower mesophilic and psychrotrophic populations. However, the rate of growth appeared to be similar for both treatments. Off odors and slime formation on refrigerated poultry samples are generally first detected when aerobic plate counts reach a population of approximately log 8.0 CFU per cm 2 (1). Shelf life of broiler carcasses is dependent on the initial level of carcass contamination with lower initial populations leading to longer shelf life. In related shelf-life studies of broiler carcasses, a small decrease in the initial carcass psychrotrophic population ( 0.5 log reduction) following exposure to chlorine dioxide extended the shelf life by 2 days when compared to the untreated control. In the present study there was a and a 1.07-log reduction in the day 0 mesophilic and psychrotrophic counts, respectively, following immersion in the nisin formulation and subsequent packaging with treated overwrap film and drip pads. However, the difference in the estimated end of shelf life (log 8.0 CFU per cm 2 ) for the nisin and control treatments was minimal (14 h shelf-life extension based on the psychrotrophic count and 10 h extension based on the mesophilic count). The results of the second shelf-life trial are summarized in Figure 5. The treatment patterns for the mesophilic and psychrotrophic populations were similar to those observed in trial 1. Throughout the storage period the nisin-treated drumsticks had significantly lower mesophilic and psychrotrophic populations than the control drumsticks. The mesophilic and psychrotrophic counts on day 0 were 1.71 and

8 1196 NATRAJAN AND SHELDON J. Food Prot., Vol. 63, No logs lower on the nisin-treated carcasses than the control carcasses. The estimated time at 4 C required for the carcass psychrotrophic population to reach log 8.0 CFU per cm 2 was approximately 7.4 days for the control and 9.6 days for the treated drumsticks (Fig. 5). Based on mesophilic counts, the projected number of days to reach log 8.0 CFU per cm 2 was 8 days for the control and 9.8 days for the nisin-treated drumsticks. Thus, the projected end of shelf life for nisin-treated drumsticks was approximately 2 days longer than the controls when psychrotrophic counts are used as the shelf-life determinant. In comparison to trial 1, the longer shelf life predicted in trial 2 for nisin-treated drumsticks may be associated with the higher initial bacterial populations detected on the control drumsticks (around a 2-log difference between trials). With higher initial populations, there may be greater opportunity for the nisin treatments to exert their effect than when starting bacterial populations are lower. These studies demonstrated that nisin treatment formulations applied as immersion dips or to packaging films and absorbent tray pads are capable of inactivating a portion of the Salmonella Typhimurium population on inoculated chicken skin plus marginally extending the shelf life of treated drumsticks. ACKNOWLEDGMENTS This research was funded in part by both the North Carolina Agricultural Research Service and the United States Poultry & Egg Association. The authors thank S. A. Hale and R. Qureshi for technical assistance in conducting these studies. REFERENCES 1. Ayres, J.C Microbiological implications in handling, slaughtering and dressing of meat animals. Adv. Food Res. 6: Barefoot, S. F., and T. R. Klaenhammer Detection and activity of Lactacin B bacteriocin produced by Lactobacillus acidophilus. Appl. Environ. Microbiol. 45: Buchman, G. W., S. Banergee, and J. N. Hansen Structure, expression and evolution of a gene encoding the precursor of nisin, a small protein antibiotic. J. Biol. Chem. 263: Council for Agricultural Science and Technology Foodborne pathogens: risks and consequences. Task Force Rep. 122: Doyle, M. P., and D. O. Cliver Salmonella, p In D. O. Cliver (ed.), Foodborne diseases. Academic Press Inc., San Diego, Calif. 6. Federal Register Nisin preparation: affirmation of GRAS status as a direct human food ingredient. Fed. Regist. 58: Fowler, G. G., and B. McCann The growing use of nisin in the dairy industry. Aust. J. Dairy Technol. 26: Hansen, N. J Nisin as a model food preservative. Crit. Rev. Food Sci. Nutr. 34: Hardaway, K. L., and C. S. Buller Effect of ethylenediaminetetraacetate on phospholipids and outer membrane function in Escherichia coli. J. Bacteriol. 137: Hurst, A., and G. J. Drig The relationship of the length of lag phase of growth to the synthesis of nisin and other basic proteins of Streptococcus lactis grown under different cultural conditions. J. Gen. Microbiol. 50: Jung, D. S., F. W. Bodyfelt, and M. A. Daeschel Influence of fat and emulsifiers on the efficacy of nisin in inhibiting Listeria monocytogenes in fluid milk. J. Dairy Sci. 75: Lieve, L The barrier function of the gram negative envelope. Ann. N.Y. Acad. Sci. 235: Lipsinka, E Nisin and its applications, p In M. Woodbine (ed.), Antibiotics and antibiosis in agriculture. Butterworth, London. 14. Nikaida, H., and M. Vaara Outer membrane, p In F. C. Neidhart (ed.), Escherichia coli and Salmonella typhimurium, vol. 1. American Society for Microbiology, Washington, D.C. 15. Ray, B Nisin of Lactococcus lactis subsp. lactis as a food biopreservative, p In B. Ray and M. Daeschel (ed.), Food biopreservatives of microbial origins. CRC Press, Inc., Boca Raton, Fla. 16. Roberts, T Microbial pathogens in raw pork, chicken, and beef: benefit estimates for control using irradiation. Am. J. Agric. Econ. 67: Rosen, J. M Surfactants and interfacial phenomena. Wiley- Interscience Publ., New York. 18. SAS Institute Inc JMP user s guide, version 2. SAS Institute Inc., Cary, N.C. 19. Shefet, S. M., B. W. Sheldon, and T. R. Klaenhammer Efficacy of optimized nisin-based treatments to inhibit Salmonella typhimurium and extend shelf life of broiler carcasses. J. Food Prot. 58: Stevens, K. A., B. W. Sheldon, N. A. Klapes, and T. R. Klaenhammer Nisin treatment for inactivation of Salmonella species and other gram-negative bacteria. Appl. Environ. Microbiol. 57: Stevens, K. A., B. W. Sheldon, N. A. Klapes, and T. R. Klaenhammer Effect of treatment conditions on nisin inactivation of gram-negative bacteria. J. Food Prot. 55: Surkeiwicz, B. F., R. W. Johnson, A. B. Moran, and G. W. Kruman A bacteriological survey of chicken eviscerating plants. Food Technol. 23: