**1. Introduction**

Ever since the U.S. Department of Agriculture Food Safety and Inspection Service (FSIS) declared *E. coli* O157:H7 and Shiga toxin-producing *E. coli* (STEC) as adulterants in non-intact beef [1], the North American beef industry has continuously evaluated and implemented the use of antimicrobial interventions during beef harvest and processing. In addition to STECs, Salmonella presence on beef has also been identified as a significant threat to public health and an economic burden to the beef industry. Just recently, Salmonella has been linked to foodborne outbreaks and millions of pounds of ground beef have been recalled for risk of Salmonella presence in ground beef [2,3]. Despite the industry efforts to implement proper sanitary dressing procedures, best practices, and use of antimicrobial interventions, hides, and endogenous extra-intestinal sources of pathogens can contaminate beef carcasses [4]. Not one single intervention has been found to render a beef product completely safe. Thus, a multi-hurdle approach of a series of targeted antimicrobial interventions can more effectively reduce the risk of possible contamination through

the slaughter process, consequently improving the microbial quality of carcasses [5]. A combination of physical and chemical interventions on beef carcasses and products may prove to be more effective than applying the same intervention at multiple stages of the slaughter and processing lines [6]. Therefore, exploring suitable and effective antimicrobial intervention alternatives may prove to be beneficial when finding synergies with already existing and implemented interventions that will further contribute to improving beef safety.

BioSecurity Technology has developed a novel ozone intervention known as Bio-Safe ™ cleaning solution [7]. Aqueous ozone's oxidation-reduction potential grants it the capacity to be used as a disinfectant by causing cell lysis and damaging nucleic acids [8]. Although the antimicrobial properties of ozone are well documented [9], previous studies assessing ozone's potential as an intervention in beef carcasses have had contradictory results, where some have significantly reduced *E. coli* O157:H7 concentration whereas others have found no significant difference than water wash (28 ◦C) treatments [10,11]. Whether an intervention works in a laboratory environment or not, does not determine its feasibility or effectiveness in the beef processing plant environment, and therefore in-plant validation studies must be conducted in a particular commercial beef processing plant to assess its real effectiveness. Lactic acid is listed in FSIS Directive 7120.1 as a safe and suitable ingredient in the production of meat products. It may be used on beef subprimals at the amount of 2 to 5 percent solution not to exceed 55 ◦C (131 ◦F). The same Directive states that ozone is safe for use on all meat products per current industry standards. There are no labeling requirements on these single-ingredient items providing the use of the substance is consistent with the FDA's definition of a processing aid, and the application on meat meets all water retention requirements of 9 CFR 441.10.

Because foodborne pathogens should not be introduced into the beef processing environment under any circumstance, *E. coli* O157:H7 and Salmonella surrogates have been developed to validate antimicrobial interventions in commercial beef processing plants without compromising safety [12]. In this study, we hypothesize that the aqueous ozone intervention will significantly reduce indicator microorganisms naturally present in beef carcasses and trim in a commercial beef processing plant environment. Furthermore, we also evaluated if this intervention significantly reduces an *E. coli* O157:H7 and Salmonella surrogate-cocktail inoculated in beef trim in a commercial beef processing facility.

### **2. Materials and Methods**

### *2.1. Intervention Parameters*

Lactic acid operation parameters as applied in the plant used for this study included a spray treatment solution with a temperature of 110–130 ◦F (43–55 ◦C), at 2–5% lactic acid concentration with a spray pressure ≥15 psi. Bio-safe by BioSecurity Technology (Ozone) intervention operating parameters included ozone generators which utilize oxygen molecules from the air (O2) and pass them through a corona field, splitting them into single atoms of oxygen (O1). These atoms combine with an O2 molecule to form a molecule of O3 (Ozone). After the intervention and immediate reaction with organic matter, it turns back into oxygen, leaving no harmful byproducts or residuals according to manufacturer's description and proprietary technology developed. Oxidation-Reduction Potential (ORP) instrumentation is used to monitor and control the reactivity and effectiveness of the sanitizing power of ozonated water. The aqueous ozone treatment spray had incoming water maintained at 50–75 ◦F (10–24 ◦C), the concentration was 1.5–2.3 ppm and the ORP was measured by an in-line meter between 700 and 900 mV with a spray pressure of ≥20 psi. Ozone application consisted of a multiple hurdle carcass intervention system with three treatment cabinets using the following specifications: 52 spray nozzles delivering 24.6 gpm with 5 s treatment time, 62 spray nozzles delivering 34.6 gpm with 5 s contact time, and 36 spray nozzles delivering 13.6 gpm with 20 s contact time for each cabinet, respectively. The cumulative application used was 72.8 gpm with a total of 30 s contact

time in carcasses. Moreover, the trim ozone intervention consisted on one treatment cabinet with 44 nozzles delivering 12.8 gpm with 18 s contact time.

#### *2.2. Evaluation of Natural Microbiota on Carcass and Trim*

For each repetition, in one production day, samples were randomly collected before and after treatment. A total of 20 carcasses were sampled before and after the final intervention. Of these carcasses, 10 were treated with lactic acid intervention and 10 with the ozone treatment intervention. Samples were taken before intervention at the harvest floor and after intervention at the hot box, for a total of 40 carcass swabs per repetition. The next day, trim was fabricated from the carcasses that were treated with the ozone intervention and lactic acid intervention, traced, and separated into different trim combos. Ten representative pieces of trim that came from the carcasses with the ozone intervention and 10 pieces of trim that came from carcasses with the lactic acid intervention were sampled before and after the trim intervention. The selected carcasses and trim were sampled on an area of 500 cm<sup>2</sup> using 25 mL buffered peptone water (BPW) EZ-ReachTM swabs (World Bioproducts, Mundelein, IL, USA). Carcasses were sampled on the foreshank area, trim was sampled on several points until reaching approximately the target area of 500 cm2. Samples were collected by Texas Tech University (TTU) trained personnel. Swab samples were immediately chilled and shipped overnight to the ICFIE-TTU Food Microbiology laboratory for microbiological analysis. Swab samples were homogenized in a stomacher (Model 400 circulator, Seward, West Sussex, UK) at 230 rpm for 1 min. Next, samples were serially diluted in 9 mL BPW (Millipore Sigma, Danvers, MA, USA) tubes and plated to determine total aerobic plate counts (APC), coliform counts, and *E. coli* counts using 3MTM PetrifilmTM (Saint Paul, MN, USA) plates. The counts of each sample were determined and converted to Log CFU/cm<sup>2</sup> for carcasses and Log CFU/sample for trim samples before statistical analysis. A total of six repetitions were conducted.

#### *2.3. Salmonella and E. coli O157:H7 Surrogate Inoculation in Trim*

### 2.3.1. Nonpathogenic Cocktail Preparation

Five non-pathogenic American Type Culture Collection (ATCC) Salmonella and *E. coli* O157:H7 surrogate strains were selected for this section of the study. These strains of non-virulent *E. coli* (BAA 1427, 1428, 1429, 1430, and 1431), when used as a cocktail, have been previously shown to mimic Salmonella and *E. coli* O157:H7 antimicrobial intervention behavior [13–16]. The use of surrogate strains to validate interventions in plant environments has been previously discussed and at times encouraged by FSIS USDA, which has allowed the use of such non-pathogenic surrogates with appropriate precautions [12]. The surrogate strains were independently propagated in a food grade biological safety level I (BSL-I) laboratory at TTU. Each ATCC strain was retrieved from a −80 ◦C freezer, separately transferred into 4 mL brain heart infusion (BHI; Becton, Dickinson and Company, Franklin Lakes, NJ, USA) tubes, and incubated at 37 ◦C for 18–24 h. Next, overnight enriched tubes were screened for *E. coli* O157:H7 and Salmonella presence using BAX® real-time *E. coli* O157:H7 Exact and Salmonella assays (Hygiena, Wilmington, DE). After found negative for both pathogen screenings, 500 μL of each enriched surrogate broth was transferred into 49.5 mL BPW tube and cleared to be used for the challenge study. Then, all five tubes were decanted onto a sprayer and mixed. The bottle sprayer was then used for trim target inoculation of 5–6 LogCFU/cm2.

### 2.3.2. Trim Inoculation and Quantification

For each repetition, chuck and shank trim were randomly selected for inoculation. A total of 15 pieces of chuck and 15 pieces of trim were inoculated using the sprayer. Each piece of trim was sprayed with the *E. coli* O157:H7 and Salmonella surrogate cocktail and allowed for 30 min of cell attachment while at ambient temperature. After attachment time, an area of 100 cm<sup>2</sup> was sampled using a 25 mL BPW EZ-ReachTM swab. Trim was next treated with the ozone treatment and immediately after intervention but before entering the production line, trim was sampled. All swabbed areas were marked with 100 cm<sup>2</sup> stamped area to ensure that the same area was not sampled repeatedly. Samples were collected by TTU trained personnel and shipped overnight to the TTU Food Microbiology laboratory for microbial enumeration. Swabs were homogenized in a stomacher at 230 RPM for 1 min. *E. coli* counts were determined using the TEMPO® system (Marcy-l'Étoile, France) following the manufacturer's instructions. TEMPO® cards were incubated at 35 ◦C for 22–28 h. *E. coli* counts were directly obtained from the TEMPO® Reader and converted to LogCFU/cm<sup>2</sup> before statistical analysis. A total of six repetitions were conducted.

### *2.4. Statistical Analysis*

All data were analyzed using R (Version 4.0.3) Statistical analysis software to evaluate differences between lactic acid and the ozone intervention and testing for a significant reduction of microbial loads after each intervention in the natural microbiota setting was performed. A two-way ANOVA was done using intervention type (ozone and lactic acid), sampling point (before and after intervention), and their interaction as fixed effects. For the surrogate study, a two-way ANOVA was performed using trim type (chuck and shank), sampling point (before and after intervention), and their interaction as fixed effects. Post hoc analysis was done using a pairwise T-test with Bonferroni p-adjustment method for multiple comparisons. If parametric assumptions were not met, the Kruskal–Wallis test was used as a nonparametric alternative for the ANOVA, with post-hoc analysis using Wilcoxon rank-sum tests with a BH p-adjustment method for multiple comparisons. Significant differences were evaluated at the 0.05 alpha level. Historical data of *E. coli* O157:H7 presumptive positives from the commercial beef processing plant where the challenge study was conducted was shared with TTU researchers for information purposes. Chi-square comparison to identify the difference in prevalence before and after the ozone intervention application by year and on a per month basis was conducted.
