**4. Discussion**

The present study shows for the first time the effect of the combined application of LA and UV-C on *L. monocytogenes* and LAB in vacuum-packed beef. A central composite design and Response Surface Methodology were used to optimize the concentration of LA and the dose of UV-C to reduce the population of *L. monocytogenes* and LAB without major changes in meat color.

The major findings of the present study were: (i) the quadratic model obtained allowed us to predict *L. monocytogenes* log reduction in vacuum-packed beef treated with LA and UVC, (ii) the maximum log reduction for both *L. monocytogenes* (1.55 ± 0.41 log CFU/g) and LAB (1.55 ± 1.15 log CFU/g) with minimal impact on meat color was achieved with the application of 2.6% LA and 330 mJ/cm<sup>2</sup> UV-C, and (iii) under these conditions, there was no increase in *L. monocytogenes* counts over 8 weeks of storage at 4 ◦C, and LAB growth was delayed by 2 weeks compared to control samples.

In the present study, the quadratic model obtained for predicting inoculated *L. monocytogenes* reduction had a good predictor value (*R*<sup>2</sup> = 0.9038). Both LA (0–5%) and UVC (0–330 mJ/cm2) were significant (*p* < 0.05) factors and had independent effects (no significative interaction (*p* >0.05)). The non-significant interaction between the factors indicated that the effects were additive, observed also for *Salmonella* inactivation in a different food matrix [26]. The reduction in *L. monocytogenes* increased as LA concentration and UV-C dose increased. According to the model, the combination of the maximum levels of LA (5.0%) and UV-C (330 mJ/cm2) reduced the 5.8 log inoculum by 1.73 log, a fraction of viable microorganisms remained in the sample, indicating the presence of a tailing effect. This is depicted in the 3D Response Surface plot (Figure 1a), where *L. monocytogenes* log reduction had an initial sharp increasing rate and then decreased at higher UV-C doses and LA concentrations. As mentioned before, there are no other studies reporting the combined action of LA and UV-C on *Listeria monocytogenes* in fresh meat, although similar inactivation patterns were observed in fresh beef treated separately with LA or with UV-C [9,15,27]. In this respect, a previous study from our group obtained a reduction of 1.13 log using 2.5% LA. DeGeer et al. 2016, using a 4% LA solution, reduced by 1.3 log a *L. monocytogenes* inoculum of 8 log and, Kalchayanand et al. 2020, using a 590 mJ/cm<sup>2</sup> UV-C dose, reduced a 6 log *L. monocytogenes* inoculum by 0.89 log. The observed tailing effect for *L. monocytogenes* inactivation in meat may be explained by the ability of the meat matrix to buffer the antimicrobial solution and to entrap *L. monocytogenes* into muscle fibers shielding the bacteria from LA and UV-C radiation [9,29].

LAB reduction by LA and UV-C was adjusted to a linear model in which the factors LA and UV-C were both significant (*p* < 0.05) and independent (no significative interaction (*p* > 0.05)). The low precision (*R*<sup>2</sup> = 0.5774) of the model for predicting the response in the design space was a consequence of the variability among the five replicas of the central point (Table 1). This variability may be attributed to the natural diversity of the LAB present in the meat samples, which may have different sensitivity to LA and UV-C [14,20]. The maximum level of LAB reduction matched with the highest LA concentration and UV-C dose used suggesting that both factors can be further increased to achieve a higher level of reduction as shown in the 3D Response Surface plot (Figure 1b).

However, it was not feasible to increase LA concentration because (i) regulations of USDA/FSIS and European Commission do not allow concentrations greater than 5%, and (ii) high LA concentrations produced unwanted color changes in meat. Regarding increasing the UV-C dose, our data suggested that an increase in UV-C dose would not achieve a larger reduction in the *Listeria monocytogenes* population. Though our model did not allow us to predict outside the design space, the level of *L. monocytogenes* reduction obtained at the +α experimental point applying 2.5% LA and 368 mJ/cm<sup>2</sup> (Table 1) was similar to the level of reduction achieved with 2.5% LA and 330 mJ/cm2. In agreemen<sup>t</sup> with this observation, McLeod, et al., 2017 [29] reported that the application of 3 J/cm<sup>2</sup> did not increase the level of *Listeria monocytogenes* reduction in chicken breast beyond the reduction level obtained with 0.3 J/cm2. The antilisterial effect of doses higher than 330 mJ/cm<sup>2</sup> combined with 2.5% LA needs to be further studied, as well as the effects on meat quality.

Meat color change, expressed as Chroma value, was only related to LA acid concentration, and was fitted to a linear model with one factor (Figure 1c). Chroma value detrimental change was mostly due to the decrease in redness value (a\*) (Table S1, Supplementary File), a well-known effect of lactic acid in beef [19]. The fact that UV-C doses applied did not have a significant effect on fresh meat color was in agreemen<sup>t</sup> with previous studies [15].

Using the models obtained for each response (*L. monocytogenes* reduction, LAB reduction and Color), RSM predicted that 2.6% LA concentration and 330 mJ/cm<sup>2</sup> of UV-C dose were the conditions that combined satisfied the constraints imposed (highest *L. monocytogenes* and LAB reduction and chroma value equal or larger than 20). The *L. monocytogenes* reduction predicted was 1.55 log. This reduction level was higher than the reduction levels obtained with the highest LA concentration or the highest UV-C dose alone.

Treatment with 2.6% LA and UV-C of 330 mJ/cm<sup>2</sup> at the time of packaging prevented the surviving fraction of the inoculated population of *L. monocytogenes* from thriving, showing a tendency to decrease with time when stored at 4 ◦C. In the untreated samples, the counts of *L. monocytogenes* increased (Figure 2a). LA and UV-C caused cellular injury in the fraction of survivors preventing them from overcoming the additional stress imposed by low oxygen and temperature. There are no reports of the combined application though a similar trend was reported for *L. monocytogenes* over time in beef treated with LA [27,30]. Regarding the behavior of *L. monocytogenes* in control samples, previous studies reported both growth [31,32] and inhibition [27,30,33] during storage at 4 ◦C. Differences in *L. monocytogenes* behavior may be due to variations in experimental conditions such as moisture and pH of meat samples, oxygen permeability of the vacuum bags and the *L. monocytogenes* strains used. In this study, the strain used was isolated from a refrigerated environment after having suffered different types of stress which, according to Skandamis et al., 2008, may affect subsequent stress tolerance.

LAB followed a sigmoidal growth curve [22]. Treatment prolonged the lag phase by two weeks, probably because a fraction of the remaining living cells were injured by UV-C and would have a slower growth rate or would be unable to replicate under stress [14]. The final LAB count in control samples reached 8 Log while in the samples treated with LA/UVC reached 7 Log. Though these results were relevant, more studies are needed to understand the impact on the combined LA/UV-C application on meat shelf life.

At time zero, redness was the only color component that was affected by the combined LA/UV-C application. As mentioned above, the decrease in initial a\* is mainly due to the application of LA [19,34]. After eight weeks of storage, treated samples had different values of L\* and a\* with respect to samples without treatment, however the chroma value was similar in both samples. The greater decrease in the value of L\* and a\* at week 8 in the treated samples may be due to the fact that both LA and UVC have the capacity to oxidize myoglobin and to cause lipid oxidation with the consequent loss of color in meat [26,35]. However, more studies need to be done to assess the color of treated meats over time.

In summary, the combined application of LA 2.6% and UV-C 330 mJ/cm<sup>2</sup> contributed to improving safety of vacuum packed beef, with a low impact on color. Although, more studies must be carried out regarding the effects on other bacteria present on meat and other physicochemical changes such as lipid and protein oxidation.
