**3. Results and Discussion**

A higher antioxidant activity was found in the PE in comparison with the CM containing ascorbic acid (554.42 ± 3.38 and 1149.02 ± 13.69 μg TE/100 g in CM and PE, respectively). The high antioxidant activity recorded in the PE is in agreemen<sup>t</sup> with a previous studies that reported the powerful antioxidant activity of olive phenolic compounds [22,41–44]. The results for the lipid oxidation and antioxidant activity of minced beef meat indicated by TBARs and ORACFL values are reported in Table 2.


**Table 2.** Lipid oxidation (TBARS) and antioxidant capacity (ORACFL) in beef burger during storage.

C = control group; A = basic recipe with addition of 10 g/kg commercial antioxidant; AP = basic recipe with addition of 5 g/kg commercial antioxidant and 350 mg/kg phenolic extract; P = basic recipe with addition of 700 mg/kg phenolic extract. Different letters in the same row (a, b, c, d) indicate differences between mean values during sampling times (*p* ≤ 0.001); different letters in the same column (W, X, Y, Z) indicate differences between mean values for different experimental groups (*p* ≤ 0.001). SEM, standard error of the mean. T = time; S = sample.

Concerning lipid oxidation, increased TBARs values were detected during storage for all experimental formulations, and, among groups, differences were reported starting from 2 days with higher values for C samples. The addition of 700 mg/kg PE in meat (P group) guaranteed the lowest level of TBARS in the second part of storage (days 5–7) with a reduction of 62% of lipidic oxidation on day 7 for P compared to C samples. Similar results, albeit with a lower magnitude, have been reported by Martínez-Zamora et al. [43], demonstrating how the incorporation of synthetic hydroxytyrosol reduced the oxidation of lamb patties by 35% with respect to the control sample at the end of shelf-life. The hamburger formulation and storage time significantly affected the lipid oxidation, and the interaction between these two factors was also significant (Table 2). A PE-concentrationdependent effect preserving lipid oxidation was also observed by other studies in the literature both in raw and grilled beef burger [20].

The ORACFL assay revealed differences in the antioxidant activity of four experimental groups, with the highest mean values recorded in the P group (Table 2). A reduction in the antioxidant activity during storage was recorded in A, AP and P. This reduction, as found in a previous study, is due mainly to the oxidative degradation phenomena of phenols or antioxidant molecule that occur during storage [20,44]. Despite the degradation of phenols, the higher level of integration in beef meat (P group) ensures considerable antioxidant activity until the end of storage time. It is well-known that phenols can act as hydrogen donors and compounds linked with a o-dihydroxyl functionality possess a high antioxidant activity, due to the formation of intramolecular hydrogen bonds observed during the reaction with free radicals [22]. Among these, the highest antioxidant activity was attributed to 3,4-DHPEA (hydroxytyrosol) and secoiridoid derivatives such as 3,4- DHPEA-EDA (oleacein) [22]. In particular, hydroxytyrosol's strong antioxidant potential is strictly related to its chemical structure: a phenol ring formed by a catechol group and three hydroxyl groups [43]. The combination of these functional groups could represent the main explanation for its preservative action in products of animal origin, as previously demonstrated in the literature [45,46].

The preliminary evaluation of the antimicrobial activity of the CM and PE performed through the agar well diffusion technique revealed that the PE possessed greater in vitro antimicrobial activity compared to the CM. Indeed, after incubation, the inhibition halos were measured for each strain, and the PE showed halos of 16, 13 and 11 mm for *P. fluorescens*, 14, 11 and 9 mm for *S. aureus*, and 9, 7 and 0 mm for *E. coli* for 750, 375 and 187 mgPE/mL, respectively. The assay's results sugges<sup>t</sup> that the CM had no effect on microbial growth as the inhibition halos were absent for all the concentrations and micro-organisms tested.

Concerning the microbial analysis of beef burger, the *Salmonella* spp. detection showed that the pathogen was absent during the entire duration of products' shelf-life in all the

experimental groups, complying with the food safety criterion of EU Regulation [37]. Similarly, the process hygiene criterion of *E. coli* in meat preparations was fully respected as the microbial count at T0 was below 2 Log CFU/g for all experimental groups [37]. This evidence confirms both the satisfactory safety and the hygiene levels of the beef burger production process.

The results of microflora evolution during storage for refrigerated beef burgers are depicted in Table 3. As shown, a significant (*p* < 0.001) increase was observed for all microbial populations and for all experimental groups studied as storage time elapsed. Immediately after production (T0), higher microbial populations were recorded for the TVC and LAB count followed by *Pseudomonas* spp. The initial (T0) TVC in all studied groups was approximately 4.4 Log CFU/g, which can be considered a characteristic value for minced meat products after manufacturing [47]. Indeed, this result is in agreemen<sup>t</sup> with the levels reported in the available literature for similar minced beef meat products [48,49], albeit other studies have found higher values [50] or lower ones [51]. It has been reported that a possible explanation for this relatively high initial TVC contamination in beef burgers may be attributed to the mincing process, which contributes to the total viable counts, likely as a consequence of the disruption of muscle structure, making nutrients easily available to micro-organisms [49]. However, the low initial value of *Enterobacteriaceae* counts (average value 1.33 Log CFU/g) confirms the optimal initial microbiological quality attributable to the good physiological status of the animal at slaughter and to proper postmortem meat acidification as well as to the good hygienic conditions during slaughter, handling and production processes [47,52].

**Table 3.** Microbial quality (Log CFU/g) of the four formulations of beef burger stored at 4 ◦C under aerobic conditions for 7 days.


C = control group; A = basic recipe with addition of 10 g/kg commercial antioxidant; AP = basic recipe with addition of 5 g/kg commercial antioxidant and 350 mg/kg phenolic extract; P = basic recipe with addition of 700 mg/kg phenolic extract. Different letters in the same row (a, b, c, d) indicate differences between mean values during sampling times (*p* ≤ 0.001); different letters in the same column (W, X, Y, Z) indicate differences between mean values for different experimental groups (*p* ≤ 0.001). SEM, standard error of the mean. T = time; S = sample.

As above mentioned, following TVC, LAB and *Pseudomonas* spp. where the two microbial population with the highest initial value (T0), with average levels among experimental groups of 4.09 and 4.05 Log CFU/g, respectively. Similar counts were recorded for analogous products by Zamuz et al. [47] and Andres et al. [53], while slightly higher values were recorded by Parafati et al. [54] and Marrone et al. [55] for *Pseudomonas* in minced beef meat products.

At the end of the storage period (day 7), significant (*p* < 0.001) differences between experimental groups were observed for TVC, *Pseudomonas* and *Enterobatteriaceae* (Table 3). Specifically, the lowest value for this microbial population corresponded to those burgers formulated with the highest amount of PE (P group), suggesting that the bioactive molecules contained in the extract affected this microbial population by limiting its growth, as preliminary suggested by an in vitro assay.

The parameters characterizing the growth curves of targeted microbial populations in the four experimental groups were obtained by modeling growth data by means of the Baranyi equation [40] and are summarized in Table 4.

Regarding the TVC, the addition of polyphenols in beef burger resulted in an extended lag phase (λ) in P samples in comparison with C and A and a reduction in the maximum growth rate (μmax) in AP compared with all other groups and the final value for P. For *Staphylococcus* spp., the experimental treatment affected the microbial growth by extending the λ in P samples in comparison with C and A and by reducing the μmax in P compared with all other groups and in AP compared to C. The final value was also affected by a significant reduction in the two experimentally manufactured burger groups (AP, P). Considering *Pseudomonas* spp., the P group recorded a longer λ compared to C and A samples while AP was not statistically different from the other experimental groups; concerning μmax, both AP and P showed lower values compared to C and A. For this microbial population, the final value was significantly reduced in the P group.

Neither λ nor the final value of the LAB population were different among groups, while the μmax was slightly higher in AP and P compared to C and A, agreeing with previous results from Servili et al. [56], who reported that olive by-product polyphenols did not affect LAB growth in fortified foodstuff. For *Enterobacteriaceae*, data show a reduction in the final value in the AP and P, while no significant differences were highlighted for λ and μmax; however, an increasing trend was recorded for λ and a decreasing one was noted for μmax (*p* = 0.06 and *p* = 0.056, respectively, data not shown).

In agreemen<sup>t</sup> with what is reported above, Mexis et al. [57] noticed a reduced growth rate in TVC, LAB and pseudomonads in ground chicken meat with the addition of Citrus spp. extracts. In Mortadella meat products with citrus fiber, thyme and rosemary, essential oil lowered the growth rate of the TVC during storage [58]. Roila et al. [25] reported that the addition of olive oil by-product polyphenols in Fior di Latte cheese brine resulted in an extended λ for *P. fluorescens* and *Enterobacteriaceae* and in a reduced μmax for *P. fluorescens*.

As shown, the result of growth data modeling appears to be related to the influence exerted by the experimental addition of olive-mill-wastewater-derived polyphenols on extending the λ and reducing the μmax and the final value. Analogous conclusions have been previously reported for similar compounds and for other preservation methods, albeit a systematic comparison is difficult as the characteristics of microbial growth curves can be affected by differences in food matrices, storage condition and duration [25,59].

Concerning the antimicrobial activity, it has been demonstrated that the dialdehydic structure of olive phenols exerts an antimicrobial effect by strongly interacting with amino acids, proteins and membrane molecules, promoting membrane permeabilization and bacterial cell lysis [60]. Indeed, studies have shown that tyrosol inhibits the activity of cyclooxygenase enzymes and hydroxytyrosol has a protein-denaturing ability [60]. Other studies report that many polyphenolic compounds are potent iron scavengers, and the lack of iron affects the growth of certain pathogenic bacteria by a reduction in the ribonucleotide precursor of DNA [61]. Besides the molecular mechanisms, the microbial growth inhibition exerted by this phenolic compound is strongly related to its chemical structure; as a

consequence, the key factor determining the antibacterial activity of the phenolic extracts is their phenolic profile [62].

**Table 4.** Output parameters estimated by the DMFit program for each microbial population in the four formulations of beef burgers.


λ = lag phase (h); μmax = maximum growth rate (Log/CFU/g/h); final value (Log/CFU/g); SE = standard error of fitting; R<sup>2</sup> = adjusted R-square statistics of the fitting. C = control group; A= basic recipe with addition of 10 g/kg commercial antioxidant; AP = basic recipe with addition of 5 g/kg commercial antioxidant and 350 mg/kg phenolic extract; P = basic recipe with addition of 700 mg/kg phenolic extract. Different letters in the same row (a, b, c) indicate differences between mean values for different experimental groups (*p* ≤ 0.001).
