Effect of the addition of microcapsules with avocado peel extract and nisin on the quality of ground beef

Abstract This study evaluated the incorporation of microcapsules containing nisin and avocado peel extract on the shelf life of ground beef. Ten treatments were studied and divided into two groups: one packaged under vacuum and the other in permeable packaging. Each group contained: (a) control, (b) extract, (c) nisin, (d) empty microcapsules (only wall microcapsule system), and (e) microcapsules with extract and nisin. The samples containing the microcapsules presented lower bacterial growth and less oxidation. On day 10, the vacuum‐packaged samples with microencapsulated preservative presented a reduction in the oxidation of proteins of approximately 45%, of 30% in the growth of mesophiles, and of 38% in the growth of coliforms, as well as a reduction in the changes in the pH and ɑ W that occur during storage, compared with the permeable control. The combination of microcapsules with vacuum packaging reduced the physicochemical and microbiological changes that occur in the controls.

tend the shelf life of meats and other foods and to reduce undesirable reactions. These strategies include the inclusion of antioxidants and antimicrobials, for which consumers encourage the use of those from natural origins (Falowo et al., 2014;Salehi, Tumer, et al., 2019;Sharifi-Rad et al., 2018), also if they have health benefits (Salehi, Sharifi-Rad, et al., 2019;Salehi, Vlaisavljevic, et al., 2019). A good source of antioxidants and antimicrobials are agro-industrial residues, such as the seed and peel of fruits (Ayala-Zavala et al., 2011). Avocado peel extract has been demonstrated to have an antioxidant capacity because of its polyphenol content, and there have been reports about its antimicrobial activity or synergic effect with other natural antimicrobials, such as nisin (Calderón-Oliver et al., 2016;Fu et al., 2011;Rodríguez-Carpena, Morcuende, Andrade, Kylli, & Estevez, 2011).
However, the addition of some preservatives, such as polyphenols in food, produce sensorial changes, could interact with other compounds, reducing their effectiveness as antioxidants, or could be degraded by proteases, which is the case for nisin in meat (Fang & Bhandari, 2010;Liu & Hansen, 1990). The encapsulation process of those preservatives decreases their interaction with the matrix of the food, masks undesirable odors or flavors, and regulates their liberation into the medium (Shahidi & Han, 1993).
Microencapsulation by complex coacervation is a method to encapsulate various compounds or probiotics with good encapsulation efficiency, ensuring the release of the bioactive compound (Comunian et al., 2013;Eratte et al., 2015). The complex coacervation involves the interaction of two biopolymers with opposite charge to form the capsule matrix (Gouin, 2004). In previous work, the elaboration and characterization of microcapsules made with this technique were reported, and the collagen-pectin system in a proportion of (1:1) with the spray dryer process had a higher encapsulation efficacy of nisin and avocado peel extract (84.66 ± 1.2 and 82.96 ± 1.25%, respectively) (Calderón-Oliver, Pedroza-Islas, Escalona-Buendía, Pedraza-Chaverri, & Ponce-Alquicira, 2017).
Therefore, this study aims to evaluate the effect of microcapsules containing a mixture of avocado peel extract and nisin (an antioxidant plus an antimicrobial) on the quality of minced meat.
This evaluation considers the study of three factors (time, type of package, and addition microcapsules) on the physicochemical and microbiological responses in minced meat.
It is important to mention that the composition of the avocado peel extract, the optimization of the nisin-extract mixture, and the microencapsulation system is already previously reported by our group (Calderón-Oliver et al., 2016.

| Chemicals
Nisin Z (2.5% w/w balanced with sodium chloride and denatured milk solids, 10 6 IU/g) was provided by Handary S.A. Avocado peel extract with antioxidant activity (ORAC value of 285.18 µg Trolox equivalents/mg) was extracted according to a previous study (Calderón-Oliver et al., 2016). Hydrolyzed collagen (Peptan 5000) was purchased from Rousselot. Partially amidated low-methoxyl pectin (standardized with sucrose, Genu LM-104 AS) was purchased from CP Kelco. The minced beef meat was purchased in a local market with TIF certification (Mexican safety and health inspection regulation certification). The meat samples were transported and stored in refrigeration (4°C) until use. The composition of meat samples was measured according to the Association of Official Analytical Chemists methods (air-oven for moisture, Kjeldahl method for crude protein, Soxhlet method with ether petroleum extraction for lipid content and ashes with weight difference after incineration at 525°C for 4 hr (AOAC, 1998)).

| Microcapsules preparation
Microcapsules were prepared by complex coacervation according to the method previously reported by our group (Calderón-Oliver et al., 2017). Briefly, a W/O emulsion (1:2 ratio) that contained 0.15 g/ ml of avocado peel extract and 0.1 g/ml of nisin was mixed with a solution of collagen 1% (w/v) and pectin 1% (w/v); then, the pH was adjusted to 3. After 24 hr of storage, the solution was spray-dried (140°C inlet air and 70°C outlet air). The encapsulation efficiencies of avocado peel and nisin in these microcapsules were 84.66 ± 1.20 and 82.96 ± 1.25%, respectively. The empty microcapsules were prepared using the method described above, without the addition of avocado peel extract and nisin.

| Incorporation of microcapsules in meat
Ten treatments were studied to compare the effects of microcapsules in three independent batches of meat. Each treatment is described in Table 1, which includes a control (C), empty microcapsules (EM), avocado extract (AE), and nisin (N), and others with microcapsules that contain the extract and nisin (MAN). Five treatments were packaged under an oxygen permeable system (code as P), and the other five were packaged under vacuum conditions (code as V). The corresponding number of microcapsules, AE or N (Table 1), were added to 200 g of meat, mixed for 3 min using a kitchen aid mixer, and then packaged and stored at 4°C for up to 10 days. Independent samples were taken for analysis on the 1st, 3th, 6th, 8th, and 10th days. The microbiological analyses include the count of mesophiles, coliforms, and lactic acid bacteria. Physicochemical analyses include evaluation of pH, ɑ W , and oxidation of lipids and proteins. Each determination was evaluated in triplicate. Only one concentration of microcapsules was used because it incorporated the maximum amount of nisin allowed by the Codex Alimentarius for meat products (Codex Alimentarius, 2019).

| Microbiological analyses
Twenty-five grams of meat from each treatment were added to 225 ml of sterile saline solution (at 0.85%) and homogenized on a

| pH
Ten grams of meat were added to 100 ml of distilled water and homogenized in an Ultra-Turrax (IKA) at 7,000 rpm for 1 min. Then, the homogenate was filtered (Whatman #4), and the resultant filtrate was used for pH measurement using a potentiometer (Orion, Versa Star, Thermo Fisher Scientific) (Schilling et al., 2018).

| Water activity (ɑ W )
Water activity determination was made in a water activity meter (Aqualab 4TE, Decagon Devices) by placing approximately 2 g of meat in the sample cup and following the manufacturer's instructions (Capita, Álvarez-González, & Alonso-Calleja, 2018).

Lipid oxidation (TBARS method)
Lipid oxidation was determined using the method described by Salih, Smith, Price, and Dawson, (1987) with some modifications. Five grams of meat were homogenized at 11,000 rpm for 2 min in 15 ml of 4% perchloric acid. Then, the homogenate was filtered and centrifuged at 1,157 g for 10 min at 4°C. Two milliliters of the supernatant was added to 2 ml of 80 mM 2-thiobarbituric acid. The reaction was developed via incubation for 30 min at 100°C. The absorbance of the samples and the standard curve (1,1,3,3-tetramethoxypropane in perchloric acid in concentrations of 1.5-24 μM) were measured at 530 nm. The results are expressed as mg of malondialdehyde (MDA)/kg.

Protein oxidation
Protein oxidation was determined by derivatization of total carbonyls with 2,4-dinitrophenylhydrazine, according to the method described by Oliver, Ahn, Moerman, Goldstein, and Stadtman, (1987) with modifications. One gram of meat was homogenized with 10 ml of 100 mM buffer phosphates pH 7.4 and then centrifuged at 1,157 g for 5 min.
The supernatant was divided into two aliquots of 0.2 ml, and 1 ml of 10% trichloroacetic acid was added to each aliquot to precipitate proteins. After centrifugation of 3,214 g for 5 min, one of the pellets was treated with 1 ml of 2 M HCl (for protein determination), and the second was treated with 1 ml of 2,4-dinitrophenylhydrazine (0.2% in 2 M HCl) for carbonyls determination. Both aliquots were incubated for 1 hr at room temperature, in the dark. Then, 1 ml of 10% trichloroacetic acid was added and incubated for 10 min at 4°C and centrifuged at 10,000 g for 10 min at 4°C. The incubation with trichloroacetic was repeated twice. Each pellet was treated with 1 ml of ethanol-ethyl acetate (1:1) and centrifuged at 10,000 g for 10 min at 4°C. This operation was repeated twice. The final pellets were dissolved with 1 ml of 6 M guanidine HCl (pH 6.5), incubated for 10 min at 37°C and centrifuged at 10,000 g for 10 min at 4°C. The absorbance was determined at 370 nm for the tubes with dinitrophenylhydrazine and 280 nm for the tubes with HCl. The results are expressed as concentration of carbonyls (nM)/mg of protein.

| Statistical analysis
For factorial design: Two-way ANOVA and Tukey's multiple comparisons test were conducted utilizing the XLSTAT software (version TA B L E 1 Treatment groups Additionally, a statistical analysis was carried out considering the 10 samples as levels of a single factor to find the differences per day, and a second analysis considering the days as levels of a factor for a single sample.

| Physicochemical characteristics of different treatments of meat
The pH variable presents a significant difference between treatments, days of storage, and type of package (Tables 2 and 3, and A-H: significant difference between treatments in the same day (p < .05), a-e: significant difference between days in the same treatment (p < .05). In contrast, ɑ W values decreased as a function of time (Tables 2   and 3), and the treatments with microcapsules, as well as the permeable package samples, were the most affected. The decrease in ɑ W values could be related to the microcapsules' capacity for holding water with some polysaccharides, such as pectin, for its functional groups (Vaclavik & Christian, 2014). These characteristics of polysaccharides, plus the presence of the extract and nisin, might be associated with the lower microbial load for treated samples.
Both pH and ɑ W , as well as other parameters such as temperature, are factors that can destabilize the microcapsules made by the coacervation method because these factors are controlled during their elaboration and allow for interaction between the polymers that make them, thus affecting the release of the encapsulated bioactive compound (Qv, Zeng, & Jiang, 2011).
agro-industrial residues that contain polyphenols decrease the oxidation in meat and meat products (Monteiro et al., 2014;Zhang, 2015).
Surprisingly, when encapsulated by other techniques (e.g., spray drying), the activity of these extracts decreases or changes. This is the case for the extract of Jabuticaba peel, which does not induce a significant decrease in oxidation levels when applied on mortadella (Baldin et al., 2018).
The decrease in the oxidation of lipids and proteins can translate into an extended shelf life for the product, as well as a reduction in sensory changes such as rancidity.

| Microbiological stability of meat with microcapsules
The storage time, type of packaging, and presence of microcapsules show a significant (p < .05) ( Table 2) effect on microbial growth.
In general, the microbial populations increased over time (Table 5 and Figure S2)

| Principal component analysis (PCA) of different meat treatments and overall overview
Principal component analysis was performed to study which variable responses were correlated or best describe each treatment. In Figure 1a,b, factor 1 (x-axis) explains 68% of the total variability in the data. This means that the variable responses, such as bacteria count, oxidation of macromolecules, and ɑ W , were correlated in 68% of the samples. The microbiological and oxidative variable responses describe the last days of storage (day 8 and 10), while the first and third day were positively correlated with the ɑ W response. The main differences between treatments were marked by differences in bacterial growth and oxidation, which are important for estimating product quality. Factor 2, which explains 14.46% of the variance, separates the samples as a function of the pH, and there is a tendency for the treatments with permeable packages to be correlated with the pH values. The correlation is more evident at the end of storage time.
The samples with microcapsules containing nisin and avocado extract present a slight tendency to behave similar to the samples from the analysis on the previous day, by decreasing the oxidation of lipids and proteins and microbial growth.

| CON CLUS ION
The addition of microcapsules with an antioxidant and an antimicrobial effect on minced meat, such as avocado peel extract and nisin, decreases the oxidation of lipids and proteins and decreases the growth of bacteria such as mesophiles and BAL. These findings indicate that microcapsules increase the effect of antimicrobial and antioxidant properties in the minced meat, and its effect is better when combined with other technology such as vacuum packaging. These microcapsules could be used as natural preservatives in the food industry to reduce the concentration of some preservatives or eliminate the use of synthetic preservatives.

ACK N OWLED G M ENTS
The authors thank CONACyT for the scholarship (230991) of Mariel Calderón-Oliver during her Ph.D. studies.

CO N FLI C T O F I NTE R E S T
The authors declare that they have no conflict of interest.

E TH I C A L A PPROVA L
The study did not involve any human or animal testing.