Combined effect of oxygen-scavenger packaging and UV-C radiation on shelf life of refrigerated tilapia (Oreochromis niloticus) fillets

This study investigated the physicochemical, instrumental and bacterial parameters of tilapia fillets subjected to oxygen-scavenger packaging, alone or in combination with UV-C radiation at two doses (0.102 and 0.301 J/cm2), stored at 4 ± 1 °C for 23 days. The oxygen scavenger, both UV-C doses, and the oxygen scavenger combined with UV-C, independently of the dose, extended the shelf life in 5, 6 and 7 days, respectively, by decreasing the bacterial growth rate and the formation of degradation compounds (e.g., TVB-N and ammonia). Oxygen-scavenger packaging, alone or in combination with UV-C at 0.102 J/cm2 and 0.301 J/cm2 showed lower amounts of free amino acids (FAA; 34.39, 34.49 and 34.50 mg L-lysine/kg fish tissue, 3.63, 3.57 and 3.61 mg L- ornithine/kg fish tissue, 27.52, 27.63 and 27.67 mg L-arginine/kg fish tissue), biogenic amines (BA; 3.81, 3.87 and 3.89 mg cadaverine/kg fish tissue, 12.88, 12.91 and 12.86 mg putrescine/kg fish tissue, 2.41, 2.44 and 2.47 mg spermidine/kg fish tissue), redness (2.53, 2.55 and 2.59), yellowness (6.65, 6.69 and 6.72), lipid oxidation (1.52, 1.53 and 1.58 mg malondialdehyde/kg fish tissue) and protein oxidation (5.06, 5.11 and 5.18 nmol carbonyls/mg protein), with higher hardness (3273.41, 2652.98 and 2687.57 g) than control (air packaging; 41.97 mg L-lysine/kg fish tissue, 4.83 mg L- ornithine/kg fish tissue, 37.33 mg L-arginine/kg fish tissue, 4.82 mg cadaverine/kg fish tissue, 16.56 mg putrescine/kg fish tissue, 3.21 mg spermidine/kg fish tissue, 4.26 of redness, 8.17 of yellowness, 2.88 mg malondialdehyde/kg fish tissue, 9.44 nmol carbonyls/mg protein and 2092.58 g of hardness), respectively, on day 13 of storage when the control fillets were unfit for consumption (7 log CFU/g) (p < 0.05). However, in the same day of storage, both UV-C doses had similar values for BA (p > 0.05), higher amounts of FAA (44.28 and 44.13 mg L-lysine/kg fish tissue, 5.16 and 5.12 mg L- ornithine/kg fish tissue, 40.20 and 40.28 mg L-arginine/kg fish tissue), redness (4.86 and 5.33), yellowness (9.32 and 10.01), lipid oxidation (3.09 and 3.52 mg malondialdehyde/kg fish tissue) and protein oxidation (10.27 and 11.93 nmol carbonyls/mg protein), as well as lower hardness (1877.54 and 1767.39 g), respectively, than control fillets (p < 0.05). The combined preservation methods were the most effective in extending the shelf life and prolonging the physicochemical quality of the refrigerated tilapia fillets and the O2 scavenger proved to be a potential alternative to prevent the negative changes induced by both UV-C doses.

Oxygen scavenger system. In the OSP, OSUV1 and OSUV3 treatments, an oxygen-scavenger sachet was placed inside the package before sealing. The sachet used was the Ageless SS-50, with O 2 absorption capacity of 50 mL (Mitsubishi Gas Chemical Co., Inc., Tokyo, Japan). This sachet reduces oxygen levels through spontaneous iron oxidation, converting ferrous oxide (Fe 2+ ) to ferric oxide (Fe 3+ ) in the presence of oxygen, resulting in an O 2 concentration <0.01% according to information from the manufacturer (Mitsubishi Gas Chemical Co., Inc., Tokyo, Japan).
Free amino acids analysis. L-lysine, L-ornithine and L-arginine were analyzed as described by Gatti et al. 23 with modifications in the sample deproteinization step. In brief, 0.1 g of sample (tissue) was mixed with 1 mL of 1.5 M perchloric acid (v/v) to remove proteins. After 2 min at room temperature, 0.325 mL H 2 O and 0.5 mL potassium carbonate were added. The tubes were centrifuged at 10,000 × g for 2 min. The sample (50 μL) was mixed with 50 μL H 2 O and 40 μL of 2,5-dimethyl-1H-pyrrole-3,4-dicarbaldehyde (DPD) reagent solution (v/v) for 10 min. 360 μL of the mobile phase (0.05 M triethylammonium phosphate buffer) was added to the derivatized solution, which was immediately analyzed by HPLC. The HPLC device was equipped with an ACE C18 3-μm reversed-phase column (250 × 4.6 mm I.D.), a 5-μm Ascentis C18 reversed-phase guard column (20 × 4.6 mm I.D.) and an RF-10AXL photodiode array detector (SHIMADZU, Kyoto, Japan), monitoring the absorbance at 320 nm. The results were expressed as mg free amino acids/kg fish tissue.
Biogenic amines analysis. Cadaverine, putrescine and spermidine were determined according to the method of Lázaro et al. 24 , using an HPLC (SHIMADZU, Kyoto, Japan) equipped with a CBM-20A controller composed of an LC-20AD pump, SPD-M20A diode-array detector, CTO-20A oven and SIL-20AC autosampler. The amines were separated using a Spherisorb ODS2 C18 column (15 × 0.46 cm I.D., 5 μm particle size) for the stationary phase, and an acetonitrile:water mixture (42:58, v/v) as the mobile phase, under isocratic conditions. The biogenic amines were detected at 198 nm, and the results were expressed as mg biogenic amines/kg fish tissue.

Determination of total volatile basic nitrogen (TVB-N) and ammonia (NH 3 ). TVB-N was deter-
mined by Conway's microdiffusion method 25 and the results were expressed as mg TVB-N/100 g fish tissue. Ammonia was quantified by the colorimetric method, using a UV-1800 spectrophotometer (SHIMADZU, Kyoto, Japan) at 425 nm according to the protocol of Rodrigues et al. 11 . Results were expressed as µg NH 3 /g fish tissue, based on a standard curve (R 2 = 0.996) constructed from seven different NH 3 concentrations (1 to 15 µg NH 3 ).
Determination of lipid and protein oxidation. Lipid oxidation was evaluated by the thiobarbituric acid-reactive substances (TBARS) assay according to the method of Yin et al. 26 adapted by Joseph et al. 27 . The absorbance values were read at 532 nm, using a UV-1800 spectrophotometer (SHIMADZU, Kyoto, Japan), and the results were expressed as mg malonaldehyde (MDA)/kg fish tissue from a standard curve (R 2 = 0.999) constructed with eight different MDA concentrations (0.5 to 400 µmol). Protein oxidation was evaluated by the carbonyl content, following the method of Oliver et al. 28 with modifications 29,30 . The absorbance values were measured at 280 nm (protein) and 370 nm (carbonyl) by a UV-1800 spectrophotometer (SHIMADZU, Kyoto, Japan), and the results were expressed as nmol carbonyls/mg protein. Protein content was determined by a standard curve (R 2 = 0.999) constructed from five different concentrations of bovine serum albumin (0.1-1.0 mg), while the carbonyl content was calculated using an absorptivity coefficient for the protein hydrazones of 21.0/ mM/cm. Instrumental color measurements. Lightness (L*), redness (a*) and yellowness (b*) values were measured with an illuminant A, 8 mm-diameter aperture, and 10° standard observer through a Minolta CM-600D portable spectrophotometer (Minolta Camera Co., Osaka, Japan). The color parameters were determined at four random locations on the surface of each fillet immediately after it was removed from the packaging 31 .
Instrumental texture profile. The texture-profile analysis (TPA) was measured utilizing a TA.XTplus Texture Analyser (Stable Micro Systems, Surrey, UK) equipped with a cylindrical P/36 R probe. Each fillet was cut transversely into four pieces (2 × 2 × 2 cm 3 ), which were compressed twice to 50% of their original height with the time of 5 s between the two compression cycles, and pre-test, test speed, and post-test of 1 mm/s following conditions established by Sun et al. 32 . The parameters determined were hardness, chewiness, cohesiveness, springiness, and resilience. Statistical analysis. The experiment was conducted in duplicate, using a fully randomized design (n = 2).
A linear regression analysis was performed separately for each treatment to investigate the relationship between each physicochemical parameter and days of storage. The area under the curve (AUC), calculated by the trapezoidal method, was used to calculate the total amount of each physicochemical parameter produced during a time interval. To identify differences in the AUC among treatments (AP, OSP, AUV1, OSUV1, AUV3 and OSUV3), a one-way ANOVA was used. An additional post-hoc test with Tukey's adjustment was performed. All analyses were performed with a 0.05 confidence level, using GraphPad Prism version 5.00 (GraphPad Software, San Diego, California, USA). The bacterial growth curves were obtained by the predictive primary model 21 through the DMFit program version 2.0 (Institute of Food Research, Norwich, UK), and the differences among treatments regarding bacterial growth parameters (lag, log and stationary phases) were identified by one-way ANOVA with Tukey post-hoc test (p < 0.05).

Results and Discussion
Bacterial growth during storage. The results for TAMC, TAPC and Enterobacteriaceae are shown in Table 1 and Fig. 1a-c. The lag phase was absent in all bacterial groups. Although the number of colonies in the stationary phase of the fillets treated with the oxygen scavenger and/or UV-C radiation (0.102 or 0.301 J/cm 2 ) was higher than in the fillets in air packaging (AP), these emerging techniques alone or in combination extended the  Table 1. Bacterial growth parameters of tilapia (Oreochromis niloticus) fillets non-and treated with oxygen scavenger packaging (OSP) and ultraviolet radiation (UV-C) stored at 4 ± 1 °C for 23 days. Results are expressed as means ± standard deviation (n = 2). a,b,c,d Different letters in the same row indicate significant differences (p < 0.05) among treatments. ¥ TAMC -Total aerobic mesophilic count; TAPC -Total aerobic psychrotrophic count. # Lag -lag phase (h); EGR -exponential growth rate (log CFU/g/h); NC -number of colonies in the stationary phase (log CFU/g). € AP (air packaging); OSP (oxygen scavenger packaging); AUV1 (air packaging + UV-C at 0.102 J/cm 2 ); OSUV1 (oxygen scavenger packaging + UV-C at 0.102 J/cm 2 ); AUV3 (air packaging + UV-C at 0.301 J/cm 2 ); and OSUV3 (oxygen scavenger packaging + UV-C at 0.301 J/cm 2 ). www.nature.com/scientificreports www.nature.com/scientificreports/ shelf life of the tilapia fillets by decreasing (p < 0.05) the exponential growth rate (EGR) of the microorganisms ( Table 1). The initial bacterial counts were 4.24 log CFU/g for TAMC, 3.45 log CFU/g for TAPC and 2.78 log CFU/g for Enterobacteriaceae. Considering the limit of 3 log CFU/g for initial counts of Enterobacteriaceae proposed by the International Commission on Microbiological Specifications for Foods 33 , the tilapia fillets showed good initial microbial quality. The limit of 7 log CFU/g for TAMC proposed by ICMSF 33 was also used as the microbiological criterion to establish the shelf life of tilapia fillets during refrigerated storage. AP exceeded the limit of 7.0 log CFU/g for TAMC on day 9, while OSP, AUV1, AUV3, OSUV1 and OSUV3 reached this limit on storage days 14, 15, 15, 16 and 16, respectively.
The microbiota of tropical freshwater fishes such as tilapia is composed predominantly of Gram-negative aerobic and facultative anaerobic bacteria, including bacteria from the family Enterobacteriaceae, and Gram-positive bacteria 34 . Molinari et al. 35 and Pakingking et al. 36 , evaluating the microbiota of tilapia, found a wide variety of bacterial genera and species, including Pseudomonas spp., Shewanella putrefaciens, Aeromonas spp., Pasteurella pneumotropica, Photobacterium damselae, Plesiomonas shigelloides, Vibrio spp., Burkholderia cepacia, Chromobacterium violaceum, and Flavimonas oryzihabitans (Gram-negative aerobic and facultative anaerobic bacteria); Citrobacter spp., Edwardsiella spp., Enterobacter cloacae, Klebsiella oxytoca, Escherichia coli (Enterobacteriaceae); and Bacillus sp. and Staphyloccocus sp. (Gram-positive bacteria). However, along with the increase in the storage time under aerobic conditions, Pseudomonas spp. became the dominant spoilage bacteria in refrigerated fish, due to low temperature 34 .
In our study, the oxygen scavenger and both UV-C doses (OSP, AUV1 and AUV3) had similar effects on EGR for TAMC and Enterobacteriaceae. However, UV-C radiation (AUV1 and AUV3) showed a higher (p < 0.05) EGR for TAPC than OSP. This fact may be explained by antimicrobial effect of the UV-C 16,17 . Gram-negative bacteria are more sensitive to UV-C radiation due to their lack of a thick cell wall, which prevents UV-C absorption by microbial DNA 37 . Nevertheless, although Pseudomonas spp. are Gram-negative, they are resistant to radiation due to their ability to form a biofilm in response to UV-C induced stress, in an attempt to repair damaged DNA 38,39 . On the other hand, obligate aerobic bacteria such as Pseudomonas spp. are highly sensitive to low oxygen concentrations from O 2 scavenger 13 . Our results demonstrated that the O 2 scavenger delayed the EGR of Enterobacteriaceae, which are facultative anaerobic bacteria. This delay can be attributed to the sensitivity of these bacteria to carbon dioxide, which increases in the package headspace due to the relative decrease in the O 2 level caused by O 2 scavengers 13,40 .
With respect to the combined preservation methods, the oxygen scavenger plus UV-C radiation, at both doses (OSUV1 and OSUV3), was the most effective in delaying the EGR in all bacterial groups, indicating a synergistic effect between the two preservation methods. While the O 2 scavenger inhibits the growth of obligate aerobic bacteria and decreases the growth rate of facultative anaerobic bacteria of family Enterobacteriaceae by removing O 2 and increasing the CO 2 level inside the package, UV-C radiation decreases the growth rate of the microorganisms, especially Gram-negative bacteria, through direct or indirect damage to microbial DNA 13,16,17 .
OSP, AUV1, OSUV1, AUV3 and OSUV3 showed more viable cells in the stationary phase than AP. This difference may be explained by sublethal injury induced by CO 2 and UV-C radiation to bacterial cells, which at first grow more slowly than intact cells, and more rapidly after recovery, mainly in an environment without natural competition 17,41,42 .
In agreement with the present results, previous researchers demonstrated that O 2 scavengers were effective in extending the shelf life of refrigerated rainbow-trout fillets 14 and ground beef 43 by 5 and 2 days, respectively. Mohan et al. 15 found an extension of 6-7 days in the shelf life of sardines packed with an O 2 scavenger. Likewise, Bottino et al. 18 reported that UV-C at 0.055 and 0.160 J/cm 2 extended the shelf life of tambacu (Colossoma macropomum × Piaractus mesopotamicus) fillets stored at 4 °C by 50% and 100%, respectively. Monteiro et al. 1 observed that the shelf life of refrigerated tilapia fillets exposed to UV-C radiation at 0.103 J/cm 2 was extended by at least 2.5-fold.
Free amino acids and biogenic amines. The levels of free amino acids (L-lysine, L-ornithine, L-arginine) and biogenic amines (cadaverine, putrescine, spermidine) increased in all treatments throughout the storage period (p < 0.05; Table 2). AUV1 and AUV3 showed higher total amounts (p < 0.05), while OSP, OSUV1 and OSUV3 had lower (p < 0.05) total amounts of free amino acids than AP throughout the storage period ( Table 2). The results of free amino acids and biogenic amines in all days of storage can be found as Supplementary Table S1. The increase of free amino acids during storage is attributed to the action of endogenous and microbial proteolytic enzymes 44 . Our results are attributable to the resistance of Pseudomonas spp. to UV-C radiation, together with the effect of UV-C in increasing the amount of oxidized proteins, which are more susceptible to proteolysis, resulting in a high level of free amino acids 17,38,39 . On the other hand, oxygen scavenger is highly effective against Pseudomonas spp. 13 and it is able to minimize ROS-induced oxidation 45 .
Regarding biogenic amines, cadaverine, putrescine and spermidine are formed mainly by bacterial decarboxylation of precursor free amino acids such as L-lysine, L-ornithine and L-arginine, respectively 46 . Metabolization of L-arginine to L-ornithine is another pathway to formation of putrescine 46 , which explains the high amount of this amine in relation to others (cadaverine and spermidine). The present study found no difference (p > 0.05) in the total amounts of cadaverine, putrescine and spermidine among AP, AUV1 and AUV3; whereas OSP, OSUV1 and OSUV3 resulted in lower (p < 0.05) total amounts of these biogenic amines than the other treatments (Table 2). Although O 2 OSP, AUV1 and AUV3 had similar effect in controlling the growth of Enterobacteriaceae, which is the main bacterial group associated with the formation of biogenic amines 47 , UV-C radiation may cause oxidative decarboxylation of amino acids by catalyzing the production of Fe 3+ 48,49 . On the other hand, O 2 absorber has the capacity to minimize the oxidative reaction pathways by oxygen scavenging 45 , explaining our results for combined preservation methods (OSUV1 and OSUV3).
www.nature.com/scientificreports www.nature.com/scientificreports/ Currently, there is little information about the effect of O 2 absorbers and UV-C radiation on the production of free amino acids and biogenic amines in fish species during refrigerated storage. Similarly to our results, an increase in the amount of free amino acids by UV-C has been previously reported in fish stored at 4 °C 11,19 . Likewise, Mohan et al. 50 observed a delay in the formation of putrescine, cadaverine and spermidine by use of an O 2 scavenger in seer fish (Scomberomorus commerson) stored under refrigeration. However, no effect on the formation of putrescine, cadaverine and spermidine by similar UV-C doses was reported in freshwater fish species during refrigerated storage 11,19 . Total volatile basic nitrogen (TVB-N) and ammonia (NH 3 ). The initial levels of TVB-N and NH 3 were 10.08 ± 0.00 mg TVB-N/100 g and 7.66 ± 0.04 µg NH 3 /g fish tissue. As expected, the TVB-N and ammonia levels increased (p < 0.05) in all treatments during the storage period, with the highest increases in the tilapia fillets under aerobic packaging (AP; Table 3). However, no treatment exceeded the limit of 25 mg TVB-N/100 g  Table 2. Free amino acids and biogenic amines of tilapia (Oreochromis niloticus) fillets non-and treated with oxygen scavenger packaging (OSP) and ultraviolet radiation (UV-C) stored at 4 ± 1 °C for 23 days. Results are expressed as means ± standard deviation (n = 2). a,b,c Different superscripts in the same column indicate significant differences (p < 0.05) among treatments. ¥ AUC -Area under curve; AUC0-13 -from day 0 to 13 among treatments AP, OSP, AUV1, OSUV1, AUV3, and OSUV3; AUC15-23 -from day 15 to 23 among treatments OSP, AUV1, OSUV1, AUV3, and OSUV3. NA -Not applicable. € AP (air packaging); OSP (oxygen scavenger packaging); AUV1 (air packaging + UV-C at 0.102 J/cm 2 ); OSUV1 (oxygen scavenger packaging + UV-C at 0.102 J/cm 2 ); AUV3 (air packaging + UV-C at 0.301 J/cm 2 ); and OSUV3 (oxygen scavenger packaging + UV-C at 0.301 J/cm 2 ).
www.nature.com/scientificreports www.nature.com/scientificreports/  Table S2). In freshwater fish species, TVB-N values are related mainly to the ammonia concentration, due to absence or low level of trimethylamine oxide in vivo 2,52 . However, there is no limit for ammonia content in freshwater fish species. In the present study, at the point when the fillets were unfit for consumption (7 log CFU/g), the ammonia levels were 10.60 ± 0.03, 10.44 ± 0.32, 11.46 ± 0.05, 9.99 ± 0.22, 11.47 ± 0.03 and 10.04 ± 0.20 µg of NH 3 /g of fish tissue for AP, OSP, AUV1, OSUV1, AUV3 and OSUV3, respectively (Supplementary Table S2).
AP had the highest (p < 0.05) total amounts of TVB-N and ammonia produced during the storage period (AUC), followed by tilapia fillets exposed to UV-C radiation alone, at both doses (AUV1 and AUV3), OSP alone, and OSP and UV-C in combination (OSUV1 and OSUV3; Table 3). These results agree with those obtained for bacterial growth parameters (Table 1) and free amino acids (Table 2), which are important substrates for ammonia formation 34 . Similarly, Bottino et al. 18 and Monteiro et al. 1 observed that, although the initial formation of TVB-N and ammonia in freshwater fish flesh was increased by UV-C radiation, it was still delayed during the storage period as a whole, compared to their control counterparts. The effectiveness of an O 2 scavenger in reducing TVB-N and ammonia levels throughout refrigerated storage of fish species was also previously reported 13-15,45 . Lipid and protein oxidation. An increase in the malonaldehyde (MDA) and carbonyl levels was observed during refrigerated storage in all treatments, especially in AUV1 and AUV3 ( Table 3). The increases in lipid and protein oxidation by UV-C radiation were dose-dependent. AUV3 showed the highest (p < 0.05) MDA and carbonyl levels during the storage period, followed by AUV1, AP, and treatments with the O 2 scavenger (OSP, OSUV1 and OSUV3), which did not differ from each other (p > 0.05; Table 3).
Scientific RepoRtS | (2020) 10:4243 | https://doi.org/10.1038/s41598-020-61293-8 www.nature.com/scientificreports www.nature.com/scientificreports/ our findings may be attributed to pro-oxidant properties of the UV-C radiation 6,17,54 and capacity of the O 2 scavenger in minimizing ROS-induced oxidation 45 . In agreement with the results of this study, an increase in the MDA and carbonyl levels by UV-C radiation was observed during refrigerated storage of sea bass fillets 19 and tilapia fillets 4 . Some previous studies also found that an O 2 scavenger retarded the lipid oxidation of rainbow trout fillets 14 , fresh cobia 13 and sardines 15 stored under refrigeration; however, only limited information is available regarding the effect of O 2 scavengers on protein oxidation of fish species.
Two milligrams of MDA/kg is considered the limit above which meat is unfit for human consumption 55 . AP, OSP, AUV1, OSUV1, AUV3 and OSUV3 exceeded this limit on days 9, 19, 6, 19, 5 and 17, respectively (Supplementary Table S2). In spite of the importance of protein oxidation to food quality, there are no regulatory limits on carbonyl levels in meat products. In our study, when the acceptable microbial limit of 7 log CFU/g was reached, the carbonyl levels were 6.47 ± 0.06, 5.12 ± 0.00, 10.68 ± 0.00, 5.61 ± 0.40, 13.12 ± 0.22 and 5.58 ± 0.40 nmol of carbonyl/mg of protein for AP, OSP, AUV1, OSUV1, AUV3 and OSUV3, respectively (Supplementary Table S2). These results indicate the effectiveness of the O 2 scavenger in retarding oxidative processes, even when oxidation-inducing treatments were used.
Instrumental color measurements. Lightness (L*), redness (a*) and yellowness (b*) increased with the increasing storage period in all treatments (p < 0.05; Table 4). Throughout the entire storage period, no difference (p < 0.05) was found for L* values between treatments. AUV3 showed the highest (p < 0.05) a* and b* values during the entire storage period, followed by AUV1, AP, and treatments containing an O 2 absorber (OSP, OSUV1 and OSUV3), which did not differ from each other (p > 0.05; Table 4). As in the lipid and protein oxidation, UV-C radiation increased the a* and b* values in a dose-dependent manner. The results of instrumental color parameters in all days of storage can be observed in Supplementary Table S3.
The increase in lightness has been reported previously in freshwater fish species stored under refrigeration 4,56 , and has been associated with changes in the reflectance of the meat surface due to protein denaturation, exposing hydrophobic groups 57 . On the other hand, the increase in the a* and b* values in refrigerated white fish species leads to darkening, which has been related to discoloration 58 . It occurs due to myoglobin autoxidation, where ferrous iron (Fe 2+ ) is oxidized to ferric iron (Fe 3+ ), resulting in the formation and accumulation of metmyoglobin (MetMb) 8 . MDA can also contribute to an increase in MetMb accumulation by inactivating the metmyoglobin-reducing system and/or by interacting with myoglobin molecules through covalent bonds, which alters their primary structure, making myoglobin more susceptible to redox reactions 8,9 . In this study, the increase in the a* and b* values agrees with and can be explained by our results for lipid and protein oxidation, including the differences found among the treatments. Similarly, Monteiro et al. 4 and Park & Ha 58 observed that UV-C radiation increased a* and b* values in tilapia fillets and fresh chicken breast, respectively, over the refrigerated  Table 4. Instrumental color parameters of tilapia (Oreochromis niloticus) fillets non-and treated with oxygen scavenger packaging (OSP) and ultraviolet radiation (UV-C) stored at 4 ± 1 °C for 23 days. Results are expressed as means ± standard deviation (n = 2). a,b,c,d Different superscripts in the same column indicate significant differences (p < 0.05) among treatments. ¥ AUC -Area under curve; AUC0-13 -from day 0 to 13 among treatments AP, OSP, AUV1, OSUV1, AUV3, and OSUV3; AUC15-23 -from day 15 to 23 among treatments OSP, AUV1, OSUV1, AUV3, and OSUV3. NA -Not applicable. € AP (air packaging); OSP (oxygen scavenger packaging); AUV1 (air packaging + UV-C at 0.102 J/cm 2 ); OSUV1 (oxygen scavenger packaging + UV-C at 0.102 J/cm 2 ); AUV3 (air packaging + UV-C at 0.301 J/cm 2 ); and OSUV3 (oxygen scavenger packaging + UV-C at 0.301 J/cm 2 ).
Instrumental texture parameters. Hardness, chewiness, cohesiveness, springiness and resilience decreased (p < 0.05) during the refrigerated period in all treatments (Table 5). OSP, OSUV1 and OSUV3 showed the highest (p < 0.05) hardness and chewiness, followed by samples submitted to air packaging (AP) and UV-C radiation at both doses (AUV1 and AUV3) during the storage period (Table 5). Cohesiveness, springiness and resilience were not affected (p > 0.05) by the O 2 absorber and/or UV-C radiation, regardless of the dose, during the refrigerated storage period. The results of instrumental texture parameters in all days of storage can be found as Supplementary Table S4.
Softening during the post-mortem period is related to the activity of endogenous and microbial proteolytic enzymes, which results in protein breakdown 59 . The results for hardness and chewiness in this study can be explained by the results for free amino acids, MDA level, carbonyl content, and TAPC. The pro-oxidant effect of the UV-C radiation increased the amount of free amino acids, indicating a higher proteolysis rate, while ROS formation at 0.102 and 0.301 J/cm 2 was mitigated by O 2 absorber. Furthermore, when compared to OSP, OSUV1 and OSUV3, both UV-C doses were less effective against growth of aerobic psychrotrophic bacteria, where Pseudomonas spp. is the dominant proteolytic spoilage bacteria in freshwater fish species 34 . There are no studies related to instrumental texture parameters in fish species packed with an O 2 scavenger. However, in agreement  Table 5. Instrumental texture parameters of tilapia (Oreochromis niloticus) fillets non-and treated with oxygen scavenger packaging (OSP) and ultraviolet radiation (UV-C) stored at 4 ± 1 °C for 23 days. Results are expressed as means ± standard deviation (n = 2). a,b,c,d Different superscripts in the same column indicate significant differences (p < 0.05) among treatments. ¥ AUC -Area under curve; AUC 0-13 -from day 0 to 13 among treatments AP, OSP, AUV1, OSUV1, AUV3, and OSUV3; AUC 15-23 -from day 15 to 23 among treatments OSP, AUV1, OSUV1, AUV3, and OSUV3. NA -Not applicable. € AP (air packaging); OSP (oxygen scavenger packaging); AUV1 (air packaging + UV-C at 0.102 J/cm 2 ); OSUV1 (oxygen scavenger packaging + UV-C at 0.102 J/cm 2 ); AUV3 (air packaging + UV-C at 0.301 J/cm 2 ); and OSUV3 (oxygen scavenger packaging + UV-C at 0.301 J/cm 2 ).