Animals and diets.

This study was performed by the laboratory of nutritional physiology of the “Unidad Académica de Agricultura”, of the “Universidad Autónoma de Nayarit” in Mexico. For this study, 24 male neutered Landrace-Yorkshire pigs were used with an initial weight of 55±3 kg, in the fattening stage were used, housing one pig per corral with individual access to food and water, fed ad libitum, according with the recommendations of the official animal welfare ^{NOM-062-ZOO-1999}. Following five days of adaptation to the assigned diet, plus 56 days of experimental diet, the pigs were sacrificed by a method approved by national regulation contained in the ^{NOM-033-SAG/ZOO-2014}, reaching an average live weight of 109±4 kg. Avocado meal obtention was done as described by ^{Lemus et al. (2017)}, using avocados of the Hass variety discarded for human consumption due to their small size or physical damage, collected from packing plants. Avocados were left at room temperature until they reached consumption maturity, whole and fresh fruits (pulp, seed and peel) were ground, producing a paste that was dried at room temperature for four days, ground a second time to obtain a meal, which was stored at room temperature in plastic containers without addition of any preservative or other substances. Pigs were distributed in groups of equal number (n=8) and were provided with one of three different diets, with different percentage of inclusion of avocado meal (AM) mixed in the diet; AM0, AM5 and AM10 corresponding to 0 (control), 5 and 10 % of dry matter (Table 1).

Samples.

At the time of sacrifice, a sample of approximately 100 g of the Longissimus dorsi muscle was collected, packaged under vacuum and frozen at a temperature of -20 °C, until its use in chemical analysis. Aditionally, at slaughter, three samples of approximately 0.5 g were taken from inside the Longissimus dorsi muscle and liver for gene expression analysis. The samples were collected in 2.0 mL cryotubes on ice, with a nucleic acid stabilizing solution (DNA/RNA Shield, Zymo Research, USA) and stored at -20 ºC until use.

Analysis of proximal chemical composition and γ-tocopherol of the Longissimus dorsi muscle.

The chemical composition was determined following the method described by the Association of Official Agricultural Chemists (AOAC, 1997) for moisture (925.10), ash (923.03), proteins (920.87) and lipids (920.39). Quantification of γ-tocopherol was performed with 500 mg of intramuscular fat (IMF) diluted in 1 mL of 2-propanol. The determination was done by High-performance liquid chromatography (HPLC). The separation was done in a Nucleosil C18 reverse phase column (250 mm x 4.6 mm, particle size 5 µm) (^{Can-Cauich et al., 2019}). The system operated by isocratic elution with a flow rate of 1 mL/min, with spikes registered at 285 and 335 nm as wavelength of excitation and emission, respectively. The volume of injection was 20 µl. Identification of retention time and quantification of the peaks was done comparing to a standard of γ-tocopherol. Results were expressed as µg of γ-tocopherol/g of fresh weight sample.

Determination of the total phenolic compounds (TPC) content, antioxidant activity of DPPH and ABTS of the Longissimus dorsi muscle.

5 g of each sample of Longissimus dorsi muscle were used, mixed with 10 mL of distilled water and then placed in a Grant XB3 ultrasound bath (Boekel Scientifi Inc, USA.), for 30 minutes in a cold-water bath. The resulting extract was centrifuged at 3,500 x g for 15 min at 4°C, the supernatant was filtered with Whatman No. 42 paper. The filtered solution was stored at -20°C, until analysis. The TPC content of the samples was analyzed according to the Folin-Ciocalteu colorimetric method (^{Moo-Huchin et al., 2014}). The results were expressed as mg equivalent to gallic acid (GA)/100 g of meat. The antioxidant activity (DPPH and ABTS) of the meat extracts was measured according to the methodology described by Moo-Huchin et al. (2014). For both analyses, a calibration curve was established using Trolox as standard, and the results were expressed as µM equivalent of Trolox/100 g of meat.

Fatty acid analysis of the Longissimus dorsi muscle.

The lipids of the Longissimus dorsi muscle were extracted following the procedure of ^{Hanson & Olley (1963)}, 500 mg of fat was converted into methyl esters of fatty acids with methanol of potassium hydroxide (KOH) and boron trifluoride (BF3) (^{Morrison & Smith, 1964}). For the chromatographic analysis a gas chromatography equipment model Trace GC Ultra from Thermo scientific (USA) was used. The chromatographic system was coupled to a flame ionization detector, set at a temperature of 250 °C and with a gas flow of helium (He) of 35 mL/min, air of 350 mL/min, With an injector temperature of 250 °C. An HP-Innowax capillary column of 60 m by 0.32 mm internal diameter and 0.25 µm particle thickness was used for the separation of fatty acids, using He as carrier gas with a flow rate of 3 mL/min. The separation of fatty acid methyl esters (FAME) was achieved using a temperature ramp starting at 60 °C for 3.5 min, increasing by 10°C/min until it reaches 200 °C and this temperature was maintained for 15 min, then another increase of 1 °C/min until a temperature of 225 °C was reached, and another of 2 °C/min until 240 °C which was maintained for 12.5 min. The fatty acid methyl esters of the samples were identified by comparing their retention times with the corresponding FAME of a 37 component standard mix (Supelco 37 Component FAME Mix 47885-U Sigma-Aldrich). The fatty acids were expressed as the ratio of each individual fatty acid to the total area of the identified fatty acids in the samples.

Analysis of gene expression in the Longissimus dorsi muscle and liver.

From the tissues collected, 75 mg were weighed, and RNA extraction was done using the nucleic acid extraction kit Direct-zol™ RNA MiniPrep (Zymo Research, USA), following the manufacturer’s instructions, the RNA concentration and purity were quantified by spectrophotometry with a Nanodrop. Subsequently, cDNA synthesis was performed, with 1000 ng RNA from each sample, using the Maxima H Minus First Strand cDNA Synthesis Kit with dsDNase (Thermo Scientific, USA), according to the manufacturer’s instructions.

Expression of the ACACA, SCD, SREBP1, ACP and FASN genes, reported by ^{Duran-Montgé et al. (2009a}, ^2009b) and associated with lipid metabolism in pigs (Table 2), were evaluated. The endogenous gene RNA18S was used to normalize gene expression in a real-time PCR (qPCR) using the SYBR Green/ROX qPCR Master Mix (2x) kit (Fermentas, USA), with a final volume of 20 uL per reaction tripled in each sample, using a Step One Plus real-Time PCR equipment (Applied Biosystems, Stockholm, Sweden). The real-time amplification was carried out in 40 cycles, considering the following conditions: initial denaturation (95 °C for 8min), cycling (95 °C for 15s and 60 °C for 30s). The specificity of the amplification of each gene was confirmed by analysis of melting curve, the temperature ramp of this analysis was 60°C to 95°C for 5s. The reading for the dissociation curve gave a single peak for each of the samples, confirming the amplification of each gene. No amplification was detected for the negative samples.

Statistical analysis

An analysis of variance was done with variables of nutritional value, total phenols and tocopherol, antioxidant activity and lipidic profile of the Longissimus dorsi muscle using a single effect model considering the diets as a fixed effect and comparing means with a Tukey test. Additionally, correlations and linear regression of levels of inclusion of AM and measured variables was done. In accordance to the method used by ^{Benítez et al. (2015}, ²⁰¹⁶⁾, the gene expression data were statistically analyzed for both tissues (Longissimus dorsi and liver) following the methodology of ^{Steibel et al. (2009)}, which consists of the analysis of the cycle threshold (Ct) values for target and endogenous genes, simultaneously using a linear mixed model. For this purpose, the following model was used in univariate analysis, of the measurements of gene expression recorded for ACACA, SCD, SREBP1, ACP and FASN. ygijkr=Ti + Pj + Ak +Wr + eijkr. Where ygijkr= -log2 (Eg-^{ct gijkr}), Eg provides the PCR efficiency of each gene, Ct ijkr is the value obtained from the thermal cycler software for the gene of the r^th well, in the j^th qPCR plate, corresponding to the k^th animal subjected to the i^th treatment. Ti is the specific effect of the i^th treatment on gene expression. Pj and Ak are the specific random effects on the expression, in the r^th qPCR well measured by triplicate in each sample of the gene, in the j^th P plate and the kth A pig. eijkr is the residual effect. Three different Ti treatments were considered in the model: dietary effects (three levels: AM0, AM5 and AM10) in finished pigs at 109 ± 4 kg of live weight. To test the differences between the classes in the rate of expression of genes of interest (diffT) data was normalized by the endogenous gene (∆Ct=Ct^h _objetive-Ct_endogenous) and differences between diets and tissues were tested. The values of adjusted statistical probability (p) were calculated using the Bonferroni correction method. To obtain values of change or proportion of fold change (FC) from the diffT estimated values the next equation was applied: FC =2^-diffT. All calculations were done using ^{SPSS, version 20 (2011)}.

Proximal chemical composition, γ-tocopherol, the content of total Phenolic Compounds (TPC), antioxidant activity DPPH and ABTS of the Longissimus dorsi muscle.

Feeding the experimental diets (AM5 and AM10) to the pigs had a significant (p < 0.01) reduction in IMF levels (between 9.7 and 26.6 % reduction) compared to the control diet (AM0) (Table 3). It decreased by -0.21 % for every 1 % inclusion of AM in the diet (Table 4).

Percentages of protein, ash and moisture were similar in all three diets. There was a significant increase (p < 0.01 a p < 0.001) in γ-tocopherol, TPC and antioxidant DPPH and ABTS activity in the Longissimus dorsi muscle including AM in the diets (Table 3); with a lineal effect in these variables (p < 0.01 a p < 0.001), for which the content increases with a higher AM inclusion (Table 4).

In commercial pigs, dorsal fat and IMF have decreased, because a very intense selection pressure has been exerted genetically, negatively affecting the organoleptic properties of the meat (^{Steven et al., 2019}; ^{Dzib-Cauich et al., 2020}). According to ^{Katsumata (2011)}, promoting the increase of IMF in pigs is important for meat quality, however, excessive IMFs may be unhealthy for humans. It is considered that IMF is related to the juiciness of the meat, which is currently low in commercial pigs, for which there are studies aimed at modulating the quantity, but also the quality of fatty acids, through the ingredients of the diets (^{Wood et al., 2008}; ^{Wood & Enser, 2017}), making an emphasis to explain the lipogenic mechanisms (^{Katsumata, 2011}). An example of manipulation by the IMF is in hams from Iberian pigs, which present above 9.5 % of this, but which diminish to 7.08 % when feed with monounsaturated fatty acids (^{Jimenez-Colmenero et al., 2010}). A similar effect is reported by ^{Hernández-López (2016b)} in Landrace-Yorkshire pigs, where the IMF in Longissimus thoracis decreases when fed avocado, with a significant increase in tocopherol; this effect is favored by the nutritional value of avocado (^{Lemus et al., 2017}), and higher antioxidant activity in shell and seed (^{Tesfay et al., 2010}).

Just as tocopherol, phenolic compounds are increased in the meat by feeding pigs with avocado as published by ^{Hernández-López et al. (2016a}, ^2016b), improving meat quality and oxidative stability of fats and protein. ^{Tejerina et al. (2011)}, found that acorns have a high antioxidant capacity and healthy fatty acids because they are composed of high tocopherol, phenolic compounds (mainly tannins) and n-9 monounsaturated fatty acids (MUFA); this description is very similar to AM. Avocado is recognized for its potential effects on human health, due to its nutrient composition and antioxidant activity (^{Méndez-Zúñiga et al., 2019}); these nutrients can be used by pigs since avocados discarded due to low weight do not demerit their nutritional properties (^{Grageola et al., 2010}), and it is a fruit that can modulate the quality of the meat, according to the considerations of ^{Wood et al. (2008)} and ^{Wood & Enser (2017)}. Considering also that the apparent digestibility of the nutrients of this type of food is relatively high when offered to pigs (^{Grageola et al., 2010}; ^{Ly et al., 2015}; ^{Hernández-López et al., 2016a}).

The use of the whole avocado fruit discarded because of its size and transformed into meal is an interesting alternative for animal feeding, since it is not practical to separate the fruit pulp from the seed and the skin (^{Grageola et al., 2010}), and important nutrients can be used, avoiding environmental pollution with the waste. AM contains a high amount of fat and total energy, with important amounts of proteins and alpha tocopherol, for which give this byproduct of avocado production an interesting perspective in animal feeding (^{Lemus et al., 2017}), by improving meat quality through delaying the oxidation of lipids, in addition to increasing the percentage of mono- and polyunsaturated fatty acids. On the other hand, it is important to consider the high content of tannins that can depress consumption, as reported by ^{Fránquez et al. (2017)}, when feeding avocado paste to pigs. Comparing AM with an acorn of Spanish, brings it closer in nutritional quality and as an alternative for feeding pigs, to produce differentiated meat (^{Lemus et al., 2017}; ^{Rey et al., 2006}; ^{Tejerina et al., 2011}).

Fatty acids in the Longissimus dorsi muscle.

The feeding of the experimental diets that included AM, modified the intramuscular composition of the fatty acids present on the Longissimus dorsi muscle (Table 5). Feeding with AM5 showed an important effect in fatty acids compared to the other two diets (p < 0.05 a p < 0.001); there was an increase in palmitic acid (16:0), palmitoleic (16:1), arachidonic (20:4 n6) and eicosapentaenoic (20:3 n3) while there was a diminishment in stearic (18:0) fatty acids with a higher proportion of 18:1/18:0 (9.66).

By feeding pigs with a higher inclusion, AM10, the most noticeable effect on the measurement of saturated fatty acids is the decrease of palmitic (C16:0), which is usually the most abundant in pork (^{Wood & Enser, 2017}). The contents of arachidic (20:0) and linoleic (18:2 n6) fatty acids increased, with a higher proportion of PUFA and PUFA/SFA and PUFA/MUFA ratios (p < 0.01 a p < 0.001). There was no significant effect on MUFA/SFA relationship or in oleic acid (18:1) content by inclusion of AM.

^{Duran-Monge et al. (2008)} when adding about 10 % sunflower oil, flaxseed oil or fish oil/linseed combination to the diet, also found a decrease in Longissimus muscle, palmitoleic(C16:1), stearic (C18:0), oleic (C18:1) and total MUFA. Similar results were found by ^{Hernández-López et al. (2016b)} when the avocado paste was included in the diet, palmitoleic acid (C16:1) and the sum value of saturated fatty acids (SFA) were decreased, with no effect on oleic acid (C18:1). When acorn-fed pigs according to ^{Rey et al. (2006)} and ^{Jimenez-Colmenero et al. (2010)}, maintain values >49 % of oleic acid (C18:1) and >57 % of MUFA. An important effect is that with AM, oleic acid (C18:1) did not decrease by 48 %, which is the case with ^{Duran-Monge et al. (2008)}, who report <38 % with food sources rich in oleic and unsaturated fatty acids.

As in this research when fed AM10, linoleic acid (C18:2) was also increased by providing diets with sunflower or flaxseed oil in the research by ^{Duran-Monge et al. (2008)}; the same happened with diets containing avocado paste (^{Hernández-López et al., 2016b}). The increase in the total proportion of PUFA in AM10, did not reach critical values of 15 %, percentage from which begin to present problems of soft and oily fats or “floppy meat” occurs (^{Wood et al., 2004}). The balance in the diet determines the balance in the tissues, the composition of the fatty acid stored in the adipose tissues, reflects a great extent that of the ingested lipids (^{Wood et al., 2008}).

The PUFA/SFA ratio is beneficial, with 0.45 % in this study when feeding AM10, it is slightly above that recommended by nutritionists, which should be greater than 0.40 %. The high proportion of PUFA is not necessarily healthy if it is not in an adequate proportion with the ratio n6/n3 (^{Fernandez et al, 2007}). Considerations of ^{Jiménez-Colmenero et al. (2010)} indicate that there are some cases of Iberian hams in which they do exceed this value of 0.4 % PUFA/ SFA, in these cases the pigs had been fed with corn oil or supplemented with MUFA and n3. ^{Hernández-López et al. (2016b)} report that pigs fed with avocado paste possessed a PUFA/SFA relation of 0.26 % in the Longissimus dorsi muscle; ^{Rey et al. (2006)} quantified 0.13 % in the Longissimus dorsi muscle feeding with acorn, ^{Fernández et al. (2007)} reported values of 0.19-0.30 in Iberic pigs, and ^{Duran-Monge et al. (2008)} 0.78 in Longissimus muscle feeding with linseed and 0.82 with sunflower oil. ^{Benítez et al. (2015)}, show an increase of PUFA/SFA relation comparing diets with PUFA (0.16) against SFA (0.08) supplemented diets.

Gene expression in Longissimus dorsi muscle and liver.

As observed in Table 6, the FC of the Longissimus dorsi muscle and liver show differences when compared (p < 0.001); situation predicted according to the reports of several researchers, pointing out that the expression is different according to the tissue and feeding (^{Duran-Mongé et al., 2009a}, ^2009b; ^{Benítez et al., 2015}; ^{Benítez et al., 2016}; ^{Fernández et al., 2017}; ^{Wang et al., 2020}).

With the AM5 diet in contrast to AM0, the content of mRNA expression in the Longissimus dorsi muscle was higher in all five genes (p < 0.001). For liver tissue, only the ACACA and SCD genes had differences, with a higher expression for AM5 in contrast to AM0 (p < 0.001).

When fed AM10 is compared with the control diet AM0, similar effects occurred in both tissues; there is no difference in the expression of SREBP1, but it decreases, as well as the rest of the genes where it was significant (p < 0.01).

With the AM5 diet the high expression of lipogenic genes (ACACA, ACP and FASN), as reported in the literature (^{Li et al., 2020}; ^{Fernandez et al., 2017}), can be associated with the increase of arachidonic (20:4 n6) and eicosapentaenoic (20:3 n3) acids, for the effects of biogenesis of long-chain fatty acids (see Table 5). Meanwhile, the esterification of fatty acids (SCD), can be associated with the decrease of stearic acid (18:0) and increase in the 18:1/18:0 relationship. SCD has been associated with the catalytic synthesis of MUFA (palmitoleic and oleic acids) using SFA as substrates (palmitic and stearic acids) (^{Fernández et al., 2017}).

Feeding with AM5 diet in contrast to AM0, in Longissimus dorsi was expressed more the biogenesis of long-chain fatty acids and esterification of fatty acids. The higher expression of SREBP1 in the Longissimus dorsi muscle and not in the liver for this AM5 diet indicates that there was an association between this gene and the rest of those evaluated in muscle. We do not agree with ^{Duran-Mongé et al. (2009b)} who found differences in tissue expression, but expressing more ACACA and SREBP1 in the liver, with more SCD and FASN in adipose tissue.

The situation was the opposite when feeding AM10 since this diet decreased the expression of the genes studied, a situation similar to that pointed out by ^{Mohan et al. (2012)}, who reported that the expression of the important lipogenic mediator SREBP1, can be affected by the type of diet, as it happened in the AM10 diet, where it is not overexpressed.

The diet with sunflower oil according to ^{Mohan et al. (2012)} and ^{Benitez et al. (2018)} also decreases the expression of FASN, SCD and SREBP1; the reduction of SREBP1 can be due to the increase of n-6 PUFA, which in turn reduces the expression of FASN and SCD. In researches such as those by ^{Duran-Montgé et al. (2009a}, ^2009b) it was observed that the expression of FASN and SCD genes were considerably reduced in liver and muscle of pigs fed with PUFA, as it happened in this work when AM was increased to 10 % (AM10). The lower activity of these lipogenic enzymes in muscle has already been demonstrated in pigs and can be related to the low amounts of fat in the muscle (^{Guillevic et al., 2009}).

In this research, IMF was decreased by including more AM level in the diet (see Tables 3 and 4), but nutritionally it is important, that there was an increase in the proportions of PUFA and arachidic (20:0) and linoleic (18:2 n6) long-chain fatty acids to higher AM conditions; considering the publications by ^{Wood et al. (2008)} and ^{Wood & Enser (2017)}, ) the incorporation of long chain fatty acids in the feed is probable through the AM proportionated. On the other hand, according to ^{Li et al. (2020)}, there are other genes such as the peroxisome proliferation-activated receptors (PPARs) that regulate the expression of genes involved in lipidic metabolism, which could influence the results and were not incorporated for analysis in this research.

In relation to diets without incorporating fat, ^{Duran-Mongé et al. (2009a)} reported that pigs had a higher expression of genes involved in stearic synthesis (ACACA and FASN), and also of the gene involved in desaturation (SCD); the same happened when comparing the AM0 diet against AM10 (with lower expression), nevertheless, the synthesis of long-chain fatty acids is not increased in the AM0 diet. According to the results of ^{Wang et al. (2020)}, the increase of FASN in AM0 diet, could influence in greater IMF deposition, since it is a predictor of the IMF content in the Longissimus dorsi muscle in Laiwu pigs.

^{Muñoz et al. (2007)} point out that in Iberian pigs ACACA is a good predictor, its overexpression is related to the increase and content of monounsaturated fatty acids (palmitoleic, vaccenic and oleic), reducing the percentage of stearic acid; when fed AM5 this effect happens when the stearic acid decreases and the palmitoleic acid increases, but it does not happen in AM10, since it had a low expression of this gene.

Considering the obtained results and the consulted literature, it can be considered that feeding pigs with AM, influenced the IMF, the presence of antioxidants and the modulation of lipids in Longissimus dorsi, with different gene expression in tissues such as the Longissimus dorsi and liver.