Increased Inclusion of Dietary Methionine Sources Improves Pork Quality in Association with the Regulation of Energy and Protein Metabolism and Muscle’s Fiber Prole

Background Intensive selection for faster growth rate and higher lean percentage led to increase in protein deposition but deterioration in meat quality of pigs, thus there is growing interest in exploring the nutritional strategies to improve meat quality. Methionine has been shown to activate mechanistic target of rapamycin complex 1 protein kinase that plays pivotal roles in the regulation of protein and lipid synthesis. However, few study reports are available regarding the effects of dietary methionine supplementation at levels beyond growth requirements on lipid and protein metabolism and thus on pork quality. The objective of this study was to assess whether pork quality was improved by increasing dietary digestible sulfur amino acids (SAA) levels, with pigs fed the control (100% SAA), DL-Methionine (125% SAA)- or OH-Methionine (125% SAA)-supplemented diets during 11–110 kg period. Increasing SAA above requirements did not signicantly affect growth performance, whereas improved pork quality as indicated by the decreased drip loss and a tendency towards decrease in shear force of longissimus lumborum muscle. Moreover, fresh muscle from barrows fed OH-Methionine showed a higher lightness value compared with the control and DL-Methionine treatments. The relatively lower shear force might be explained by the decrease in crude protein and increase in glycolytic potential, while the decreased drip loss was associated with down-regulation of genes (like fast glycolytic IIx) regulating ber types. The increased lightness value of fresh muscle from barrows fed OH-Met diets appeared to be associated with the increased lactate level, which can be further explained by the increased plasma short-chain fatty acids concentrations, up-regulated G-protein coupled receptor 43 activation and enhanced glucagon-like peptide 1 secretion. Increased SAA consumption appeared to improve pork water-holding capacity and tenderness likely through regulation of energy and protein metabolism and muscle’s ber prole, which provides new insights into the nutritional strategies to improve meat quality. OH-Met-fed glycolytic potential in level glycolytic potential drip loss and lead to pale, soft and exudative (PSE) pork. drip loss, comparison control, the occurrence of PSE pork in the DL-Met and OH-Met treatment. increased glycolytic potential benecial for the development of tenderness dietary DL-Met or OH-Met supplementation growth meat tenderness through changing protein and energy metabolism. with the mRNA level of GLUT-4 and AMPKα2 in LM signicantly increased in the DL-Met or OH-Met groups. AMPK can activate phosphorylase kinase, which then activates glycogen the primary transporter of glucose uptake is up-regulated in AMPK found the up-regulated expression of FATP-1 whereas the down-regulated expression of HSL in the DL-Met and OH-Met compared with the CON treatment, further suggesting the activation of AMPK pathway [26] . In contrast, it has been reported that pigs (initial weight 105 kg) fed diets supplemented with Met at 5 times the level of the control diet, for the last 14 days before slaughter, showed a tendency of reduced lipid content and glycolytic potential in the LM [27] . The long-term (20 weeks) supplementation and the lower increase (25%) of Met applied in the current study might have led to different mechanisms of response. In support of this, following increased SAA consumption, plasma taurine levels were elevated from 11– kg period whereas it remained unchanged from 70–110 kg period, indicating the different responses of pigs varying in body weight to supplemental Met of rapamycin complex 1; MyHC, Myosin Heavy Chain; OH-Met, DL-2-hydroxy-4 (methylthio) butanoic acid; SAA, sulfur amino acids; SAM, S-adenosylmethionine; SCFA, short-chain fatty acids; SID-SAA, standard-ileal-digestible sulfur amino acids; T-GSH, total glutathione.


Introduction
Intensive selection for greater growth rate and lean percentage led to increase in protein deposition rates but deterioration in meat quality of pigs. Methionine (Met) is considered the second or third limiting amino acids (AA) for the corn-soybean-based swine diets (NRC, 2012) [1] . Met is not only a precursor of protein synthesis, but also involved in the methylation cycle, providing essential metabolites or precursors. For example, S-adenosylmethionine (SAM), which acts as mediators or signal molecules, plays important roles in numerous biological processes. A study has shown that the mechanistic target of rapamycin complex 1 (mTORC1) protein kinase can be activated by SAM through SAMTOR sensor [2] . The mTORC1 plays pivotal roles in the regulation of anabolism including protein and DNA synthesis. Moreover, there is evidence that SAM stimulates the expression of the genes linked to fatty acid metabolism by suppressing Wnt/ β-catenin and Hedgehog signaling pathways in porcine muscle satellite cells [3] . Simultaneously, a previous in vitro study has shown that a severe de ciency in Met and cysteine (Cys) decreases both the proliferation and differentiation, whereas a mild de ciency only alters the differentiation in porcine preadipocytes [4] .
These results suggested the potential regulation of dietary Met on protein and lipid metabolism. However, few study reports are available regarding the effects of dietary Met supplementation at levels beyond growth requirements on lipid and protein metabolism and thus on the meat quality in growing-nishing pigs.
DL-Met and its hydroxy analogue, DL-2-hydroxy-4 (methylthio) butanoic acid (OH-Met), have long been used as supplemental Met sources in commercial livestock production. Interestingly, our previous studies suggested that OH-Met might be considered as more than a Met precursor, given that pigs fed the OH-Met had a higher acetate concentration in plasma [5] than DL-Met-fed pigs. Acetate has been shown to play an important role in insulin signaling and lipid metabolism [6] . In the present study, we hypothesized that methionine levels and sources potentially modi ed meat quality through regulation of protein and lipid metabolism. To test this hypothesis, both genders of pigs (castrated males vs. females) throughout the growing-nishing period were fed diets supplemented with DL-Met or OH-Met to contain digestible sulfur amino acids (SAA) levels above the growth requirement, and biochemical indices for blood and muscle samples and expression of genes in relation with lipid metabolism and muscle's bers pro le for muscle samples were measured.

Study design, diets and animals
The study was conducted in a randomized complete block design. A total of 144 (half gilts) crossbred (Duroc × Large White × Landrace) barrows and gilts (11.16 ± 0.23 kg, ~ 5 weeks of age) selected from a single weanling batch were blocked by gender and bodyweight (BW) and allocated to 36 pens (4 piglets/pen) located in the same room.
Piglets in each block were randomly fed one of the three experimental diets (Additional le 1-1). The basal diet, fed to the control (CON) group of pigs, was formulated to meet swine nutrition requirements of NRC (2012) [1] . The DL-Met and OH-Met diets were formulated to contain 125% of standard-ileal-digestible (SID) SAA, 25% of which, namely, the portion of methionine equivalence beyond the CON level, was supplied by DL-Met (99%) and OH-Met (88%, Rhodimet® AT88 from Adisseo Life Science, Shanghai, China), respectively. The experimental period consisted of ve phases with pigs fed the experimental diets from SID-SAA as recommended by NRC (2012) [1] . Each diet was replicated 12 times with 6 pens of gilts and barrows, respectively. All pigs had free access to pelleted feed and water throughout the experimental period.

Analysis of dietary composition
The dry matter content of the experimental diets was determined by drying at 105 °C for 4 h. This procedure was repeated again until diets were dried to constant weight. Crude protein, ether extract, ash and crude ber in diets were analyzed according to procedures described by AOAC (2000) [7] . Amino acids in hydrolyzed feed were analyzed by ion-exchange chromatography using an L8900 high-speed AA analyser (Hitachi, Tokyo, Japan). The Cys and Met contents were determined as cysteic acid and Met sulphone, respectively, after performic acid oxidation before hydrolysis. OH-Met in diet was determined by using the HPLC method as previously described [8] . The analyzed nutrient levels in experimental diets are shown in Additional le 1-2. Blood and tissues sampling At the end of each phase, one pig per pen was used for blood sample collection. After overnight fasting (12 h), blood samples were collected via jugular venipuncture by using disposable sterilized syringe and then injected into disposable vacuum tube. Samples were centrifuged at 2550 × g for 15 min at 4℃, plasma and serum were collected and stored at -20 °C until analyzed.
Samples of longissimus lumborum muscle (LM) were taken within 20 min postmortem from the left side carcass at the last rib, cut into small pieces and snap frozen in liquid nitrogen, and then stored at − 80 °C for analysis of genes expression, glycogen content and glycolytic potential. In addition, the LM and liver samples (~ 100 g) were taken and frozen at − 20 °C for chemical analysis.
Free AA in plasma samples were analyzed as described [10] . Brie y, 350 µl of the sample and 700 µl of 10% (w/v) sulfonyl salicylic acid solution were mixed thoroughly and centrifuged at 12 000 g for 15 min. Then, the supernatant was collected and analyzed for AA by ion-exchange chromatography using an L8800 high-speed AA analyser (Hitachi, Tokyo, Japan).

Measurement of carcass traits and meat quality
At the end of the experiment, one pig with average BW (~ 110 kg) was selected from each pen and slaughtered for measurement of carcass and meat quality.
After 24 h fasting, pigs were anesthetized by electrical stunning and killed by exsanguination. Live weight at slaughter, carcass weight, dressing percentage, average backfat depth, LM area and carcass length were measured immediately postmortem according to the Chinese guidelines on technical regulation for testing of carcass traits in lean-type pig (NY/T 825-2004, 2004) [11] . Backfat depth is the average of measurements at three points: the rst rib, the last rib, and the last lumbar vertebra. The percentage of lean meat was calculated as previously described [12] .
The LM samples were used to measure marbling score, drip loss, pH value, shear force, muscle color and cook loss. The measurements were performed in the following order: 1) 4.0-cm-thick chop used for drip loss measurement between the 3rd and 4th last rib; 2) Assessment of pH value, marbling score and muscle color at the last rib; 3) 6.0-cm-thick chop used for shear force measurement between the rst lumbar and the last lumbar.
Initial (pH 45 min) and nal pH (pH 24 h) were measured in triplicate at 45 min and 24 h after slaughter, respectively, by using a hand-held pH meter (pH-STAR, SFK-Technology, Denmark). Meat color was measured using a CR-400 Chroma Meter (MinoltaLtd., Milton Keynes, Japan). Meat color values (L*=lightness, a*=redness, b*=yellowness) determined at 45 min and 24 h postmortem. Marbling score was determined according to National Pork Producers Council guidelines (NPPC, 1999) [13] . Drip loss, cooking loss and shear force were determined as previously described [14] . Drip loss and cooking loss were expressed as the weight change percentage. Warner-Bratzler shear force of cooked meat was determined using a Texture Analyzer (TA.XT. Plus, Stable Micro Systems, Godalming, UK).

Muscle and liver chemical composition
To determine the moisture content in LM and liver tissues, samples were freeze-dried for 72 h and then dried to a constant weight at 105 °C. Crude protein and ether extract were analyzed according to procedures as described AOAC (2000) [7] . Intramuscular fat (IMF) in LM was determined as ether extract and calculated by weight loss.
About one gram of frozen LM samples were homogenized for measurement of glycogen content and glycolytic potential. The glycolytic potential was calculated as the following equation: glycolytic potential = 2 ⋅ (glycogen content + glucose content + glucose-6-phosphate content) + lactate content. Glycogen and lactate content were measured by using commercial kits (Nanjing Jiancheng Biochemical Institute, Nanjing, China). Glucose content in muscle was determined by using enzymatic methods (GAHK-20, Sigma-Aldrich Inc., St. Louis, MO). The glycolytic potential was expressed as micromoles of equivalent lactate per gram of wet tissue. According to the procedures described in detail [14] , the LM samples for CSA determination were prepared by classic hematoxylin and eosin (HE) staining, and the CSA of myo ber was measured by using image processing software (Image-Pro Plus 6.0, Silver Spring, MD, USA).

Antioxidant capacity
The content of total glutathione (T-GSH; number S0053), glutathione disul de (GSSG, number S0053), glutathione peroxidase (GPX, number S0056) and malondialdehyde (MDA; number S0131) in muscle were analyzed by using commercial kits (Beyotime Biotechnology, Shanghai, China). The concentrations of protein carbonyl (PC; number BC1275) were measured by using commercial kits (Beijing Solarbio Science & Technology Co., Ltd). Frozen LM samples (~ 150 mg) were quickly removed and homogenized (wt/vol) with ice-cold extracting solution using homogenizer and then centrifuged at 4 °C. The supernatant was then collected and immediately used for further analysis. All operations were carried out according to the instructions provided with the kits.
RNA extraction and real-time PCR LM samples were used to determine the expression of genes in association with energy and lipids metabolism, the IGF-1 pathway and the muscle ber type.
These genes include the fatty acid transport protein − 1 (FATP-1), the fatty acid synthase (FAS), the hormone-sensitive lipase (HSL), the AMP-activated protein kinase α2 (AMPKα2), the glucose transporter-4 (GLUT-4), the G protein-coupled receptor (GPR) 43, IGF-1, IGF-1 receptors (IGF-1R) and the Myosin Heavy Chain (MyHC) , a, x, b. The RNA extraction and real-time PCR were performed as we previously described [10] . Beta-actin was ampli ed for each sample to verify the presence of cDNA and as an internal control to calculate the relative level of target gene expression using the 2 −ΔΔCT method [15] . The primer sequences are shown in Additional le 2.

Statistical analysis
Data were analyzed by using the mixed model procedure of SAS statistical package (V8.1, SAS Institute Inc., Cary, NC, USA), according to the following equation where Y is the response variable, µ is the intercept, α i is the effect of treatment (i = 1, 2, 3), β j is the effect of gender (j = 1, 2), (αβ) ij refers to the interaction between treatment and gender, and ε ij represents the residual error. The pen was used as the experimental unit. When an effect was signi cant (P < 0.05), the Tukey test was used to determine speci c differences between means.

Growth Performance
The BW at the beginning and at the end of each experimental phase was signi cantly (P < 0.05) lower in gilts than in barrows ( Table 1). The average daily gain (ADG) and average daily feed intake (ADFI) were also lower (P < 0.05) in gilts than in barrows at each of the experimental phases from 25 kg to 110 kg. The ratio of gain to feed (G:F) was higher (P < 0.05) in gilts than in barrows during the 45-70 kg, 70-95 kg, and 11-110 kg periods. The increase of Met above requirements did not signi cantly (P > 0.05) affect the ADG, ADFI and G:F throughout the experimental period. Plasma IGF-1, GLP-1 and short-chain fatty acids concentrations Plasma IGF-1 and GLP-1 concentrations both were signi cantly (P < 0.05 and P < 0.01, respectively) affected by dietary treatment. Speci cally, plasma IGF-1 concentrations were lower (P < 0.05) in the OH-Met than in the CON treatment, whereas plasma GLP-1 concentrations were higher (P < 0.05) in the OH-Met than in the CON and DL-Met treatments ( Table 2). The OH-Met treatment showed higher (P < 0.05) serum propionate concentrations and a tendency (P = 0.086) towards increase in serum acetate concentrations compared with the CON and DL-Met treatments ( Table 2).

Plasma free amino acids concentrations
Increasing dietary SAA contributed to variance of plasma amino acids pro le in different experimental periods (Table 3, Additional le 3). Notably, plasma taurine concentration showed an increase (P < 0.05 or P < 0.10) following increased SAA consumption as compared with the CON at 11-70 kg period, whereas showed no difference (P > 0.10) among treatments at 70-110 kg period (Table 3). In contrast, plasma phenylalanine concentration was higher (P < 0.05) in CON than in increased SAA treatment at 70-95 kg period, whereas was lower (P < 0.05) in CON than in increased SAA treatment at 11-70 kg period (Table 3).

Carcass characteristics and meat quality measured at slaughter
Dietary treatment showed no effect on pH values, color parameters or marbling score, whereas signi cantly (P < 0.05) affected drip loss and tended (P < 0.10) to affect shear force (Table 4). Compared with the CON treatment, drip loss was decreased (P < 0.05) while shear force tended (P < 0.10) to decrease in both the DL-Met and OH-Met treatments. The pH value at 45 min, the color parameter L 45min and drip loss were signi cantly (P < 0.05) affected by the interaction of sex and diet. Interestingly, pH 45min value was higher in the OH-Met-fed gilts than in the DL-Met-and CON-fed gilts, while L 45min value was higher in the OH-Metfed barrows than in the DL-Met-and CON-fed barrows. Drip loss was higher (P < 0.05) in gilts than in barrows in the CON treatment, whereas it showed no difference (P > 0.10) between sex in the DL-Met and OH-Met treatments. Overall, the DL-Met-and OH-Met-fed gilts had lower (P < 0.05) drip loss than the CONfed gilts. In addition, compared with gilts, barrows had higher (P < 0.05) backfat depth and marbling score values, whereas had lower (P < 0.05) lean meat percentage, LM area and shear force. Protein and fat content in muscle and liver The sex of pigs showed a signi cant (P < 0.05) effect on crude protein and IMF content in LM muscles. Overall, the gilts had a higher (P < 0.05) crude protein and a lower (P < 0.05) IMF content than barrows. In addition, crude protein content in fresh liver tended (P < 0.10) to be lower in the DL-Met and OH-Met treatments than in the CON treatment (Table 5). 1 T-GSH = total glutathione; 2 GSH = reduced glutathione; 3 GSSG = oxidized glutathione; 4 GPX = glutathione peroxidase; 5 MDA=malondialdehyd

Glycolytic potential in muscle
Dietary treatment showed a signi cant (P < 0.05) effect on lactate, free glucose and glucose-6-P, and glycolytic potential of LM muscle. Speci cally, compared with the CON treatment, the OH-Met treatment had higher (P < 0.05) lactate content, whereas the DL-Met treatment had increased (P < 0.05) free glucose and glucose-6-P content ( Table 5). As a result, both OH-Met and DL-Met treatments showed higher (P < 0.05) glycolytic potential than the CON treatment.

Oxidation indices in muscle
The oxidation indices, including T-GSH, GSSG, GSH, GSSH/GSSG, GPX, MDA and protein carbonyl in muscle, were not affected (P > 0.10) by diet, sex or diet × sex interaction ( Table 5).
The CSA of myo bers and expression of genes in muscle effect on AMPKα2 expression and tended (P < 0.10) to affect FAS expression in LM muscles. Compared with barrows, gilts had signi cantly (P < 0.05) lower AMPKα2 mRNA abundance while tended (P < 0.10) to have lower FAS mRNA abundance. Dietary treatment had a signi cant (P < 0.05) effect on MyHC x, GPR43, IGF-1, AMPKα2, GLUT-4 and FATP-1 expression, while tended to affect (P < 0.10) HSL expression in LM muscles. Compared with the CON treatment, the DL-Met and OH-Met treatments had decreased (P < 0.05) mRNA abundance of MyHC x and IGF-1 while increased (P < 0.05) mRNA abundance of FATP-1, GLUT-4 and AMPKα2. The OH-Met and DL-Met treatments tended (P < 0.10) to decrease HSL mRNA abundance compared with the CON treatment. The GPR43 mRNA abundance was higher (P < 0.05) in the OH-Met than in the CON treatment. In contrast, MyHC , MyHC a, MyHC b and IGF-1R expression were not affected (P > 0.10) by sex, diet or their interactions.  In the present study, barrows showed higher body weight from the beginning until the end of the experiment, which might in part explain the lower growth rate and feed intake in gilts than in barrows during the growing-nishing period. Meanwhile, feed e ciency as indicated by G:F was higher in gilts than in barrows during the 45-70 kg, 70-95 kg, and 11-110 kg periods. These results were in agreement with the previous reports that barrows had higher ADFI and ADG, but lower feed e ciency than gilts [16] . The higher feed effeciency of gilts could be explained by the greater protein deposition. In support of this, carcass analysis indicated that barrows had lower lean meat percentage and produced fatter carcass compared with gilts. Consistently, it was reported that improvement of feed e ciency in pigs could be achieved by increasing lean growth rate, which resulted in lower feed intake [17] . Increasing dietary SAA showed no effect on carcass traits, which agreed well with previous studies [18] .

Effects of sex and increased dietary SAA on meat quality
In the present study, barrows showed lower shear force (a mechanical indicator of tenderness) than gilts, which agreed well with previous studies [19] .
Meanwhile, it was also reported that tenderness scores were positively correlative with IMF content (P = 0.008) [19] . Consistent with these results, barrows also showed higher IMF in LM than gilts. Besides, it has been suggested that an increase in IMF leading to a decrease in the shear force could potentially be due to a decrease in the density of muscle which is positively correlated with protein content [20] . Consistent with this, barrows showed a lower crude protein content in LM compared with gilts, which also suggested the potential changes in the myo brillar and cytoskeletal proteins correlating with shear force [21] . Mechanistically, IGF-1 has been shown to stimulate muscle protein synthesis and inhibit protein degradation via the ubiquitin-proteasome and autophagylysosome pathways [22] . Thus, the lower plasma IGF-1 concentration in barrows than in gilts provided physiological explanation for the reduced protein deposition and shear force.
The increase in dietary SAA as DL-Met or OH-Met tended to (P = 0.06) decrease the shear force of LM, which might also be explained by a tendency of a lower crude protein content in the LM of pigs fed increased SAA diets. As addressed above, the numerically lower plasma IGF-1 concentrations and, moreover, the down-regulation of IGF-1 expression in LM might in part account for the lower protein deposition in DL-Met-and OH-Met-fed pigs than in CON-fed pigs.
Corresponding to the relatively lower crude protein deposition in LM, there was a reduction in crude protein of liver in pigs fed increased SAA diets, which might be associated with the limitation of circulating essential amino acids levels. In support of this, pigs fed DL-Met and OH-Met showed lower plasma histidine levels at 95-110 kg period compared with the CON (Additional le 3-5), but the mechanism for increased SAA-induced change of amino acids pro le remains to be elucidated.
Remarkably, in comparison to the CON-fed pigs, the DL-Met-and OH-Met-fed pigs had increased glycolytic potential in LM. Generally, the higher level of glycolytic potential would increase drip loss and lead to pale, soft and exudative (PSE) pork. However, the decreased drip loss, in comparison to the control, excluded the occurrence of PSE pork in the DL-Met and OH-Met treatment. A moderately increased level of glycolytic potential was even considered to be bene cial for the development of tenderness [23] . Based on these results, we speculated that dietary DL-Met or OH-Met supplementation above growth requirements improved meat tenderness through changing protein and energy metabolism. Consistent with this assumption, the mRNA level of GLUT-4 and AMPKα2 in LM signi cantly increased in the DL-Met or OH-Met groups. AMPK can activate phosphorylase kinase, which then activates glycogen phosphorylase and promotes glycogenolysis [24] . GLUT-4, as the primary transporter of glucose uptake in the muscle, is up-regulated in the muscle upon activation of AMPK [25] . We also found the up-regulated expression of FATP-1 whereas the down-regulated expression of HSL in the DL-Met and OH-Met treatments compared with the CON treatment, further suggesting the activation of AMPK pathway [26] . In contrast, it has been reported that pigs (initial weight 105 kg) fed diets supplemented with Met at 5 times the level of the control diet, for the last 14 days before slaughter, showed a tendency of reduced lipid content and glycolytic potential in the LM [27] . The long-term (20 weeks) supplementation and the lower increase (25%) of Met applied in the current study might have led to different mechanisms of response. In support of this, following increased SAA consumption, plasma taurine levels were elevated from 11-70 kg period whereas it remained unchanged from 70-110 kg period, indicating the different responses of pigs varying in body weight to supplemental Met sources.
A decreased drip loss was observed in DL-Met-and OH-Met-fed pigs compared to the CON-fed pigs. As indicated by previous studies [27] , Met supplementation in pig diets improved the water-holding capacity (WHC) of LM, which was associated with the increased antioxidant capacity as manifested by the lower level of MDA in LM. However, the content of MDA in LM showed no difference among treatments in the present study. This might be associated with that Met supplementation had no signi cant effect on glutathione synthesis, which agreed well with previous studies [28] . It was even observed that reduced liver glutathione concentration following dietary supplementation with excessive cysteine [29] . These results suggested that the antioxidant capacity in response to SAA supplementation varied with the supplemental levels. Physically, drip loss originates from the spaces between muscle ber bundles and the perimysial network, and the spaces between muscle bers and the endomysial network [30] . Myo ber is constituted by four different ber types including slow-oxidative or type I, fast oxido-glycolytic or type IIa, and fast glycolytic IIx or IIb. Glycolytic muscles had larger extra-myo brillar uid spaces and higher CSA than oxidative muscles [31][32] . Moreover, drip loss was negatively related to ber area percentages of type I and IIa, while positively related to type IIx and IIb percentage [33] . Interestingly, we found that the increase in dietary SAA decreased the CSA of myo bers and down-regulated mRNA levels of MyHC IIx. Mechanistically, AMPK can directly regulate the transcriptional activity of PPARδ in skeletal muscles and thus increase the proportion of type I bers (by 38%) [34] . Consistently, AMPKα2 in LM was up-regulated in the DL-Met and OH-Met groups in comparison to the CON group. Thus, it would appear that the decreased drip loss following increased SAA consumption was associated with the change of ber types. In addition, we also found that drip loss was signi cantly affected by the interaction of diet and sex. Speci cally, for the CON-fed pigs, drip loss of LM was higher in gilts than in barrows. As mentioned above, both IMF and marbling score were higher in barrows than in gilts. It was reported that marbling score was negatively correlated with drip loss (r = -0.459; P = 0.001) [35] . In the present study, the drip loss tended (P = 0.08) to be negatively correlated (r = -0.52) with the IMF content. Thus, the higher IMF and marbling score in LM might account for the lower drip loss in the CON-fed barrows than in the CON-fed gilts.
Interestingly, the lightness of fresh muscle as indicated by L 45min value was higher in the OH-Met-fed barrows than in the DL-Met-and CON-fed barrows. It has been reported that muscle lightness and plasma lactate and glucose re ect increased muscle glycogenolysis which, in turn, can explain the change of muscle lightness and plasma lactate and glucose levels [36] . Consistent with the higher lightness value, the OH-Met treatment had higher degree of muscle glycogenolysis as suggested by the higher glycolytic potential. Moreover, the increase in the glycolytic potential in the OH-Met treatment was mainly associated with the increased content of lactate, the end-product of glycolysis. In contrast, the increase in the glycolytic potential in the DL-Met treatment was mainly ascribed to the increase in free glucose and glucose-6-P content, the initial product of glycolysis. These results indicate that a higher degree of complete glycolysis process did occur in muscle of the OH-Met treatment rather than in that of the DL-Met treatment. The highest lightness value, L 45min, observed in the barrows of the OH-Met treatment was accompanied with the highest muscle lactate level, further suggesting the contribution of lactate to lightness value. Pigs fed OH-Met had higher serum acetate and propionate, both of which can stimulate GLP-1 secretion via activating the GPR43 receptors in intestinal L-cells [37] . Consistently, GPR43 expression was upregulated in LM and plasma GLP-1 concentration was increased in the OH-Met treatment. GLP-1 can increase glycogen synthase activity and stimulate both glucose oxidation and lactate formation [38] . Taken together, SCFA-induced GPR43/GLP1 signaling might account for the higher level of lactate in the OH-Met treatment.

Conclusions
In summary, the increase in dietary SAA levels by 25% over NRC recommendation showed a positive effect on pork quality. Speci cally, the increased SAA consumption decreased drip loss and tended to decrease shear force. The trend of lower shear force of LM was associated with the decrease in crude protein and increase in glycolytic potential. Changes in drip loss was associated with the types of muscle ber. The increased lightness value of fresh muscle from barrows fed OH-Met diets appeared to be associated with the increased lactate level, which can be further explained by the increased plasma SCFA

Consent for publication
Not applicable Availability of data and materials The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.