Effects of concentrate supplementation on fatty acid composition and expression of lipogenic genes of meat and adipose tissues in grazing lambs

Abstract Effects of natural grazing (NG) or grazing with supplementary feeding (GS) of Hulunbeier lambs (HL) and Hulunbeier × Dorper crossbred lambs (HZ) on fatty acid (FA) profile and lipogenic gene expressions in the longissimus thoracis (LT) and subcutaneous adipose tissue (SAT) were determined. The study was conducted as a 2 × 2 factorial arrangement using four-month old lambs. Thirty animals were divided into each group. The FA composition and the expression of lipid metabolism gene of sheep were affected by the feeding regimens and animal breeds. Compared with NG, GS increases de novo FA synthesis of LT and SAT, which decreases GS lambs’ FA nutrition value. Meat or fat from NG lambs present more n-3 long-chain polyunsaturated FA that is beneficial to human health. Under the two feeding regimens, de novo synthesis of FA and fat deposition seems to higher in HZ lambs compared with those in HL as the expression of genes (SREBP1c, FAS, ACC, C/EBPα, PPARγ) that are associated with correlative metabolism was increased in HZ lambs. Meat from HZ lambs displays a less favourable level of saturated FA as containing higher content of C14:0, but it also presents higher profile of EPA than HL lambs. Highlights Feeding ways and breeds affect fatty acids and gene. Meat from grazing lambs is beneficial to human health. Supplementary feeding increases fat deposition.


Introduction
Hulunbeier lambs (HL), the local meat breed, which are mainly distributed in the Hulunbeier prairie of China and are given priority to traditional grazing for a long time, are known for cold-resistance, crude-feeding tolerance, fast-growth and delicious meat. Dorper lamb is a foreign meat sheep well known for their growth potential and good carcase quality (Kariuki et al. 2010). The hybrid generation of Hulunbeier ewes and Dorper ram (known as HZ) has high adaptability, with the meat production efficiency apparently higher than that of HL. However, with the continuous increase in grassland grazing capacity, the degradation and desertification of grassland is of concern (Dong et al. 2010). Consequently, grazing with supplementary feeding (GS) has become a measure to alleviate grassland carrying pressure and to improve lamb fattening effects.
Fat and fatty acids (FA) in muscle and adipose tissue are the major factors significantly affecting meat quality, particularly its nutritional value and palatability (Wood et al. 2008). Attention has been given to lipids that contain different types of FA associated with health problems. The feeding strategy can modify the FA profile in meat through altering the gene expression of enzymes related with fat metabolism (Dervishi et al. 2011). On the other hand, breed is a determinant factor affecting the FA profile of lamb's fat (De Smet et al. 2004). Our previous study found that mutton production efficiency of GS group is higher than that of natural grazing (NG) group, and mutton production efficiency of HZ is also significantly higher than that of HL (Wen et al. 2017). However, we do not know if the FA composition in muscle and fat of pasture-fed lambs will be affected by supplementing concentrate. Therefore, the aim of this study was to determine that the effects of NG or GS on FA profiles and lipid metabolism genes of the longissimus thoracis (LT) muscle and subcutaneous adipose tissues (SAT) from HL and HZ.

Experimental design, animals and diets
All procedures involving animals were evaluated and approved by the guidelines of the Animal Care and Use Committee of the Inner Mongolia Agriculture University (Hohhot, Inner Mongolia, China). This study was conducted in Inner Mongolia Hulunbeier Sheep Breeding Farm, Hulunbeier City, Inner Mongolia Autonomous Region, China (49 12 N,119 39 E). Four-month-old male Hulunbeier (HL) (29.02 ± 0.80 kg) or Hulunbeier Â Dorper crossbred (HZ) (37.84 ± 0.80 kg) lambs subjected to either natural grazing (NG) or grazing with supplementary feeding (GS) were studied in a 2 Â 2 factorial arrangement of treatments for 60 days. Thirty animals were divided into each group. A total of 40 hectares of grassland in Inner Mongolia Hulunbeier Sheep Breeding Farm were used for lambs grazing, 20 hectares for NG lambs and another 20 hectares for GS lambs. The grazing method was continuous stocking and rest period was included from March to June. No fertiliser had been used on this site. The botanical composition of the two pastures was herbage (mainly Stipa capillata and Cleistogenes Keng), compositae (mainly Artemisia aurata) and shrub (Caragana, Ammopiptanthus mongolicus, Ceratoides lanata, Salsola collina). In our experimental period, the pasture grass was in the heading stage and maturity stage. Lambs in all the groups were grazed on the natural rangeland from 7 a.m. to 6 p.m. each day. In addition to grazing, lambs in the GS groups received a measured quantity of concentrate mixture 0.26 kg/lamb (air dry basis) per day during and early period (1-30 d) and 0.53 kg/lamb (air dry basis) per day during a late period (31-60 d). Concentrate mixture was fed daily at night and was completely consumed. Fresh water was freely available at all times.

Sampling and slaughtering procedures
Six lambs from each group were identified for faecal collection. The same animals were used in all the next period to avoid individual variations. They were harnessed with faecal bags at the first five consecutive days used for collection of diet and faeces samples, which following seven days allotting to acclimatisation of animals for harness in each period. Faecal bags were emptied in the morning (07:00 h) and evening (18:00 h) to estimate daily faecal output. Representative samples (10%) of faeces were pooled separately for 5 days collection period and stored at -20 C to use for future chemical analyses.
Samples of concentrate were collected in separate vale bags at the beginning of each period and stored at -20 C until chemical analysis. The vegetation picked up by the sheep harnessed with faecal bags during grazing was snatched by the operator (daily 60 bites from the mouth of each sheep, 30 bites in the morning and 30 bites in the afternoon) before it was contaminated by saliva and masticated (Sankhyan et al. 1999) at the first five consecutive days in each period. Samples of vegetations were mixed and pooled for 5 days collection period according to the forage species and stored at -20 C to use for future chemical analyses. The contribution of each forage species was calculated according to the weight of each forage and the total weight of all forage (% dry matter basis). Percentage of contribution for each forage species, ingredients of concentrate and diet nutrient levels are presented in Tables 1-3, respectively. The consumption of FA was calculated for each 30day period as a proportion of metabolic weight, and results showed that GS lambs ingested more C12:0, C14:0, C15:0, C16:0, C18:0, C18:1n9c, C18:2n6c, C20:4n6, C22:6n3, SFA, MUFA, n-6PUFA and n-6/n-3, while less C16:1, C18:3n3, C20:5n3, PUFA and n-3PUFA than NG lambs (P < .05); On the other hand, HZ lambs ingested more C12:0, C14:0, C16:0, C16:1, C18:0,  C18:1n9c, C18:2n6c, C22:6n3, SFA, MUFA, n-6PUFA and n-6/n-3, while less C18:3n3, C20:5n3, PUFA and n-3PUFA than HL lambs (P < .05).
At 60 days, 10 lambs from each treatment were randomly selected and slaughtered by exsanguination. Before slaughter, the lambs were prevented from consuming food for 24 h and from drinking for 2 h. Immediately after death, LT and SAT samples were collected from the left side of carcase at the fifth lumbar vertebra levels under sterile conditions, snap frozen in liquid N 2 and stored at À80 C until RNA extraction for gene expression analysis. Fifty gram samples of LT and SAT were vacuum-packed and stored at À20 C for lipid extraction and determination of FA composition.

Proximate analysis of diets and faeces
The grass, concentrate mixture and faeces samples were analysed for the proximate chemical composition by oven drying at 60 C till constant weight according to the method of the Association of Official Analytical Chemists (AOAC 2004). Nitrogen was determined by Kjeltec Auto Analyzer and converted to crude protein (CP ¼ N Â 6.25). Ether extract (EE) was determined by extracting the sample with diethyl ether (60-75 C) using a Soxhlet extractor. The neutral detergent fibre (NDF) and acid detergent fibre (ADF) were determined according to Van Soest et al. (1991) with an Ankom 220 Fiber Analyser (Ankom Co., USA) and expressed inclusive of residual ash. Heat-stable amylase was not used in the NDF determination. Samples were ashed in a muffle furnace at 550 C for 4 h to determine the ash content. Gross energy (GE) was measured using an adiabatic calorimeter bomb (IKA C7000, Staufen). Digestible energy (DE) was determined by subtracting energy excreted in faeces from GE intake. Meanwhile, representative samples of faeces, grass and concentrate mixture were dried in oven at 105 C for 48 h to measure dry matter (DM). The diet ingredients and faecal samples were also analysed for Acid Insoluble Ash (AIA) using the 4N HCI procedure described by Vogtmann et al. (1975).

Calculation of dry matter intake
Dietary intake of animals was assessed by 4N-HCL AIA marker in combination with total faecal collection, according to the following equation: DMI ¼ faecal output Â % AIA in faeces/% AIA in diets, where faecal output is expressed as dry matter amount. We calculated the total forage intake (presented in Table 1) by subtracting the concentrate mixture intake from the dry matter intake.

Measurement of fatty acid
Fatty acid methyl esters were produced from 0.5 g samples of feed, muscle or 0.05 g sample of adipose tissue according to O'Fallon et al. (2007). For determination of FA concentration, a GC-2014 gas chromatograph (Shimadzu International Trading CO., LTD) with a HP-88 column (Agilent, 100 m Â 0.25 mm ID Â0.2 mm film thickness) was used. The injector temperature was 250 C, and the column was programmed to run at 120 C for 10 min, warmed to 230 C at 3.2 C/min, and held for 35 min to achieve satisfactory separation. Nitrogen was the carrier gas with a flow rate of 1.75 ml/min and split ratio of 1:50.

Statistical analysis
ANOVA was performed to compare the effects of pasture vs. supplementary feeding on lamb muscle and adipose tissue FA and lipogenic gene expression. The data means were considered significantly different at P < .05, and tendencies were considered at .05 P < .10.

Gene expression in LT muscle
The relative expression of genes in the LT muscle is presented in Table 7. Compared to GS lambs, HSL gene expression showed a significant increase in the NG lambs, while other genes expression showed the opposite result (P < .05) with one exception, ACC, whose expression did not show difference between NG and GS lambs. Between breeds, HSL gene expression showed a significant increase in the HL lambs compared with HZ lambs, but other genes expression showed the opposite result (P < .05).
Interactions among fattening ways and breeds had significant effects on mRNA expression of lipid metabolism genes in the LT muscle. In LT muscle, PPARc and C/EBPa were highest from HZ GS treatments and lowest from HL NG treatments (P ¼ .0001, P ¼ .0001). Similarly, expression of ACC and LPL were highest from HZ GS and HZ NG treatments, with no difference between HZ GS and HZ NG treatments, and lowest from HL NG treatments (P ¼ .0001, P ¼ .04). HSL was highest from HL experienced NG, and lowest from HZ experienced GS (P ¼ .0001). Expression of FAS was highest from HZ GS treatments, and lowest from HL Table 7. Effect of feeding system on relative expression of lipogenic and lipolysis genes in the LT muscle of lambs.  NG and HZ NG treatments (P ¼ .0001). Expression of SREBP1c was highest from HZ GS and HL GS treatments, and lowest from HL NG treatments (P ¼ .0001).

Gene expression in SAT
The relative expression of genes in SAT is presented in Table 8. ACC, FAS, LPL, PPARc, SREBP1c and C/EBPa showed lower expression levels in NG lambs compared to GS lambs (P < .05). In reverse, HSL showed higher expression level in NG lambs compared to GS lambs (P < .05). Compared with HL, there was a significant increase in mRNA expression of FAS, ACC, SREBP1c, C/EBPa, PPARc and LPL in SAT from HZ, but HSL gene expression was significantly reduced in HZ compared to HL (P < .05). The mRNA expression of genes related to lipid metabolism of the SAT was significantly influenced by the interaction between fattening ways and breeds. In SAT, HSL was highest from HL experienced NG and lowest from HZ experienced GS (P ¼ .0002); The relative expressions of SREBP1c, ACC, FAS and LPL were highest from HZ experienced GS and lowest from HL experienced NG (P ¼ .0001, P ¼ .0001, P ¼ .002, P ¼ .008); C/EBPa and PPARc were highest from HZ experienced GS and lowest from HL experienced NG and HZ experienced NG (P ¼ .0001, P ¼ .0001).

Discussion
We measured FA profiles and levels of mRNA expression of genes implicated in de novo FA synthesis (ACC, FAS), hydrolysis of the triacylglycerol (TG) core of circulating TG-rich lipoproteins (LPL), hydrolysis of stored triacylglycerol (HSL) as well as transcriptional factors (SREBP1c, PPARc, C/EBPa) in LT and SAT of lambs experienced NG and GS. In both feeding regimens, the predominant FA in LT muscle and SAT of HL and HZ were C16:0 and C18:0 as SFA, C18:1n9c as MUFA and C18:2n6c as PUFA which are similar to those reported by Lee et al. (2008). Most SFA are presumed to increase the risk of heart disease, while PUFA and MUFA are generally regarded as beneficial to human health (Scollan et al. 2001). Especially, consumption of n-3PUFA has been shown to positively influence immune function, blood pressure, cholesterol and triglyceride levels in humans. These benefits are almost certainly a consequence of the conversion of the C18:3n3 to C20:5n3 (EPA) and C22:6n3 (DHA) (n-3 long chain polyunsaturated fatty acids, LCPUFA; Tu et al. 2010). The levels of C18:3n3, EPA, n-3PUFA and PUFA in LT muscle and SAT of NG lambs were significantly higher, indicating NG and GS had significant effects on the FA profiles in tissues of lambs, and the causes of these differences were associated with the ingested diets. Concentrate was supplemented in GS groups. Therefore, CP, EE, DE and P were indeed slightly higher in GS than in NG, while NDF, ADF and Ca were slightly lower. However, Castro et al. (2005) showed that C18:3n3 in meat and adipose tissue did not change with an increase in dietary CP, EE, NDF and ADF. Ebrahimi et al. (2013 and studied the effects of dietary FA on FA and gene expression in goat meat and adipose tissue and observed increases in meat and adipose C18:3n3 with increases in dietary C18:3n3. The only variable in their experiments was dietary FA, and dietary ME, CP, EE, NDF, ADF, Ca and P were same. It is clear from their results that dietary C18:3n3 exerts a major influence over tissue C18:3n3. Therefore, we contented that it is dietary FA rather than dietary energy, fat, protein, etc. exerts a major influence over tissue FA and genes expression. This dietary FA effect may be related to a change in the FA composition of adipocyte membranes, which is strongly influenced by the relative abundance of PUFA in the diet (Hulbert et al. 2005). Such changes could alter membrane fluidity, which is involved in the mechanisms of FA transport across membranes Table 8. Effect of feeding system on relative expression of lipogenic and lipolysis genes in the SAT of lambs. PPARc: Peroxisome proliferator-activated receptor gamma; SREBP1c: Sterol regulatory element binding transcription factor 1c; C/EBPa: CCAAT/enhancer-binding proteins alpha; FAS: Fatty acid synthetase; LPL: Lipoprotein lipase; ACC: Acetyl-CoA carboxylase; HSL: Hormone-sensitive lipase; SEM: standard error of the mean; NG: natural grazing; GS: grazing with supplementary feeding; HL: Hulunbeier lambs; HZ: Hulunbeier Â Dorper crossbred lambs; SAT: subcutaneous adipose tissues. P1: P value of feeding regimens. Means within a row with different superscripts (a, b) are different between the two fattening ways (P < .05). P2: P value of breeds. Means within a row with different superscripts (a, b) are different between the two breeds. P1 Â P2: P value of interactions of fattening ways and breeds. Means within a row with different superscripts (A, B, C, D) are different.
(Bojesen and Bojesen 1999), leading to the different concentrations of FA in adipocytes. Ebrahimi et al. (2014) reported that the reduction of SFA in intramuscular tissue of Boer kid goats obtained by feeding flaxseed oil was mainly due to the inhibitory effect of a-linolenic acid (C18:3n3) and/or its biohydrogenation products on de novo FA synthesis. Studies have shown that dietary PUFA regulates lipid metabolism through inhibiting FA synthesis de novo by restraining the expression of FAS and ACC (Iritani et al. 1998) and increasing the gene expressions of enzymes associated with FA oxidation (Nakamura and Nara 2004). In our study, NG lambs ingested more 18:3n3 and PUFA as primary FA than for GS lambs, and the LT and SAT of lambs from NG groups displayed lower expression value of FAS, SREBF1c and ACC (only SAT) than lambs from GS groups which could be the explanations for some of the decreases in individual and total SFA in tissues of the pasture-fed lambs. DHA content in LT muscle and SAT of NG lambs were significantly higher than that of GS lambs, but intake of DHA of NG lambs was lower than that of GS lambs in our study, which was agree with Ebrahimi et al. (2014) who found that higher proportions of ingested C18:3n3 could result in C18:3n3, EPA, DHA and total n-3PUFA increase, and decrease the profile of total n-6 PUFA as well as n-6/ n-3 ratio in muscle and adipose tissue of lambs. This may be due to that chain elongation of C18:3n3 could escape biohydrogenation (Vatansever et al. 2000). The more C18:3n3 that lambs consuming, the more DHA and EPA (elongation products of C18:3n3) produced in the rumen. As a result, more DHA and EPA deposited into muscle and fat tissues. In addition, members of the n-6 and n-3 families compete for the elongation-desaturation pathways (Sprecher 1981). The enzymes required to synthesise n-3 LCPUFA from C18:3n3 are also used to synthesise n-6 LCPUFA from C18:2c6. Bioconversion of C18:3n3 into higher homologues depends on the ratio of ingested n-6/n-3 PUFA, a ratio of 1/1 being proposed to lead to the highest formation of n-3 LCPUFA (Harnack et al. 2009). In our study, the n-6/n-3 ratio in the LT muscle and SAT of NG lambs was closer to 1, a context, which might favour the beneficial C18:3n3 conversion (Gruffat et al. 2011). Therefore, GS lambs presented more C18:3n6 or C20:3n6 (elongation products of C18:2n6c) in LT or SAT. On the other hand, our results presented that HZ lambs ingested less C18:3n3 and EPA, but EPA and DHA contents in LT and SAT did not less than those of HL lambs, which showed heterosis. C18:2n6c present in the tissue is derived entirely from its presence in the diet (Romero-Bernal et al. 2017); this FA is particularly abundant in food concentrates (grains and oil seeds) and is degraded to MUFA and SFA in the rumen by rumen microbial biohydrogenation, with only a small proportion of around 10%, compared to the amount ingested with the diet, being available for incorporation into tissue lipids (Wood et al. 2008). This is the main reason for the higher proportion of C18:2n6c from tissues of GS and HZ lambs than NG and HL lambs, respectively. Significantly, HZ lambs ingested and deposited more C18:2n6c in LT and SAT than HL lambs, but the proportion of elongation products of C18:2n6c including C18:3n6, C20:3n6 and C20:4n6 was significantly decreased in SAT, and the proportion was significantly increased in LT compared to HL lambs. This indicated that FA profile was varied according to different breeds or tissues. Interestingly, in our study, C18:2n6c and C18:3n3 in both GS and NG lambs were at higher proportions in muscles than adipose tissues. For example, the proportions of 18:3n3 in total muscle lipid and adipose tissue lipid from NG lambs were 1.890 and 1.523 g/100 g. For GS lambs, these were 1.570 and 0.894 g/100 g. This implies that ruminants preferentially incorporate essential fatty acids, into muscle rather than storing them in adipose tissue. The nutritional value of food fat is assessed by considering that the n-6/n-3 ratio should be less than 4 (Enser et al. 1999). The PUFA/ SFA ratio (P/S) fixed for human nutrition should be around 0.7 or lower (Raes et al. 2003). Therefore, in our study, meat from all lambs was suitable as human food. We did not find significant differences in C18:0 proportions among the treatments which has a neutral effect towards humans (Bonanome and Grundy 1988), although GS lambs and HZ lambs ingested more C18:0 than NG lambs and HL lambs, respectively. According to Doreau and Ferlay (1994), a greater inhibition of rumen biohydrogenation could be displayed in the lambs consuming concentrate. On the other hand, due to the hydrogenation of rumen microbes, the majority of PUFA in pasture are converted into SFA in the rumen, which only a small percentage of PUFA escapes (Wood et al. 2008). Therefore, GS lambs and HZ lambs may produce less C18:0 than NG lambs and HL lambs, respectively. As a result, no differences of C18:0 proportions in LT and SAT was observed among the treatments. In the ruminants, D9 desaturase is an important enzyme that has been shown to control the degree of unsaturation. It can convert C16:0 and C18:0 into C16:1 and C18:1. In our study, more C18:1 was present in the lamb muscles of GS groups, but C16:1 showed the opposite result and it seems to be related to the ingested FA. The D9 desaturase activity is down regulated by PUFA (Ntambi and Bene 2001), which is more common in the tissues of lambs with pasture-based diets. C18:1n9 was the major MUFA in the lamb muscle fat, and our lambs that received supplemented concentrate displayed greater proportions of this FA than the NG lambs; in agreement with the FA proportions of the diets.
Because the SFA C12:0, C14:0 and C16:0 increase synthesis of cholesterol and favour the accumulation of low density lipoproteins, they are a major risk factor for cardiovascular diseases in humans (Moloney et al. 2001). Moreover, Yu et al. (1995) suggested that C14:0 is 5-6 times more atherogenic or hyper-cholesterolaemic than either C12:0 or C16:0. Thus, meat and fat from HZ lambs displayed a less favourable level of SFA as containing higher content of C14:0. The higher levels of total SFA and C12:0, C14:0 and C16:0 detected in HZ lambs than in HL lambs, would seem to confirm that breed is a determinant factor affecting the fatty acid profile of lamb's fat, which was similar to other studies (De Smet et al. 2004).
The effect of feeding regimens on the higher concentration of C15:0 found in NG lambs could be attributed to the development of ruminal microflora because odd-chain fatty acids are generated from bacterial lipids (Jenkins 1993), and also could be attributed to the higher C15:0 intake. However, the effect of breeds on the higher concentration of C15:0 found in HL lambs could only be attributed to the development of ruminal microflora, because HL lambs ingested less C15:0 compared to HZ lambs.
Compared with NG, GS improves growth performance and slaughter performance of HL and HZ lambs; compared with HL, HZ lambs have higher values of growth performance and slaughter performance under the same feeding modes, including NG and GS (Wen et al. 2017). The present study revealed that GS lambs had less HSL expression level in LT and SAT than NG lambs, which indicates that the pasture grass energy level does not meet their growth requirements. On the other hand, mRNA expression of this gene in both LT and SAT of HZ NG treatment was lower than that of HL GS treatment, which showed heterosis. LPL-catalysed reaction products, FA and monoacylglycerol are taken up in part by adipose tissue and skeletal muscle and stored as neutral lipids (Wang and Eckel 2009), which suggests that the LPL gene may be a genetic marker predictive of fat deposition. Guo et al. (2014) noted that C/EBPb is induced early to transactivate the expression of C/EBPa and PPARc, two master transcription factors for terminal adipocyte differentiation during 3T3-L1 adipocyte differentiation. The present study also revealed that GS lambs had higher LPL, C/EBPa and PPARc expression levels as well as FAS, ACC and SREBP1c expression levels as mentioned above in SAT or/and LT than NG lambs, indicating more fat deposition in GS than NG lambs as also found by Wang et al. (2015), who demonstrated that concentrate supplementation increased IMF content in the LT muscle. At the same time, we found that some genes were influenced by the interaction of feeding regimens and breeds, including ACC, LPL, PPARc and C/EBPa in LT as well as FAS, ACC, LPL and SREBP1c in SAT. mRNA expression of all these genes of HZ NG treatment was higher than that of HL GS treatment, which showed heterosis. Interestingly, supplementation of concentrate only increase the expression of ACC and LPL in LT of HL lambs, but did not have effect on HZ lambs, indicating that diets influence genes expression may be associated with animal breeds. Further studies are needed.

Conclusions
Feeding regimens (NG vs. GS) and breeds (HL vs. HZ) have significant effects on the FA profiles and gene expression of enzymes related with fat metabolism of the intramuscular fat and subcutaneous adipose tissues of HL and HZ. Compared with NG, GS increases de novo FA synthesis of LT and SAT, which decreases GS lambs' FA nutrition value. Meat or fat from NG lambs present more n-3 LCPUFA that is beneficial to human health. Under the two feeding regimens, de novo synthesis of FA and fat deposition seems to higher in HZ lambs compared with those in HL as the expression of genes that are associated with correlative metabolism was increased in HZ lambs. Meat from HZ lambs displays a less favourable level of SFA as containing higher content of C14:0, but it also presents higher profile of EPA than HL lambs.

Ethical approval
All procedures involving animals were evaluated and approved by the guidelines of the Animal care and Use Committee of the Inner Mongolia Agriculture University (Hohhot, Inner Mongolia, China).

Disclosure statement
No potential conflict of interest was reported by the authors.