The study group consisted of 30, single-born, male Norduz lambs. The 30 lambs (age: 160–185 days; mean: 171 days; live weight: 35.01 ± 0.338 kg) were assigned to one of the two dietary regimes: 15 lambs grazed on pasture (PS) and 15 lambs were fed with concentrate ration in stall (CO). The grazing pasture, populated with multiple species of native plants, was located at 38∘18′ N and 42∘49′ E and 2460 m above sea level. The pasture has also been documented by Beyis (2009), who stated that 14.3 % Poaceae (grasses) and 13.4 % Fabaceae (legume) taxa contributed to the floristic composition, with 72.3 % taxa from other plant families, from which a hay harvest yielded 955 kg ha-1. For the nutrient composition of the pasture, given in Table 1, the values determined by Karaca (2010) in the following year for the same pasture were used. Each lamb on the concentrate diet was fed with 200 g of alfalfa hay and ad libitum concentrate (barley 72.5 %, cotton seed pulp 24.0 %, calcium carbonate (CaCO3) 2.4 %, salt (NaCl) 0.5 %, vitamin supplement 0.5 %, and mineral supplement 0.1 %). The chemical compositions of the diets of the lambs are given in Table 1.

All lambs were slaughtered after 12 h of fasting at the end of the fattening period of 84 days. After slaughter, the carcasses were stored at 4 ∘C for 24 h. The jointing of carcasses was based on the method given by Colomer-Rocher et al. (1987). The longissimus thoracis (LT) muscle (6th–13th ribs) was removed from the left side of the carcass at 24 h postmortem and used for color, water-holding capacity, proximate and FA analyses. In addition to this, the semimembranosus (SM) and triceps brachii (TB) muscles, tail fat (TF), and subcutaneous (SC) fat samples were removed from each carcass for FA profile determination.

Carcass pH was measured at 45 min post-slaughter (pH45) and at 24 h post-slaughter (pH24) using a digital pH meter (Hanna HI 99163N, Hanna instruments, Romania) equipped with a penetrating electrode and thermometer. The pH was directly measured on the LT muscle between the 12th and 13th thoracic vertebrae. In order to determine meat color, each sample had five measurements, from the fat-free areas, using a Lovibond RT-300 portable spectrophotometer (The Tintometer Limited, UK) (CIELAB-Illuminant D65/10∘) at 24 h post-slaughter. The measurements were performed on a freshly cut surface of 2.5 cm thick LT (12th–13th ribs) after allowing the muscle surface to bloom in the chiller for 45 min. Water-holding capacity (WHC) was determined according to Wierbicki and Deatherage (1958), where a 0.5 g sample of muscle tissue was placed on filter paper and pressed at 500 psi min-1 between two plexiglass plates. The results were expressed as the percentage of free water.

Longissimus muscle samples were vacuum-packed and frozen to -18 ∘C for use in chemical analysis. The samples were then thawed overnight, preceding the start of the analysis, and minced and homogenized. The nutrient content was analyzed according to AOAC (2000) by homogenizing the samples and measuring dry matter, ash, fat, and protein content (moisture: 950.46; ash: 920.153; fat: 960.39 (determined using ether extraction); protein: 928.08 (determined using the Kjeldahl method)). Neutral detergent fiber (NDF) and acid detergent fiber (ADF) contents of feed samples were determined according to Van Soest and Robertson (1979).

Fatty acid analysis was performed on samples that had been stored at -18 ∘C for 1 month and then allowed to thaw overnight at 4 ∘C. Fat samples that were used included SC (from above the 6th–12th chop) and TF, and intramuscular fat taken from the LT, SM, and TB muscle samples. Muscle (25 g) or fat (2 g) were homogenized with chloroform–methanol (2 : 1, v/v), and total lipids were extracted according to the procedure described by Folch et al. (1957); methylation was performed as described previously by Basturk et al. (2007). FA composition was determined by gas chromatography (Agilent 6890 N, Agilent Technology, USA) equipped with a flame ionization detector and a polar capillary column (DB-23, Agilent Technology, USA; 60 m × 0.25 mm I.D., 0.25 µm). Helium was used as the carrier gas (1.5 mL min-1). The oven temperature was programmed at 120 ∘C for 5 min, increased to 240 ∘C at a rate of 15 ∘C min-1 and held at 240 ∘C for 20 min. FA methyl esters were identified by matching their retention times with those of their relative standards (Supelco 37 component FAME Mix, Sigma-Aldrich Co., USA). Moreover, the fatty acid composition of feed samples was determined according to the official EEC (1991) method (2568/91).

The saturated fatty acids (SFAs), monounsaturated fatty acids (MUFAs) and polyunsaturated fatty acids (PUFAs) were calculated with the following Eqs. (1–3): SFA=C10:0+C12:0+C14:0+C15:0+C16:0+C17:0+C18:0+C20:0, MUFA=C14:1+C15:1+C16:1+C17:1+C18:1n-9+C20:1, PUFA=C18:2n-6+C18:3n-6(γ)+C18:3n-3(α).

The desirable fatty acids (DFAs) were calculating according to Huerta-Leidenz et al. (1991) with the following Eq. (4): DFA=Total unsaturated fatty acids(TUFA)+C18:0.

Activities of desaturase activity indices were calculated according to Juárez et al. (2008) with the following Eqs. (5) and (6): Δ9DS(C16)=100×C16:1/(C16:0+C16:1),Δ9DS(C18)=100×C18:1/(C18:0+C18:1).

One-way analysis of variance was performed using the Minitab 13.0 software program. Except for FAs, data were analyzed according to the following Eq. (7): yij=μ+ai+eij. The FA composition of different anatomical regions, in both PS and CO groups, was examined using the Eq. (8): yijk=μ+ai+bj+(ab)ij+eijk, where yijk is the value of the examined characteristic for the kth animal in the jth anatomical region from the ith feeding system; μ is the overall mean; ai is the fixed effect of feeding system (ai: concentrate or pasture); bj is the fixed effect of anatomical region (j: SC, TF, LT, SM, and TB); (ab)ij is the interaction of the effects; and eijk is the random error. The principal component analysis was performed using the multivariate subsection of Minitab 13.0.

The results of slaughter traits for lambs raised under different feeding systems are given in Table 2. It was determined that the CO group had a higher slaughter and carcass weight than the PS group at the end of the fattening period. The dressing percentages were found to be lower in the PS lambs rather than the CO lambs. It has been frequently reported (Karaca and Kor, 2015; Priolo et al., 2002) that pasture-fed lambs have lower dressing percentages than concentrate-fed lambs, and this is thought to be related to the differences in the gastrointestinal content and fattening level of the feeding groups. Papi et al. (2011) reported that lambs fed low-energy and high-fiber diets had a higher digestive system content than those fed high-energy diets. The CO lambs had significantly higher omental–mesenteric fat (226.3 vs. 118.49 g; p< 0.05) and kidney-knob and channel fat (175.0 vs. 114.4 g; p< 0.05) than the PS lambs (data not shown in the table), whereas the percentages of fat recorded at the slaughter and cold carcass weights were found to be similar between the two groups (Table 2). The weights of the non-carcass parts, such as feet, which correlate with poor body growth, were greater in the PS lambs than the CO lambs and are consistent with those reported by Moron-Fuenmayor and Clavero (1999).

Meat quality traits of CO and PS lambs are presented in Table 3. Although pH45min results were similar between groups, PS lambs had significantly higher pH24h than CO lambs. The ultimate high pH in PS lambs can be related to higher muscle activity and a low-energy diet in these lambs compared to CO lambs. Some studies have also shown that increases in metabolic energy of the diet result in parallel increases in the muscle glycogen content (Immonen et al., 2000). Further results (Duckett et al., 2013; Karaca and Kor, 2015) also support the abovementioned findings, and the ultimate pH was higher in the lambs fed on pasture than lambs fed with a high-energy diet. However, it was reported by different researchers that the different feeding systems had limited effects, particularly on pH, as well as on the WHC and tenderness of the meat (Diaz et al., 2002; Sanudo et al., 2007).

The ultimate pH of meat has a determining role in the meat color, and high ultimate pH results in a darker color compared to a lower pH (Sanudo et al., 2007). Comparable to these reports, PS lambs in our study, with a high ultimate pH, had lower luminosity (L*) (p< 0.001) compared to CO lambs. Similar to our findings, research has shown that the meat color in PS lambs is darker than that in CO lambs (Diaz et al., 2002; Karaca and Kor, 2015; Priolo et al., 2002). In addition, Minchin et al. (2009) reported that the brighter appearance of meat from cows fed with high-energy diets can be due to the changes caused in reflectance values by increased fat deposition. However, Priolo et al. (2001) reported that the meat color was darker in pasture-fed lambs, even in conditions where fat deposition was high, when compared to concentrate-fed lambs.

Another important consideration in the pasture-fed lambs is the redness of meat or the a* value. It was reported that the a* value was higher in pasture-fed lambs than in concentrate-fed lambs (Carrasco et al., 2009a; Ripoll et al., 2008). The high a* value is associated with raised pigmentation depending on increased muscle activity and live weight of the lambs (Ripoll et al., 2008; Sanudo et al., 2007; Carrasco et al., 2009a). In our study, no significant difference with respect to the a* value could be found between the CO and PS lambs. Thus, it can be suggested that increases in the a* values due to high slaughter weight of CO lambs may lead to similar results in these groups. In addition to this, meat yellowness (b*) and saturation (h∘) values were significantly higher in CO lambs than in PS lambs (p< 0.05) (Table 3). The high ultimate pH of PS lambs had a negative effect on meat color parameters and plays an important role in the variance between b* and h∘ values.

It was determined that the percentage of free water was lower and WHC was better in the CO lambs than in the PS lambs (Table 3). Santos-Silva et al. (2002b) reported similar results. However, Diaz et al. (2002) found no difference in WHC in concentrate- and pasture-fed lambs. It is also known that in meats with a higher ultimate pH, the WHC increases (Sanudo et al., 2007). However, in our study, the PS lambs had a higher ultimate pH than the CO lambs, which had a lower WHC. It seems that the higher fat deposition recorded in the CO group affected the results of WHC.

The production system was also found to have an important effect on the nutritional content of the meat in the present study (Table 3). The moisture content in meat was found to be higher in PS lambs than in CO lambs (p< 0.001), while the protein and fat content was higher in CO lambs. It should be noted that the dry matter content also varied between the groups due to differences in the moisture content of the nutrients. When these results were evaluated on the basis of dry matter (DM %), protein and ash were found to be higher in the PS lambs. In similar studies, it was determined that the moisture content was higher in the pasture-fed lambs than in the concentrate-fed lambs, while the fat content was lower (Karaca and Kor, 2015; Rowe et al., 1999).

In ruminants, most dietary unsaturated FAs are hydrolyzed and then form saturated FAs, and this process varies significantly depending on the biohydrogenation activity (Wood et al., 2008; Zervas and Tsiplakou, 2011). Comparing the effects of the feeding system in terms of the percentages of C16:0 and C18:0 FA in depot fats, it can be observed that the percentage of C16:0 was higher in the CO lambs, whereas the percentage of C18:0 was higher in the PS lambs (Table 4). In terms of the intramuscular fat in CO and PS lambs, the percentage of C18:0 was found to be higher in the PS lambs (Table 5). Similar results were also obtained in some previous studies (Diaz et al., 2002; Guler et al., 2011; Rowe et al., 1999). The amount of C12:0, C14:0, and C16:0 found in meat is important due to the hypercholesterolemic effects, which may increase the risk of atherosclerosis and thus have adverse effects on human health (Zock et al., 1994). The amount of C18:0, which is one of the primary FAs in ruminant meats, depends on the biohydrogenation of unsaturated C18 FAs. The biohydrogenation activity of polyionic FAs decreased due to the concentrate diet reducing rumen pH levels and the time spent by the concentrate in the rumen when compared to that of forage (Jenkins et al., 2008; Wood et al., 2008; Diaz et al., 2002). However, it was reported that there were similarities in the percentages of C16:0 and C18:0 in SC and TF (Nuernberg et al., 2008; Velasco et al., 2001) and intramuscular fats (Cividini et al., 2014; Popova, 2007) from pasture and concentrate lambs. In the present study, no significant difference in the percentage of saturated FAs (SFA) in depot fats was determined. However, the percentage of the SFA in intramuscular fats was higher in PS lambs than in CO lambs (p< 0.001).

The percentages of some of the monounsaturated fatty acids (MUFAs) in depot fats, such as C14:1, C16:1, and C17:1, was found to be higher in CO lambs than in PS lambs (Table 4). Regarding intramuscular fats, CO lambs were found to have a higher C18:1n9 level than PS lambs (Table 5). These results may be associated with the differences in dietary FA profile and rumen biohydrogenation activity. In a concentrate-based diet, the percentage of the C18:2n6 is quite high compared to that in pasture (Table 1). There is also an observed increase in the synthesis of oleic acid from stearic acid, depending on an increasing level of Sterol-CoA desaturase (Δ9 desaturation) enzyme activity combined with an increased fat deposition (Velasco et al., 2001; Cividini et al., 2014). The Δ9 desaturation (C18) index, both in depot and intramuscular fats, was found to be higher in CO lambs than in PS lambs (p< 0.001; Tables 4 and 5). In similar studies, the C16:1 and/or C18:1 content was high in the concentrate-fed lambs (Cividini et al., 2014; Fisher et al., 2000; Santos-Silva et al., 2002a; Nuernberg et al., 2008), but these results were not found to be significant in other studies (Diaz et al., 2002; Popova, 2007). Moreover, the effect of feeding system on C16:1, C18:1n9 and C18:3n3 varied depending on muscle type, and significant interactions were present for these FAs (p< 0.001). The percentages of C18:1n9 and C18:3n3 were similar in CO and PS lambs for LT, while PS lambs had a lower percentage of C18:1n9 and a higher percentage of C18:3n3 than CO lambs for SM and TB (Table 5). The differences in muscle type in terms of activity and low intramuscular fat deposition in these muscles might cause LT to have different profile for these fatty acids than SM and TB in PS lambs. Hence, Moreno et al. (2006) suggested that low intramuscular fat deposition could alter the FA profile of calves fed on pasture and concentrate. In addition to this, Rogowski et al. (2013) reported that increasing Δ9-desaturase activity of muscle increases triglyceride PUFA content.

The results obtained in this study were similar to those of previous studies (Diaz et al., 2002; Popova, 2007; Rowe et al., 1999), and the percentage of linoleic acid (LA, C18:2n6) was found to be higher in depot and intramuscular fats of CO lambs than of PS lambs, whereas alpha-linolenic acid (ALA, C18:3n3) levels were higher in the PS lambs except for in LT (Tables 4 and 5). ALA is elongated and desaturated into longer-chain omega-3 FAs, such as C20:5n3 (EPA) and C22:6n-3 (DHA), and therefore, the availability of ALA is highly important (Bessa et al., 2015). In general terms, the effect of the feeding system on polyunsaturated FAs (PUFA) in tissues can be associated with the high amount of LA found in concentrate feed and of ALA in pasture (Table 1).

The results of the principal component analysis (PCA) are presented in Fig. 1a. The first component (PC1) explained 52.7 % of the total variation. SC and TF were located on the opposite side of PC1 and were clearly separated from each other. It can be seen that SFA and SC were located on the left side of the PC1 and were associated with each other, whereas PUFA, MUFA, Δ9 DS 16/18, n6, and TF were located on the right side. The results confirmed the differences between TF and SC and were in agreement with Table 4. The second component (PC2) explained 27.3 % of variability and was characterized by CO, PS, and n3. CO was located on the opposite side of PS and n3, in PC2, and the relationship between PS and n3 FAs can be clearly seen. This relationship is in agreement with results reported by other researchers (Fisher et al., 2000; Özcan et al., 2015).

The results of PCA analysis in intramuscular FA composition were presented in Fig. 1b. PC1 and PC2 explained 41.7 and 21.1 % of the total variation, respectively. Unlike fat depots, PS was associated with SFA and located on the left side of the first component, opposite to CO, which was related to MUFA and Δ9 DS 16/18 on the right side of the PC1, in the intramuscular fat. PC2 was characterized by n3 and PUFA and closer to SM than LT and TB, which were located opposite to PC2.

Some of the total FAs found in depot and intramuscular fats and their ratios, which are used as criteria in healthy nutrition have been shown in Tables 4 and 5. Although there was no significant difference regarding the percentages of the SFA, MUFA, and PUFA in depot fats, both higher SFA (p< 0.001) and lower MUFA (p< 0.001) levels were found in the intramuscular fats of PS lambs than in that of CO lambs. Similar results were also reported in some previous studies (Diaz et al., 2002; Guler et al., 2011; Özcan et al., 2015; Popova, 2007). The British Department of Health (1994) has recommended the PUFA / SFA ratio to be above 0.45 and the n6 / n3 ratio to be below 4.00. Despite a similar PUFA / SFA ratio in depot fats for both groups (Table 4), the ratio was found to be higher in the intramuscular fats of CO lambs than in that of PS lambs (p< 0.01) (Table 5). Moreover, the n6 / n3 ratio in both depot and intramuscular fats was found to be lower in PS lambs than in CO lambs (p< 0.001) except for LT.

With the results associated with FA composition in SC and TF, it was determined that the percentage of the SFA was higher in SC, whereas MUFA and PUFA were determined to be higher (p< 0.01) in TF (Table 4). It must be noted that the effect of the feeding system on some FAs varies according to individual SC and TF and that interactions related to these FAs were found to be significant (p< 0.05). Despite the fact that the percentages of the C14:1, C15:1, C16:1, and C17:1 MUFA as well as Δ9 desaturation (C16) index in SC and TF of CO lambs showed similarities, the same FAs in the SC of PS lambs were found to be significantly lower when compared to the TF values (p< 0.05). In accordance with these results, Gallardo et al. (2014) reported that FA composition, in both SC and TF in lambs, varied according to pasture type. The researchers noted that the tissue-specific response, depending on the pasture type, may be associated with gene and protein expression or lipogenic enzyme activity (stearoyl-CoA desaturase; SCD). Moreover, TF had a higher PUFA / SFA ratio than SC (p< 0.001), and the n6 / n3 ratio of TF was found to be lower than that of SC (p< 0.05). It can be suggested that TF has more beneficial FAs than SC.

Ruminants tend toward storing essential FAs in intramuscular fats instead of depot fats (Wood et al., 2008). Alpha-linoleic acid was found to be the most important source of diversity in intramuscular fats (SM > TB > LT; p< 0.05), and SM was the muscle tissue with the highest PUFA / SFA and lowest n6 / n3 ratios (Table 5). Similar results were reported by Coutinho et al. (2014), and SM and TB had higher PUFA / SFA ratios than LT.