1. Introduction
The Assaf breed has successfully spread in the Mediterranean area during the last four decades, and it is currently the main sheep breed in intensive dairy production systems in several countries [
1,
2,
3]. In these systems, milk sales represent the main income component, and most male lambs are sold as milk-fed lambs [
4,
5]. Nevertheless, heavy fattening male lamb production is increasing in response to growing demand from African countries. It represents an important opportunity to compensate, in the short term, for seasonal fluctuations in local demand and, in the long term, for the reduction in red meat consumption in developed countries [
6].
In practice, the feeding of fattening lambs very much depends on protein supplements since dietary protein content could affect feed intake, animal growth, and carcass and meat characteristics [
7,
8]. Nevertheless, protein over-feeding in ruminant diets projects a negative image of the production system because of both the environmental cost and the competition for these resources with humans and other livestock species that are more efficient than ruminants [
9]. Moreover, Europe is dependent on overseas land for soybean meal, the source of protein most used in livestock feeding.
Different feeding strategies could be used to reduce protein over-feeding, but the first option should be to accurately meet animal requirements to avoid over-feeding protein. Most studies carried out in heavy fattening Assaf lambs have used diets with 17–20% of crude protein (CP; dry matter (DM) basis) [
10,
11,
12,
13,
14]. However, this protein content is greater than that estimated as optimum by different feeding systems [
15,
16]. Likewise, Abo Omar and Naser [
17] did not observe differences in either feed intake or average daily gain (ADG) in Assaf lambs fed diets containing 140 or 180 g of CP/kg DM. On the contrary, Fernandez et al. [
18] reported differences in ADG in Assaf male lambs fed diets with 136 or 205 g CP/kg DM, and Rodríguez et al. [
19] observed similar results when comparing the animal performance of light fattening Assaf lambs fed a conventional diet (166 g CP/kg DM) or raised on a free-choice feeding system where animals selected a diet containing 230 g CP/kg DM. In addition to these controversial results, most of these studies were focused on animal performance, ignoring effects on meat quality. Therefore, there is a need to establish the optimal level of protein in male Assaf lambs while considering its effect on animal performance and meat quality. It is important to note that most published studies regarding meat quality in Assaf lambs have been carried out with milk-fed and light fattening lambs [
19,
20], and, to the best of our knowledge, the scientific information published on heavy fattening meat Assaf lambs is scarce [
14].
The objective of the present experiment was to evaluate the effect of dietary protein content on animal performance, feeding costs, ruminal fermentation pattern, blood acid–base status, biochemical profile, and carcass and meat quality in heavy fattening male Assaf lambs.
2. Materials and Methods
Experimental conditions and handling practices followed the recommendations of the Directive 2010/63/EU of the European Parliament and were evaluated and approved by the CSIC Animal Experimentation Committee (Protocol number 100102/2017-4).
2.1. Animals and Diets
Thirty male Assaf lambs were distributed in three experimental groups, equilibrated for weight (30 ± 1.9 kg) and age (89 ± 0.8 days), and they were randomly allocated to one of the three experimental diets differing in CP contents (134, 157, and 173 g of CP/kg DM for low protein (LP), medium protein (MP), and high protein (HP) diets, respectively). The ingredients and chemical composition of the experimental diets are shown in
Table 1.
Animals were individually penned during the whole experimental period (70 days) and fed the corresponding experimental diet ad libitum. Lambs had free access to fresh water and were able to see and hear the other animals.
2.2. Experimental Procedure
Feed intake was recorded daily, with the amount offered adjusted to allow for refusals of a minimum of 10%. Samples of the feed offered and refusals for each animal were collected every day, and a composited sample per week was stored for subsequent analyses.
Body weight (BW) was recorded every 2 weeks before daily feeding. On days 0, 35, and 70 before morning feeding and after having removed the refusals, a blood sample was taken by jugular venipuncture into lithium heparin tubes, placed in ice, and then centrifuged at 3520× g for 20 min at 4 °C. Then, plasma samples were frozen at −20 °C until the analysis of the biochemical profile. On days 35 and 70, an extra blood sample was obtained, and samples were assayed for acid–base parameters at approximately 1 h after collection.
On day 71, all animals were slaughtered. After 8 h without feed, lambs were weighed, stunned, slaughtered by exsanguination from the jugular vein, eviscerated, and skinned. In 4 representative lambs of each group, the rumen was removed, its content was mixed, and then it was straightened through 4 layers of cheesecloth. Then, pH was measured, 0.8 mL of fluid were added to 0.5 mL of deproteinizing solution (20 g/L metaphosphoric acid and 0.6 g/L of crotonic acid) for volatile fatty acid (VFA) determination, and 4 mL were added to 20 µL of 20% H
2SO
4 (vol/vol) for ammonia-N analysis. The rest of ruminal fluid was used for the in vitro trial that is described in
Section 2.5.
2.3. Carcass and Meat Characteristics
Carcass weight was recorded immediately after slaughter and after refrigeration at 4 °C for 24 h (cold carcass weight—CCW) in order to calculate chilling losses and dressing percentage. At 24 h, pH was measured on the longissimus thoracis muscle at the level of the 6th rib using a pH meter equipped with a penetrating electrode and a temperature probe (Metrohm, Switzerland). Subcutaneous fat color was determined using the CIELAB system and by calculating the chroma and hue values from the L* (lightness), a* (redness), and b* (yellowness) coordinates [
21]. Color measurements were carried out in duplicate on the subcutaneous fat in the lumbar region near the tail base using a Minolta CM-2002 chroma meter (Konica-Minolta Sensing, Osaka, Japan) operating with the D65 illuminant, SCI mode, with a 10° visual angle, an 11 mm aperture for illumination, and an 8 mm aperture for measurement. Carcass internal length (L), chest width (Th), and pelvic limb length (F) were measured, and the compactness index of carcass (CIC—CCW/carcass internal length) was calculated [
22].
The left sides of the carcasses were divided into commercial cuts (legs, foreribs, loin, shoulder, breast, neck, and tail), as described by Colomer-Rocher et al. [
23], and these were weighed. After jointing, a transversal cut of the loin just after the 13th rib was carried out, and subcutaneous fat depth and l. lumborum width and depth were measured with a caliper [
23]. Afterwards, the l. thoracis and l. lumborum muscles were removed from the corresponding joints for each half carcass and weighed together, and the weight percentage of the muscles in the half-carcass was calculated as 100 × (weight of l. thoracis and l. lumborum muscles/sum of the commercial cut weights). Leg tissue composition was determined by following the method of Fisher and de Boer [
24].
Two 2.5 cm thick slices from the distal end of each l. thoracis muscle were cut and placed on a polypropylene tray, which was wrapped with polyvinylchloride cling film and refrigerated (4 °C) in darkness. At days 0 (2 h after storage), 5, and 10 of storage, the tray was unwrapped, the slices were weighed, and the color was determined in duplicate on the upper surface of the slices following the procedure previously described for the color of subcutaneous fat. Furthermore, in order to assess meat color stability during refrigerated storage, the following ratios of reflectance were calculated: 630 and 580 nm (R
630/R
580), 610 and 525 nm (R
610/R
525), and 572 nm and 525 nm (R
572/R
525) [
25]. A third 2.5-cm thick slice was cut from the distal end of the l. thoracis muscle, and this slice was packaged under vacuum, cooked in water at 80 °C for 40 min, and cooled. Half of the slice was immediately analyzed for thiobarbituric acid reactive substances (TBARS), and the other half was wrapped into polyvinylchloride cling film, refrigerated (4 °C) for 48 h, and then analyzed for TBARS. The rest of the l. thoracis was used for chemical (proximate composition) analysis.
Cooking losses and meat hardness (shear force) were determined using the muscle l. lumborum, both at 24 h postmortem (day 0) and after 5 additional days of refrigerated (4 °C) storage. The muscle from the half carcass was transversely cut to obtain two similar portions (distal and caudal) that were randomly assigned to the day 0 or day 5 groups. The day 0 muscle portions were immediately weighed, vacuum-packaged and cooked in a water bath at 80 °C for 40 min, then cooled with tap water, removed from the packaging bag, dried on its surface with filter paper, and weighed again. Water loss was then determined as the difference in weight between cooked and raw meat, and it is expressed as percentage of raw meat. The day 5 portions were weighed and refrigerated for 5 days under the same conditions as described for the l. thoracis (color determination) before being weighed, vacuum-packaged, cooked, cooled, and weighed again. Prism-shaped (3 × 1 × 1 cm) samples from the cooked muscle portions (3–4 prisms per portion) were obtained and subjected to Warner–Bratzler shear force analysis according to Santos et al. [
26].
2.4. In Vitro Fermentation
An in vitro trial was carried out to compare rumen fermentation pattern using starch and neutral detergent fiber (NDF—obtained from barley straw) as substrates. Rumen fluid was mixed with the culture medium, as described by Goering and Van Soest [
27], in a 1:4 proportion. Incubations were performed in 120 mL serum bottles in which 300 mg of DM substrate were weighed. Rumen fluid from each lamb was used as a separate inoculum. Two bottles per substrate and animal and blanks were included in the incubation trial. After anaerobically dispensing 30 mL of the diluted rumen fluid, each bottle was sealed with rubber stoppers and aluminum seals, and all the bottles were placed in an incubator at 39 °C. At the end of the 24 h incubation, the total gas production was determined using a pressure transducer (Delta Ohm DTP704-2BGI, Herter Instrument SL, Barcelona, Spain) and a calibrated syringe, bottles were swirled in ice to stop fermentation, and then the bottles were opened to measure pH in the incubation medium. Samples for ammonia-N and VFA analysis were taken as described for samples taken at slaughter.
2.5. Protein Degradability
Protein degradability was determined in sacco using three ruminally fistulated Assaf ewes (68 ± 3.7 kg) that were fed the HP diet ad libitum. Nylon bags (125 × 100 mm size with 45 µm-sized mesh), containing 7 g of the corresponding diet that were ground to pass a 4 mm screen, were incubated in the rumen for 2, 6, 12, 24, and 48 h. Three bags (one per animal) were used for each time interval and diet. After incubation, bags were removed, rinsed slightly with tap water, and frozen for at least 5 days at −30 °C. Finally, all bags, including 2 extra bags per diet to estimate disappearance at zero time, were machine-rinsed using a cold water program for 30 min. Dried bags were weighed, residues were ground to pass a 1-mm screen, and the CP content of each was determined.
2.6. Chemical Analysis
The DM content of feeds, refusals, and in sacco residue samples were determined according to ISO 6496:1999 [
28]. Feed samples were analyzed for ash (ISO 5984:2002, [
29]) and protein (ISO 5983:2009, [
30]). Crude fat was determined by the Ankom fiber bag technique [
31]. aNDFom (neutral detergent fiber expressed without residual ash) and ADFom (acid detergent fiber expressed without residual ash) were analyzed as described by Van Soest et al. [
32] using a fiber analyzer (Ankom Technology Corp., Macedon, NY, USA). Sodium sulfite and heat-stable amylase were used in the analysis of aNDFom. Ruminal fluid samples were analyzed for ammonia and VFA concentration, as previously described by Carro et al. [
33].
Muscle samples were trimmed to eliminate connective tissue, homogenized, and freeze-dried to determine dry matter content. Afterwards, l. thoracis samples were analyzed for ash (AOAC official method 920.153), CP (AOAC official method 981.10), and fat (AOAC official method 960.39) contents. TBARS were determined following the method of Nam and Ahn (2003). Blood samples were assayed in a VetStat blood gas and electrolytes analyzer (Idexx, Barcelona, Spain) for pH, CO2 pressure (pCO2), total CO2 (tCO2), bicarbonate, and other parameters indicative of acid–base status, such as anion gap, Na, K, and Cl concentrations. Frozen plasma samples were defrosted overnight at 4 °C, and aspartate aminotransferase (AST), alanine aminotransferase (ALT), total cholesterol, glucose, triglycerides, lactate, calcium, phosphorus, urea, creatinine, total protein, and albumin contents were determined with an ILAB 650 biochemical profile autoanalyzer (Instrumentation Laboratory, Lexington, MA, USA).
2.7. Calculations
Protein degradability (dg) values were fitted with the model described by Orskov and McDonald [
34]:
. The potential rumen-degradable protein (RDP) content of the experimental diets was calculated from the values estimated from the model (
a + b) for each feed. The effective rumen degradation of CP was calculated following the method of AFRC [
16] while assuming a rumen outflow rate (
kp) of 0.083 h
−1.
The ADG (g/d) was estimated as the regression coefficient (slope) of body weight against time, and the feed-to-gain ratio was obtained by dividing the dry matter intake (DMI) by the estimated ADG. Residual feed intake was calculated as described by Koch et al. [
35].
2.8. Statistical Analysis
Degradability parameters a, b, and c were estimated by an iterative least squares procedure using the NLIN procedure of SAS (SAS Inst. Inc., Cary, NC, USA).
The data of the feed intake, ADG, feed efficiency, carcass characteristics, in vivo rumen fermentation, and tissue and chemical composition of the meat were analyzed using the MIXED procedure of SAS (SAS Inst. Inc., USA). The model included the fixed effect of the diet and the animal nested within the diet as residual error. The initial BW was also included as a covariate in the model, but it was removed because the effect was mainly non-significant (p > 0.05). Orthogonal polynomial contrasts were used for linear and quadratic trend comparisons.
Data from in vitro trial were averaged before statistical analysis (two values per animal), and they were analyzed with the above-described statistical model.
Data from blood gases and biochemical parameters, texture, cooking losses, and color of the meat were analyzed as a repeated measures model using the MIXED procedure of SAS. This procedure included the fixed effects of the diet, the experimental day, and their interactions, and the random effect of animal nested to the diet was used as the error to test the diet effect. The mean square of the day × animal (diet) interaction was used as the residual error to test the effects of the day and the interaction between the diet and the day. Different covariance matrixes were evaluated on the basis of Schwarz’s Bayesian information criteria. Plasma values at day 0 were used as a covariate, and they were removed from the model when their effect was not significant (p > 0.05). The level of significance was determined at p < 0.05, and means were separated using the least significant difference procedure.
3. Results
3.1. Feed Intake, Animal Performance and Fermentation Pattern
The DMI was not affected (
p > 0.05) by dietary protein level, although when linear and quadratic effects were evaluated, the former was significant (
p = 0.0422). The average daily gain and the feed-to-gain ratio increased (
p < 0.01) and decreased (
p < 0.01), respectively, with increasing levels of dietary protein. Nevertheless, the residual feed intake was unaffected by dietary treatments (
p > 0.05,
Table 2).
Table 3 shows the values of the in vivo and in vitro ruminal fermentation parameters. There were no differences (
p > 0.05) among dietary treatments, either in the pH and ammonia concentration or in the total VFA concentration and the proportions of the individual VFAs in the ruminal fluid samples taken at slaughter. However, the HP group showed higher (
p < 0.05) values of total VFA concentration and lower (
p < 0.051) values of acetate proportion than the LP and MP groups when NDF from straw was used as the substrate in the in vitro fermentation assay. Only the ammonia concentration and the proportion of valerate plus caproate were significantly (
p < 0.05) affected by dietary treatments when starch was used as the substrate, the values being higher in the HP group than in the LP group.
3.2. Blood Acid–Base Status and Metabolic Profile
Table 4 shows the values of parameters related to the acid–base status of the lambs, as well as their plasma biochemical profile. None of the parameters related to acid–base status were affected (
p > 0.05) by dietary treatments. However, sampling day had a significant effect (
p < 0.05) on blood tCO
2, HCO
3−, Na, K, and Cl concentration, the lowest values being observed at the end of the experimental period (day 70), and on anion gap (
p < 0.05), with highest values on day 70.
The plasma concentrations of protein, albumin, glucose, and Ca were affected (p < 0.05) by dietary treatments, with LP lambs showing lower values than HP lambs. Lambs from the MP group showed intermediate values that were significantly (p < 0.05) different from LP values for albumin and calcium. The LP and MP groups had the lowest and highest urea concentration (p < 0.05), respectively, with HP lambs showing intermediate values. Concerning the sampling day, it had effect on all the parameters except for AST and lactate. The values were higher (p < 0.05) on day 35 than on day 70 for ALT and triglycerides, and the opposite was observed for the rest of the parameters.
3.3. Carcass and Meat Characteristics
The mean values of the carcass characteristics are displayed in
Table 5. The effect of dietary protein level was not significant (
p > 0.05) for any of the recorded quality traits, although a linear effect was observed for CCW (
p linear = 0.0402).
The effect of dietary treatment by storage time interaction was not significant (
p > 0.05) for any of the meat quality characteristics, and, hence, only the mean values for main factors are shown in
Table 6. Dietary protein did not affect muscle composition, color, cooking losses, hardness, or lipid oxidative stability, as assessed by TBARS determination in cooked meat. Storage time in raw meat affected the color values related with meat discoloration, i.e., a* and a*/b* and the three reflectance ratios [
25]. In cooked meat, storage time increased cooking losses and decreased shear force. The TBARS of meat just after cooking were below 0.1 mg of malondialdehyde per kg of meat; therefore, the increments in TBARS concentration during the two-day refrigerated storage were equivalent to the TBARS levels shown by the stored meat.
3.4. Feeding Costs
The feeding costs per kg of ADG were 1.90, 1.98, and 2.14 €/kg for the HP, MP, and LP groups, respectively. However, the feeding cost decreased as protein content decreased when it was expressed in relation to CCW (1.51, 1.48, and 1.42 €/kg CCW for the HP, MP, and LP groups, respectively).