Supplementation with Sunflower/Fish Oil-Containing Concentrates in a Grass-Based Beef Production System: Influence on Fatty Acid Composition, Gene Expression, Lipid and Colour Stability and Sensory Characteristics of Longissimus Muscle

Beef contains an array of conjugated linoleic acid (CLA) isomers for which positive effects have been reported in animal models of human disease. The objectives were to develop a CLA-enriched beef production system and to assess its quality. Sixty Spring-born heifers were housed in Autumn and offered unwilted grass silage and a barley/soyabean concentrate or wilted grass silage and a concentrate containing sunflower oil and fish oil. In May, both groups were offered either pasture for 22 weeks, restricted pasture and sunflower oil and fish oil for 22 weeks, or pasture for 11 weeks and restricted pasture and sunflower oil and fish oil for the final 11 weeks. The predominant CLA isomer in beef was cis9, trans11 representing on average, 80% total CLA. The modified winter diet followed by supplementation for 22 weeks resulted in beef that had a CLA concentration that was higher, at a comparable intramuscular fatty acid concentration, than previously reported. The lipid and colour stability (over 10 days in modified atmosphere packaging) and sensory characteristics were generally not negatively affected. There were minor effects on the expression of candidate genes involved in lipid metabolism. Consumption of this beef would make a substantial contribution to the quantity of CLA suggested to have a positive effect on consumer health.


Introduction
The interest of consumers in the relationships between diet and well-being has resulted in a growing preference for foods which are perceived as offering health benefits [1]. While beef is perceived to have a high proportion of saturated fatty acids (SFA), it also contains polyunsaturated fatty acids (PUFA), particularly omega-3 PUFA, which are beneficial to human health [2]. Beef also contains an array of conjugated linoleic acid (CLA) isomers, products of ruminal biohydrogenation of dietary PUFA [2] of which the cis9, trans11 isomer is most prominent. Positive effects of CLA are widely reported in animal models of human disease and supported, albeit not to the same extent, by human studies [3,4]. A desire to enhance the nutritional value has focused on strategies to increase the concentration of CLA in beef. The most effective strategy has been to modify the diet of the animal by changing phase of the experiment progressed, the allowance of the CLA concentrate was increased to ensure similar growth rate in both groups while maintaining the oil allowance at a maximum of 50 g/kg dietary dry matter (DM). The animals were also pre-assigned to post-winter treatments of (1) unsupplemented grazed grass from turnout for 22 weeks (G0), (2) unsupplemented grazed grass for 11 weeks and then a CLA supplement for 11 weeks (G11), or (3) grazed grass and a CLA supplement for 22 weeks (G22).
During the winter, the heifers were penned in groups of 3 or 4 according to winter diet (9 pens/diet) balanced for block as far as was practicable. Fresh feed was offered daily and the weight of any uneaten feed recorded and removed.
From early-May, the G0 and G11 animals rotationally grazed a predominantly Lolium perenne L. pasture as 4 groups of 10 animals/group (two groups from each winter diet). The daily DM allowance of 25 g kg −1 BW was achieved by measuring pre-grazing grass mass using a rising plate meter (Filips Folding Plate Meter, Jenquip, New Zealand) and calculating the amount of grass available. The area required to supply the grass allowance (without access to the previous grazing area) was then calculated. Animals were offered a fresh allowance every 2 to 3 days or more frequently depending on rainfall. The two groups of G22 animals separately grazed a smaller area to compensate for the high energy concentrates, and were offered a daily concentrate allowance of 2.5/425 kg bodyweight. The concentrate initially contained per kg, 133 g sunflower oil, 67 g fish oil (derived from a mix of mackerel and herring oil), 50 g cane molasses, 725 g pollard and 25 g mineral + vitamin (20,000 IU vitamin E/kg) mixture. Due to slow consumption, after 1 month 200 g of pollard was replaced by rolled barley and by a further 100 g 10 days later. The grass allowance was adjusted according to growth of the animals to ensure a similar rate of carcass growth. The concentrates were offered individually using an auto-locking feeding trailer and uneaten feed removed and weighed. Samples of grass and concentrates were taken daily and twice weekly, respectively. All samples were stored at −20 • C for chemical and fatty acid analysis.

Slaughter and Sample Collection
Animals were slaughtered at Meadow Meats Ltd., Rathdowney, County Laois, Ireland. The animals were weighed the day before slaughter and slaughtering commenced 45 min after arrival at the abattoir. All slaughter and dressing procedures complied with Regulations (EC) No. 1099/2009 and No. 853/2004 and electrical stimulation was not applied. Post slaughter, carcasses were weighed and graded for conformation and fatness. Approximately 30 min post-mortem, a sample (50 g approx.) of longissimus muscle (LM) and subcutaneous adipose tissue was removed from each carcass from above the 11th and 12th rib for subsequent gene expression analysis. The samples were dissected aseptically into smaller pieces and stored in RNALater™, (Ambion Ltd., Cambridge, UK) for 24 h. Subsequently, the RNALater™ was removed and the samples were stored at −80 • C. Carcasses were then placed in a chill.
At 24 h post-mortem, the right LM from the 10th rib to the posterior end (3 rib striploin commercial cut,) was excised from each carcass. The muscle was vacuum packaged (Webomatic ® vacuum-packaging systems Super Vax, ThyssenKrupp Schulte GmbH, Düsseldorf, Germany) and transferred to Teagasc Food Research Centre, Ashtown, Dublin and stored at 2 • C. At 48 h post-mortem, pH was measured at the 10th rib area by making a scalpel incision approximately 2 cm into the LM and inserting a pH electrode (EC-2010-11, Reflex Sensors, Ltd., Dublin, Ireland) connected to a portable pH meter (Model No. 210A, Thermo Electron Corp., Orion Products, Beverly MA, USA) set to record at 5 • C. The pH electrode was calibrated using buffers of pH 7.00 and 4.00 and rinsed between measurements. A section of LM (15 cm) was then vacuum packaged, aged for 14 days (2 • C) and stored at −20 • C pending sensory assessment. A further section (20 cm) was removed, vacuum packaged and aged for a further 19 days (21 days aging in total) for colour stability measurement. Finally, two steaks (25 mm thickness) were removed, vacuum packaged and stored at −20 • C pending chemical and fatty acid analysis, respectively.

Chemical Analysis
Fat was extracted from 2 g homogenised LM, separated into the neutral lipid (NL) and polar lipid (PL) fractions and fatty acids methylated as previously described [14]. Fatty acid methyl esters (FAME) were analysed using a Varian 3500 GLC (Varian, Harbor City, CA, USA) and a 100 m CP-Sil 88 column (100 m × 0.25 mm i.d., 0.2 µm film thickness, Supelco, Bellefonte, PA, USA). Hydrogen was the carrier gas and GC conditions were as described in [15]. Individual FAME were identified by retention time with reference to external standards (Supelco 37 component FAME Mix, Supelco Inc., Bellefonte, PA, USA). Individual standards from Matreya (Matreya Inc., Pleasant Gap, PA, USA) were used for identification of FAME not contained in the mix. Fatty acids for which no commercial standards were available had been identified in identical chromatographic conditions as in the present study, by Shingfield et al. [16] using 4,4-dimethyloxazoline derivatives and analysed by GC-MS. The appropriate retention times were used to identify these fatty acids in the present analysis. Individual FAME were quantified by using C23:0 as the internal standard.
For CLA methyl ester analysis, FAME were evaporated under nitrogen, dissolved in heptane and analysed by HPLC using four silver-impregnated silica columns (ChromSpher 5 lipids, 250 × 4.6 mm; 5 µm particle size, Varian Ltd., Walton-on-Thames, UK) coupled in series and 0.1% (v/v) acetonitrile in heptane as the mobile phase [17]. Isomers were identified using an authentic CLA methyl ester standard (O-5632, Sigma-Aldrich St. Louis, MO, USA) and chemically synthesised trans-9, cis11 CLA [17]. Identification was verified by cross-referencing with the elution order reported in the literature [18] using cis9, trans11 CLA as a landmark isomer. Vitamin E and TBARS concentrations were measured as described in [19].
The general composition of feeds was determined as previously described [20]. The fatty acid composition of feeds was determined using the procedure described in [21] with the minor modification that toluene was used instead of benzene.

RNA Purification, cDNA Synthesis and Quantitative Polymerase Chain Reactions (QPCR)
RNA was extracted and purified from 100 mg of tissue in Tri-Reagent (Sigma-Aldrich, St. Louis, MO, USA), followed by a DNase step (Promega, Madison, WI, USA). The total RNA was quantified and assessed for purity on a NanoDrop Spectrophotometer ND1000 (Thermo Scientific, Wilmington, DE, USA). All cDNA synthesis was carried out using 1 µg of total RNA using Superscript™ III First-Strand Synthesis kit for RT-PCR (Invitrogen Life Technologies, Carlsbad, CA, USA) and random hexamers according to the manufacturers' instructions.
Normalised relative quantities were obtained using the software, qbase PLUS (Biogazelle, Ghent, Belgium) from stable reference genes; GAPDH and RPL0. These genes were confirmed to have M values (<1.5) as calculated by the GeNorm algorithm within qbase PLUS.

Sensory Assessment
Samples were arranged according to the blocking structure of the experimental design. The day before sensory assessment, loin sections were thawed in a refrigerator set at 1 • C. On the morning of sensory assessment, loin sections were removed from their packs and steaks 1.9 cm thick were cut. Steaks were cooked under a conventional grill, turning every three minutes, until the internal temperature of the muscle reached 74 • C as measured by a thermocouple probe (Testo Limited, Alton, UK). Ten samples, (2 cm × 2 cm × 1.9 cm) were then cut from the approximate centre of the steaks avoiding incursions of connective tissue where present, wrapped in pre-labelled foils, placed in a heated incubator at 65 • C and served hot to a 10-person trained professional taste panel, using the same people for the duration of each experiment. Assessors scored individual flavours using 0-100 mm unstructured intensity line scales where 0 = nil and 100 = extreme. Assessments took place in a purpose-built panel room illuminated by red light. Each booth contained a computer screen and optical mouse as part of the computerised sensory system (Fizz, Version 2.20 h, Biosystemes, Couternon, France), for direct entry of sensory responses. At each session, assessors tasted 6 samples of loin steaks in balanced order such that first order carry over effects were reduced [22].

Data Calculations and Statistical Analysis
Daily grass DM intake was estimated based on the growth of the animals and their associated energy requirement [23]. The fatty acid concentrations in total muscle were calculated as the sum of the fatty acid concentrations in the NL and PL fractions. Selected nutritionally relevant fatty acid indexes were calculated, according to Ulbricht and Southgate [24]. Data were analysed according to a split plot design using Genstat (19th edition, VSN International, Hemel Hempstead, UK). The model had block and winter ration in the main plot and summer ration and all winter by summer interactions in the subplot. The effect of duration of concentrate supplementation was examined using orthogonal polynomials. For colour data relating to retail display, the design was a split-split plot. The split plot was as described above with time of display and all time-related interactions in the sub subplot. Gene expression data that were not normally distributed were transformed using the appropriate lambda function determined through the Box Cox transformation (ABOXCOX procedure in Genstat) and analysed as a split-plot design as described above. Multiple analysis of variance using SAS was used to calculate partial correlation coefficients (p), from the error sum of squares and cross products (SSCP) matrix, between selected concentrations of fatty acids and transformed gene expression data in muscle. (Table 1) The wilted silage tended to have a higher DM and a lower oil concentration but the chemical composition was generally similar to that of the unwilted silage. The grass tended to have a higher digestibility and tended to have the highest proportion of C18:3 compared to the silages. The oil-rich concentrates averaged 233 g oil/kg DM which had a higher proportion of C18:2 than the standard concentrate. The inclusion of fish oil was reflected in the proportions of C20:5 and C22:6 detected in the oil-rich concentrates compared to the standard concentrate. All stated differences in this and subsequent sections were statistically significant (p < 0.05).
Increasing the duration of concentrate supplementation linearly increased the PUFA: SFA (also quadratic) and omega-6: omega-3 PUFA ratios and linearly decreased the concentrations of C16:1trans12 (quadratic only) and CLA trans10,cis12 and the thrombogenic index (also quadratic).
There was an interaction (linear) between the winter ration and the duration of concentrate supplementation such that the concentration of C16:1cis9 and C18:1trans13 decreased with the increase in duration of concentrate supplementation for the US diet but increased for the WO diet. For the concentration of C18:3cis9,cis12,cis15 (linolenic acid LNA), while the mean value was higher for the US compared to the WO winter ration and linearly decreased with the increase in the duration of concentrate supplementation, the decrease was greater for US ration.   (Table 4) The predominant CLA isomer was cis9, trans11 representing on average, 80% total CLA. The trans11, cis13 was the next prominent isomer, followed by trans 7, cis9 and trans11, trans13. Of those CLA isomers present at >1% of total CLA, the proportion of trans7, cis9 tended (p = 0.09) to be lower while the proportions of trans9, trans11 and trans11, trans13 were higher in LM from cattle offered the US ration during the winter. Increasing the duration of concentrate supplementation linearly increased the proportion of cis9, trans11 and trans7,cis9 and linearly decreased the proportion of trans11, cis13, trans9, trans11, trans11, trans13 and trans12, trans14. There was an interaction between the winter ration and concentrate supplementation (linear) for trans8, cis10 such that the decrease with increase in the duration of concentrate supplementation was greater in LM from animals offered the WO ration during the winter.  1 sed. is the standard error of the difference for the W × D interaction with n = 10/group; L, Q are linear and quadratic effects of duration of supplementation, respectively, + = p < 0.1, * = p < 0.05, ** = p < 0.01, *** = p < 0.001. (Table 5) Only those fatty acids detected at >0.1 g/100 g fatty acids (for all treatment means) are summarised in Table 5. The proportion of C16:0, C16:0iso, C16:1cis9, C17:0 and C18:1cis11 were higher in LM from cattle offered the US ration during the winter, while the proportions of C18:1trans11, C18:1trans10, C18:1trans12, C18:1trans9, C20:0, CLA cis9, trans11 and total CLA isomers were lower. Increasing the duration of concentrate supplementation linearly increased the proportions of C17:0iso + C16:1trans9 (also quadratic), C18:1trans11, C18:1cis16, C18:1trans10, C18:1trans12, C18:1trans9, C18:1trans16, C18:2 10,13, LA, C18:2trans11,cis15, C20:0, C20:4cis8,cis11,cis14,cis17 (quadratic only), C22:0, DHA (also quadratic),CLA cis9, trans11, total CLA isomers and PUFA (also quadratic).
There was an interaction between the winter ration and concentrate supplementation (linear) for the proportion of C16:1trans7 + trans8 such that it increased with the duration of concentrate supplementation in LM from animals offered the US but deceased with the duration of concentrate supplementation in LM from animals offered the WO ration during the winter. There was an interaction between winter ration and concentrate supplementation (linear) for the proportion of C20:1such that the increase with the duration of concentrate supplementation was greater in LM from animals offered the US ration during the winter.

Muscle Colour Stability and Sensory Characteristics
The TBARS values (high pH samples excluded) at the start of retail display averaged 0.35 mg malonaldehyde/kg meat and did not differ between treatments. After 10 days of display, the TBARS values were 1.20, 1.52, 1.74, 0.89, 1.32 and 0.96 mg malonaldehyde/kg meat (sed = 0.329) and did not differ between treatments.
Colour variables during retail display are shown in Figure 1. There were no effects of supplementation at pasture and no interactions between the pasture phase and the winter ration for any colour-related variables. The redness (Figure 1a), yellowness (Figure 1b), chroma ( Figure 1c) and R 630 -R 580 (Figure 1e) of LM decreased and hue ( Figure 1d) and percentage metmyoglobin (Figure 1f) increased during aerobic display. There was a winter ration by time of display interaction for all variables in Figure 1, which mainly reflected the differences at day 10 of display where LM from animals offered the oil-enhanced winter ration was more colour stable.
Muscle sensory characteristics are shown in Table 6. Muscle from cattle offered the US ration during the winter tended (p = 0.06) to be rated more abnormal than muscle from cattle offered the WO ration. Increasing the duration of concentrate supplementation linearly increased the ratings for juiciness, beef (also quadratic), greasy and overall liking and linearly decreased the ratings for fishy and cardboard (quadratic only). There was an interaction between winter ration and concentrate supplementation (quadratic) such that the rating for rancid was lower in LM from cattle offered the concentrate for 11 weeks after receiving the US ration in the winter but was higher in LM from cattle offered the concentrate for 11 weeks after receiving the WO ration in the winter.   1 sed is the standard error of the difference for the W × D interaction with n = 10/group; L, Q are linear and quadratic effects of duration of supplementation, respectively, * = p < 0.05, ** = p < 0.01. 2 0 = nil, 100 = extreme, 0-100 mm unstructured intensity line scale.    (Table 7) The expression of SREBP1 was lower and the expression of PPARγ tended (p = 0.067) to be lower in LM from cattle offered the US ration during the winter. Increasing the duration of concentrate supplementation linearly increased the expression of the SCAP and tended (quadratic, p = 0.098) to decrease the expression of the SREBP1.
There was an interaction between the winter ration and concentrate supplementation (linear) for the proportion of C18:1cis 11. Thus, while the proportion of C18:1cis11 was higher in LM from cattle offered the US rations during the winter, the increase with the duration of concentrate supplementation was greater in LM from those animals. For LA, the increase with the duration of concentrate supplementation was greater in LM from cattle offered the US ration during the winter. For C16:1trans 7 + trans8, increasing the duration of concentrate supplementation linearly increased the proportion in LM from cattle offered the US ration during the winter, but linearly decreased the proportion in LM from cattle offered the WO ration during the winter.

Context
Beef from grass-based production systems is appreciated by some consumers based in part on the perception of a superior fatty acid profile compared to beef produced in other systems [1,27]. Given the putative health benefits of CLA [4,28] and in particular the cis9, trans11 CLA isomer which is found predominantly in ruminant-derived foods, an increase in the CLA concentration would enhance the value of grass-fed beef further. Since dietary manipulation is the most effective strategy to increase the concentration of CLA in beef [29], all the dietary critical control points must be optimised in an enhanced beef production system. We choose one particular production system, Spring-born suckled heifers slaughtered from pasture in autumn at approximately 20 months of age [30] in which the diet of the animals is predominantly grass-based. The dietary critical control points for increasing the concentration of CLA in beef from these animals are the diet of the mother when the calves are suckling (not considered in this study), the post-weaning indoor diet before turnout to pasture and the pasture finishing phase. Since many of the studies reported to date have focused on enhancing the CLA in the finishing phase of cattle and few have considered a grazing scenario, our ambition was to optimise the latter two critical control points in this production system. The dietary manipulation in the indoor phase was based on Noci et al. [6,7] who demonstrated that sunflower oil inclusion in the concentrate supplement to grass silage-fed steers increased the CLA proportion in muscle and that wilted grass silage increased the concentration of CLA in muscle when compared to unwilted silage, respectively. The dietary manipulation in the pasture finishing phase was based on [8,9] who demonstrated that supplementing grazed grass with sunflower oil alone or with fish oil, respectively, also increased the concentration of CLA. The pathway of biohydrogenation of LA produces CLA directly but biohydrogenation of LA and LNA also includes C18:1trans11 as an intermediate (see [31]). Since most of the CLA found in milk fat resulted from desaturation of C18:1trans11 by the action of ∆9 desaturase [32], the increase in CLA concentration likely reflects an increase in C18:1trans11 accumulation in the rumen and subsequent tissue desaturation. Fish oil has been demonstrated to inhibit the terminal reaction in the biohydrogenation pathway in the rumen and to increase the outflow of C18:1trans11 to the small intestine [33]. In the present study, fish oil was used as a tool to manipulate biohydrogenation rather than to supply longer carbon chain PUFA, although a small proportion of EPA and DHA escaped rumen biohydrogenation and was deposited in tissue.
The feeding strategy was for the cattle to consume similar quantities of energy such that intramuscular fat (IMF) concentration would be similar across treatments and thus avoid confounding treatment effects with fatness [14]. Growth during the indoor phase was close to the target for this system (0.6 kg/day, [30]). Growth at pasture was higher for the supplemented groups reflecting the challenges in implementing the supplementation strategy particularly during a long grazing season. Nevertheless, the differences in the total lipid concentration between treatments only approached significance. It is recognised that if the supplemented grazing groups had unrestricted access to pasture these animals would likely have grown faster which would be closer to a more commercial production system. The implications of this in terms of the fatty acid composition of beef merit examination in a future study.

General Fatty Acid Composition
For individual and classes of fatty acids, presentation of the fatty acid data expressed as a proportion has merit and can allow a more complete comparison with the literature. From a product labelling and ultimately a marketing perspective, concentration data are more relevant for some variables while for others, proportional data are more relevant. Accordingly, the fatty acid data are presented in both forms. Thus, based on the SFA concentration in LM, beef from each production system could be labelled "low in saturated fat" i.e., SFA concentration < 1.5/100 g solid [34]. For MUFA, EU [35] states that "a claim that a food is high in monounsaturated fat may only be made when at least 45% of the fatty acids present" are monounsaturated. Based on the proportion of MUFA in LM, beef from all of the production systems in the present study would meet this claim. For PUFA, EU [34] states that "a claim that a food is high in polyunsaturated fat may only be made when at least 45% of the fatty acids present" are polyunsaturated. Despite the beneficial increase in PUFA observed due to supplementation at pasture, beef from none of the production systems in the present study would meet this claim.
With regard to ratios of fatty acid classes, there is a recommendation [36] on a desirable ratio of total PUFA: total SFA on a whole diet basis (>0.45), but it does not relate to individual foods. All LM in the present study was below this ratio. From a human nutrition perspective, LM from the supplemented groups has a more desirable PUFA: SFA ratio but the effect is small. Similarly, there is a recommendation [36] on a desirable ratio of total omega-6 PUFA: total omega-3 PUFA on a whole diet basis (<4). While the lower omega-6 PUFA: omega-3 PUFA ratio in LM from the un-supplemented groups may be viewed as positive for "Grass-Fed" beef, the difference is unlikely to be important in this regard.
The concentrations in 100 g of tissue for beef to be labelled as a "source" of omega-3 fatty acids are 300 mg LNA or 40 mg EPA + DHA [37]. In the present study, the highest concentration of LNA was 33 mg/100 g muscle (control indoor diet and no supplementation at pasture) reflecting the higher grass consumption by this group. The highest concentration of EPA + DHA was 12 mg/100 g muscle (oil-rich diet in winter and supplementation for 22 weeks at pasture) reflecting the consumption of fish oil, some of which clearly escaped biohydrogenation in the rumen. While either, on its own, might be viewed as positive for that treatment, and make a contribution to meeting human dietary recommendations, none of the beef in the present study could be labeled a "source" of omega-3 fatty acids as defined by EFSA [37].

Conjugated Linoleic Acid
Manipulating the composition of the indoor ration increased the concentration of total CLA isomers, albeit not significantly (52 vs. 58 mg/100 g muscle). When expressed on a proportional basis, thereby adjusting for the difference in the total fatty acid concentration mentioned above, the effect was statistically significant (1.77 vs. 2.05 g/100 g fatty acids) supporting the observations in [6,7]. When compared with the "control" production system, the concentration and proportion of the cis9, trans12 CLA isomer, the isomer most reported in the literature, was 2.4-and 2.2-fold higher in LM from the oil-enriched followed by long-term supplementation at pasture group (86 mg/100 g muscle or 2.5 g/100 g fatty acids). These values are higher than in most previous reports (reviewed by [5,38]). In two studies identified in these reviews where a higher concentration of CLA was reported (156 mg and 134 mg/100 g tissue), the IMF concentration was considerably higher (6.6/100 g and 10.5/100 g tissue, respectively) than in the present study. This reflected the preferential deposition of CLA in the triacylglycerol or NL fraction as demonstrated in the present study (below). As of yet, there is no reference intake for CLA. In a review of the literature, Siurana and Calsamiglia [39] concluded that with respect to human health, "an effective dose would be 0.8 g per day for the anti-carcinogenic effect, 0.6 g per day for the anti-atherosclerotic effect and 3.2 g per day for the reduction of body fat. For other effects, no specific dose has been recommended". If the contribution to the CLA concentration due to the desaturation of C18:1trans11 to CLAcis9, trans11 in human tissue, (20-25%, [40]) is considered, an average LM steak (200 g) from the oil-enriched followed by long-term supplementation at pasture group could supply approximately 0.3 g CLA. This would make a substantial contribution to the effective dose reported by Siurana and Calsamiglia [39].
While it is recognised that an array of CLA isomers arise from ruminal biohydrogenation of dietary lipids, there are relatively fewer reports on isomers other than cis9, trans11 and trans10, cis12 because they cannot be separated using conventional GC. Because of the potential bioactivity of other CLA isomers [31], and the nature of the supplements, it was considered important to measure their concentration under the dietary scenarios of the present study. Cis9, trans11 was the major isomer and trans11, cis13 and trans7, cis9 the second and third most abundant isomers in the control group in agreement with previous findings for cattle slaughtered from pasture-based production systems [41,42]. The main impact of the dietary modification in the present study was to enrich the LM with the cis9, trans11 isomer (77.4 to 84.8 g/100 g total CLA). While there was some re-ordering of the abundance of the isomers due to the dietary treatments imposed, the concentration of the next abundant isomer was <4.6/100 g total CLA and therefore unlikely to be of relevance to human health.

Trans Fatty Acids
While the ruminal biohydrogenation pathway was successfully modified to increase the flow of C18:1trans11 from the rumen in the present study, biohydrogenation of dietary PUFA by the rumen microbial system also results in a range of other cis-and trans C18:1isomers. The human health implications of the consumption of trans fatty acids, and in particular their origin, are of current interest [43]. Recent research on ruminally derived trans fatty acids, of which C18:1trans11 predominates (72% of detected C18:1trans isomers in LM from the oil-enriched followed by long-term supplementation at pasture group was C18:1trans11) suggests a positive or neutral effect on human health compared to the detrimental effects of industrially derived trans fatty acids which have a higher proportion of C18:1trans10 and a more diverse profile [44]. However, the duration and daily amount of ruminally derived trans fatty acid consumption required to cause significant effects on human health are still unclear. There is no reference intake value for C18:1trans11 currently.
Acknowledging the expense and additional management required to supplement cattle for a full grazing season, a shorter period of supplementation was also examined. With regard to the fatty acids of primary interest in this study, CLA and C18:1trans 11, the linear response with duration of supplementation indicates that the full grazing season was necessary to reach the maximum concentrations. The decision on whether a producer might implement this strategy therefore becomes a balance between the added cost of production and the premium achievable in the marketplace.
To explore the site of deposition of fatty acids, the extracted intramuscular lipids were separated into NL and PL fractions. Differences in total intramuscular fatty acid concentrations tend to mainly reflect differences in the size of the NL fraction as the size of the PL fraction is generally quite constant. The preferential incorporation of CLA and C18:1trans10 and C18:1trans11, in particular, into the NL fraction and the preferential incorporation of the longer carbon chain PUFA LA, LNA, EPA and DHA into the PL fraction was also reported by Moreno et al. [14]. This suggests that had a higher target carcass weight, with an associated increase in IMF deposition been chosen, the concentration of CLA in LM would have been even higher. With respect to CLA isomers, the profile was generally similar in the NL and PL fractions, with some evidence that the proportion of the trans11, trans13 and trans12, trans14 isomers are preferentially deposited in the NL fraction (Supplementary Table S2).

Gene Expression
Using a candidate approach, we examined genes directly involved in lipogenesis, i.e., FAS and LPL or in the regulation of fatty acid metabolism. Dietary lipids act as ligands for a range of receptors which in turn regulate genes coding for transporters and enzymes involved in lipid metabolism [12]. These can result in alterations in the concentration and/or profile of tissue lipids [12]. The ligands for PPARs (key fatty acid metabolic regulators and sensors) encompass a range of exogenous and endogenous lipids, including various fatty acids [45]. The general lack of effect of the production system modifications on the expression of these genes in muscle tissue and their relationships with the concentrations of fatty acids reflects, in part, the minor differences in the latter. Sterol regulatory element-binding proteins (SREBPs), membrane-bound transcription factors that are essential in the regulation of cholesterol, fatty acid and triglyceride biosynthesis and its chaperone SCAP, are essential for promoting nuclear translocation of SREBP1 and activation of FAS gene transcription. The decrease, albeit a quadratic response pattern, in SREBP1 due to supplementation at pasture is consistent with [46]. The opposite patterns for SCAP and SREBP1, an increase and decrease with duration of supplementation at pasture, respectively, indicate an uncoupling between these two regulatory elements. The negative correlations between SCAP and the concentration of EPA and fatty acid indices of relevance to human nutrition are a novel finding and suggest that downregulation of this gene could be a target if the objective was to increase those variables in muscle.
The gene coding for the desaturase enzyme that catalyses the conversion of C18:1trans11 to CLA cis9, trans11 and also C18:0 to C18:1cis 9 (SCD) was of particular interest since part of the dietary strategy sought to increase the supply of C18:1trans11 for subsequent tissue desaturation. While there was a trend towards a decrease in SCD activity based on the desaturase index (calculated from the fatty acid concentrations), due to supplementation at pasture, this was not reflected in SCD gene expression which supports the conclusion that the index is not a good proxy for enzyme activity or indeed gene expression [47]. In contrast, Waters et al. [46] reported a significant decrease in SCD gene expression upon dietary omega-3 PUFA intervention. The expression of the gene coding for the desaturase enzyme that catalyses the first and rate limiting step in the conversion of LA and LNA, FADS2, to highly longer carbon chain unsaturated fatty acids was similarly not affected by the modifications of the production system examined. Overall, the changes in gene expression and therefore their role, if any, in the observed fatty acid profile of muscle were rather modest.
Gene expression was also measured in the subcutaneous lipid as a proxy for the NL fraction for muscle (Supplementary Table S3). As with muscle, there was little effect on SCD expression. While there was an effect on the expression of PPARα, the quadratic response pattern indicates that the timing of sample collection is important in attempting to unravel the relationship between gene expression and the fatty acid profile, i.e., expression tended to be decreased after 11 weeks of supplementation and increased after 22 weeks. As with muscle, SCAP expression was increased by supplementation but only in adipose tissue from animals offered the control ration during the winter which corresponded with an increase in SREBP1 expression. The opposite pattern was observed in adipose tissue for animals offered the oil-rich diet during the winter. These data highlight the challenges in seeking to unravel the role of gene expression in a production system context.

Lipid Oxidation and Colour Stability
Dietary supplementation with PUFA and fish oil in particular can increase lipid oxidation during retail display of beef when compared with beef from un-supplemented cattle [11]. The scale of this effect is influenced by the concentration of long carbon chain PUFA, the concentration of vitamin E (and other antioxidants) and whether display is aerobic or in MAP. The TBARS values for muscle from all groups were below the 2 mg malonaldehyde/kg threshold value for the detection of rancidity in meat by consumers [48]. That the panellists could detect rancidity, albeit at a low level, reflects the training they received prior to asssessment of the beef from the present study. The lack of an effect on lipid oxidation in the present study likely reflects the high vitamin E supply from the oil-rich rations that maintained LM vitamin E concentration, since supplementation with oil-rich feeds without fortification, frequently results in depletion of vitamin E in muscle [49]. Colour stability can be stabilised when muscle has a vitamin E concentration of 3.0-3.5 mg/kg [50] as was the case in this study. CLA has been proposed to have antioxidant properties [51] which may also have contributed to the lipid stability. In support of this, there was relatively little effect on colour stability with the LM from the animals fed the oil-enriched winter ration being more stable after prolonged retail display. Since colour is an important influence on the purchasing decision of the consumer [52], these findings are positive from the perspective of marketing CLA-enhanced beef.

Sensory Characteristics
From a consumer perspective, deleterious effects on the sensory characteristics would diminish the benefits of nutritional enhancement of beef. The higher score for greasiness and beef flavour may be related to the differences in total fatty acid concentrations [53] while the lower score for "fishy" likely reflects the decrease in LNA concentration [48]. That panellists rated LM from the group offered the control winter ration and supplemented for 11 weeks highest for rancidity was unexpected in view of the TBARS values discussed above. Nevertheless, overall, there were relatively minor effects on the measured sensory characteristics of beef which can be viewed as a positive finding as confirmed in the hedonic score for overall liking. It is recognised that while this is an indication of preference by the panel, the assessors cannot be considered typical consumers because of the training they have received in meat assessment and this finding needs confirmation using untrained consumers.

Conclusions
Modification of the diet of cattle within a grass-based 20-month heifer beef production system resulted in beef that had a CLA concentration that was higher, at a comparable intramuscular fatty acid concentration, than previously reported. When also accounting for the conversion of the enhanced C18:1trans11 concentration, consumption of this beef would make a substantial contribution to the quantity of CLA suggested to have a positive effect on consumer health. That the lipid and colour stability and sensory characteristics of this beef were generally not negatively affected, is a positive result from the perspective of marketing such nutritionally enhanced beef. Since the concentration of both CLA and C18:1trans11 was linearly increased with the duration of supplementation, the decision for the beef producer on the implementation of the strategies explored in this study is based on the added cost of production and the premium achievable in the marketplace. Identification of the latter is an important subject for future research on the topic of CLA-enhanced beef.
Supplementary Materials: The following supporting information can be downloaded at https: //www.mdpi.com/article/10.3390/foods11244061/s1: Table S1: Panel of Bovine oligonucleotide primers used for real-time PCR. Table S2: The CLA isomer profile of intramuscular lipid fractions of the longissimus muscle of beef heifers. Table S3: Gene expression in subcutaneous adipose tissue of beef heifers.

Institutional Review Board Statement:
This study was carried out under licence from the Irish Government Department of Health and Children and with the approval of Teagasc, the Agricultural and Food Development Authority (RMIS 5409). All procedures used complied with national regulations concerning experimentation on farm animals.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.