The perirenal fat from sixteen male Tudanca-breed calves was used for a volatile compound analysis. The animals were reared following two different production systems, a semi-extensive suckling system and an intensive system, on the experimental farm “Finca Aranda”, Cóbreces, northern Spain, with an elevation of about 84 m above sea level.

In the semi-extensive suckling system production (SE), eight calves, born in January, were maintained on pasture with their mothers until slaughter at 10 months old (163±12 kg of body weight). These animals were reared in a rotational grazing system in an area of 4.9 ha divided into three paddocks and supplemented with crushed barley ad libitum from six months of age onwards. The botanical composition of the grassland was estimated by recording at 600 points, located 1 m apart, on two transects per paddock. The botanical composition and the percentage of the main species with regard to total species (in brackets) was as follows: monocotyledons, Agrostis capillaris (31), Lolium perenne (19.2), Holcus lanatus (17.8), other (4.8); and dicotyledons, Trifolium repens (15.6), Trifolium pratense (2.5), Lotus corniculatus (2.4), Plantago lanceolata (2.0).

Animals reared under the intensive system were born in March. They were weaned at 5 months of age and then allocated to a feedlot pen, where they were fed ad libitum with a commercial concentrate (15.5% crude protein, 2.5% crude fat, 7.0% crude fiber and 7% ash; composed of corn, wheat, corn distillers grains with solubles, wheat middling, wheat bran, decorticated soybean meal, sugar beet pulp, sunflower meal, sugarcane molasses, calcium carbonate, sodium bicarbonate, sodium chloride and monocalcium phosphate) and rye grass (Lolium perenne) silage (pH 4.14, 27.5% dry matter, 3.0% crude fat and 9.1% ashes) until slaughtered at 9 months of age (144±5 kg of body weight).

All animals were transported to a commercial slaughterhouse. The slaughter took place immediately after arrival and was performed according to the European legislation on animal welfare [Council Regulation (EC) No. 1099/2009]. Total perirenal fat of each carcass was taken 45 minutes post-slaughter, placed in individual polyethylene bags and maintained under refrigeration for approximately 24 hours. Samples of 100 g were then wrapped in aluminum foil, sealed in polyethylene bags with a vacuum packaging machine and stored frozen (−80 °C) until analysis (up to three months). Before analysis, samples were thawed at 4 °C overnight and then homogenized in a food processor.

The volatile compounds of samples were extracted using the simultaneous distillation-extraction technique and then analyzed using gas-chromatography coupled mass spectrometry (GC/MS). For the extraction, 40 g of homogenized fat were added to 200 mL of ultrapure deionized water in a flask (500 mL). Volatile constituents were extracted for 3 hours in a Likens-Nikerson apparatus (J&W scientific, Folsom, CA, USA). The sample flask, with the fat, was immersed in a glycerin bath maintained at 150 °C, and the solvent flask (250 mL), containing 50 mL of diethyl ether, was immersed in a water bath at 45 °C. The distillate was concentrated to 1 mL using a Makro Kuderna-Danish concentrator (Sigma-Aldrich, St. Louis, MO, USA) in a water bath at 45 °C. After that, anhydrous sodium sulfate was added to remove traces of water and the concentrate was transferred to chromatographic vials.

Volatile compounds were analyzed in duplicate using a GC 7890A equipment coupled to an MS 5975C detector (Agilent System Zwingen, Switzerland). One μl of the concentrate was injected into the injection port operating at 260 °C in the split mode (5:1 split ratio). Compounds were separated using a DB-5MS column (60 m×250 μM×0.25 μM; J&W Scientific) and detected as described by Vieira et al. (2012). Briefly, helium was used as carrier gas at a constant flow rate of 1.5 mL·min⁻¹. After injection, the oven was kept at 35 °C for 1 min, heated at 10 °C·min⁻¹ to 50 °C, then the temperature was raised at 4 °C·min⁻¹ to 200 °C, and afterwards at 50 °C·min⁻¹ to 250 °C, which was kept for 11 min. The temperature of the transfer line and source were both set at 250 °C. The mass spectrometer operated in electron impact mode with an electron energy of 70 eV and an emission current of 50 μA. Detection was carried out in full-scan mode, the scan range was 40–350 mass-to-charge and the scan rate 3.94 scans/s. Compounds were identified by comparing their mass spectra with those contained in the NIST/EPA/NIH mass spectral database together with personal interpretation. Moreover, a series of n-alkanes (Hydrocarbons/C5-C20; Sigma-Aldrich) was used to calculate the experimental linear retention indexes (LRI) for each volatile and, whenever possible, compound identities were confirmed by comparison of their experimental LRI with those from the literature.

Volatiles were extracted from the fat samples in duplicate, based on the method described by Machiels and Istasse (2003) with some modifications. A 2-g homogenized fat aliquot was placed in a 15-mL vial which was sealed with a cap with a silicone/polytetrafluoroethylene septa (Agilent Technologies). The vial was then sonicated to equilibrium in a 200-W ultrasonic water bath (JP Selecta, Barcelona, Spain) for 20 min at 50 °C. Ultrasounds were then switched off and a 75 μm carboxen/polydimethylsiloxane-coated fused silica SPME fiber (SPME; Supelco, Bellefonte, PA, USA) was exposed to the head-space of the vial with the sample for 50 min at 50 °C. Afterwards, the compounds adsorbed by the SPME fiber were desorbed in the gas chromatograph injection port for 2 min at 260 °C in the splitless mode.

The chromatographic conditions regarding column and oven were the same as those described above. The detection was carried out with the detector operating in selected ion monitoring (SIM) mode to increase the sensitivity in the detection of terpenes. Ions m/z 93 and 136 were monitored for monoterpenes and ions m/z 93, 136, 161, 189 and 204 for sesquiterpenes (Viallon et al., 2000).

Presumptive terpenes were first detected from the chromatogram peaks by spectral interpretation, i.e., considered as presumptive (detected, although unidentified) monoterpenes or sesquitepenes in the compounds associated to peaks showing the selected monitored ions in appropriate ratios (expected ratios for those compounds). Furthermore, when possible, presumptive terpenes were identified by comparing and contrasting the retention times and spectral data of the previously identified terpenes (identified from the simultaneous distillation-extraction analysis of volatiles) with those from the present SPME-extraction analysis.

A single factor analysis of variance (one-way ANOVA using F distribution) was carried out in order to determine the effect of the production system on the volatile composition. The production system was the factor, and fat samples from each animal were the experimental units. Moreover, a principal component (PC) analysis was also carried out. In this PC analysis model, only the content of the terpenes showing significant differences in the ANOVA analysis (P<0.05) were considered as variables. Analyses were performed using the STATISTICA for Windows software (Release 6.0; StatSoft, Tulsa, OK, USA).

Sampling of the adipose tissue from calf carcasses was considered more suitable than sampling of muscle tissue for the purpose of this study, i.e., the first represents a less destructive sampling and is richer in terpenoids. Furthermore, perirenal fat was selected among different fat depots following the recommendations of Serrano et al. (2011). Volatile compounds of perirenal fat samples extracted with the Likens-Nickerson technique are shown in Table 1. They were classified into ten chemical families: aliphatic-alicyclic hydrocarbons (7 compounds), aliphatic aldehydes (13), aliphatic ketones (7), aliphatic alcohols (1), aliphatic acids (2), esters (1), furans (1), benzene compounds (16), sulfur compounds (1), and terpenoids (9). Moreover, four peaks could not be identified (unknown compounds). The chemical families showing the highest concentrations were, in order of abundance, aliphatic-alicyclic hydrocarbons, aliphatic aldehydes, terpenoids and aliphatic ketones.

The predominance of aliphatic hydrocarbons and aldehydes in fat agrees with previous studies into beef fat volatiles (Watanabe et al., 2008; Watkins et al., 2012). The majority of the hydrocarbons, aldehydes and ketones detected could be considered as fatty acid degradation/oxidation products (Frankel, 1982; Mottram, 1998; Narváez-Rivas et al., 2014).

Aldehyde levels were lower in fat from the semi-extensive production system (SE-fat) than in fat from the intensive production system (I-fat). Statistical differences were found for the levels of octanal, 2-octenal and 2,4-decadienal (P<0.05), and statistical trends for the levels of decanal, 2-decenal and the sum of aldehydes (P<0.1) (Table 1). Aliphatic aldehyde levels in cooked meat are indicative of lipid oxidation/degradation (Shahidi, 2001). Similar to aldehydes, the levels of 2-heptanone and 1-octen-3-one, which are compounds also derived from lipid oxidation/degradation (Resconi et al., 2012), were lower in the SE-fat. Therefore, SE-fat would have been more stable to lipid oxidation/degradation. This is supported by a previous study by Soto et al. (2014) using Longissimus dorsi muscle samples from the same animals, where six-day refrigerated-stored semi-extensive beef showed lower levels of thiobarbituric acid reactive substances than intensive beef (0.25 vs 0.80 mg of malonaldehyde per kg of beef). The higher stability of the lipids from pasture-fed cattle in comparison to those from intensively grain-fed cattle can been attributed to higher amounts of vitamin E in the former (Descalzo et al., 2005; Humada et al., 2014).

In contrast to that observed for the above-mentioned ketones, the levels of 2,3-octanedione were higher in SE-fat (P<0.05). Several studies have reported higher levels of 2,3-octanedione in meat or fat from ruminants fed on green-forage based diets than in those from ruminants fed conserved forage- and concentrate-based diets (Vasta and Priolo, 2006; Sivadier et al., 2010; Serrano et al., 2011). In these studies, it has been suggested that 2,3-octanedione could be used as an indicator of a green herbage diet in ruminant meat. Young et al. (1997) proposed that 2,3-octanedione could originate from the action of lipoxygenase (an enzyme abundant in green leafy tissue) on linoleic and linolenic acids. On the other hand, this volatile can also be originated from lipid oxidation in meat (Elmore et al., 2004).

Terpenoids, the third group in abundance, could originate from feeding since they are thought to be directly transferred from the diet, mainly grass, to animal tissues (Vasta and Priolo, 2006; Narváez-Rivas et al., 2012). In fact, different terpenes, together with other volatiles, have been used or proposed as biomarkers to distinguish between intensive and extensive production systems in pigs (Narváez-Rivas et al., 2008 and 2011) or ruminants (Sivadier et al., 2010; Serrano et al., 2011).

Most of the terpenoids identified in this study have been previously found in beef (Moon et al., 2004; Insausti et al., 2005; Serrano et al., 2011; Vasta et al., 2012). However, to our better knowledge, fenchene, eucalyptol and β-gurgujene have not been reported in previous studies on beef volatiles. Qualitative and quantitative variations among studies on the terpenoid levels in beef are expected because of their dependence not only on the extraction method but also on cattle production systems and the botanical composition of grass and forage which animals can graze on (Cornu et al., 2001; Prache et al., 2005; Vasta and Priolo, 2006).

In this study, terpenoids were more abundant (P<0.05) in the SE-fat than in the I-fat (Table 1). The levels of α-pinene and aromadendrene were higher (P<0.05) in SE-fat, and α-phellandrene, eucalyptol and α-gurjunene were found only in SE-fat. These results agree with previous studies (Vasta et al., 2012; Serrano et al., 2011; Vasta and Priolo, 2006) which reported that several specific terpenes, i.e., α-pinene and aromadendrene (found in this study) or α-terpinolene, β-copaene, β-caryophyllene, α-ylangene, germacrene D, α- and γ-cadinene (not found in this study) were present at higher concentrations in beef from pasture-fed animals than in those fed on conserved forages and/or high-concentrate diets. On the other hand, also in agreement with these studies, limonene and β-gurjunene were not related to a green herbage or concentrate-based diet.

Finally, statistical differences were also detected in two benzene compounds (Table 1) so as the levels of indole were higher in the I-fat and those of skatole were higher in SE-fat (P<0.05). Among these two compounds, skatole seems to be more markedly influenced by the ruminant dietary regimen (Vasta and Priolo, 2006). The skatole concentration in ruminant fat or meat has been found to be inversely related to the proportion of concentrate in the diet (Calkins and Hodgen, 2007; Vasta et al., 2012; Serrano et al., 2011). Skatole is a lyposoluble compound deriving from the degradation of tryptophan by a microbial action in the rumen (Young et al., 2003; Vasta and Priolo, 2006) and, according to Sheath et al. (2001), tryptophan degradation tends to be higher in forage-based diets as a consequence of a higher protein/non-fibrous carbohydrate ratio.

The terpenes of perirenal fat as analyzed using SPME extraction are shown in Table 2. Chromatograms from a sample of both SE-fat and I-fat are shown in Fig. 1. A total of 25 compounds (16 monoterpenes and 9 sesquiterpenes) were detected. Among them, only six (3 monoterpenes: α-pinene, α-phellandrene and limonene; and 3 sesquiterpenes: α-gurjunene, β-gurjunene and aromadendrene) could be positively identified by comparing their retention times and the spectral data obtained from the SPME-extraction analysis with those obtained from the previous Likens-Nickerson-extraction analysis. The rest of compounds were considered as presumptive terpenes. Regarding the identified terpenes, in agreement with the results presented in Table 1, aromadendrene levels were higher in SE-fat, and α-pinene, α-phellandrene and α-gurjunene were only detected in SE-fat.

No significant effect of production system on the sum and most of the presumptive monoterpenes (11 out of 13) was observed. However, contrary to the above-mentioned trend (higher levels of terpenes in meat from grazing cattle) monoterpenes X and XII showed lower values in the SE-fat than in the I-fat. On the other hand, the levels of sesquiterpenes were clearly higher in SE-fat than in I-fat: total presumptive sesquiterpenes and compound XVII showed significant differences, and compound XVI was not detected in I-fat whereas it was an important compound in SE-fat.

These findings regarding terpenes confirm those observed by Serrano et al. (2011), who found that among the monoterpenes detected in beef fat, only the levels of three of them were affected by grass feeding, and α-pinene and γ-terpinene were the only two monoterpenes which presented higher levels in the fat from grass-fed animals. In contrast, sesquiterpenes were more strongly affected by feeding, and thus would have higher potential than monoterpenes to be used as biomarkers of grass feeding. Moreover, it should be taken into account that the method used in this study for detecting presumptive terpenes could cause higher interference, and thus a higher occurrence of false positive identification rates, in monoterpene identification than in sesquiterpenes. This is due to the fact that only two ions were monitored (93 and 196) for monoterpene identification whereas five ions were monitored for sesquiterpene identification.

A PC analysis was performed from the results of the SPME analysis by including the detected terpenes showing significant differences in the ANOVA (Table 2; Fig. 2). The first and second PCs accounted for 69.1% and 14.7% of the variation, respectively. α-Phellandrene, α-pinene, α-gurjunene, aromadendrene and terpene XVI had the highest scores in the first PC, while the terpene X and XVII did in the second. The highest loadings on the PC1 were shown by those terpenes not detected in I-fat (detected only in SE-fat). As can be seen in the lower part of Fig. 2, fat samples from each system were separated using the PC1 at the level of the coordinate value of 1.0 approximately, so as I-fat samples are located in the left hand, SE-fat samples in the right hand; furthermore, I-fat samples were more dispersed than SE-fat samples. Regarding the PC2, most of I-fat samples were in the negative part of the PC2 axis, while most SE-fat samples were in the positive section. Results indicate that the levels of selected terpene obtained by the SPME-GC/MS method could be considered as potential pasture-feeding biomarkers in Tudanca calves.