Article pp.497-511 du Vol.23 n°4 (2003)

(1)

L’ALIMENTATION ET LA VIE

Analysis of the volatile fraction of lamb fat tissue:

influence of the type of feeding

Isabel Sebastián, Christine Viallon-Fernandez, Philippe Berge, Jean-Louis Berdagué

ABSTRACT

The volatile compounds of subcutaneous fat tissue of lambs were analysed by dynamic headspace sampling-high resolution gas chromatography-mass spectrometry (DHS-HRGC-MS). The animals were produced in six European countries according to local traditional rearing conditions and they were classified into two groups according to the predominant type of feed consumed (pasture or concentrate). The analysis showed that the pasture-fed lambs clearly differed from the concentrate-fed lambs. Fat samples from the pasture-fed lambs desorbed greater quantities of compounds of green leaf tissue origin such as long-chain alkanes, C7 aldehydes and 2,3-octanedione. The fat samples from the concentrate-fed lambs desorbed greater quantities of short-branched and non-branched acids and methyl ketones.

The present results suggest that 2,3-octanedione or long-chain alkanes, and 4-heptanone or 2-octanone in the lamb subcutaneous fat tissue are specific of feedings consisting predominantly of pasture and concentrate, respec-

Dans chaque numéro, cette rubrique met en avant un article traitant d’un des aspects de la nutrition, du rôle des technologies agroalimentaires sur la qualité des aliments jusqu’à la « cuisine

», en passant par les problèmes nutritionnels, la toxicologie alimentaire, et plus généralement les conséquences sur la santé des pratiques alimentaires. Les articles retenus sont soit des travaux de synthèse de haut niveau faisant le point sur une question, soit des publications originales ren- dant compte de travaux de recherche appliquée récents apportant un regard nouveau.

La Société scientifique d’hygiène alimentaire (SSHA), société savante créée en 1904 pour contri- buer à la diffusion des connaissances en nutrition et sécurité sanitaire, est aujourd’hui formée de deux départements : l’Institut supérieur de l’alimentation (ISA) développe des actions de formation, d’information et de conseil ; l’Institut supérieur d’hygiène alimentaire (ISHA) propose un catalogue complet d’analyses (composants nutritionnels, contaminants, analyse sensorielle, microbiologie…).

Les propositions d’articles, remarques et suggestions peuvent être envoyées à : Claude Bourgeois

SSHA

Rue du Chemin Blanc, BP 138, Champlan F-91163 Longjumeau cedex Tél. : + 33 (0)1 69 79 31 50

Fax : + 33 (0)1 64 48 82 49 http://www.ssha.asso.fr [email protected]

1. Laboratoire Flaveur, Station de Recherches sur la Viande, INRA, Centre de Clermont-Ferrand/Theix, 63122 Saint-Genès-Champanelle, France ; e-mail address: [email protected].

(2)

tively. These compounds could be used as markers of the feeding background of the animal.

Key words

lamb, feed traceability, fat tissue, volatile compound.

1 – INTRODUCTION

The characterisation of animal food products is a major market issue in the developped countries. Consumers are increasingly concerned by diverse aspects of the quality of the foodstuffs they buy, e.g. visual appearance, hygiene, sensory properties, nutritional value, convenience. Most quality traits of raw meat cannot, or simply are not objectively evaluated before the product is sold, and consequently this information is not made available to the consumer. Thus the consumer makes his choice on the basis of a subjective appraisal of the product quality based on criteria such as visual appearance, geographical origin and conditions of production (e.g. breed, age, sex condition, type of feeding, housing). For instance, the meat producing animals reared on pasture have a more positive image in European consumers than those reared indoors and fed concentrate based diets (PRACHE and THÉRIEZ, 1999). Also, the consumers often detect specific organoleptic properties in the meat of animals diffe- ring by their feeding background, e.g. pasture lambs vs. concentrate lambs (WONG et al., 1975 ; SUZUKI and BAILEY, 1985 ; HA and LINDSAY, 1990 ; YOUNG et al., 1997 ; ROUSSET-AKRIM et al., 1997 ; PRIOLO et al., 2002). For meat products, as for other food products, different official quality signs exist at national and European levels that meet the consumer’s demand for information on product quality (Protected Designation of Origin, Protected Geographical Indication, Traditional Speciality Guaranteed, Red Label, Agriculture Biologique, Certifica- tion de Conformité). These quality certification schemes ensure a product added value for the benefit of producers. Conversely, the guarantee they are supposed to bring to the consumer must rely as far as possible on an objective assessment of quality based on analytical measurements. However, in the con- text of meat, traceability of the type of feed consumed by the animals is made difficult by the complexity of the mecanisms involved in the growth and compo- sitional changes of body tissues throughout the animal’s life.

The aim of this work was to test on a limited number of animals the potential of the analysis of the volatile fraction of the fat tissue to discriminate lambs dif- fering in their feeding background. For this purpose, commercial lambs were obtained from six European countries, representing a wide range of lamb types and/or rearing conditions including breed, sex, age, type of feeding and growth rate. The fat tissue was chosen because it can be sampled easily at the abattoir without affecting the commercial value of the carcass.

(3)

2 – MATERIAL AND METHODS

2.1 Animals

Thirty four commercial lambs from six different European countries, repre- sentative of 11 different lamb types (or production systems) were studied. The description of the animals for each lamb type and the corresponding rearing conditions are given in table 1. The lamb types were distributed into two groups according to the predominant type of feed consumed before slaughter, namely grass at pasture (group P) and concentrate (group C).

Table 1

Description of the lamb types used and the corresponding types of feeding.

2.2 Analysis of volatile compounds

2.2.1 Preparation of samples and collection of volatile fractions

Extraction of volatile compounds by dynamic headspace was performed on the subcutaneous fat tissue from the 34 lambs at the level of the longissimus thoracis (12th and 13th ribs) and longissimus lumborum muscles. The samples were collected 24 h post slaughter and kept frozen at – 20°C until analysis. A mass of 0.2 g of fat, wiped clean of adhering meat, was introduced into a cylin- drical glass extractor (diameter 40 mm, height 90 mm) and placed in a heating aluminium block. In order to limit the formation of lipid oxidation compounds, the cartridge was subjected to a protective helium flow (60 ml min^–1) for 5 min before heating. The samples were equilibrated to 100°C by preheating for 10 min. The volatile compounds were extracted using a 60 ml min^–1 helium flow for 1 h. Then, they were passed through a deactivated stainless steel transfer line (diameter 1/16) heated to 150°C and adsorbed on a Tenax trap (Tenax TA 60-80 mesh, Alltech, Deerfield, Illinois 60015, USA) of a Desorbtion Concentra-

Treatment (predominant type of feed before slaughter)

Country

(no. of animals) Breed

Age at slaughter

(weeks) Sex condition CCW¹

(kg)

Type of feeding²

Group C (concentrate, n = 13)

France (n = 3) France (n =1) Greece (n = 3) Spain (n = 3) Italy (n = 3)

Lacaune Texel Karagouniko Rasa Aragonesa

Appenninica

13.5 30.3 15.0 11.5 10.1

female female entire male entire male entire male

15.9 16.7 15.7 10.2 11.1

concentrate, straw concentrate, hay concentrate, hay concentrate, straw concentrate, milk, straw Group P

(pasture, n = 17)

France (n = 2) Iceland (n = 3) Iceland (n = 3) Italy (n = 3)

UK (n = 3) UK (n = 3)

Texel or Ile-de-France Icelandic Icelandic Bergamasca Suffolk ✕ Mule Welsh Mountain

27.1 18.0 18.0 49.2 15.2 29.9

female entire male

female castrated male castrated male entire male

16.1 16.6 16.4 29.6 18.0 15.4

pasture pasture, milk pasture, milk pasture³ pasture, milk

pasture 1.CCW = cold carcass weight.

2. predominant feed in bold type.

3. under conditions of transhumance.

(4)

tion Injection (DCI) dynamic headspace apparatus (Delsi Instruments, 92150 Suresnes, France).

2.2.2 Gas chromatography and mass spectrometry

Injection of volatile compounds into a gas chromatograph (Hewlett-Packard 5890 Series II) coupled to a mass-spectrometer (Hewlett-Packard 5971A) was achieved by thermal desorption of the Tenax trap at 200°C. Chromatographic conditions were as follows: Supelco capillary column (60 m x 0.32 mm; CH- 1196 Gland, Switzerland); stationary phase, SPB5 (1 µm); carrier gas, helium at 1 ml min^–1 (N55 quality, Air Liquide, France); injection split, 1 ml min^–1; oven temperature, 35 to 230°C with a slope of 3°C min^–1. The mass spectra were obtained in electronic impact at 70 eV. The scan range of mass spectra acquisi- tion was 33 to 250 atomic mass unit (a.m.u.).

2.2.3 Identification and semi-quantification

Identification of the volatile compounds was conducted by comparison of the experimental retention indices with those of the data bank compiled by Kondjoyan and Berdagué, (1996) and by comparison of the experimental mass spectra with those contained in the NIST/EPA/MSDC (version 3.01, 1990, Royal Society of Chemistry, Milton Road, Cambridge CB4 4WF, UK) and the NIST/

EPA/NIH (version 1.6, 1998, National Institute of Standards and Technology, Gaithersburg, MD 20899, USA) Mass Spectral Databases. All chromatographic peak areas were calculated by estimating the total ionic current from integra- tions performed on specific ions using the MS Chemstation software (Hewlett- Packard, 91947 Les Ulis, France).

2.3 Statistical analyses

The effect of type of feeding on volatile compounds was studied by a one- way analysis of variance using the General Linear Model procedure (SAS Insti- tute Inc, 1988). Least Square Means were used to test differences between treatment means. The model of analysis of variance was: Y = µ + TF + ε where Y is the volatile compound, µ is the constant term, TF is the main effect of type of feeding and ε is the residual variation. For each variable, the percentage of variance explained by the type of feeding was calculated (% variance = SSD of the factor / SSD of the model ✕ 100, with SSD = Sum of the Squares of the Differences). The volatile compounds significantly influenced by the type of feeding were studied by Principal Component Analysis (STATISTICA, 1999).

(5)

3 – RESULTS AND DISCUSSION

3.1 Nature and origin of volatile compounds

More than 144 compounds were detected among which 104 were identified (table 2). These volatile compounds, exhibiting molecular weights between 86 (hexane) and 240 (heptadecane), were of several chemical families: aldehydes (n = 20, including 2-methyl- branched aldehydes of 4 carbon atoms), ketones (n = 20, including 8-methyl ketones), alkanes and alkenes (n = 13), alcohols (n = 13), carboxylic acids (n = 10, including 3-methyl-branched acids of 3 and 4 carbon atoms), benzenic hydrocarbons (n = 10) and terpenes (n = 6). A smaller number of lactones (n = 3), nitrogen compounds (n = 3), furans (n = 2), sulphur compounds (n = 2), one pyrazine and one ester were also identified.

Table 2

Effect of type of feeding on the volatile compounds extracted from lamb fat tissue.

Name of compound Retention

index

Reliability of identifi-

cation

Effect of type of

feeding F Signifi- cance

C P

(1) (2) (3) (4) (4) (5) (6)

Hexane 600 a 7.3 20.8 11.0 **

2,3-butanedione 602 a 18.9 51.6 3.9 NS

Butanal, 3-methyl- 657 a 65.2 50.1 0.7 NS

Butanal, 2-methyl- 666 a 4.7 3.8 0.6 NS

Propanoic acid 681 a 21.8 9.3 9.0 **

1-penten-3-ol 686 a 18.3 26.5 2.4 NS

Heptane 700 a 55.5 26.0 2.9 NS

2-butanone, 3-hydroxy- 703 a 17.8 45.1 2.4 NS

1,2- propanediol 732 a 49.5 57.8 0.1 NS

Propanoic acid, 2-methyl- 741 a 52.8 26.8 10.1 **

Dimethyl disulphide 744 a 1.7 0.6 1.1 NS

Toluene 768 a 7.3 22.5 4.4 NS

Butanoic acid 769 a 16.6 7.6 12.2 **

2,3-butanediol 783 a 61.1 75.0 0.2 NS

2-hexanone 786 a 2.2 2.0 0.0 NS

Hexanal 797 a 60.2 52.0 0.6 NS

1-octene 799 a 59.7 121.2 0.7 NS

Octane 800 a 51.8 50.1 0.1 NS

2-octene 806 a 6.3 9.0 0.2 NS

Butanoic acid, 3-methyl- 828 a 12.3 3.0 20.0 ***

Butanoic acid, 2-methyl- 839 a 9.5 2.3 25.3 ***

1H-pyrrole, 2-methyl- 839 a 5.1 5.4 0.2 NS

2-hexanone, 3-methyl- 851 a 6.2 6.4 0.1 NS

2-propanone, 1-(acetyloxy)- 861 a 16.9 0.4 1.7 NS

Pentanoic acid 867 a 11.5 5.6 4.4 NS

p-xylene 868 a 14.1 14.2 0.0 NS

4-heptanone 871 a 3.1 0.03 22.5 ***

(6)

m-xylene 875 a 39.9 40.2 0.0 NS

2-heptanone 889 a 37.7 21.7 7.2 *

4-heptenal 899 a 5.9 17.7 9.0 **

o-xylene 901 a 12.5 12.9 0.0 NS

Heptanal 902 a 221.9 355.7 4.2 *

Ethanol, 2-butoxy- 905 a 214.7 39.9 2.0 NS

Gamma butyrolactone 909 a 89.3 55.4 4.6 NS

Pyrazine, 2,6-dimethyl- 910 a 2.5 1.9 0.5 NS

Dimethyl sulfone 911 a 162.1 163.1 0.0 NS

alpha-pinene 943 a 4.4 5.6 3.1 NS

Gamma pentalactone 950 a 32.2 7.9 16.4 **

2-heptenal (isomere A) 953 b 20.1 18.1 0.5 NS

2-heptenal (isomere B) 958 a 3.4 1.0 21.2 ***

Benzene, propyl - 962 a 4.0 4.3 0.4 NS

Hexanoic acid 966 a 60.1 23.2 8.2 **

Benzene, 1-ethyl, 3-methyl- 967 b 31.4 33.9 0.2 NS

Benzene, 1-ethyl, 2-methyl- 974 b 5.8 6.0 0.0 NS

Phenol 974 b 21.0 22.5 0.4 NS

1-octen-3-ol 978 b 125.0 74.7 2.5 NS

2,3-octanedione 981 a 80.1 966.1 11.6 **

5-hepten-2-one, 6 methyl- 983 a 142.9 159.5 2.2 NS

Benzene, 1,2,4-trimethyl 988 a 7.4 8.1 0.2 NS

2-octanone 990 a 170.5 59.5 36.3 ***

Furan, 2-pentyl- 992 a 13.8 11.9 0.6 NS

2,4-heptadienal (isomere A) 996 b 9.1 27.9 6.3 *

Benzene, 1,2,3-trimethyl- 1000 a 40.7 44.7 0.3 NS

Octanal 1003 a 202.7 165.6 4.3 NS

2,4-heptadienal (isomere B) 1010 a 28.5 57.3 7.7 *

1-hexanol, 2-ethyl- 1026 b 37.6 37.7 0.0 NS

p-cymene 1031 a 3.4 8.8 4.7 NS

Benzene, 1,3,5-trimethyl- 1034 a 6.4 7.3 0.5 NS

Benzyl alcohol 1036 a 25.9 33.4 0.2 NS

Limonene 1036 a 16.9 9.6 2.0 NS

Benzeneacetaldehyde 1049 a 30.8 21.0 4.6 NS

Heptanoic acid 1051 a 4.2 0.6 6.2 *

Gamma-hexanolactone 1055 a 8.0 7.6 0.2 NS

2-octenal 1057 b 25.6 20.8 1.8 NS

2-octen-1-ol 1066 b 36.0 27.3 0.9 NS

1-octanol 1067 b 32.8 30.5 0.1 NS

Phenol, 4-methyl- 1071 a 41.8 43.7 0.0 NS

Acetophenone 1072 a 18.7 19.4 0.1 NS

Unidentified compound (43, 87, 142) 1087 - 0.2 3.3 9.2 **

2-Nonanone 1090 a 192.1 45.6 12.3 **

Unidentified compound (81,119,134) 1094 - 0.8 1.5 8.0 *

Unidentified compound (55, 83) 1095 - 2.0 10.4 3.1 NS

Unidentified compound (71-99-101) 1097 - 1.8 10.2 3.1 NS

Undecane 1100 a 11.0 13.6 3.9 NS

index

cation

C P

(1) (2) (3) (4) (4) (5) (6)

(7)

Nonanal 1104 a 62.7 55.6 0.9 NS

Hexanoic acid, 2-ethyl- 1105 a 6.8 2.8 3.9 NS

Unidentified compound (57, 83, 60) 1127 - 4.1 3.1 0.3 NS

Unidentified compound (59,100) 1130 - 1.0 1.0 0.1 NS

Octanoic acid 1156 a 7.4 1.8 5.8 NS

2-nonenal 1161 b 215.1 309.5 2.8 NS

Ethanol, 2-(2-butoxyethoxy) 1188 a 298.1 269.8 0.5 NS

2-decanone 1191 a 40.3 33.6 1.9 NS

Dodecane 1200 a 22.0 22.7 0.1 NS

Decanal 1204 a 382.3 381.2 0.0 NS

Naphthalene 1208 a 15.2 16.2 0.3 NS

Unidentified compound (85, 86, 45) 1216 - 2.6 2.3 0.1 NS

Ethanol, 2-phenoxy- 1229 b 37.7 29.5 1.3 NS

Benzothiazole 1246 - 8.0 9.0 1.4 NS

2-decenal 1263 a 25.2 21.8 0.6 NS

2-undecanone 1293 a 51.7 29.4 7.7 *

2,4-decadienal (isomere A) 1297 b 134.7 140.0 0.1 NS

Tridecane 1299 a 43.4 33.9 6.0 *

Undecanal 1309 a 31.7 30.8 0.2 NS

2,4-decadienal (isomere B) 1319 b 11.6 7.6 2.8 NS

2-undecenal 1368 a 8.4 8.5 0.0 NS

Decanoic acid, ethyl ester 1390 a 5.6 0.9 9.5 *

Tetradecane 1400 a 67.3 81.0 2.7 *

o-copaene 1404 a 1.4 41.4 1.5 NS

1H-indole, 3-methyl- 1408 a 7.8 7.0 0.1 NS

Unidentified compound (69, 95,123,163) 1412 - 3.9 14.2 3.3 NS

Furan, 2-heptyl- 1412 b 7.1 37.6 4.4 NS

Isocaryophyllene 1439 b 0.0 0.2 1.1 NS

Acetophenone, p-acetyl (isomere A) 1451 b 11.7 24.4 1.4 NS 5,9-undecadien-2-one, 6,10-dimethyl- 1455 a 419.2 442.1 0.7 NS

Beta-caryophyllene 1455 a 35.5 105.3 1.7 NS

Decane, 2,3,7-trimethyl- 1466 b 94.5 202.1 4.7 *

Acetophenone, p-acetyl (isomere B) 1469 b 6.9 16.2 1.6 NS

2-tridecanone 1494 a 77.0 45.8 7.5 ***

Pentadecane 1500 a 598.5 1391.4 22.9 ***

Hexadecane 1600 a 156.3 235.6 8.2 ***

Heptadecane 1700 a 93.8 204.0 20.4 ***

2-pentadecanone 1702 a 47.8 23.0 15.6 ***

(1) For the unidentified compounds, the major ions are indicated in brackets.

(2) The retention indices are calculated for the SPB5 stationary phase of a capillary column.

(3) The reliability of the identification is indicated by the following letters : (a) identify by both mass spectrum and retention indice and (b) identify by mass spectrum; (-) unidentified compound.

(4) Mean values, expressed in arbitrary units, for each type of feeding: concentrate (C), pasture (P).

(5) F value (df 1, 28)

(6) Significance of the difference between means: NS (non significant, P ≥ 0.05), * (P < 0.05), ** (P < 0.01) and ***

(P < 0.001).

index

cation

C P

(1) (2) (3) (4) (4) (5) (6)

(8)

A large proportion of the volatiles identified have already been reported in the literature as typical compounds of the lamb fat tissues (SHAHIDI et al., 1986;

SUZUKI and BAILEY, 1985). Methyl-ketones, aliphatic aldehydes and acids, alcohols, alkanes, alkenes, furans and lactones are the predominant end-products of the oxidation of fatty acids (FRANKEL, 1991; MOTTRAM, 1998). Amino acid catabolism leads to the formation of methyl-branched aldehydes and acids (HA

and LINDSAY, 1990), phenol, sulphur compounds and 3-methyl indole or skatole (JENSEN et al., 1995). The degradation of carbohydrate produces compounds such as 2,3-butanediol, 2,3-butanedione (diacetyl) and 3-hydroxy-2-butanone (acetoin) (KANDLER, 1983). Maillard reactions which occur between amino acids and reducing sugars produce hydroxy-ketones, short chain methyl branched aldehydes, furans, pyrroles, pyrazines and sulphur compounds (MOTTRAM, 1998). Decanoic acid ethyl ester arises from esterification of ethanol and decanoic acid. The forage materials consumed by the animals explain the presence of 2,3 octanedione, terpenes (SUZUKI and BAILEY, 1985; YOUNG et al., 1997), long-chain alkanes (KING et al., 1993; URBACH and STARK, 1975), some aliphatic aldehydes (BROWN et al., 1979), phenolic and aromatic compounds (HA and LINDSAY, 1990).

3.2 Effect of type of feeding on volatile compounds

The type of feeding had a significant effect on 31 of the volatile compounds quantified (table 2). The class of chemical compounds most affected by the type of feeding were ketones (e.g. tridecanone), alkanes (e.g. heptadecane, pentadecane), linear and branched fatty acids (e.g. 2-methyl and 3-methyl butanoic acids), aldehydes (e.g. 2-heptenal) and a lactone (g-pentalactone).

The results of the PCA performed on the volatile compounds significantly influenced by the type of feeding are presented in figures 1a and 1b. The volatile compounds could be clearly separated into two groups according to component 1 (figure 1a). The first group was mainly composed of C₇ aldehydes (heptanal, 4-heptenal and two isomers of 2,4-heptadienal), alkanes (C₆ and C₁₄ to C₁₇), 2,3-octanedione and trimethyl-decane and characterised the fat samples from pasture-fed lambs (figures 1a & 1b). The second group mainly inclu- ded ketones (4-heptanone, 2-heptanone to 2-nonanone) and fatty acids (linear acids such as C₃, C₄, C₆ and C₇ ; branched acids such as 2-methyl propanoic, 2-methyl and 3-methyl butanoic), γ-pentalactone, 2-heptenal and tridecane (figure 1a). This group of volatiles was specific of fat samples from concentrate- fed lambs (figures 1a and 1b).

(9)

Figure 1

Principal component analysis of the 31 volatile compounds of lamb fat affected by the type of feeding.

(a) Variable plot

(b) Sample plot (P = pasture, C = concentrate, M* = milk – *see discussion) Canonical variable 1 (36.6 %)

Canonical variable 2 (15.0 %)

– 1.0 – 0.5 0.0 0.5

1.0

–1.0 – 0.6 – 0.2 0.2 0.6 1.0

Decane trimethyl Tetradecane

Hexane Pentadecane Heptadecane 2,3-Octanedione Unidentified 1

4-HeptenalHeptanal Hexadecane Unidentified 2

2,4-Heptadienal a 2,4-Heptadienal b Butanoic acid, 2m

2-Pentadecanone 2-Tridecanone Butanoic acid, 3m

2-Heptenal b4-HeptanonePentalactone2-Nonanone2-Undecanone 2-Heptanone

Decanoic acid, EE Heptanoic acid

Tridecane Butanoic acid

Propanoic acid, 2m Hexanoic acid

2-Octanone Propanoic acid

(a)

(b)

P

P C

P M

P C

– 3 – 2

– 2.0 – 1.5 – 1.0 – 0.5 0.0 0.5 1.0 1.5 2.0 2.5

– 1 0 1 2 3

C

C C C

CC C CC

C C C

C M

M M M

P P

P P P PP

P P P

P P P P

P P

P

(10)

In agreement with our results, JOHNSON et al. (1977), LARICK et al. (1987), and WONG et al. (1975) reported a greater desorption of short and medium chain fatty acids from fat samples of ovine and bovine species fed concentrate-based diets when compared with their pasture-fed counterparts. In the same way, SUZUKI and BAILEY, (1985) and YOUNG et al., (1997) also observed a higher desorption of methyl-ketones and lactones by the fat tissues of concentrate-fed lambs. The concentrate based diets are known to increase the concentration of unsaturated fatty acids such as linoleic acid in the fat tissues of ruminants (MEL- TON, 1990; MUIR et al., 1998; ROWE et al., 1999). These acids may be precursors of oxidation products such as ketones (e.g. methyl-ketones and lactones), and also of the 2-heptenal found in concentrate-fed lambs (BOYSTON et al., 1996:

FRANKEL, 1982; FRANKEL, 1991; YOUNG et al., 1997). The occurrence of greater concentrations of linoleic acid in the concentrate-fed lambs of the present experiment has been confirmed by the fatty acid analysis of the neutral lipid fraction of the muscle tissue conducted in a companion paper by ENSER et al., (2000). In agreement with MELTON (1990), ROWE et al., (1999) and ENSER et al., (2000) found that the fat tissue of pasture-fed lambs contained greater concentrations of linolenic acid than that of lambs fed other types of feed. Green leaf tissues contain high concentrations of linolenic acid which oxidation results in the formation of volatile compounds such as the 4-heptanal and the two isomers of 2,4-heptadienal found in the pasture-fed lambs (BROWN et al., 1979;

FRANKEL, 1982; PAQUETTE et al., 1985). Formation of 2-3 octanedione is linked to the enzyme lipoxygenase and linolenic acid, both abundant in green leaf tissue (YOUNG et al., 1997). Several authors have already reported that 2-3 octa- nedione is a good marker of pasture feeding in cattle (KING et al., 1993 ; LARICK

et al., 1987) and sheep (SUZUKI and BAILEY, 1985; YOUNG et al., 1997; PRIOLO et al., 2004). Our results also confirm that 2,3 octanedione is a good marker of a pasture diet as shown by a desorption around twelve times greater than that observed in concentrate-fed lambs (figure 2a). Although they are not as specific as 2,3 octanedione, the long-chain alkanes (C₁₄ to C₁₇), or their total abundance, could also be interesting markers of pasture feeding (figure 2b); for example pentadecane was detected in the fat of pasture-fed lambs at a desorption level three times greater than that observed in concentrate-fed lambs. Con- versely, the desorption level of 2-octanone was 3 times greater in the concentrate fed lambs in comparison with that found in the pasture-fed lambs (figure 2c). Also, 4-heptanone was found in the fat of the concentrate fed lambs while it was virtually absent from that of pasture fed lambs (figure 2d). The latter compounds could thus be proposed as markers of the concentrate feeding background. Subcutaneous fat was also sampled from milk-fed lambs (4 to 7 weeks old; n = 4) and analysed at the same time as the samples from pasture- and concentrate fed lambs (results not shown but illustrated in figure 1). The desorption of 2-tridecanone in these lambs was 2- to 4-fold greater than that observed in the lambs fed the two other types of feed. This compound could thus be proposed as a marker of the milk feeding background, but further investigation is required to confirm this hypothesis.

On a sensory point of view, the two companion papers by FISHER et al., (1999) and ENSER et al., (2000) based on a larger number of animals concluded that meat samples from pasture-fed lambs generally developed more intense

“livery” and “sheep meat” aromas. The “liver aroma” attribute has been descri- bed previously by PRIOLO et al. (2002) in 5-month old lamb, and by ROUSSET-

(11)

AKRIM et al., (1997) and YOUNG et al., (1997) in meat from 8.5-year old ewes, but the authors did not correlate it with a particular volatile compound. It appeared that some samples from pasture-fed lambs which developed a strong “sheep meat” aroma had also a high total content of branched chain fatty acids. ENSER

et al., (2000) reported that meat from concentrate fed lambs can also have a high total content of these fatty acids. In the present analysis, the 4-methyloctanoic and 4-methyl-nonanoic acids could not be detected because of the young age of the lambs used, the concentration of these two acids being most probably lower than the detection threshold of the mass spectrometer.

Figure 2

Simple statistics of selected volatile compounds of lamb subcutaneous fat as affected by the type of feeding (arbitrary units of abundance).

(a) 2,3 octanedione

(a)

Type of feeding

2,3-octanedione

0 500 1000 1500 2000 2500 3000

Median 25 %-75 % Min-Max

Concentrate Pasture

(12)

Figure 2

(b) sum of long-chain alkanes (C₁₄ to C₁₇) (c) 2-octanone

(b)

Type of feeding 0

500 1000 1500 2000 2500 3000 3500

Concentrate Pasture

(c)

Type of feeding 0

50 100 150 200 250 300

Pasture Concentrate

2-octanone

(13)

.

Figure 2 (d) 4-heptanone

4 – CONCLUSION

This study using lambs representing a large diversity of geographical origins and rearing conditions showed that the animals could be successfully discrimi- nated according to the predominant type of feed consumed, namely pasture or concentrate, by analysis of volatile compounds of subcutaneous fat. The desorption of 2,3-octanedione or long-chain alkanes, and that of 4-heptanone or 2-octanone from the subcutaneous fat tissue was much greater in lambs predominantly fed pasture and concentrate, respectively. These compounds can thus be proposed as markers of these two types of feeding. Further investigation however is required to validate these results on a larger number of lambs and on other animals species.

(d)

Type of feeding 0

2 4 6 8 10 12

Pasture Concentrate Median

25 %-75 % Min-Max

4-heptanone

(14)

ACKNOWLEDGEMENTS

This work was part of a collaborative research programme (FAIR 3CT96- 1768) involving research teams from France, Greece, Iceland, Italy, Spain and United Kingdom, and financed by the European Commission (DG VI).

REFERENCES

BOYSTON T. D., MORGAN S. A., JOHNSON K. A., W. W. R., BUSBOOM J. R., REE- VES J. J., 1996. Volatile lipid oxidation products of Wagyu and Domestic breeds of beef. Journal of Agricultural and Food Chemistry, 44, 1091-1095.

BROWN H. G., MELTON S. L., RIEMANN M.

J., BACKUS W. R., 1979. Effects of energy intake and feed source on chemical changes and flavor of ground beef during frozen storage. Journal of Animal Science, 48, 338-347.

ENSER M., NUTE G., WOOD J., SANUDO C., BERGE P., ZYGOYIANNIS D., THOR- KELSSON G., PIASENTIER E., FISHER A., 2000. Effects of production systems on the fatty acids and flavour of lamb from six european countries. In: 46th Internatio- nal Congress of Meat Science and Tech- nology, Buenos Aires, 27 August-1 September 2000, 186-187.

FISHER A., NUTE G., BERGE P., DRANS- FIELD E., PIASENTIER E., MILLS C., SANUDO C., ALFONSO M., THOR- KELSSON G., VALDIMARSDOTTIR T., ZYGOYIANNIS D., STAMATARIS C., 1999. Variation in the eating quality of lamb from divers European sheep types:

assessment by trained taste panels in six countries. In: 45th International Con- gress of Meat Science and Technology, Yokohama, Japon, 1-6 August 1999, 26- 27.

FRANKEL E. N., 1982. Volatile lipid oxidation products. Progress in Lipid Research, 22, 1-33.

FRANKEL E. N., 1991. Review - Recent advances in lipid oxidation. Journal of Science and Food Agriculture, 54, 495- 511.

HA J. K., LINDSAY R. C., 1990. Distribution of volatile branched-chain fatty acids in peri- nephric fats of various red meat species.

Lebensmittel Wissenschaft und Technolo- gie, 23, 433-440.

JENSEN M. T., COX R. P., JENSEN B. B., 1995. Microbial production of skatole in the hind gut of pigs given different diets and its relation to skatole deposition in backfat. Animal Science, 61, 293-304.

JOHNSON B., WONG E., BIRCH E. J., 1977.

Analysis of 4-methyloctanoic acid and other medium chain-length fatty acid constituents of ovine tissue lipids. Lipids, 12, 340-347.

KANDLER O., 1983. Carbohydrate metabo- lism in lactic acid bacteria. Antonie van Leeuwenhoek, 49, 209-224.

KING M.-F., HAMILTON B. L., MATTHEWS M. A., D.C. R., FIELD R. A., 1993. Isolation and identification of volatiles and conden- sable material in raw beef with supercriti- cal carbon dioxide extraction. Journal of Agricultural and Food Chemistry, 41, 1974-1981.

KONDJOYAN N., BERDAGUÉ J. L., 1996. A compilation of relative retention indices for the analysis of aromatic compounds, Laboratoire Flaveur, INRA, Clermont- Ferrand.

LARICK D. K., HEDRICK H. B., BAILEY M. E., WILLIAMS J. E., HANCOCK D. L., GAR- NER G. B., MORROW R. E., 1987. Flavor constituents of beef as influenced by forage- and grain-feeding. Journal of Food Science, 52, 245-251.

MELTON S. L., 1990. Effects of feeds on flavor of red meat : a review. Journal of Ani- mal Science, 68, 4421-4435.

(15)

MOTTRAM D. S., 1998. Flavour formation in meat and meat products: a review. Food Chemistry, 62(4), 415-424.

MUIR P. D., DEAKER J. M., BOWN M. D., 1998. Effects of forage-and grain-based feeding systems on beef quality : a review.

New Zealand Journal of Agricultural Research, 41, 623-635.

PAQUETTE G., KUPRANYCZ D. B., VAN DE VOORT F. R., 1985. The mechanisms of lipid autoxidation. I. Primary oxidation products. Canadian Institute of Food Science and Technology Journal, 18, 112-118.

PRACHE S., THERIEZ M., 1999. Traceability of lamb production systems: carotenoids in plasma and adipose tissue. Animal Science, 69, 29-36.

PRIOLO A., MICOL D., AGABRIEL J., PRA- CHE S., DRANSFIELD E., 2002. Effect of grass or concentrate feeding systems on lamb carcass and meat quality. Meat Science, 62, 179-185.

PRIOLO A., CORNU A., PRACHE S., KROG- MANN M., KONDJOYAN N., MICOL D., BERDAGUÉ J.L, 2004, Fat volatile tracers of grass feeding in sheep. Meat Science, in press.

ROUSSET-AKRIM S., YOUNG O. A., BERDA- GUÉ J. L., 1997. Diet and growth effects in panel assessment of sheepmeat odour and flavour. Meat Science, 45, 169-181.

ROWE A., MACEDO F. A. F., VISENTAINER J. V., SOUZA N. E., MATSUSHITA M., 1999. Muscle composition and fatty acid profile in lambs fattened in drylot or pasture. Meat Science, 51, 283-288.

SAS INSTITUTE INC, 1988. Statistical Analy- sis Systems. In: SAS /STAT User’s Guide,

549-640, Statistical Analysis Systems Ins- titute, Inc., Cary, NC 27513.

SHAHIDI F., RUBIN L. J., D’SOUZA L. A., 1986. Meat flavor volatiles: a review of the composition, techniques of analysis, and sensory evaluations. Critical Review in Food Science and Nutrition, 24, 141-243.

STATISTICA, 1999. A comprehensive system for statistics, graphics and application development, Statsoft, Maisons-Alfort.

SUZUKI J., BAILEY M. E., 1985. Direct sampling capillary GLC analyis of flavor volatiles from ovine fat. Journal of Agricultural and Food Chemistry, 33, 343-347.

URBACH G., STARK W., 1975. The -20 Hydrocarbons of butterfat. Journal of Agricultural and Food Chemistry, 23, 20- 24.

WONG E., NIXON L. N., JOHNSON C. B., 1975. Volatile medium chain fatty acids and mutton flavour. Journal of Agricultu- ral and Food Chemistry, 23, 495-498.

YOUNG O. A., BERDAGUÉ J. L., VIALLON C., ROUSSET-AKRIM R., THERIEZ M., 1997.

Fat-borne volatiles and sheepmeat odour.

Meat Science, 45, 183-200.

YOUNG O. A., BRAGGINS T. J., WEST J., LANE G. A., 1999a. Animal production origins of some meat color and flavor attributes. In: Quality Attributes of Muscle Foods, 11-29, Kluwer Academic/Plenum, New York.

YOUNG O. A., PRIOLO A., LANE G., FRAZER K., KNIGHT T., 1999. Causes of pastoral flavour in ruminant fat. In: 45th Internatio- nal Congress of Meat Science and Tech- nology, Yokohama, Japon, 1-6 August 1999, 420-421.

(16)