Saturated oxo fatty acids in cheese

Identification of the saturated oxo fatty acids in cheese

Identification of the saturated oxo fatty acids in cheese By ELIZABETH Y. BRECHANY AND WILLIAM W. CHRISTIE Hannah Research Institute, Ayr, Scotland KA6 5HL

(Received 15 April 1991 and accepted for publication 10 July 1991)

SUMMARY. The nature and composition of the saturated oxo fatty acids in cheddar cheese have been re-evaluated following isolation by modern chromatographic procedures, including low-pressure column chromatography and high-performance liquid chromatography (adsorption and silver ion modes). The oxo fatty acids were identified and quantified (for the first time) by gas chromatographymass spectrometry in the form of the methyl ester derivatives and many of the structures were confirmed similarly by preparing picolinyl ester derivatives. Thirty six saturated oxo fatty acids were identified, ranging in chain-length from 9 to 22 with an oxo group on carbons 4 to 17 (except for 12), of which twenty one had been identified in a previous study and fifteen were new. Fifteen fatty acids tentatively identified by others did not in fact appear to be present.

More different fatty acids have been identified in cow's milk fat than from any other natural source (Patton and Jensen, 1975). Among these are 36 saturated and 11 unsaturated oxo fatty acids (Weihrauch et al. 1974) in addition to several 3-oxo fatty acids (Van de Ven et al. 1963; Parks et al. 1964). Although they are minor constituents, they may have some importance as precursors of methyl ketones which contribute greatly to the flavour of cheddar and other cheeses (Evans and Mabbit 1974; Foda et al. 1974; Seth and Robinson, 1988). Such oxo fatty acids are not common constituents of animal lipids and their origin is unknown. Weihrauch et al. (1974) were only able to make tentative identifications of many of the oxo fatty acids in milk because of the limited resolution of the chromatographic techniques available to them, and the proportions of the different components were not determined. In recent years, there have been substantial improvements in chromatography as applied to lipids, especially highperformance liquid chromatography (HPLC) (Christie, 1987a) and gas chromatography-mass spectrometry (GC-MS) (Christie, 1989). This has prompted a re-evaluation of the nature of the oxo fatty acids and the compositions of the saturated components are described here. The lipids from cheddar cheese were studied because oxo fatty acids may be precursors of flavour compounds.

EXPERIMENTAL Reagents TM

or HPLC grades and were obtained from Fisons Scientific All solvents were Analar Equipment PLC (Loughborough, U.K.). 5-, 7- and 12-Oxostearic acids were supplied by Aldrich Chemical Co. Ltd (Gillingham, U.K.). Other reagents and standards were obtained from Sigma Chemical Co. Ltd (Poole, U.K.). Isolation of a fraction enriched in the oxo fatty acids Lipids were extracted by the procedure of Folch et al. (1957) from cheddar cheese (100 g) made at the Institute. They were fractionated by column chromatography in portions of about 5 g on a column (300 x 13 mm) containing FlorisilTM (10 mL). The bulk of the triacylglycerols were eluted with hexane-diethyl ether (4:1, v/v; 50 mL) and were discarded. A more polar fraction was then eluted with a futher portion of this solvent (50 mL) and with diethyl ether alone (100 mL).

Brechany,E.Y. and Christie, W.W. J. Dairy Res., 59, 57-64 (1992)


The combined polar fractions were transesterified with 1% methanolic sulphuric acid (Christie, 1989). Aliquots of the methyl esters (50 mg) were chromatographed on Bond Elut-NH2 solid phase extraction columns (Jones Chromatography, Hengoed, Wales), eluted with hexane-dichloromethane (99:1, v/v; 10 mL) and then with hexane-dichloromethane (85:15, v/v; 10 mL); the second fraction (156 mg in total) contained the required oxo fatty acids together with some normal acids, hydroxy acids and cholesterol. High-Performance Liquid Chromatography A Spectra-Physics Model 8700 solvent delivery system (Spectra-Physics Ltd, St. Albans, U.K.) was used in all HPLC separations, together with a Cunow Model DDL21 Light-Scattering detector (Severn Analytical Ltd, Hitchin, U.K.). A stream-splitter (approximately 10:1) was inserted between the column and the detector to permit collection of fractions. The crude oxo fraction (1-2 mg aliquots) was further purified by HPLC on a column (250 x 4.6 mm) of HypersilTM 5 silica gel (HiChrom Ltd, Reading, U.K.) eluted with a gradient of hexane to hexane-1,2-dichloroethane-2-propanol (85:13.5:1.5 by volume) at a flow-rate of 1 mL/min. Fractions were collected manually as they were seen to elute, and corresponding fractions from several runs were pooled. For silver ion HPLC, a column (250 x 4.6 mm) of NucleosilTM 5SA, converted to the silver ion form as described previously (Christie, 1987b), was donated by Chrompack BV (Middleburg, Netherlands). Samples (0.5 mg) were applied to the column in dichloroethane solution. The mobile phase was dichloromethane for the first 7 minutes then a gradient to dichloromethane-acetone (7:3, v/v) was generated over a further 33 min, at a flow-rate of 1 mL/min. Fractions were collected manually via the stream splitter and methyl nonadecanoate (0.1 mg) was added to each as an internal standard. Gas Chromatography-Mass Spectrometry The fatty acids were subjected to GC-MS in the form of the methyl esters and in addition, they were converted to the picolinyl ester derivatives, essentially by the method of Balazy and Nies (1989). The methyl esters were analysed on a fused-silica column (50 m x 0.25 mm i.d.) coated with FFAPTM CB (Chrompack UK Ltd, London, U.K.) with helium as carrier gas in a Hewlett Packard Model 5890 gas chromatograph attached to a Hewlett Packard Model 5970 Mass Selective Detector (Hewlett Packard Ltd, Wokingham, U.K.); the temperature was held at 180°C for 3 min then it was raised at 1°C/min to 240°C. The picolinyl esters were analysed on a fused-silica capillary column (25 m x 0.25 TM mm i.d.), coated with a cross-linked (5% phenylmethyl) silicone (CP-Sil 8 , Chrompack UK Ltd, o London, U.K.), temperature-programmed from 60 C to 220°C at 50°C/min then to 250°C at 1°C/min; they were also analysed on a similar column coated with the polar phase BPX 70 (50 m x 0.25 mm; SGE (UK) Ltd, Milton Keynes, U.K.) held isothermally at 260°C. In each instance, the column outlet was connected directly into the ion source of the mass spectrometer, which was operated at an ionization energy of 70 eV.

RESULTS AND DISCUSSION The oxo fatty acids in milk and dairy products are present in esterified form in the triacylglycerols in very low concentrations and they must initially be concentrated by some means. It was convenient to first obtain a polar triacylglycerol fraction by adsorption column chromatography before conversion to the methyl ester derivatives, which were also subjected to column chromatography. This yielded a fraction (0.4 % of the total lipids), which was greatly enriched in the oxo fatty acids. Further purification was achieved by HPLC on silica gel. Preliminary GC analyses showed that the mixture of saturated and unsaturated oxo fatty acids was highly complex, so they were simplified by HPLC in the silver ion mode as illustrated in Figure 1, i.e. saturated and unsaturated fractions were collected for separate analysis (Christie et al. 1988). Peaks 1 and 2 represented the saturated oxo fatty acids, while the remaining peaks were unsaturated oxo and hydroxy acids with sterols (probably cholesterol). Data for the saturated compounds only are considered here. In total, they were present at a level of 102 mg per Kg of cheese. The methyl esters of the saturated oxo acids were resolved into a large number of components on a GC capillary column coated with a polar stationary phase for identification by mass spectrometry



Fig. 1. Silver ion HPLC of methyl ester derivatives of a fraction enriched in oxo fatty acids. Peaks 1 and 2 are saturated oxo acids. The experimental conditions are given in the text.

Fig. 2. Total ion trace from GC-MS of the methyl ester derivatives of the saturated oxo fatty acids in cheese. The experimental conditions are given in the text.



92

Abundance %

O CH2OOC

192

90 N

80 70 60

108

164

50 55

40

151 262 249

30 206

65

20 10

60

Fig. 3.

+

234

M

304 346 389 290 332 125 374 80 100 120 140 160 180 200 220 240 260 280 300 320 340 360 m/z

Mass spectrum of a synthetic standard, picolinyl 7-oxostearate.

Abundance %

234

90 80 92

CH2OOC

70

O

N

60 50 40

108 164 151

30 55 20

220

71

10

206

248

+

M

291 360

389

80 100 120 140 160 180 200 220 240 260 280 300 320 340 360

m/z

193 60

304 276 332

Fig. 4 Mass spectrum of picolinyl 10-oxostearate, the most abundant oxo acid in cheese.

as illustrated in Figure 2. The characteristic fragmentation patterns in the mass spectra of methyl esters of oxo fatty acids that permit the location of the oxo group are well documented (Ryhage and Stenhagen, 1960). Ions formed by cleavage of the molecule beta to the oxo group on each side of it are usually of greatest diagnostic value, although ions formed by alpha cleavage are usually detected also. With complex mixtures when components are imperfectly resolved, the presence of other ions can lead to difficulties in interpretation, however.



Table 1.

Component

standard 4-oxo 9:0 5-oxo 10:0 4-oxo 12:0 5-oxo 12:0 7-oxo 12:0 4-oxo 13:0 4-oxo 14:0 5-oxo 14:0 6-oxo 14:0 9-oxo 14:0 11-oxo 14:0 5-oxo 15:0 4-oxo 16:0 5-oxo 16:0 7-oxo 16:0 8-oxo 16:0 11-oxo 16:0 13-oxo 16:0 14-oxo 16:0 15-oxo 16:0 9-oxo 17:0 9-oxo 18:0 10-oxo 18:0 13-oxo 18:0 15-oxo 18:0 16-oxo 18:0 17-oxo 18:0 11-oxo 19:0 4-oxo 20:0 5-oxo 20:0 11-oxo 20:0 15-oxo 20:0 4-oxo 22:0 5-oxo 22:0 7-oxo 22:0 13-oxo 22:0 a b c

Composition of the saturated oxo fatty acids found in Cheddar cheese, expressed as mg/Kg of cheese and as weight percent of the fraction.

Retention time a (min)

Concentration mg/Kg

Wt %

29.94 10.25 12.19 17.17 17.45 18.49 20.84 25.24 25.61 26.94 27.35 27.79 30.08 36.04 36.41 37.44 38.95 39.20 39.46 41.32 43.46 44.57 50.38 53.40 53.56 53.67 54.92 56.86 58.01 61.64 62.24 64.05 64.42 76.04 76.80 77.99 79.21

trace 0.29 0.77 0.26 0.28 0.09 0.21 0.51 1.42 1.66 0.19 trace trace 0.49 trace 16.59 3.24 0.59 0.59 0.63 0.44 trace 66.07 2.99 0.27 0.64 0.27 2.67 0.12 0.14 0.52 0.21 0.08 trace trace 0.22

trace 0.30 0.79 0.27 0.29 0.09 0.22 0.52 1.45 1.70 0.19 trace trace 0.50 trace 17.00 3.32 0.60 0.60 0.65 0.45 trace 67.69 3.06 0.28 0.66 0.28 2.74 0.12 0.14 0.53 0.22 0.08 trace trace 0.23

b

Confirmed

Found c previously

* * *

* * * *

* * *

* * *

* * * * * * * * * * * * * * * * *

* * * *

* * * * *

* *

* *

*

*

Taken from the GC trace in Figure 2 Structure confirmed from the mass spectrum of the picolinyl ester derivative. Weihrauch et al. 1974

Rather simpler mass spectra are in general obtained when picolinyl as opposed to methyl ester derivatives of fatty acids are prepared (Christie, 1989), and this also proved to be true of oxo fatty acids. The mass spectrum of a standard compound, picolinyl 7-oxoostearate, is illustrated in Figure 3. This particular compound was not present in the cheese, but a definitive interpretation of the spectrum is possible as the structure is known absolutely. Ions for fragmentations around the picolinyl ester group at m/z = 92, 108, 151 and 164 were abundant as is usual, but the molecular ion (m/z = 389) was small. A distinctive ion at m/z = 192 represented cleavage on the carboxyl side of the molecule beta to the oxo group. The ion representing cleavage beta to the oxo group on the hydrocarbon side carried an additional hydrogen atom and thus was odd-numbered (m/z = 249); this is unusual with picolinyl esters and therefore may have diagnostic value. In the region of the spectrum of higher molecular weight, a



regular series of ions 14 amu apart were apparent, because of smooth radical-induced cleavages of successive methylene groups down the chain, until the ion at m/z = 234 when there was a gap of 28 amu for loss of the oxo carbon and its oxygen atom. Regular series of ions 14 amu apart were then seen on the carboxyl side of the oxo group. The spectrum of picolinyl 10-oxostearate (Figure 4), the most abundant component found in this study, exhibited similar features, except that the ions representing cleavage beta to the oxo group are at m/z = 234 and 291, while the gap of 28 amu for loss of the oxygenated carbon is between m/z = 248 and 276. With spectra of isomers in which the oxo group is closer to the carboxyl group, e.g. picolinyl 5-oxostearate (spectrum not illustrated), similar features were apparent in the mass spectrum but all the ions of higher molecular weight were less intense. Some analogous features were found by Tulloch (1985) in the mass spectra of pyrrolidide derivatives of oxo fatty acids. Picolinyl esters do not give as good resolution as methyl esters when subjected to GC, although the problem was ameliorated by using two liquid phases of different polarity. On the other hand, when both types of derivative were examined, valuable complementary information was obtained. The mass spectra of methyl esters were used for the primary identifications, and mass spectra of picolinyl esters derivatives confirmed the presence of many of the components. Those saturated oxo fatty acids found in this study are listed in Table 1 together with the relative proportions of each and the absolute amounts in the cheese. The GC retention times of components are listed merely so that they can be related to specific peaks in Figure 2. 36 Distinct fatty acids were positively identified with 9 to 20 carbon atoms. These included a number of positional isomers with oxo groups in one of positions 4 to 17 (except for 12). Of the fatty acids identified here, 21 had been found previously by Weihrauch et al. (1974), but the remainder are new. We were not able to detect 15 of the fatty acids identified (some only tentatively) previously. While it is possible that some of these may have been present in the milk used by Weihrauch et al. (1974) and not in the cheese analysed here, it seems more probable that they made erroneous assignments from the mass spectra because of the poor resolution of the chromatographic techniques then available. No 3-oxo fatty acids were detected, presumably because these were rapidly converted to methyl ketones (Van der Ven et al. 1963; Parks et al. 1964). It is doubtful whether the major oxo acids detected here could serve as precursors of those odd-chain methyl ketones found in cheese.No quantitative data were supplied in the earlier study. Although 36 different oxo acids were detected in the present work, two of these, 10-oxo-stearate and 8-oxo-palmitate, comprised more than 80 % of the total. 13-Oxo-stearate and 11-oxo-pamitate were reported to be the most abundant components by Weihrauch et al. (1974). Fatty acids containing oxo substituents are not believed to be common in nature, although recent work suggests that they may be more widely distributed in foods of plant and animal origin than has been believed, albeit at very low concentrations (Schwartz & Rady, 1990). Their origin is not known. The structures of the compounds reported here from cheese do, however, permit some speculation. Thus, the more abundant fatty acids of each chain-length group, i.e. 10-oxo-18:0, 8-oxo16:0, 6-oxo-14:0 and 4-oxo-12:0, could be derived from the first by partial beta-oxidation. 13-Oxo-18:0, 11-oxo-16:0, 9-oxo-14:0, 7-oxo-12:0 and 5-oxo-10:0 form a similar sequence and could have arisen in the same way. As enzymes systems would be expected to be too specific in their properties to be able to insert oxo groups in a wide range of positions, especially at both odd- and even-numbered carbon atoms in an aliphatic chain, it is possible that the precursors are hydroperoxides of polyunsaturated fatty acids in the diet of the animals, that are subjected to biohydrogenation in the rumen or are metabolized, perhaps as part of a detoxification mechanism, in the tissues of the animal. Betaoxidation would certainly be expected to be a tissue process. Recently 9- and 13-keto-octadecadienoic acids were found in the lipids of biological membranes from rabbits, and they were presumed to arise by the action of cellular lipoxygenases (Kuhn et al. 1990). Further information may come from a study of the composition of the unsaturated oxo fatty acids that is in progress.



REFERENCES BALAZY,M. & NIES,A.S. 1989 Characterization of epoxides of polyunsaturated fatty acids by mass spectrometry via 3-pyridinylmethyl esters. Biomedical & Environmental Mass Spectrometry 18 328-336 CHRISTIE,W.W. 1987a High-Performance Liquid Chromatography and Lipids. Oxford : Pergamon Press CHRISTIE,W.W. 1987b A stable silver-loaded column for the separation of lipids by high performance liquid chromatography. Journal of High Resolution Chromatography & Chromatography Communications 10 148-150 CHRISTIE,W.W. 1989 Gas Chromatography and Lipids. Ayr: The Oily Press CHRISTIE,W.W., BRECHANY,E.Y. & STEFANOV,K. 1988 Silver ion high-performance liquid chromatography and gas chromatography-mass spectrometry in the analysis of complex fatty acid mixtures: application to marine invertebrates. Chemistry & Physics of Lipids 46 127-135 EVANS,E.W. & MABBIT,L.A. 1974 The flavour of cheddar cheese. Biennial Reviews. National Institute for Research in Dairying (Shinfield, Reading) pp. 59-72 FODA,E.A., HAMMOND,E.G., REINBOLD,G.W. & HOTCHKISS,D.K. 1974 Role of fat in flavour of cheddar cheese. Journal of Dairy Science 57 1137-1142 FOLCH,J., LEES,M. & STANLEY,J.S. 1957 A simple method for the isolation and purification of lipides from animal tissues. Journal of Biological Chemistry 226 497-509 KUHN,H., BELKNER,J., WIESNER,R. & ALDER,L. 1990 Occurrence of 9- and 13-ketooctadecadienoic acid in biological membranes oxygenated by the reticulocyte lipoxygenase. Archives of Biochemistry & Biophysics 279 218-224 PARKS,O.W., KEENEY,M., KATZ,I. & SCHWARTZ,D.P. 1964 Isolation and characterization of the methyl ketone precursors in milk fat. Journal of Lipid Research 5 232-235 PATTON,S. & JENSEN,R.G. 1975 Lipid metabolism and membrane functions of the mammary gland. Progress in the Chemistry of Fats and Other Lipids 14 163-277 RYHAGE,R. & STENHAGEN,E. 1960 Mass spectrometric studies VI. Methyl esters of normal chain oxo-, hydroxy-, methoxy- and epoxy-acids. Arkiv for Kemi 15 545-573 SETH,R.J. & Robinson,R.K. 1988 Factors contributing to the flavour characteristics of mould-ripened cheese. In Developments in Food Microbiology - 4 pp. 23-46 (Ed. R.K. Robinson) London : Elsevier Applied Science SCHWARTZ,D.P. & Rady,A.H. 1990 Determination and occurrence of oxofatty acids in fats and oils. Journal of the American Oil Chemists' Society 67 635-641. TULLOCH,A.P. 1985 Mass spectra of pyrrolidides of oxo, hydroxy and trimethylsliyloxy octadecanoic acids. Lipids 20 652-663. VAN DER VEN,B., HAVERKAMP BEGEMANN,P. & SCHOGT,P.J.M. 1963 Precursors of methyl ketones in butter. Journal of Lipid Research 4 91-95 WEIHRAUCH,J.L., BREWINGTON,C.R. & SCHWARTZ,D.P. 1974 Trace constituents in milk fat: isolation and identification of oxo fatty acids. Lipids 9 883-890
