A

fraasccm©

K

1

bclhirac

(ta

0d

Journal of Chromatography A, 1169 (2007) 175–178

Fast analysis by gas–liquid chromatographyPerspective on the resolution of complex

fatty acid compositions

Frederic Destaillats ∗, Cristina Cruz-HernandezNestle Research Centre, Vers-chez-les-Blanc, Switzerland

Received 19 July 2007; received in revised form 27 August 2007; accepted 31 August 2007Available online 5 September 2007

bstract

Separation of fatty acids as methyl ester (FAME) derivatives has been carried out using short and highly polar capillary column developedor fast gas–liquid chromatography (GLC) applications. The GLC parameters have been optimized in order to achieve separation of FAMEanging from 4:0 (butyric acid) to 24:1 in less than 5 min. Milk fat that has by far the most complex fatty acid composition among edible fatsnd oils has been used to optimize the method. The volume of the oven has been reduced in order to allow for a heating rate of 120 ◦C/minnd to rapidly cool-down to the initial temperature (50 ◦C) of the GLC program. The GLC conditions developed are not suitable to achieveeparation of positional and geometrical isomers of octadecenoic acid but are useful to perform separation of major fatty acids in milk fat. The

onditions developed could be used to analyze edible fats and oils or biological samples such as plasma or red blood cell lipids. The resultsonfirmed that short and highly polar fast columns operating under optimal conditions could be used to separate the fatty acids in variousatrices. 2007 Elsevier B.V. All rights reserved.

eywords: Fast gas chromatography; Fat; Fatty acid methyl ester; Gas–liquid chromatography; Milk fat; Plasma; Red blood cell; Tuna oil

atttaftllGad

. Introduction

Fast analysis by gas–liquid chromatography (GLC) coulde achieved by optimizing key operating parameters such asarrier gas type and velocity, length and diameter of the capil-ary column [1]. The reduction of the capillary column lengthas a dramatic impact on peak resolution and shows interest-ng results for the class separation of triacyglycerols [2]. Theeduction of the diameter of the column led to the emergence ofnew generation of open-tubular capillary column named fast

olumns.The use of fast GLC to analyze fatty acid methyl esters

FAMEs) has been reported in different studies [3–11]. One ofhe most interesting applications of fast GLC separation of fattycid methyl esters (FAME) has been published by Masood et

∗ Corresponding author. Tel.: +41 21 785 8937; fax: +41 21 785 8553.E-mail address: frederic.destaillats@rdls.nestle.com (F. Destaillats).

rwbp[

F

021-9673/$ – see front matter © 2007 Elsevier B.V. All rights reserved.oi:10.1016/j.chroma.2007.08.073

l. [3]. In their study, the authors optimized GLC conditionso increase the throughput of plasmatic fatty acid composi-ion in the frame of large clinical trials. In addition, reducinghe time of the chromatographic run, these authors developednd automated the procedure with a robot to prepare FAMErom human plasma [3]. Fast GLC has been also applied tohe analysis of the fatty acid composition of plasma phospho-ipids [4] and to the quantification of conjugated isomers ofinoleic acid in human plasma [5,6]. Mondello et al. used fastLC to analyze the fatty acid composition of fish oils [7,8]

nd different edible oils and fats [7–9]. These improvements areue to the commercial availability of fast columns. From theesults obtained so far, it is clear that the use of fast columnsill impact significantly the domain of lipid analysis. However,esides the work conducted on fish oil, few studies have been

erformed on complex natural lipid matrices such as milk fat7–9].

The present study aims to explore the separation of complexAME preparation using short and highly polar fast columns.

1 J. Ch

Tms

2

2

iriSMb

2

udapa

2

lU(A(ho5ia41

2

aia(a(tala(ea

nlat[

3

tyatspeed heating rate could be achieved by decreasing the volumeof the oven. In the present study, we used a commercial device(Agilent) to limit the volume to about 5400 cm3.

Fig. 1. Resolution of fatty acid methyl esters (FAMEs) derived from milk fatachieved by fast gas–liquid chromatography using a 10-m BPX-70 column (seeSection 2 for experimental details). Identification of the chromatogram regions:

76 F. Destaillats, C. Cruz-Hernandez /

he operating conditions developed to analyze this complexatrix could be applied to other edible fats and oils or biological

amples.

. Materials and methods

.1. Samples and reagents

Milk fat, cocoa butter and tuna oil samples were collectedn different countries by Nestle collaborators. Rat plasma anded blood cell samples used in the present study were collectedn the frame of an animal experiments taking place on site.odium methoxide (30% in methanol) has been obtained fromerck (Darmstadt, Germany). Standard FAME mixtures have

een obtained from Nu-Chek-Prep (Elysian, MN, USA).

.2. Preparation of fatty acid methyl esters (FAMEs)

FAMEs have been prepared as described previously [12]sing sodium methoxide as a catalyst in order to avoid degra-ation or loss of fatty acids from milk fat (e.g. short chain fattycids or conjugated isomers of linoleic acid). Derivatization oflasma and red blood cell fatty acids has been carried out undercidic conditions as described by Masood et al. [3].

.3. Fast analysis of FAMEs by gas–liquid chromatography

Analysis of total FAMEs were performed on a 7890 Agi-ent gas chromatograph (Agilent Technologies, Palo Alto, CA,SA), equipped with a fused-silica BPX-70 capillary column

10 m × 0.1 mm I.D., 0.2 �m film thickness; SGE, Melbourne,ustralia). Split injector (500:1) and flame-ionization detection

FID) system were operating at 250 ◦C. The volume of the ovenas been reduced to about 5400 cm3 using a commercial devicebtained from Agilent. Oven temperature programming was0 ◦C isothermal for 0.2 min, increased to 180 ◦C at 120 ◦C/min,sothermal for 1 min at this temperature then increased to 220 ◦Ct 20 ◦C/min and then to 250 ◦C at 50 ◦C/min (total run time.9 min). The carrier gas (H2) flow was maintained constant atmL/min and the acquisition of the FID signal at 100 Hz.

.4. Identification of FAMEs

A mixture was used to identify FAME in milk fat sampless described previously [12]. The mixture contained the follow-ng methyl esters: butyric acid (4:0), caproic acid (6:0), capryliccid (8:0), capric acid (10:0), undecanoic acid (11:0), lauric acid12:0), tridecanoic acid (13:0), myristic acid (14:0), myristoleiccid (14:1 n-5), pentadecanoic acid (15:0), pentadecenoic acid15:1 n-5), palmitic acid (16:0), palmitoleic acid (16:1 n-7), hep-adecanoic acid (17:0), heptadecenoic acid (17:1 n-7), steariccid (18:0), elaidic acid (trans-18:1 n-9), oleic acid (18:1 n-9),inolelaidic acid (all trans-18:2 n-6), linoleic acid (18:2 n-6),

rachidic acid (20:0), �-linoleic acid (18:3 n-6), eicosenoic acid20:1 n-9), linolenic acid (18:3 n-3), heneicosanoic acid (21:0),icosadienoic acid (20:2 n-6), behenic acid (22:0), eicosatrienoiccid (20:3 n-6), erucic acid (22:1 n-9), eicosatrienoic acid (20:3

(hfga

romatogr. A 1169 (2007) 175–178

-3), arachidonic acid (20:4 n-6), docosadienoic acid (22:2 n-6),ignoceric acid (24:0), eicosapentanoic acid (20:5 n-3), nervoniccid (24:1 n-9) and docosahexaenoic acid (22:6 n-3) was usedo identify FAME in milk fat samples as described previously12].

. Results and discussion

GLC analysis of FAMEs could not be performed using allypes of gas chromatograph apparatus. To reduce time of anal-sis the first requisite is to use hydrogen as carrier gas [1]. Thepparatus should be able to provide enough pressure to main-ain a constant flow of at least 1 mL/min (290 kPa at 50 ◦C). High

Part A) represents the chromatographic region ranging from butyric (4:0) toeptadecenoic (17:1) acids and (Part B) the chromatographic region rangingrom stearic (18:0) to docosahexaenoic (DHA, 22:6 n-3) acids. An amplifiedraph of the chromatographic region from stearic (18:0) to linoleic (18:2 n-6)cids is provided in Fig. 2.

F. Destaillats, C. Cruz-Hernandez / J.

Fig. 2. Enlarged view of the chromatographic region from stearic (18:0) tolmd

dtcpb

FbB

baApibuwtiiol1pAmtahcmers of cis- and trans-18:1 elute in two distinct peaks [18]. Theseparation obtained in the present study represents neverthelessa significant result since this resolution is usually achieved with a50–60 m highly polar capillary column. However, accurate sep-

inoleic (18:2 n-6) acids. Separation has been achieved by fast gas–liquid chro-atography using a 10-m BPX-70 column (see Section 2 for experimental

etails).

Fig. 1 illustrates the results obtained with the operating con-itions developed in the present study. The chromatogram shows

hat resolution of methyl butyrate (4:0) from the solvent peakould be achieved even with a 10 m column by starting tem-erature programming at 50–60 ◦C. The resolution of the peaksetween stearic (18:0) and linoleic (18:2 n-6) acids is not perfect

ig. 3. Resolution of fatty acid methyl esters (FAMEs) derived from (A) cocoautter and (B) tuna oil achieved by fast gas–liquid chromatography using a 10-mPX-70 column (see Section 2 for experimental details).

Fpcdlt

Chromatogr. A 1169 (2007) 175–178 177

ut sufficient to quantify the sum of the different geometricalnd positional of octadecenoic acid (total 18:1 acid isomers).n amplified view of this critical chromatographic region isrovided in Fig. 2. Trans isomers of octadecenoic acid rang-ng from trans-4 to trans-11 are resolved from oleic acid. Aase resolution between trans-4 to trans-11 18:1 acid isomers isseful to calculate the total trans-18:1 isomers when combinedith other chromatographic techniques [13]. However, trans-12

o trans-16 18:1 acid isomers overlaps with their positional cissomer homologues. There is a small overlap between minor cissomer of octadecenoic acid and positional/geometrical isomersf octadecadienoic acid also occurred (see Fig. 2). Better reso-ution and therefore accurate quantification of almost all trans8:1 fatty acid (TFA) isomers could be achieved using highlyolar capillary column having at least 100 m length [13–17].pre-separation by Ag-TLC or high-performance liquid chro-atography is required to avoid overlap between some cis- and

rans-18:1 acid isomers [17]. Class separation between transnd cis isomers of octadecenoic acid could also be achieved atigh temperature (up to 230 ◦C) under specific chromatographiconditions (long and very polar capillary column), positional iso-

ig. 4. Resolution of fatty acid methyl esters (FAMEs) derived from (A) ratlasma lipids and (B) rat red blood cell lipids achieved by fast gas–liquidhromatography using a 10-m BPX-70 column (see Section 2 for experimentaletails). DMA stands for dimethylacetates. DMA are formed during the methy-ation of plasmalogens; tentative identification of the DMA residues accordingo retention time gives: DMA 16:0, DMA 18:0 and DMA 18:1.

1 J. Ch

ar

utihthdf(staaaa

fof(fms

4

aTormr

bd

A

p

R

[

[[

[

[[

78 F. Destaillats, C. Cruz-Hernandez /

ration of cis/trans isomers in milk fat is not required to conductoutine fatty acid analysis.

Dairy fat is one the most complex natural dietary fat and hasnique nutritional and physical properties [19]. It was reportedhat dairy fat contains about 30–40 major fatty acids and a myr-ad of minor fatty acids such as branched-chain, oxo-, keto-,ydroxy-fatty acids for a total of 400 fatty acids [19]. Separa-ion of the major fatty acids present in milk fat in few minutesas never been achieved so far. The experimental conditionseveloped for milk fat are suitable to analyze almost all edibleats and oils. As example, cocoa butter (Fig. 3A) and tuna oilFig. 3B), two very different matrices having composition muchimpler than milk fat have been analyzed. The resolution ofhe FAMEs, including long-chain polyunsaturated FAMEs suchs methyl eicosapentaenoate (EPA), docosapentaenoate (DPA)nd docosahexaenoate (DHA), obtained are suitable to performccurate integration of the signals for milk fat which was hardlyccomplished previously [7,8].

In the framework of clinical and pre-clinical research,atty acid profiles of plasma samples are often performed inrder to follow fatty acid metabolism [2]. Analysis of theatty acid profile of plasma (Fig. 4A) and red blood cellsFig. 4B) were performed using the GLC condition optimizedor milk fat. As illustrated in Fig. 4, the results show that theethodology developed could also be used to analyze such

amples.

. Conclusion

The emergence of polar fast columns allows performing GLCnalysis of complex fatty acid composition such as milk fat.hese columns could be used also to perform routine analysis

f edible fats and oils and biological samples such as plasma ored blood cell. It is predictable that more columns will be com-ercially available worldwide in the near future. The obtained

esults support the idea that fast GLC analysis of fatty acid will

[[

[[

romatogr. A 1169 (2007) 175–178

e the standard procedure used in lipid laboratory in the nextecade.

cknowledgements

We are grateful to I. Masserey-Elmelegy for technical sup-ort.

eferences

[1] C.A. Cramers, P.A. Leclercq, J Chromatogr. A 842 (1999) 3.[2] F. Destaillats, M. de Wispelaere, F. Joffre, P.A. Golay, B. Hug, F. Giuffrida,

L. Fauconnot, F. Dionisi, J Chromatogr. A 1131 (2006) 227.[3] A. Masood, K.D. Stark, N. Salem, J. Lipid Res. 46 (2005) 2299.[4] Bondia-Pons, S. Morera-Pons, A.I. Castellote, M.C. Lopez-Sabater, J.

Chromatogr. A 1116 (2006) 204.[5] Bondia-Pons, C. Molto-Puigmartı, A.I. Castellote, M.C. Lopez-Sabater, J.

Chromatogr. A 1157 (2007) 422.[6] I. Bondia-Pons, A.I. Castellote, M.C. Lopez-Sabater, J. Chromatogr. B 809

(2004) 339.[7] L. Mondello, P.Q. Tranchida, R. Costa, A. Casilli, P. Dugo, A. Controneo,

G. Dugo, J. Sep. Sci. 26 (2003) 1467.[8] L. Mondello, P.Q. Tranchida, P. Dugo, G. Dugo, J. Pharm. Biomed. Anal.

41 (2006) 1566.[9] L. Mondello, A. Casilli, P.Q. Tranchida, R. Costa, B. Chiofalo, P. Dugo, G.

Dugoa, J. Chromatogr. A 1035 (2004) 237.10] W. Boers, C. Bauer-Plank, T. van de Broek, G. Duchateau, Lipid Technol.

12 (2000) 87.11] P. Sandra, F. Davis, J. Chromatogr. Sci. 40 (2002) 248.12] P.A. Golay, F. Dionisi, B. Hug, F. Giuffrida, F. Destaillats, Food Chem. 101

(2007) 1115.13] C. Cruz-Hernandez, Z. Deng, J. Zhou, A.R. Hill, M.P. Yurawecz, P. Del-

monte, M.M. Mossoba, E.R. Dugan, J.K.G. Kramer, J. AOAC Int. 87 (2004)545.

14] F. Ulberth, M. Henninger, J. Dairy Res. 61 (1994) 517.15] R.L. Wolff, C.C. Bayard, J. Am. Oil Chem. Soc. 72 (1995) 1197.

16] D. Precht, J. Molkentin, Int. Dairy J. 6 (1996) 791.17] F. Destaillats, P.A. Golay, F. Joffre, M. de Wispelaere, B. Hug, F. Giuffrida,

L. Fauconnot, F. Dionisi, J. Chromatogr. A 1145 (2007) 222.18] R.H. Thompson, J. Chromatogr. Sci. 35 (1997) 536.19] R.G. Jensen, J. Dairy Sci. 85 (2002) 295.