ANALYST. MAY 1990. VOL. 115 521

Identification of Triacylglycerols by High-performance Liquid Chromatography - Gas - Liquid Chromatography and Liquid Chromatography - Mass Spectrometry*

N. W. Rawle Dalgety PLC, Group Research Laboratory, Station Road, Cambridge CB7 2JN, UK R. G. Willis and J. D. Baty Department of Biochemical Medicine, Ninewells Hospital and Medical School, Dundee DD 7 SSY, UK

Triacylglycerols from rat adipose tissue were chromatographed by high-performance liquid chromatography (HPLC), with a gradient of propan-2-01 in acetonitrile as the mobile phase. Fractions of the material eluting from the column were collected and analysed by automated gas - liquid chromatography of the fatty acid methyl esters obtained after transmethylation. Triacylglycerols were identified by using a combination of their fatty acid content and elution time from the HPLC column. Fractions corresponding to whole peaks or groups of peaks were also collected and re-chromatographed on a liquid chromatography - mass spectrometry system equipped with a belt interface. For most triacylglycerols, good agreement was obtained between the two methods, although mass spectrometric identification of the early eluting peaks was complicated by poor resolution of the triacylglycerols on the HPLC system.

Keywords: Triacylglycerol; high-performance liquid chromatography; gas - liquid chromatography; mass spectrometry

The study of triacylglycerols poses several problems for the analyst. In natural triacylglycerol mixtures, the number of possible molecular species is large, so that a single-step analysis is not usually sufficient for complete analysis. A triacylglycerol mixture may be analysed by hydrolysis fol- lowed by analysis of the component fatty acids by gas - liquid chromatography (GLC). 1-2 This provides no information about the distribution of the fatty acids between the triacyl- glycerols.

Some workers have used GLC to separate intact triacyl- glycerols,"4 but high temperatures are required to volatilise them. This can cause an excessive amount of stationary phase bleed from the column, particularly when relatively polar phases are used. Less polar phases are more thermally stable, but are less selective towards unsaturated triacylglycerols. When GLC is used in conjunction with mass spectrometry, this may cause problems in the identification of triacyl- glycerols which co-elute.5

The most popular method for the scparation of triacyl- glycerols has been reversed-phase high-pcrformance liquid chromatography (HPLC) with non-aqueous solvents used for the mobile phase.611 Thc major problem in the analysis of triacylglycerols by HPLC is that of detection. I ? Triacylglyccrol molecules do not posscss a strong chromophore and cannot be dcrivatiscd without disruption of the original structure and consequent loss of information. Ultraviolet (UV) dctectionl3 at short wavelengths (1YO-230 nm) and infrared detection14 have been used to monitor triacylglycerols. In both instances, detection is based on absorbance by carbonyl and isolated double bonds. Both detection systems limit the choice of solvents that can be used in the mobile phase and exhibit different responses towards saturated and unsaturated triacylglycerols. Base-line drift may occur during the opera- tion of gradients. The refractive index detector suffers from poor sensitivity towards triacylglycerols and cannot be used in conjunction with gradient elution. Systems which involve the removal of solvent, such as flame ionisation" and mass detection, '(3.17 can be applied to triacylglycerol analysis

-i' Prcccnted at SAC 89, the 8th SAC International Conferencc on Anal) tical Chemistry. Cambridge. UK, 30 July-S August, 1989.

without the occurrence of base-line drift during gradient elution. Further information about the structure of the triacylglyccrols may be obtaincd by transmethylating triacylglycerols that havc been separated by HPLC and analysing thc resulting fatty acid mcthyl esters by GLC.11

Gas chromatography - mass spectrometry (GC - MS)"lX and liquid chromatography - mass spectrometry (LC - MS)lY.*O have been used in the analysis of triacylglycerol mixtures. In the mass spectrometer, under electron impact conditions, triacylglyccrols produce characteristic ions,zI corresponding

128]+ and [RCO]+, where R represents the acyl chains of the component fatty acids of the triacylglycerol.

We arc interested in studying the distribution of the fatty acids amongst the triacylglyccrols in adipose tissue. Triacylglycerols which have been separated by HPLC are analysed further by two separate procedures. In onc instance the triacylglyccrols are analysed, as their component fatty acid methyl esters, by GLC. In the other they are analysed by LC - MS. In this paper the results obtained by using thc two mcthods are compared.

to [A4 - RCOO]+, [M - RCOOH]+, [RCO + 74]+, [RCO +

Experimental Reagents

Triacylglycerol and fatty acid methyl ester reference standards were purchased from Sigma (Poole, Dorset, UK). All solvents wcre of analytical-reagent or HPLC grade and supplied by BDH (Poole, Dorsct, UK) or Rathburn Chemicals (Walkcr- burn, Peeblesshire, UK).

HPLC

High-performancc liquid chromatography was performed with an Altcx l0OA dual-piston pump and Altex l1OA pump, controlled by an Altex 420 programmer (Altcx Scientific, Bcrkcley, CA, USA). A Rhcodyne 7126 injection valve (Rhcodync I , Cotati, CA, USA) with a 20-4 loop and a column (25 cm x 4.6 mm i.d.) of 3 pm Spherisorb ODS 2 were used. The mobile phase was propan-2-01 - acetonitrile (32 + 68) for SO min, then to propan-2-01 - acetonitrile (60 + 40) in 30 min, at a flow-rate of 0.8 ml min-1. Ultraviolet detection

Publ

ishe

d on

01

Janu

ary

1990

. Dow

nloa

ded

by M

cMas

ter

Uni

vers

ity o

n 30

/10/

2014

12:

51:1

3.

View Article Online / Journal Homepage / Table of Contents for this issue

522 ANALYST, MAY 1990. VOL. 115

was with a Kratos Spectroflow 757 detector (Kratos Analytical Instruments, Ramsey, NJ, USA) at a wavelength of 213 nm. Fractions eluting from the column were collected with a Gilson 203 fraction collector (Gilson Medical Electronics, Villiers-le-Be1 , France), which was controlled by a Hewlett- Packard 3388A integrator.

GLC

A Hewlett-Packard 7672A autosampler was used to inject fatty acid methyl ester solutions into a Hewlett-Packard 5710A gas chromatograph, equipped with a flame-ionisation detector. A glass column (2 m X 2 mm i.d.), packed with 10% SP-2330 on 100-120 Chromosorb W AW (Supelchem, Saw- bridgeworth, Hertfordshire, UK) was used. The carrier gas was nitrogen at a fow-rate of 25 ml min-1 with an oven temperature of 155 "C. Autosampler control and data handling were performed by a Hewlett-Packard 3388A integrator.

LC - MS

Liquid chromatography - mass spectrometry was carried out with a PU 4003 HPLC system (Philips Analytical, Cambridge, UK) containing a column (25 cm x 2.0 mm i.d.) of 5 pm Spherisorb ODS 2, with acetone - acetonitrile (63.7 + 36.3) at a flow-rate of 300 pI min-1 as the mobile phase. Injection was via a septum injector and the HPLC system was linked to a VG 7070F double focus mass spectrometer (VG Organic, Altrin- cham, Cheshire, UK) via a belt interface. Resolution was 1000 (10% valley). A source temperature of 260 "C, an electron energy of 20 eV and an accelerating voltage of 4 kV were used. Ions were scanned from 640 to 195 u at 1 s decade-'.

Procedure

A sample (ca. 50 mg) of adipose tissue was removed from an epididymal fat pad of a male Sprague - Dawley rat. The sample was homogenised in 15 ml of chloroform - methanol (2 + 1) and passed through a 0.45-pm nylon filter (Gelman Sciences, Northampton, UK). The extract was dried under a stream of nitrogen and re-dissolved in chloroform (2 ml). A portion (400 pl) of the extract was applied to a silica gel 60 thin-layer chromatography plate (BDH) which was developed with hexane - diethyl ether - acetic acid (85 + 25 + 1). The triacylglycerols were recovered from the plate with chloro- form - methanol (2 + 1) and reconstituted in acetone. Approximately 0.5 mg of triacylglycerol, dissolved in 20 pl of

acetone, was injected into the HPLC system. Fractions eluting from the detector were collected for further analysis by GLC or LC - MS.

GLC Fractions of the material eluting from the HPLK column were collected at 0.2-min intervals. This interval was increased in stages to a final value of 1 .O min, to take band broadening into account. After addition of triheptadecanoin (16 pg) as internal standard for the GLC, each fraction was re-dissolved in diethyl ether (1 ml) and transesterified with sodium methoxide according to the method of Christie.2' The resulting fatty acid methyl esters wcre dissolved in dodecane (150 PI), 1 p1 of which was analysed by automated packed-column GLC.

L C - M S Fractions corresponding to whole peaks or groups of peaks were collected from the HPLC column and dried under a stream of nitrogen. Each fraction was re-dissolved in 200-300 p1 of acetone and a portion (10 p1) of this solution injected into the LC - MS system.

Results and Discussion The HPLC trace of the triacylglycerols obtained with U V detection at 213 nm is shown in Fig. 1. The base-line drift occurring after 60 min, caused by the operation of the gradient elution, demonstrates the difficulty of quantification of triacylglycerols when short wavelength UV detection is used.

With HPLC - GLC, triacylglycerols wcre identified by a combination of their elution time from the HPLC column, compared with those of standard compounds, and their fatty acid content. For the identification of triacylglycerols by LC - MS, a reference library of 165 triacylglycerol mass spectra was compiled from published data21 for the fragment ions corresponding to [ M - RCOO]+, [ M - RCOOH]+, [RCO + 74 I + , [RCO + 128]+ and [RCO]+. Triacylglycerols were therefore identified by a combination of their elution time from the HPLC and a comparison of their mass spectra with the library data.

The identities assigned by the two methods to the major triacylglycerols contained in the fractions are listed in Table 1. For fractions 2-8, good agreement between the two methods was achieved. For fraction 1, the LC - MS method wrongly identified the triacylglycerols contained in this fraction. This occurred because the second peak of fraction 1 contained a mixture of four triacylglycerols, namely LLL, PeLL, PePeL and PePePe. The resulting mixed spectrum caused the peak to

I I I I I

0 20 40 60 Time/m in

80

Fig. 1. 1-8 were collected for further analysis by LC - MS

Chromatogram of rat adipose tissue triacylglycerols obtained by HPLC with UV detection at 223 nm. For conditions see text. Fractions

Publ

ishe

d on

01

Janu

ary

1990

. Dow

nloa

ded

by M

cMas

ter

Uni

vers

ity o

n 30

/10/

2014

12:

51:1

3.

View Article Online

ANALYST. MAY 1990, VOL. 115 523

Table 1. Identities given to the major triacylglycerols present in fractions 1-8 by further analysis of the component fatty acid methyl cstcrs by GLC or intact triacylglycerols by LC - MS

GLC of fatty acid LC - MS of intact Fraction methyl esters* triacylglycerols

2

I MMM PePePe PcPcL PCLL LLL OLL PLL

PeOL PPeL

3 OOL 4 PLO 5 PPL 6 000

P O 0 7 PPO 8 PPP

POS

MMM PPePe OLLn

PeOL PPeL

OOL PLO PPL 000 P O 0 PPO PPP POS

* M = Myristic acid. P = palmitic acid. Pe = palmitoleic acid, S = stearic acid, 0 = oleic acid, L = linoleic acid, Ln = linolenic acid and POS is a triacylglyccrol composed of palmitic. oleic and stearic acids. No distinction is made here between the three fatty acid positions of the triacylglycerols.

be wrongly identified as OLLn or PPePe. It is known that these triacylglycerols do not elute at this point, as the HPLC - GLC analysis of this fraction shows palmitic and oleic acids to be absent from these triacylglycerols.

This work shows the advantages of HPLC - GLC and LC - MS for the analysis of triacylglycerols over those methods which use more traditional modes of detection. Both methods allow identification of triacylglycerols, such as tripalmitin, which have only a very small response when UV detection is used. Because the mobile phase is removed prior to detection, base-line drift is not a problem when gradient elution is used. These methods provide structural information about the triacylglycerols, although no distinction is made between those molecules which are positional isomers.

When LC - MS is used, caution must be exercised in the interpretation of mass spectra, as incorrect identities can be assigned to triacylglycerols which elute together.

1.

2.

3.

4. 5 .

6. 7. 8.

9.

10.

11. 12.

13.

14.

15.

16.

17.

18. 19.

20.

21.

22.

References Jacobsen, B. K.. Trygg, K., Hjermann, I . . Thomassen. M. S. , Rcal, C., and Norum, K. R., A m . J . Clin. Nutr., 1983,38,906. Plakke, T., Berkel. J . , Beynen, A. C., Hermus, R. J . J., and Katan, M. B.. Hum. Nutr. Appl . Nutr., 1983. 37A, 365. Mares. P., Skorepa, J . , Sindclkova, E., and Tvrzicka. E. , .I. Chromutogr., Biomed. Appl., 1983, 273, 172. Hamilton, R. J. , J . Chromatogr. Sci.. 1975, 13, 474. Oshima, T.. Yoon, H.-S., and Koizumi. C., lipid^, 1989, 24, 53.5. Aitzetmiiller, K., Prog. LipidKes. , 1982, 21, 171. Plattner, R. D. , Methods Enzymol . , 1981, 72, 21. El-Hamdy, A. I]. , and Perkins, E. G., J . A m . Oil Chem. Soc., 1981, 58, 49. El-Hamdy, A . H. , and Perkins, E. G . , J . A m . Oil Chem. Soc., 1981. 58, 867. Dong, M. W., and DiCesarc, J . I , . . J . A m . Oil Chem. Soc., 1983, 60, 788. Baty, J. D., and Rawle, N . W., J . C'hromatogr., 1987,395.395. Robinson, J . L. , and Macrae, R., J . Chromutogr., 1984, 303, 386. Singleton, J . A , . and Pattee, H. W.. J. A m . Oil Chem. Soc., 1984, 61, 761. Hamilton, R. J . , Mitchell, S . F., and Sewell, P. A.. J . Chromatogr.. 1987, 395, 33. Phillips, F. C . , Erdahl, W. L., Nadenicek, J. D., Schmit, J . A., and Privctt, 0. S., Lipids, 1984, 19, 142. Stolyhwo, A., Colin, H. , and Guiochon, G.. Anal. Chem., 1985, 57, 1342. Robinson, J . L., Tsimidou, M., and Macrae, R., J . Chromat- ogr., 1985, 324, 35. Murata. T., and Takahashi, S. , Anal. Chem., 1973, 45, 1816. Myher, J. J . , Kuksis, A. , Marai, L. , and Manganaro. F., J . Chromatogr., 1981, 283, 2x9. Marai, L., Myher, J . J., and Kuksis, A . , Can. J. Biochem. Cell Biol., 1983. 61, 840. Barber. M., Merren, T. O., and Kelly, W., Tetrahedron Lett.. 1964, 18, 1063. Chriqtie, W. W., J. Lipid Res., 1982, 23. 1072.

Paper 9104758A Received November 6th, 1989

Accepted December 19th, 1989 Publ

ishe

d on

01

Janu

ary

1990

. Dow

nloa

ded

by M

cMas

ter

Uni

vers

ity o

n 30

/10/

2014

12:

51:1

3.

View Article Online