Development of an Off-line LC×LC/MS Method for Identification of Triacylglycerols (TAGs) in Marine Organisms using Different Extraction Procedures

By: Marco Beccaria1 and Luigi Mondello1,2,
1Chromaleont s.r.l., c/o “Scienze del Farmaco e Prodotti per la Salute” Department, University of Messina, viale Annunziata, 98168 Messina, Italy
2“Scienze del Farmaco e Prodotti per la Salute” Department, University of Messina, viale Annunziata, 98168 Messina, Italy


The sea and its living organisms constitute samples of high interest in environmental analysis. Knowledge of the chemistry of the sea and marine species can provide reliable information about pollution (through direct analysis of the water or measurement of bioaccumulated contaminants), nutritional value and toxicology of marine organisms that enter the human food chain, and metabolic pathways of biological interest or that impact the biosystem1.

Fish fat includes different classes of lipid, including triacylglycerols (TAGs), phospholipids (PLs), sphingolipids (SPLs), waxes, and sterols. Pioneer epidemiological studies on the low incidence of heart disease in Eskimo populations presented scientific evidence of beneficial effects of fish lipids on human health2. The impact of lipid intake and metabolism on human health, especially with regard to cardiovascular disease, continues to be an important area of biomedical research.

Goals of the Study

The objective of this work was to develop suitable extraction methodologies for the isolation of lipids from fish, that could also be applied to mollusks from the Mediterranean Sea, and their successive analysis by means of advanced chromatographic instrumentation. More specifically, three different sample preparation methodologies were studied: Bligh and Dyer's, Folch's, and solvent maceration.

Overview and Approach

Triacylglycerols were analyzed by an off-line combination of silver-ion liquid chromatography with non-aqueous reversed-phase liquid chromatography (NARP). Due to the large number of possible combinations of fatty acids on the glycerol backbone, the determination of TAG profile represents a very challenging task. An effective solution has been demonstrated to be the use of both silver ion liquid chromatography (Ag+-LC), which provides separation mainly on the basis of the degree of unsaturation (number of double bonds, DBN), and nonaqueous reversed-phase liquid chromatography (NARP), which provides separation mainly on the basis of hydrophobicity according to increasing partition number (PN). PN is defined as the total number of carbon atoms (CN) minus, twice the DBN. Mass spectrometric (MS) detection greatly supported the identification procedure. In particular, with respect to HPLC, MS with atmospheric pressure chemical ionization (APCI) is considered the most convenient and effective detector for the separation of TAGs3.

Materials and Methods

Fish Samples
Fish under investigation were sea bass (Dicentrarchus labrax) and gilthead bream (Sparus aurata), either purchased at the local fish market or collected from the wild in the Strait of Messina (Mediterranean Sea).

First Dimension (1D) LC: Ag+-HPLC-UV-ELSD
For the first dimension analysis, a Shimadzu™ Prominence liquid chromatography system (Shimadzu, Japan), equipped with two LC-20AD pumps, a CBM-20 A controller, a DGU-20A5 degasser, a SIL-20AC autosampler, a FRC-10ADvp fraction collector, was used. The diode array detector was a Shimadzu SPD-M20A, set at λ = 210 nm. The Shimadzu evaporative light scattering detector (ELSD) was set at the following conditions: 50 °C; gain: 7 mV; sampling frequency: 2.5 Hz, nebulizer gas: nitrogen (250 kPa). The column used was a Nucleosil® SA (sulfonic acid), 25 cm × 4.6 mm I.D., 5 µm particles. For the argentation procedure, the method reported by Christie was followed4. The following linear gradient of increasing acetonitrile (B) percentages in A (0.5% butyronitrile (BCN) in hexane) was run at a mobile phase flow rate of 0.8 mL/min: 0 min, 0% B; 100 min, 100% B. All of the lipidic extracts obtained by means of each extraction methodology were filtered through 0.45 µm nylon membranes prior to Ag+-HPLC-ELSD analysis.

Second Dimension (2D): NARP-HPLC-APCI-MS
For the second dimension analysis, a Shimadzu Prominence LC-20A System with CBM-20A controller, two LC-20AD dual-plunger parallel-flow pumps, a DGU-20A5 degasser, was used. The detector was an LCMS-2020 mass spectrometer, APCI(+). Mass range: 400-1100 m/z; scan speed: 4000 amu s-1; nebulizing gas (nitrogen) flow rate: 2.5 L min-1; event time: 0.2 s; detector voltage: 1.5 kV; interface voltage: 4.5 kV; interface temperature: 400 °C; CDL: 250 °C; heat block temperature: 200 °C. The column used was a Fused-Core® Ascentis® Express C18, 15 cm × 4.6 mm I.D., 2.7 µm particles. The following linear gradient of increasing IPA (B) percentages in acetonitrile (A) was run at a mobile phase flow rate of 1 mL/min: 0 min, 0% B; 50 min, 70% B (hold for 5 min); 56 min, 0% B. The Shimadzu LabSolution software ver. 5.10.153 was used for data collection and handling. The 1D dried fractions were dissolved in 200 µL of acetone and 5.0 µL of this solution was injected in the second dimension.

Sample Preparation

Extraction of lipids from fish tissue took place by means of three different extraction procedures.

1) Bligh and Dyer Method
A 10 g sample of fish tissue were ground in a mortar with 10 mL of chloroform and 20 mL of methanol. The ground mixture was suspended again in 10 mL of chloroform and 10 mL of distilled water and stirred. The mixture was filtered through filter paper and allowed to stand. Finally, the solution was centrifuged for 15 min at 3,000 rpm. The bottom layer was collected and transferred into a rotary evaporator. For an exhaustive extraction, the upper layer was subjected again to all the steps described above.

2) Folch Method
A 10 g sample of fish tissue was ground in a mortar with 67 mL (20:1 solvent/tissue parts) of chloroform:methanol (2:1). The mixture was then placed in an ice bath and stirred for 30 minutes. The suspension was transferred to a separatory funnel and agitated for a five minutes. The upper layer was collected and centrifuged for 15 min at 3,000 rpm. Finally, the bottom layer was evaporated to dryness and the upper layer subjected again to the extraction process two additional times.

3) Solvent maceration
A 10 g sample of fish tissue was ground in a mortar with 100 mL of n-hexane and allowed to stand for two hours. The hexane layer was collected. This procedure was carried out three times. The three hexane fractions were combined and evaporated to dryness in the rotary evaporator.

Results and Discussion

The lipidic extracts, previoulsy analyzed by GC, were subjected also to HPLC analysis in order to investigate the triacylglycerol (TAG) composition and some preliminary data are shown in this report. To this aim, the primary column eluate was fractionated every five minutes prior to being injected onto the secondary column. Twodimensional LC chromatography, by means of an off-line combination of silver ion chromatography with non-aqueous reversed-phase (NARP) chromatography, was applied to the determination of TAGs. In argentation liquid chromatography the elution order of TAG species depends upon increasing number of double bonds. Retention occurs because of the stability of the complex formed between Ag+ and π electrons of double bonds present in acyl chains. On the other hand, NARP-LC has demonstrated to be the most suitable elution technique for the separation of TAGs. In this case, separation of TAGs occurs on the basis of hydrophobicity, namely by partition number (PN). The combination of these two elution techniques provides high orthogonality, which leads to a dramatic increase of the separation power of the LC system. In the case of lipidic matrices from marine species, boosting the system’s selectivity becomes mandatory due to the very high complexity of this type of matrix. As shown in Figure 1, Ag+-LC was used as first dimension to achieve a rough separation of the fish extract. Sixteen fractions were then isolated from the first dimension separation. Each fraction was then evaporated to dryness and further re-injected into the second dimension (NARP-LC). Compared to GC, the LC evaluation of the different extracts, obtained with Folch, Bligh and Dyer, and maceration procedures, did not highlight any relevant differences in the corresponding chromatographic profiles, neither for fish nor for mollusks (data not shown). Once injected into the second dimension, each fraction generated unique but still complex 2D chromatograms.

1D-Ag+-LC-ELSD Analysis of Gilthead Bream (Sparus aurata) Extracted by Folch, Bligh and Dyer, and Maceration Methods

Figure 1. 1D-Ag+-LC-ELSD Analysis of Gilthead Bream (Sparus aurata) Extracted by Folch, Bligh and Dyer, and Maceration Methods

CONDITIONS: column: Nucleosil® SA, 25 cm × 4.6 mm I.D., 5 µm (50174-U); mobile phase: [A] 0.5% BCN in hexane; [B] acetonitrile; gradient: 0 to 100% B in 100 min; flow rate: 0.8 mL/min; detector: ELSD; sample: fish tissue extracted by Folch, Bligh and Dyer, and maceration methods


Mass spectrometry with APCI (APCI+) in positive ionization mode, along with the predictability given by the retention behaviour of analytes in NARP, permitted the successful identification of all the lipid species separated. In the analysis of TAGs, mass spectrometry is considered the technique of choice for unravelling such a highly complex matrix. It has also been reported to be the best tool for the analysis of regioisomers. Moreover, within mass spectrometry, the most convenient ionization technique is represented by APCI+, due to the nonpolar nature of analytes and its sensitivity. The application of APCI+ to the column eluate produces pseudomolecular ions, plus diglyceride ions, that result from the cleavage of the acyl chains from the glycerol backbone. Peak assignment is finally performed through a combination of interpretation of mass spectra, evaluation of fragment masses and their possible combinations as di- or triacylglycerols, and predictability given by the elution pattern in NARP5. The off-line comprehensive approach, beyond increasing the system’s peak capacity, produced clearer mass spectra thus a more reliable LC/MS analysis.

Figure 2 shows how a preliminary separation in the first dimension (Ag+-LC) improves noticeably the quality of the LC analysis of TAG species. At the top, the NARP-LC-APCI-MS chromatogram of a whole gilthead bream lipidic extract is presented. Below this the 2D chromatograms obtained from some of the isolated fractions are displayed. The complexity of the original sample is evident in the upper chromatogram, showing likely numerous co-elutions that are to great extent resolved in the successive 2D runs of the single fractions.


NARP-LC-APCI-MS Analysis of Gilthead Bream Extract following the Folch Method

Figure 2. NARP-LC-APCI-MS Analysis of Gilthead Bream Extract following the Folch Method
CONDITIONS: column: Ascentis Express C18, 15 cm × 4.6 mm I.D., 2.7 µm (Product No. 53829-U); mobile phase: [A] acetonitrile; [B] 2-propanol (IPA); gradient: 0 to 70% B in 50 min; held at 70% B for 5 min; to 0% B in 1 min; flow rate: 1 mL/min; detector: MS, APCI(+), mass range 400–1,100 m/z; injection: 5 µL; sample: 1D dried fractions were dissolved in 200 µL of acetone


In Figure 3, a mass scan (#14,174) at 35.43 min of 2D NARP-LC-APCI-MS analysis of Sparus aurata is reported. More components with the same PN number (44) are clearly coeluting at this elution time; in fact, as can be seen in the mass spectrum, five different m/z values, namely 830 (3 DB), 856 (4 DB), 882 (5 DB), 908 (6 DB), and 934 (7 DB), corresponding to the protonated molecules, are present. From these, a great number of fragment ions are generated, corresponding to different diacylglycerol species (Table 1). On this basis, more than 20 different TAGs could be tentatively identified, through all the possible combinations of such fragments with the observed protonated molecules, not considering the occurrence of positional isomers. In order to elucidate this issue, Table 1 reports all of the TAGs and DAGs derived from the combination of parent and daughter ion information with data obtained from GC investigation. TAGs and their fragmentation DAGs reported in Table 1 were determined in fractions from #2 to #5 of the fish extract. The multidimensional LC/MS approach allowed to discard 19 false positive TAGs.

The beneficial effects of the gain in resolution attained by the off-line Ag+-LC×NARP-LC are pointed out in Figure 4 which shows the same scan (#14,174) in more fractions (from #3 to #5). TAGs coeluted in monodimensional NARP-LC/MS were separated in the off-line multidimensional approach. However, the off-line LCxLC data relative to TAGs determination are only preliminary and part of a wider study still underway.


Mass Scan Number 14,174 (35.43 min) from Monodimensional NARP-LC-APCI(+)-MS Analysis of Gilthead Bream

Figure 3. Mass Scan Number 14,174 (35.43 min) from Monodimensional NARP-LC-APCI(+)-MS Analysis of Gilthead Bream


Table 1. TAGs and DAGs Derived from the Combination of Parent and Daughter Ion Information with Data Obtained from GC Investigation


Table 2. Fatty Acid Methyl Esters Determined in Fish (Sparus aurata and Dicentrarchus labrax)



Mass Scan Number 14,174 (35.43 min) from 2D NARP-LC-APCI-MS Analysis of Fractions 3, 4, and 5 of Gilthead Bream Isolated by 1D Silver Ion

Figure 4. Mass Scan Number 14,174 (35.43 min) from 2D NARP-LC-APCI-MS Analysis of Fractions 3, 4, and 5 of Gilthead Bream Isolated by 1D Silver Ion

Concluding Remarks and Future Perspectives

Three main conclusions can be derived from this study:

  • The application of the three different extraction lipid methodologies, Folch, Bligh and Dyer, and solvent maceration, did not result in any relevant qualitative differences in the fish tissue TAG profiles by using LC/MS
  • Due to their significantly different selectivites and high orthogonality, the exploitation of Ag+-LC (on a sufonic acid stationary phase) and NARP-LC (on a C18 phase) as two dimensions of a triacylglycerol separation, permitted the determination of a very complex matrix with a considerably high number of coelutions
  • Applying a combination of retention data with atmospheric pressure chemical ionization (APCI) mass spectral information succeeded in identifying numerous distinct TAG species with high predictability

Legal Information

Acrodisc is a registered trademark of Pall Corporation
Ascentis, Nucleosil are registered trademarks of Sigma-Aldrich Co. LLC
Fused-Core is a registered trademark of Advanced Materials Technology, Inc.





  1. van der Oost, R.; Beyer, J.; Vermeulen, N. P. Environ. Fish bioaccumulation and biomarkers in environmental risk assessment: a review. Toxicol. Pharmacol. 2003, 13, 57-149.
  2. Bang, H. O.; Dyerberg, J.; Nielsen, A. B. Plasma Lipid and Lipoprotein Pattern inGreenlandic West-Coast Eskimos. The Lancet, 1971, 297, 1143-1146.
  3. Holcapek, M.; Dvorakova, H.; Lisa, M.; Giron, A. J.; Sandra, P.; Cvacka, J. Regioisomeric analysis of triacylglycerols using silver-ion liquid chromatography–atmospheric pressure chemical ionization mass spectrometry: Comparison of five different mass analyzers. J.Chromatogr., A. 2010, 1217, 8186-8194.
  4. Christie, W. W. A stable silver-loaded column for the separation of lipids by high performance liquid chromatography. J. High Resolut. Chromatogr. Chromatogr. Commun. 1987, 10, 148–150.
  5. Dugo, P.; Beccaria, M.; Fawzy, N.; Donato, P.; Cacciola, F.; Mondello, L. Mass spectrometric elucidation of triacylglycerol content of Brevoortia tyrannus (menhaden) oil using non-aqueous reversed-phase liquid chromatography under ultra high pressure conditions. J.Chromatogr., A. 2012, 1259, 227–236.
  6. FAMEs, Fatty Acid Methyl Esters: Mass Spectral Database, Shimadzu Europe, Duisburg, Germany, 2013.


Related Links