HPLC is the most efficient method to separate  and estimate the amount of each glycoglycerolipid in an unknown sample. The use of a light-scattering detector allows the direct quantification of lipid compounds without  trouble from mobile-phase solvents as it could be observed with UV detection.

Only a few application of HPLC to separate glycolipids in plant extracts have been reported. The first one (Christie WW et al., J Chromatogr 1988, 436, 510) described a method to separate all polar lipids from cereal grains. Another study used a complex ternary solvent gradient system to separate glycolipids but also all lipid classes of wheat flour (Conforti FD et al., J Chromatogr 1993, 645, 83).

An efficient procedure was recently reported using optimal HPLC conditions with a silica column, a simple binary gradient system and a light-scattering detection (Sugawara T et al., Lipids 1999, 34, 1231). Although the described method is efficient to separate glycolipid classes, the phospholipids present in plant extracts may be also estimated. To shorten the analysis time and to prevent the interference of other polar lipids it appears useful to concentrate previously the whole glycoglycerolipid class on a silica gel column by acetone elution


Extraction: fresh plant samples are cut into pieces and treated to inactivate hydrolytic enzymes. After homogenization the residues with a chloroform / methanol mixture, the lipid extract is redissolved in chloroform and subjected to a separation on a silica gel column. The glycolipid fraction is evaporated and dissolved in a minimum volume of chloroform.
Column: Lichrospher Si 60 from Merck (120x4 mm, 5
mm) maintained at 40C.
Mobile phase: solution A: chloroform, solution B: methanol/water (95/5, v/v). 
Flow rate: 1 ml/min
Gradient: linear and continuous gradient from time 0 (99% A and 1% B) to time 15 min (75% A and 25% B) followed by an isocratic step of 5 min and a re-equilibration time of about 10 min with the conditions of time 0 prior the next analysis.
Detection:  glycoglycerolipids can be detected between 210 and 230 nm but a better sensitivity with a fine baseline is obtained with a light-scattering detector

A calibration curve is run with solutions allowing to inject 5 to 50 mg of each glycolipid. Calibration curves are expressed as linear or power equation according to the best obtained correlation coefficient.
If sterol glycosides or ceramide monohexosides (equivalent to cerebrosides) are present they elute at position 1or 2, respectively. With mild alkaline treatment, only two peaks corresponding to these compounds remain present.
In some lipid extracts a peak at about 16 min can be found and corresponds to a trigalactosyl diglyceride (TGDG).

An efficient HPLC separation of glycolipid fractions from oilseeds has been described using a Zorbax-Sil column and an isocratic elution of isooctane/2-propanol (1/1, v/v), the detection was done at 206 nm for 30min. A 10 min regeneration with the same solvent was done before the next analysis (Ramadan MF et al., Food Chem 2003, 80, 197-204). Glycolipids were quantified by isolation of the individual subclasses, followed by hexose measurement using the phenol/sulfuric acid method (Southgate DRT 1976, Dtermination of food carbohydates, pp 108-9, London, Applied Sc Pub). The elution times were 8 min for acylated steryl glucoside, 12 min for MGDG, 13 min for steryl glucoside, 15 min for cerebroside, 19 min for DGDG, and 21 min for sulfoquinovosyldiacyl glycerol.

The simultaneous quantification of glycoglycerolipids including sulfoquinovosyldiacylglycerol may be efficiently done by HPLC using a diol column and a light scattering detector (Yunoki K et al., Lipids 2009, 44, 77). The detection limit for SQDG was 0.1
mg per injection.