Downloaded from
`
`www.jlr.org
`
` by guest, on November 8, 2020
`
`notes on methodology I
`
`Complete separation of lipid
`classes on a single thin-layer plate
`
`C. P. FREEMAN and D. WEST
`Unilever Research Laboratory, Colworth House, Sharn-
`brook, Bedford, England
`SUMMARY A double-development procedure employing
`first a polar and then a nonpolar solvent system is described
`for the complete separation by thin-layer chromatography of
`the main lipid classes encountered in natural lipids. For better
`quantification, long plates (34 cm) are employed. Diglycerides
`were separated from cholesterol, 1,2- from 1,3-diglycerides,
`and monoglycerides from phospholipids.
`KEY WORDS
`double-development
`. quantification . lipid classes
`raphy
`
`thin-layer chromatog-
`
`A SINGLE SOLVENT system capable of resolving a mixture
`of cholesterol ester, triglyceride, diglyceride, free fatty
`acid, free cholesterol, monoglyceride, and phospholipid
`has not been reported. Nonpolar solvents, though capable
`of separating hydrocarbons, cholesterol esters, and alkyl
`esters as classes, leave more polar lipids largely unresolved
`at or near the origin. More polar lipids are generally
`separated by solvent systems of petroleum hydrocarbon
`and diethyl ether in various ratios, together with 1-2Oj,
`of glacial acetic acid (1). In the less polar range of this
`system (e.g., diethyl ether-hexane-acetic acid 25 :75 : l),
`cholesterol ester, triglyceride, fatty acid methyl ester, free
`fatty acid, and diglyceride are well resolved, but mono-
`glyceride and phospholipid remained unresolved at the
`origin. Increasing the proportion of diethyl ether in the
`solvent mixture displaces monoglyceride from phospho-
`lipid slightly, but at the expense of cholesterol ester/tri-
`glyceride resolution. The system, moreover, suffers from
`the serious disadvantage that throughout the polarity
`range free cholesterol migrates with either 1,2- or 1,3-
`diglyceride. We describe here a one-dimensional, double-
`development system which gives complete separation of
`the lipid classes most frequently encountered in biological
`systems, when long (34 cm) plates are used.
`The separated lipids have been determined by the
`dichromate-reducing method described by Amenta (2).
`A single reagent is employed and elution is unnecessary.
`All chromatographic solvents
`Materials and Methods.
`were redistilled before use. The lipid standards used
`(cholesteryl oleate, tripalmitin, 1,3-dipalmitin, palmitic
`acid, cholesterol, monopalmitin, and lecithin) were puri-
`
`fied before use and their purity was checked by TLC.
`Cholesterol after initial purification contained a very
`small amount of free fatty acid (apparent in Fig. l a and
`IC) and was rechromatographed for incorporation in the
`standard and synthetic mixtures. Standards and lipids for
`estimation were made up in chloroform-methanol 1 : 1
`(v/v) for application to the plate. The thin layers (250 p )
`of Silica Gel G (E. Merck, A.G., Darmstadt, Germany)
`were prepared on 34 X 20 cm glass plates, and the ad-
`sorbent was activated for 40 min at 105OC before use.
`The glass plates were cut from 3 mm standard plate
`glass ; the developing tanks used were specimen jars (5 X
`8% X 14% inches) fitted with ground-glass lids.
`The plates were marked out into 3 cm lanes (Fig. 2).
`The standard lipid solution was applied to one lane,
`another lane was left blank, and the unknown lipid solu-
`tions were applied to two other lanes. The solutions were
`applied to give 50-250 pg of each lipid class on the plate,
`and were pipetted onto the origin as a series of discrete
`spots across the appropriate lane. Reference spots of
`each lipid class were applied in the two outer lanes. The
`plates were developed in saturated tanks as follows.
`Solvent system 7: Diethyl ether-benzene-ethanol-acetic
`acid 40 : 50 : 2 : 0.2. The solvent front was allowed to run a
`distance of 25 cm from the origin. Development time was
`approximately 60 min. The plates were removed, air-
`dried for a few minutes, and heated briefly to remove
`traces of acetic acid. They were then transferred to
`another tank containing the second solvent.
`Solvent system 2: Diethyl ether-hexane 6 : 94. This sol-
`vent was run to within 1 cm of the top of the plate. The
`time taken for development in this system was 100-120
`min. The plates were then air-dried again and given a
`final drying at 60°C for 30 min. The development times
`given are for normal operating temperatures of about
`22°C.
`The outside reference lanes were examined under UV
`light after spraying them with a 1% aqueous solution of
`Ultraphor (Badische Anilin und Soda Fabrik A.G,,
`Ludwigshafen-am-Rhein, Germany), and zones corre-
`sponding to lipid spots were marked in the unexposed
`lanes. These zones, together with a corresponding area
`from the blank strip, were scraped from the plate on to
`squares of cellophane paper, and transferred from the
`latter to 10-ml stoppered tubes for estimation.
`The reagent and general procedure of the colorimetric
`estimation were as described by Amenta (2). We found 2
`ml of reagent, and dilution of 1 ml of the developed d u -
`tions in 10 d of water for the colorimetric reading, to be
`most suitable for the range of 50-250 pg of lipid. Ab-
`sorbances were determined at 350 mp on a Unicam SP
`500 spectrophotometer against a distilled water blank.
`
`Abbreviation : TLC, thin-layer chromatography.
`
`324
`
`OF LIPID RESEARCH VOLUME 7, 1966 Notes on Methodology
`JOURNAL
`
`AKER EXHIBIT 2002 PAGE 0001
`
`

`

`IC
`
`Downloaded from
`
`www.jlr.org
`
` by guest, on November 8, 2020
`
`FIG. 1.
`Chromatoplates illustrating lipid class separation in two solvent systems, individually
`( a and b) and in successive combination (c).
`( a ) Solvent system 1, diethyl ether-benzene-ethanol-acetic acid 40:50:2:0.2.
`(b) Solvent system 2, diethyl ether-hexane 634.
`( c ) Solvent system 1 followed by second development in solvent system 2.
`CE, cholesterol ester; TG, triglyceride; DG, diglyceride; C, free cholesterol ; FA, free fatty acid;
`MG, monoglyceride; and PL, phospholipid.
`Arrow indicates position of first solvent front.
`The four mixtures applied across each plate, from left to right, were (i) monopalmitin and
`1,3-dipalmitin, (ii) lecithin, palmitic acid, and tripalmitin, (iii) cholesterol and cholesteryl
`oleate, and (iv) a mixture of (i), (ii), and (iii).
`Load:
`approx. 50 pg per spot.
`Detection by charring after spraying with saturated solution of K2Cr20, in 80% (by wt)
`H&O+
`
`Results and Discussion. The separation achieved in
`solvent system 1 is shown in Fig. la. Cholesterol is clearly
`separated from diglyceride and monoglyceride from
`phospholipid, which remains at the origin; cholesterol
`ester and triglyceride migrate together near the solvent
`front. Solvent system 2 resolves this combination while
`scarcely affecting the other lipid classes (Fig. 16). Hydro-
`carbons, when present, migrate with the solvent front in
`the second solvent system and as a result are separated
`from cholesterol esters, though hydrocarbon/cholesterol
`ester resolution is slightly inferior to cholesterol ester/
`triglyceride resolution. Development in the reverse order
`of solvent systems gave a slightly inferior separation of
`cholesterol ester from triglyceride and a less discrete
`fatty acid zone.
`A typical chromatogram, run for the analysis of intes-
`tinal and plasma lipids of a pig, is shown in Fig. 2. The
`lipid classes separate as narrow bands and are almost
`equidistantly spaced ; location of the lipid bands that
`would normally not be sprayed with detecting reagent
`is consequently unequivocal. It is felt that this justifies the
`lengthy development times required for the long plates.
`The total development time can be considerably reduced
`by running the plates at elevated temperatures or by
`employing the more normal 20 X 20 cm plates, but
`resolution is impaired. The 20 X 20 cm plates are very
`
`useful, however, for qualitative purposes with these sol-
`vent systems.
`The linearity of response to dichromate reduction was
`verified over the working range 50-250 pg, for each of
`the standard lipids run through the complete chroma-
`tographic procedure, in agreement with Amenta (2).
`Amounts of lipid below (down to 15 pg) or above this
`range can be estimated by adjusting the amount of
`dichromate reagent added (2) ; the proportions of reagent
`used in the method described here were chosen to cover
`a working range suitable for general application.
`A further check was made of the effect of unsaturation
`on the extent of dichromate reduction. With palmitic,
`oleic, and linoleic acids, a small but consistent decrease
`in reduction of dichromate with degree of unsaturation
`was observed. The decrease was of the order saturated:
`monoene :diene 1 : 0.98 : 0.91 , these figures being the means
`of ten determinations. When highly unsaturated lipids
`are being examined, either a standard should be chosen
`whose degree of unsaturation is similar to that of the
`lipid being estimated or the sample should be completely
`hydrogenated prior to analysis and estimated against a
`saturated standard.
`The accuracy and reproducibility of the technique
`was checked using synthetic mixtures. The results of a
`series of determinations on two different synthetic mix-
`
`JOURNAL OF LIPID RESEARCH VOLUME ?, 1966 Noies on Methodology
`
`325
`
`AKER EXHIBIT 2002 PAGE 0002
`
`

`

`I
`
`2
`
`3
`
`4
`
`5
`
`TABLE 1 T C L ANALYSES OF STANDARD LIPID CLASS MIX-
`TURES
`
`”n-
`thetic
`Mixture
`No.
`
`1
`
`2
`
`Lipid Class
`Cholesterol ester
`Triglyceride
`Dig1 yceride
`Cholesterol
`Free fatty acid
`Monoglyceride
`Phospholipid
`Cholesterol ester
`Triglyceride
`Dig1 yceride
`Cholesterol
`Free fatty acid
`Monoglyceride
`Phospholipid
`* Mean f SD.
`
`Per Cent
`Composition
`Known
`Found*
`17.1 f 0 . 9
`16.9
`18.4 f 2 . 4
`1 7 . 4
`17.5 f 1 . 6
`20.0
`13.3 f 1 . 8
`13.5
`16.6 f 0 . 9
`15.6
`17.1 f 1 . 0
`16.5
`0
`0
`11.4 f 1 . 6
`11.4
`11.8 f 1 . 3
`11.7
`18.0 f 2 . 4
`18.6
`9 . 7 f 1 . 6
`9.1
`16.9 f 1 . 3
`17.6
`21 .O 21.3 f 2 . 0
`10.8 f 1 . 6
`10.8
`
`No. of
`Deter-
`minations
`
`6
`
`10
`
`I
`1111
`
`.
`m
`
`CE
`
`TG
`
`I
`
`DG
`
`C
`
`FA
`
`M G
`
`PL
`
`II
`
`-
`
`FIG. 2. Tvpicd Livout of aiialvtical chroin,itoplatr, charred as
`described in Fiq. 1 to illustrate separation within
`lanes. Nor-
`mally. lanes 1 and 6 onlv would be spraved, as described in the text.
`1 and 6, reference strips: 2, standard lipid class mixture (100 pq
`each lipid clasc); 3, piq intestinal lipids after fat feeding (400
`pg applied); 4, blank lane; 5, pig plasma lipids (500 JIB applied).
`Abbreviations as for Fiq. 1.
`1,2diglyceride (the predominant isomer in pig intestinal content
`followinq the action of pancreatic lipase on triglyceride) is seen
`to be resolved from the faster movinq 1,3-isomer (lane 3).
`
`tures (with and without phospholipid) are shown in
`Table 1. The analyses on mixture 1 were carried out
`using solvent system 2 followed by solvent system 1 ; the
`normal order was employed for mixture 2.
`The “absolute error” (3) recorded for the analyses of
`mixture 2 (about fo.4y0) compared with the analyses of
`mixture 1 (about f 1 .OY0) reflects the improved chroma-
`tographic resolution observed when solvent system 1 is
`followed by solvent system 2, as opposed to the reverse
`
`Downloaded from
`
`www.jlr.org
`
` by guest, on November 8, 2020
`
`order. The standard deviation about the mean was quite
`consistent throughout.
`The absolute recoveries, from application to the plate
`to scraping off the silicic acid areas, were checked with
`palmitic acid-lJ4C and glyceryl tripaltnitate-l-I4C. The
`method of Snyder and Stephens (4) was used for scintil-
`lation counting of the silicic acid areas. The mean re-
`coveries (four determinations) were 95.6 and 93.6y0 for
`the labeled palmitic acid and tripalmitin respectively.
`Application of the sample to the plate from chloroform-
`methanol 1 : 1 was more reproducible and quantitative
`than application from diethyl ether solution, presuniably
`because of evaporation and consequent loss of lipid on
`the pipette tip when ether is used. As expected, the
`scraping-off operation accounted for most of the small
`loss recorded. This virtually quantitative recovery of
`label has been utilized in metabolic studies with labeled
`lipids to determine the specific activity of lipid classes;
`duplicate samples, one for mass analysis by the dichro-
`mate method and the other for counting as described
`above, were chromatographed on the same plate.
`When this procedwe is applied to extracts of naturally
`occurring lipids, caution should be used in attributing
`the dichromate reduction of material scraped from the
`origin entirely to phospholipid, since any very polar ma-
`terial in the extract will remain in this region. Marinetti
`( 5 ) subjected silicic acid impregnated papers to a very
`short preliminary development with chloroform-meth-
`anol 1 :1 (containing 2% water) in order to cause the
`polar lipids to migrate away from other material at the
`origin. In those cases where only a total phospholipid
`value is required, it may be preferable to carry out a
`phospholipid-phosphorus estimation ( 6 ) on the material
`at the origin in place of the dichromate reaction. Where a
`quantitative analysis of individual phospholipids is re-
`
`326
`
`JOURNAL OF LIPID RESEARCH VOLUME i’, 1966
`
`.Voles on Melhodolqy
`
`AKER EXHIBIT 2002 PAGE 0003
`
`

`

`Downloaded from
`
`www.jlr.org
`
` by guest, on November 8, 2020
`
`quired, it is suggested that the lipid extract is first run in
`a system suitable for phospholipid separation. Individual
`phospholipid fractions can then be scraped from the
`plate and quantified by the dichromate reduction
`method. Neutral lipids, which will be located near the
`solvent front in such systems, can then be eluted as a
`group, and rechromatographed and estimated by the
`procedure as described here.
`Manuscr ipt ieceiued 7 7 August 7965; accepted 77 November 7965.
`
`REFERENCES
`1. Mangold, H. K.. and D. C. Malins. J. Am. Oil Chemists’ Sac.
`37: 383, 1960.
`2. Amenta, J. S. J . Lipid Res. 5: 270, 1964.
`3. Blank, M. L., J. A. Schmit, and 0. S. Privetr. J. Am. Oil
`Chemists’ Sac. 41: 371, 1964.
`4. Snyder, F., and N. Stephens. Anal. Biochm. 4: 128, 1962.
`5. Marinetti, G. V. In Neu! Btochmical Separations, edited by
`A. T. James and L. J. Morris. Van Nostrand Publishing
`Co., New York, 1964, p. 347.
`6. Skipski, V. P., R. F. Peterson, and M. Barclay. Biochm. J.
`90: 374, 1964.
`
`JOURNAL OF LIPID RESEARCH VOLUME 7 , 1966 Notes on Methodology
`
`327
`
`AKER EXHIBIT 2002 PAGE 0004
`
`