ANALYSIS OF FATTY ACIDS BY HIGH-PERFORMANCE LIQUID CHROMATOGRAPHY
Abstract: Although gas chromatography is the
dominant technique for fatty acid analysis,
high-performance liquid chromatography has an important role to play in
applications such as the handling of less usual samples, avoidance of degradation of
heat-sensitive functional groups, and for micro-preparative purposes. Several
approaches for development of improved methods are suggested, especially for reversed-phase applications.
There can be little doubt that gas chromatography (GC) is the only technique
that need be considered for routine analysis of most fatty acid
samples. The flame-ionization detector is robust and has an
enormous dynamic range, so accurate quantification is rarely a
problem. Therefore, is there any place for high-performance
liquid chromatography (HPLC) for the analysis of fatty acids? The
answer is an undoubted yes, perhaps not for mainstream samples
but certainly for the less usual. A major advantage can be that
HPLC operates at ambient temperature so there is relatively
little risk to sensitive functional groups. It should also be
remembered that HPLC is not merely an analytical technique, but
can be used equally easily for micro-preparative purposes. For
example, it is easy to collect fractions for analysis by other
techniques such as by chemical degradation or by mass
spectrometry (MS) [1] or nuclear magnetic resonance spectroscopy [2].
Indeed, direct analysis by HPLC-MS with electrospray ionization may offer an advantage in terms of sensitivityTHE EVAPORATIVE LIGHT-SCATTERING DETECTOR FOR HIGH-PERFORMANCE LIQUID CHROMATOGRAPHY OF LIPIDS
Abstract: Evaporative
light-scattering detectors have many benefits for lipid analysis and a
number of commercial instruments are now available. Analysts should not
let the need for careful calibration deter them.
Lipid analysts were initially slow to come to terms
with the potential of high-performance liquid chromatography (HPLC),
largely because of the non-availability of a sensitive universal
detector. In contrast the flame-ionisation detector, commonly used in
gas chromatography (GC), is highly sensitive and exhibits a linear
response over a wide range of sample sizes. Transport-flame ionisation
detectors for HPLC have always looked promising, but have never been a
commercial success. We may wait for ever if we want an HPLC detector
with the simplicity, ease of operation and linearity of GC, and it may
cause us to overlook an extremely useful, proven and versatile HPLC
detector - the evaporative light-scattering or "mass" detector (ELSD).
Principles
With this instrument, the solvent emerging from the end of the column is evaporated in a stream of air in a heated chamber (see
Figure 1); the solute does not evaporate, but is
nebulized and passes in the form of minute droplets through a light
beam, which is reflected and refracted. The amount of scattered light
is measured and bears a relationship to the concentration of material
that is eluting. The commercial detector based on this principle with
which the author was first familiar was available at a cost comparable
to that of other optical detectors from Applied Chromatography Systems
(A.C.S.) Ltd, later taken over by Polymer Labs. Now, several different
manufacturers produce excellent instruments, including Alltech and
Sedere. There are no special wavelength requirements for the light
source, and in some commercial instruments, it is simply a projector
lamp. Such a detector can be considered to be universal in its
applicability, in that it will respond to any solute that does not
evaporate before passing through the light beam. The instrument gives
excellent results under gradient elution conditions, and it is simple
and rugged in use. The sensitivity is comparable to that of a
refractive index detector, but the evaporative light-scattering
detector is not affected by changes in the mobile phase or small
variations in the room temperature or in the flow-rate of the mobile
phase, as is the former.
Figure 1. Schematic diagram of the ACS evaporative-light-scattering detector
The latest generation of such instruments have much
greater sensitivity and improved linearity. Some of these detectors
have distinctive design features, and for example in the Cunow and
Sedere detectors, the larger droplets in the spray from the nebulizer
are condensed out before they reach the heater chamber. A consequence
is reported to be a more uniform particle size and improved linearity.
The Alltech detector has a laser light source and a photodiode detector
instead of a photomultiplier tube. Thus, there is now a good choice of
commercial instruments.
Once the instrument has warmed up and is running,
there is little base-line drift during continuous operation even with
abrupt changes in solvent composition. Most organic solvents, including
acetone and chloroform, for example, can be used, and these can contain
up to 20% water and small amounts of ionic species even. The minimum
detection limit is dependent to a certain extent on the nature of the
mobile phase, the nature of the sample, and the specific instrument,
but it is certainly less than one microgram.
As with all detectors, there are disadvantages. A
source of dry, filtered compressed air, that is capable of delivering 5
litres/min, is required and in practice, this means that an air
compressor must be used; a standard cylinder of air or nitrogen is
emptied in about 4 to 8 hours. The stream of air containing the
evaporated solvent must be conducted to the outside of the laboratory
or into a fume cupboard.
Although the detector is destructive in that the
sample is lost, it is possible to insert a stream splitter between the
end of the column and the detector to divert much of the sample to a
collection device. To my knowledge only one company offers a
stream-splitter commercially, and then at a very high price. My own is
home-made and consists merely of a low-dead-volume T-piece with
narrow-bore (0.25 mm) HPLC tubing to the detector and wider-bore (1 mm)
tubing to the collection vessel. I can add a tap to the outlet end if
required. This is simple and apparently effective, but it would be
helpful if instrument manufacturers would offer a more professional
splitter as a standard item with their equipment. I used to receive
more correspondence on this than on any other single topic in relation
to the detector.
When used in this way, the evaporative
light-scattering detector is a splendid research tool, since samples
can easily be collected for analysis by other means, e.g. for
determination of fatty acid compositions by GC. A further advantage in
research applications is that it is easy to make rapid adjustments, as
is often required during method development, since changing the solvent
has virtually no effect on the base line.
Detector Response
Many analysts appear to be deterred from using ELSDs
because of the well-established fact that the response is not
rectilinear for purely analytical applications. I find this attitude to
be short sighted. The first requisite for any analysis is that the
required compounds are adequately resolved. With an ELSD, we can use
the whole range of solvent groups described by Snyder and colleagues
[1] to optimise the selectivity of the mobile phase, and complex
gradients can be used, if need be, to achieve the desired separation.
With most other types of HPLC detector, we are restricted to a limited
number of solvents and isocratic elution. Secondly, it is essential
that we can monitor the separation. With UV detection at low
wavelengths, for example, it is not at all easy to detect saturated
lipids; there is no such problem with an ELSD.
Having obtained and proved that we have a suitable
separation, only then do we need to consider quantification. It is
known that the detector response follows the equation:
- where A is the peak area, and a and b are
numerical constants, i.e. a plot of peak area against the log of sample
concentration is linear and has the slope b. The value of b lies
between 1 and 2, and appears to depend on the design of the specific
detector, especially the nebulizer system. When the value of b is close
to 1, there can be a useful linear range of 1 to 2 decades. Righezza
and Guiochon [2] showed that the size distribution of solute droplets
formed in the aerosol is variable and depends on the nature of the
solvent (probably on the rate of vaporisation). This obviously has an
effect on response. In addition, they found that the amount of
scattered light depended strongly on the molar absorbtivity of the
solute.
Calibration
In practice, this means that it is essential to
calibrate the detector carefully for each analytical system. Once the
conditions have been optimised for a particular separation, these must
be kept constant while the calibration is carried out with lipid
standards as similar as possible in nature to the samples for analysis.
This is of course particularly important when absolute amounts of
particular components must be determined. On the other hand, when
relative proportions only of different lipids are required, as in
molecular species analysis for example, small changes in the
chromatographic conditions or in the nature of the fatty acid
constituents of the lipids have little effect on calibration. I would
not recommend an ELSD for quantitative analysis of simple fatty acid
derivatives, such as methyl esters, because of their relatively high
volatility, though this may be less of a problem with some commercial
instruments than others.
Lipid analysts have perhaps been spoiled by the ease
of using the flame ionization detector in gas chromatography, and
expect a perfect linear response in all circumstances. Most workers in
other fields realise that careful calibration is an essential part of
any analytical procedure, and that non-linear calibrations are not
unusual. What the lipid analyst must consider is which detector is any
better than an ELSD for his purpose. All alternatives are dependent on
the nature of the lipid, some greatly so, and few can be used with
gradients.
Applications
The first important applications to the analysis of
lipids were published in the early 1980s, so this is now a mature
technology. Two review articles on the use of the detector have
appeared [3,4] and a comprehensive bibliography of applications
is available on this website. To consider a few examples, one of the
more important tasks for HPLC and lipids is the separation and
quantification of the different lipid classes in tissues. Ideally, this
should be accomplished on a small scale, e.g. 0.2 to 0.4 mg, and in as
short a time as can conveniently be managed. I made use of the A.C.S.
evaporative light-scattering detector with a ternary solvent delivery
system and a short (5 x 100 mm) column, packed with SpherisorbTM
silica gel (3 micron particles) for the purpose [5]. In selecting a
mobile phase, the choice of solvents was constrained at first by the
need for sufficient volatility for evaporation in the detector under
conditions that do not cause evaporation of the solute, and by the
necessity to avoid inorganic ions, which would not evaporate. Similar
restrictions apply to detectors operating on the transport-flame
ionization principle. It was necessary to use a complicated
ternary-gradient elution system with eight programmed steps, starting
with isooctane to separate the lipids of low polarity and ending with a
solvent containing water to elute the phospholipids; a mobile phase of
intermediate polarity was then needed to effect the transfer from one
extreme to the other, and mixtures based on isopropanol gave
satisfactory results. At the end of the analysis, a gradient in the
reverse direction was generated to remove most of the bound water and
to re-equilibrate the column prior to the next analysis. A relatively
high flow-rate (2 mL/min) assisted the separation greatly.
In later work [6], it was observed that much better
resolution of the minor acidic components was obtained by adding small
amounts of organic ions to the aqueous component of the eluent. The
lifetime of the column was also greatly extended by this simple step.
In practice, the optimum results were obtained with 0.5 to 1 mM serine
buffered to pH 7.5 with triethylamine. In addition, hexane was used in
place of isooctane in the mobile phase, in order to reduce the maximum
operating pressure required.
The nature of the separation achieved with a lipid extract from rat liver is shown in
Figure 2. In spite of the abrupt changes in solvent
composition at various points, little base-line disturbance is
apparent, and each of the main simple lipid and phospholipid classes is
clearly resolved in only 20 minutes. Only the highly acidic lipids,
such as phosphatidic acid and to a lesser extent phosphatidyl-serine,
do not give satisfactory peaks. There is no "solvent peak" at the start
of the analysis, as is often seen with other detectors, and BHT added
as an antioxidant evaporates with the solvent so does not interfere.
After a further 10 minutes of elution to regenerate the column, the
next sample can be analysed. While this work has been superseded by
that of others in recent years, the principle has not changed.
Figure 2. Separation of rat lipids
by HPLC with evaporative light-scattering detection (CE = cholesterol
esters, TG = triacylglycerols, C = cholesterol, PG, PE, PI, PC and SPH
are various phospholipids).
Another important application of HPLC in lipid
analysis is to the separation of molecular species of lipids,
especially triacylglycerols. For example, silver ion HPLC separates
solely on the basis of the degree of unsaturation of the molecules has
been used with the ELSD, and there is further information here.
Much more use has been made of HPLC in the reversed phase mode in which
separation depends on the combined chain-lengths and the number of
double bonds in the fatty acid constituents (reviewed in some detail
elsewhere [7] and briefly here).
Very many different stationary phases of the octadecylsilyl type have
been utilised, with acetonitrile and a modifier solvent such as
acetone, dichloromethane or tetrahydrofuran as the mobile phase. With
the evaporative light-scattering detector, the choice of the mobile
phase has little effect on the response and gradients can be used, a
feature that is especially important with such complex natural fats as
fish oils or milk fat. Indeed with the latter, there may even be
virtues in hydrogenating the sample prior to analysis so that
resolution is based solely on chain-length and is not complicated by
double bond effects.
With such separations, there has also been some
debate about the efficacy of the evaporative light-scattering detector
in quantification. However, Herslof and Kindmark [8] obtained good
reproducibility for the relative proportions of different molecular
species in analyses of the triacylglycerols of soybean oil. When the
technique is used in research with triacylglycerols differing widely in
composition, the best approach to quantification consists in collecting
fractions and adding an odd-chain fatty acid as an internal standard
prior to transesterification and GC analysis, i.e. the technique long
used with thin-layer chromatography. The fatty acid composition and the
amount of each fraction are thereby obtained simultaneously.
In summary then, if you expect the evaporative
light-scattering detector in HPLC to be the equivalent of the
flame-ionisation detector in GC, you will be disappointed. If you look
for its virtues in terms of ease of use, universality and its capacity
to handle any combination of solvents and gradients, you will find much
to commend. For example, it can be used with all classes of lipid
separations and it is highly flexible in that it can be rapidly changed
from one mode of analysis to another (adsorption, reversed-phase,
silver ion, etc.). It is my opinion that, when combined with a stream
splitter, it is the most useful HPLC detector in its price range
currently available to lipid analysts. The charged aerosol detector,
which is new to the market, has the potential to change this view, but
we must await further work [9]. Similarly, mass spectrometry is
becoming more affordable as a detection/identification system, but it
is still expensive and requires a high degree of technical skill to
operate the instrument, not to mention interpretation of the data.
References
Rutan, S.C., Carr, P.W., Cheong, W.J., Park, J.H. and Snyder, L.R.
Re-evaluation of the solvent triangle and comparison to solvatochromic scales of solvent strength and selectivity.
J. Chromatogr. A, 463, 21-37 (1989).Righezza, M. and Guiochon, G. Effects of the nature of the solvent and solutes on the response of a light-scattering detector.
J. Liqu. Chromatogr., 11, 1967-2004 (1988).Christie, W.W. Detectors for high-performance liquid chromatography of lipids with special reference to evaporative light-scattering detection. In:
Advances in Lipid Methodology - One, pp. 239-271 (edited by W.W. Christie, Oily Press, Ayr) (1992).Moreau, R.A. and Christie, W.W. The impact of evaporative light-scattering detectors on lipid research.
INFORM, 10, 471-478 (1999).Christie, W.W. Rapid separation
and quantification of lipid classes by high performance liquid
chromatography and mass (light-scattering) detection.
J. Lipid Res., 26, 507-512 (1985).Christie, W.W. Separation of lipid classes by high-performance liquid chromatography with the 'mass detector'.
J. Chromatogr. A, 361, 396-399 (1986).Nikolova-Damyanova, B. Reversed-phase
high-performance liquid chromatography: general principles and
application to the analysis of fatty acids and triacylglycerols. In:
Advances in Lipid Methodology - Four, pp. 193-251 (edited by W.W. Christie, Oily Press, Dundee) (1997).Herslof, B. and Kindmark, G. HPLC of triacylglycerols with gradient elution and mass detection.Lipids, 20, 783-790 (1985).
Moreau, R.A. The analysis of lipids via HPLC with a charged aerosol detector.
Lipids, 41, 727-734 (2006).
This article is based on two previous publications (Lipid Technology,
1, 23-25 (1989) and Lipid Technology, 5, 68-70 (1993)) (by kind permission of
P.J. Barnes & Associates (The Oily Press Ltd)).
When amalgamating the two, they were substantially updated.
HPLC separation of lipid
classes
With the previously described techniques, the
quantification of the separated lipid classes represent a serious drawback since each
fraction may need a separate treatment. Considerable progress has been made to,
simultaneously, separate by HPLC and quantify with an efficient and near universal
detector, the evaporative light-scattering detector
(LSD).
The separation of the various simple and complex lipids present in natural extracts is
easily managed through an isocratic or better with a gradient elution procedure. The
non-specific LSD enables the quantification of non-polar and polar lipids in the same run.
It is recommended to first separate the crude lipid extract in two or three fractions by low pressure chromatography and analyze the fractions by HPLC .
Thus, lower complex gradients and analysis time will be required.
Among several published procedures devoted to the separation of all lipid classes in one
run, we have selected a simple and efficient procedure initially applied to the separation
of lipid classes from plant lipids (Christie WW et al, J High Resol Chromatogr 1995, 18,
97).
Apparatus:
column: YMC PVA-Sil (250 x 4.6 mm, 3 µm from
Hichrom), the phase is prepared by bonding a layer of polymerized vinyl alcohol to silica.
Ternary HPLC pump
Evaporative light-scattering detector
Reagents:
solvent A: isooctane/methyl tert-butyl ether
(98/2, v/v)
solvent B: isopropanol/acetonitrile/chloroform/acetic acid (84/8/8/0.025, v/v)
solvent C: isopropanol/water/triethylamine (50/50/0.2, v/v)
The author has replaced isooctane by isohexane for safety reasons but with similar
results.
Procedure:
A ternary gradient was generated during 40 min
with a flow rate of 1 ml/min followed by a 10 min reequilibration time.
The optimum gradient is described below.
Time
(min)ABCFlow rate (ml/min)0100001580200115445241403452141.440.1307001.44510000250100002
SE: sterol esters, S: sterols,
DAG: diacylglycerols, MGDG: monogalactosyldiglycerides, SG: sterol glycosides, CERE:
cerebrosides, DGDG: digalactosyldiglycerides, PE: phosphatidylethanolamine, PI:
phosphatidylinositol, PC: phosphatidylcholine
The quantification of lipid compounds is made using appropriate standards, the
relationships between sample size and response being dependent of the instrument used and
the concentration ranges.
An HPLC method with LSD detection was optimized and
validated for the simultaneous quantitation of cholesteryl esters, triglycerides,
cholesterol and phosphatidylcholine in human plasma. A silica Spherisorb column
was used with a multistep gradient system. The calibrations were made at levels
of 0.14-14 mg
lipid/injection (Seppanen-Laakso T et al., J chromatogr B 2001, 754, 437).
A reliable method was established to evaluate the lipid composition of plants.
The procedure focused on the polar lipid distribution of rapeseed oil but was also applied to
the estimation of waxes, triacylglycerols, and sterols (Beermann C et al.,
JAOCS 2003, 80, 747). The eluent system was modified from the method
described above and the water eluent was supplemented with 1 mM ammonium sulfate
to improve reproducibility. The gradient system was adapted to be suitable for
the separation of major lipid classes of plant materials. A precise
quantification was made on about 30 mg
of total lipids using an evaporative light scattering detector.