Professional Documents
Culture Documents
Fame & Cholesterol GC
Fame & Cholesterol GC
Swiss Federal Institute of Technology, Institute of Food Science, ETH-Zentrum, CH-8092 Zürich (Switzerland)
(Received April 22, 1996; accepted June 7, 1996)
A rapid, reliable and accurate gas chromatographic method for the determination of fatty acid composition in food was
developed and evaluated. The procedure consists of direct saponification of glycerides and hexane extraction of sterols from the
nonsaponifiable matter, followed by acidification of the residual soap and methylation of the free fatty acids. The substantial
advantage of this procedure is the possibility of simultaneous determination of the fatty acid composition and the cholesterol of
phytosterol content in food. In addition, it eliminates the use of chlorogenic chemicals and minimizes solvent use. This method
was compared to a slightly modified standard procedure, involving fat extraction prior to saponification and transmethylation of
the fatty acids. For both methods the coefficients of variation for analysed fatty acids (C ≥ 8) ranged from 0.6 to 7.4%. For the
majority of fatty acids no significant difference was observed between mean values at a 95% confidence level.
202
lwt/vol. 30 (1997) No. 2
Experimental Details
saponification
Chemicals
All reagents were of analytical reagent grade unless
esterification
specified otherwise. Pyrogallol, methyl tridecanoate,
boron trifluoride ( ~ 1.3 mol/L) in methanol, n-hexane,
ethanol, methanol, chloroform, anhydrous sodium sul- fatty acid methyl esters
phate, seasand (acid purified) and Clara-Diastase
(No.10065) were obtained from Fluka Chemie AG Fig. 2 Modified extraction method for determination of fatty
(Buchs, CH). 5α-Cholestane, cholesterol and trideca- acids in foods. * = for starch-rich food samples only
noic acid were supplied by Sigma Chemical Co.
(Munich, D). GLC Standard Mixtures of fatty acid
methyl esters (FAME) were purchased from Matreya Solutions
Inc. (Pleasant Gap, PA, U.S.A.). These products are Calibration solution for FAME was prepared by
equal weight (10 or 12.5 mg) mixtures of FAME. transferring the GLC Standard Mixtures No. GLC-101,
Thermoresistent α-amylase preparation Termamyl GLC-30, GLC-50, GLC-60 and GLC-100 into a 100 mL
120L was supplied by Novo Nordisk A/S. (Bagsvaerd, volumetric flask and diluting to the mark with hexane.
DK). The final concentration of the single FAME ranged
from 10 to 22.5 mg/100 mL. The solution was stored at
4 °C. Internal standard solution for FAME determina-
Food sample tion was prepared by dissolving approx. 100 mg ( ± 0.1
mg) tridecanoic acid (C13:0) in 50 mL n-hexane.
Internal standard solution for cholesterol determina-
*enzymatic degradation of starch tion was prepared by dissolving approx. 25 mg ( ± 0.1
mg) of 5α-cholestane in 50 mL n-hexane.
Potassium hydroxide solution was made by carefully
saponification
dissolving 50 g KOH pellets in 50 mL deionized
water.
separation
Pyrogallol solution was prepared by dissolving 5 g
pyrogallol in 500 mL ethanol, methanol or water,
respectively, as required by the different procedures
(hexane phase) (water phase) (see below).
nonsaponifiable fraction saponifiable fraction
203
lwt/vol. 30 (1997) No. 2
Instrumentation and operating conditions U.K.), heating block Reacti-Therm III™, heating/
A gas chromatograph (GC) Hewlett Packard Model stirring module (Pierce Chemical Co., Rockford, IL,
5890 Series II with a split/splitless capillary injection U.S.A.) and evaporator Rotavapor RE 111 with Water
system and a flame ionization detector (FID) equipped Bath 461, (Büchi Laboratoriums-Technik AG, Flawil,
with an autosampler HP 7673 (Hewlett Packard, CH).
Avondale, PA, U.S.A.) was used. Operating conditions
were: helium carrier gas at 1.7 mL/min, head pressure
100 kPa, 1:40 split ratio, injection port temperature 240 Procedure of the direct method (Fig. 1)
°C, detector temperature 280 °C, initial oven tem-
perature 160 °C held for 4 min, then increased up to 230 Homogenized food samples of 1.5 to 2.5 g, accurately
°C at a rate 4 °C/min and held for 3.5 min at this weighed, were placed in a 50 mL centrifuge tube. The
temperature. Injection volume was 2 µL. The GC sample weight of fat-rich food like butter or mayon-
separation was accomplished by using a fused silica naise must be reduced to approx. 0.2 g (the limiting
capillary column Supelcowax™ 10, 30 m 3 0.32 mm factor is the concentration of KOH). One millilitre
i.d. and 0.25 µm film thickness (Supelco, Inc., Belle- each of both internal standard solutions (5α-chol-
fonte, PA, U.S.A.). Peak areas were integrated and estane and C13:0), 9 mL pyrogallol in ethanol (1
calculated by Hewlett Packard 3365 ChemStation g/100 mL), 1 mL potassium hydroxide solution (50
software. g/100 g), a magnetic stirring bar, and approx. 1.5 g of
Saponification and separation of layers were carried seasand for a better fat dispersion were added and
out in 50 mL screw capped centrifuge tubes Cortex® the sample vortexed for 20 s. The capped tube was
Brand (Corning Inc., Corning, NY, U.S.A.). placed in a water-bath at 60 °C and saponified for 1 h
The following apparatus were also used: water-bath under continuous stirring. After cooling to room
Julabo U3 (Julabo Labortechnik GmbH, Seelbach, D), temperature, 5 mL of deionized water was added
magnetic stirrer/hot plate Heidolph MR 2002 (Hei- followed by 10 mL hexane and the sample vortexed
dolph GmbH, D), shaker Vortex Genie 2™ (Scientific at the highest setting for 2 min. The sample was then
Industries Inc., Bohemia, NY, U.S.A.), centrifuge MSE centrifuged for 3 min at 2200 3 g and the upper
Mistral 1000 (MSE Scientific Instruments, London, hexane layer transferred with a Pasteur pipette into a
C20:1 (11)
C10:0
C8:0
C12:0
C16:0
C20:0
C13:0 (I.S.)
C18:1 (9)
C18:0
C21:0
C19:0
C22:1 (13)
C24:1 (15)
C17:0
C22:0
0 5 10 15 20 25
Retention time (min)
Fig. 3 Gas chromatogram of the standard mixture of a fatty acid methyl esters. Column: Supelcowax™ 10, 30 m 3 0.32 mm
i.d., film thickness 0.25 µm, carrier gas helium at 1.7 mL/min; oven temperature programme: 160 °C (4 min), to 230 °C (4
°C/min), 230 °C (3.5 min), injection 2 µL, split ratio 1:40
204
lwt/vol. 30 (1997) No. 2
clean 50 mL centrifuge tube. Another 10 mL hexane L). The tube was capped tightly and placed in a
was added to the water phase and the extraction and heating block for 4 min at 100 °C. Tubes were cooled
centrifugation step repeated. The combined hexane in cold tap water to room temperature and 4 mL
extracts can be directly used for cholesterol determi- hexane added, followed by 2 mL water. The sample
nation as described by Paterson et al. (15). was vortexed vigorously for 2 min and centrifuged for
The water phase was acidified with 3 mL hydrochloric 3 min at 2200 3 g. Approx. 1.5 mL of the hexane
acid (HCl) (25 g/100 g) to pH ≤ 1.5, and shaken extract was transferred into the GC-sample vials and
briefly. Released free fatty acids were extracted with 2 µL injected into the gas chromatograph.
10 mL hexane. The sample was vortexed at the In order to prevent incomplete extraction of fat and/
highest setting for 2 min and centrifuged for 3 min at or cholesterol in starch-rich food samples like cooked
2200 3 g. The upper hexane layer was transferred pasta it is necessary to perform an enzymatic degra-
with a Pasteur pipette into a clean 50 mL centrifuge dation of the gelatinized starch prior to saponifica-
tube and another 10 mL hexane added to the water tion: 0.25 mL of thermoresistent α-amylase prepara-
phase. The extraction and centrifugation steps were tion Termamyl 120L was mixed with 4.75 mL
then repeated. The first and second hexane extracts pyrogallol in deionized water (1 g/100 mL), and
were combined using the same pipette. In order to added to the sample prior to saponification. The tube
dry the extract, approx. 4 g of anhydrous sodium was capped and held for 15 min at 95 °C in a water-
sulphate was added to the centrifuge tube and shaken bath under continuous stirring. For extraction of the
vigorously for a short time. The extract was centri- unsaponifiable matter addition of water was omitted.
fuged for 3 min at 2200 3 g and 4 mL of the clear
extract (containing 10 to 50 mg of fatty acids) trans-
ferred into a 15 mL culture tube fitted with a screw Procedure of the modified extraction method (Fig. 2)
cap. The extract was concentrated in a rotatory evap-
orator under reduced pressure of 200 mbar for 3 min Well homogenized food samples of 1.5 to 2.5 g,
nearly to dryness. The residue was redissolved in 1 accurately weighed, were placed in a 50 mL cen-
mL of hexane and then a magnetic rod placed in the trifuge tube. Sample weight of fat-rich food like
tube. Fatty acids were converted into methyl esters by butter or mayonnaise must be reduced to approx. 0.5
adding 5 mL boron trifluoride in methanol (1.3 mol/ g. One millilitre of internal standard solution for
C13:0 (I.S.)
C18:1 (9)
C16:0
C15:0
C17:0
0 5 10 15 20 25
Retention time (min)
Fig. 4 Representative gas chromatogram of fatty acid methyl esters in food sample (pasta with meat sauce). Chromatograph
conditions as in Fig. 3
205
lwt/vol. 30 (1997) No. 2
Table 1 Fatty acid composition of minced beef, Parmesan cheese and mayonnaise in mg/100 g fresh sample
Minced Beef Parmesan Cheese Mayonnaise
Ext Dir Ext Dir Ext Dir
Fatty acid Meana CV Meana CV Meana CV Meana CV Meana CV Meana CV
C 8:0 – – 202.8 5.2 213.4 4.1 – –
C 10:0 – – 487.4 3.0 497.1 3.1 – –
C 12:0 – – 640.1 5.6 642.8 3.2 – –
C 14:0 163.1 2.2 161.0 2.6 2261.4 4.5 2235.2 3.2 59.4 2.4 62.0 4.9
C 15:0 27.1 1.9 26.4 2.3 268.5 4.0 263.1 3.3 – –
C 16:0 1507.9 2.7 1490.2 2.3 5950.9 3.9 5761.9 3.4 5275.3 0.6 5300.7 2.9
C 16:1(9) 214.7 3.3 211.4 1.9 267.5 4.5 261.7 3.7 125.4 1.2 125.1 5.8
C 17:0 73.3 4.3 69.7 4.0 170.6 4.2 166.1 3.7 – –
C 18:0 939.2 2.6 926.0 2.4 2308.2 4.2 2214.6 4.1 3736.9 0.7 3679.7 3.0
C 18:1(9) 2398.5 2.9 2350.8 2.4 3974.6 3.8 3821.2 3.8 16410.6 0.8 16214.2 2.8
C 18:2 (9,12) 214.2 1.0 220.9 2.9 400.4 3.6 386.6 3.8 45397.1 1.0 44818.4 2.2
C 18:3 (9,12,15) 54.8 2.1 55.6 1.8 131.6 3.2 125.7 3.6 169.6 1.1 171.7 4.9
C 19:0 – – 21.6 3.4 22.3 3.4 – –
C 20:0 – – 52.0 5.2 49.1 4.4 223.0 1.1 214.5 3.1
C 20:1(11) – – 48.4 5.8 48.8 5.6 100.0 0.9 99.2 3.6
C 20:2 (11,14) – – – – – –
C 20:3 (11,14,17) 37.4 3.7 39.0 3.5 32.2 5.9 34.2 3.3 – –
C 22:0 – – 24.6 6.8 23.1 3.1 581.7 1.6 551.4 2.4
C 22:1(13) – – – – – –
C 24:1(15) – – – – 156.3 1.5 142.7 2.8
Total fatty acids 5630.2 5551.0 17242.8 16766.9 72235.3 71379.6
a n= 4;CV = coefficient of variation in %.
– = content of fatty acid less than 10 mg/100 g fresh sample.
Ext = modified extraction method; DIR = direct method.
FAME (C13:0) was added and mixed with 8 mL filtered with slight suction through a sintered glass
chloroform and 8 mL pyrogallol in methanol (1 g/100 funnel coated with approx. 2 g of Celite and rinsed
mL). The sample was shaken vigorously for 2 min with 4 mL mixture of chloroform/methanol 2:1. For
and another 8 mL chloroform added. The sample was separation into two phases, z mL water was added,
Table 2 Fatty acid composition of cooked pasta in mg/100 g fresh sample determined without and with enzymatic starch
degradation
Ext Ext Ext Dir Ext Dir
Enzyme Clara-Diastase Termamyl 120 L Termamyl 120 L Termamyl 120 L
(concentration (0.15 g in 15 mL (0.25 mL in 5 mL (1.0 mL in 5 mL (0.25 mL in 5 mL
and reaction sodium acetate water, water, water,
conditions) none buffer, 40–45 °C, 1 h) 95 °C, 15 min) 95 °C, 30 min) none 95 °C, 15 min)
Fatty acid Meana CV Meana CV Meana CV Meana CV Meana CV Meana CV
C 8:0 – – – – – –
C 10:0 – – – – – –
C 12:0 – – – – – –
C 14:0 – – – – – –
C 15:0 – – – – – –
C 16:0 89.3 7.6 216.6 2.6 192.0 2.6 256.6 2.9 166.6 8.8 262.2 2.3
C 16:1(9) 6.9 8.3 13.7 3.2 11.6 3.2 15.7 3.3 13.4 10.8 15.1 2.4
C 17:0 – – – – – –
C 18:0 21.4 8.5 44.1 2.9 40.0 2.9 49.1 3.0 38.0 6.5 51.3 4.3
C 18:1(9) 121.5 9.0 248.7 2.8 216.0 2.8 288.7 2.6 238.6 10.7 286.0 2.3
C 18:2 (9,12) 162.8 5.4 356.1 2.7 302.5 2.7 426.1 2.8 341.3 9.1 428.3 3.2
C 18:3 (9,12,15) 8.4 4.1 19.0 5.3 14.0 5.3 19.3 4.8 16.1 8.6 19.8 1.6
C 20:0 – – – – – –
C 20:1(11) – – – – – –
C 20:2 (11,14) – – – – – –
C 20:3 (11,14,17) – 7.4 4.1 8.0 3.7 9.1 3.5 8.5 10.5 9.5 3.0
C 22:0 – – – – – –
C 22:1(13) – – – – – –
C 24:1(15) – – – – – –
Total fatty acids 410.3 905.6 784.1 1064.6 822.5 1072.2
a n=4;CV= coefficient of variation in %.
–= content of fatty acid less than 7 mg/100 g fresh sample.
Ext=modified extraction method; Dir=direct method.
206
lwt/vol. 30 (1997) No. 2
Table 3 Fatty acid composition of the prepared meal (pasta with meat sauce) in mg/100 g fresh sample determined without
and with enzymatic starch degradation with Termamyl 120L
With starch degradation
Ext Dir
Without starch degradation 1.0 mL in 5 mL 0.25 mL in 5 mL
water, water,
Ext Dir 95 °C, 30 min 95 °C, 15 min
Fatty acid Meanb CV Meanb CV Meana CV Meana CV
C 8:0 19.1 5.5 21.1 5.9 20.5 5.2 21.0 4.7
C 10:0 43.3 5.2 45.7 5.0 43.9 4.4 44.7 5.7
C 12:0 58.3 3.6 62.7 4.6 61.2 2.8 59.4 3.0
C 14:0 205.4 3.7 213.9 3.7 217.6 2.5 208.4 3.1
C 15:0 23.3 3.6 23.7 2.7 24.8 2.4 23.3 3.7
C 16:0 984.6 4.4 1020.8 2.5 1051.9 2.2 1005.8 3.7
C 16:1(9) 81.8 4.4 85.3 2.6 86.4 2.3 83.0 3.2
C 17:0 23.6 4.2 24.4 6.2 24.6 2.8 25.6 4.6
C 18:0 397.9 5.9 396.1 3.7 412.1 2.1 393.2 2.2
C 18:1(9) 1842.3 5.3 1888.5 2.1 1909.7 2.2 1896.5 3.4
C 18:2 (9,12) 324.1 6.6 339.3 6.2 310.3 3.2 337.0 3.3
C 18:3 (9,12,15) 36.9 5.2 40.3 5.4 38.6 2.3 39.0 4.6
C 20:0 – – – –
C 20:1(11) – – – –
C 20:2 (11,14) – – – –
C 20:3 (11,14,17) 10.2 4.3 10.7 5.4 10.6 7.4 12.2 2.1
C 22:0 – – – –
C 22:1(13) – – – –
C 24:1(15) – – – –
Total fatty acids 4050.8 4172.5 4212.2 4149.1
a n= 4; bn = 8;
CV=coefficient of variation in %.
– = content of fatty acid less than 10 mg/100 g fresh sample.
Ext = modified extraction method; Dir=direct method.
where z = 6 – (P 3 W/100), P = sample weight in g continuous stirring. For separation into two phases
and W = water content of sample in g/100 g. The z – 5 mL of water was added.
sample was vortexed for 2 min and centrifuged for 3
min at 2200 3 g. The upper layer was removed with a
Pasteur pipette and the remaining chloroform phase Optimization of the Termamyl treatment
dried by adding approx. 2 g of anhydrous sodium In order to optimize the lipid extraction for starch-rich
sulphate. Four millilitre of the dried chloroform phase food samples, incubation experiments with the α-amy-
(containing 10 to 50 mg of fatty acids) was trans- lase preparation Termamyl 120L were carried out at
ferred into a 15 mL culture tube fitted with a screw different enzyme concentrations and for different
cap and concentrated in a rotatory evaporator under reaction times. Various amounts of Termamyl 120L
reduced pressure of 300 mbar for 5 min nearly to (100, 200, 500 and 1000 µL) were added to the samples
dryness. The residue was redissolved in 1 mL hexane and incubated at 95 °C for 15, 30 and 60 min. The
and a magnetic rod placed in the tube. The content concentration of liberated D-glucose at the end of the
was saponified by adding 4 mL of NaOH (0.5 mol/L) incubation period was chosen as an indicator of optimal
in dry methanol and placing the capped tube for 5 to conditions for the Termamyl action. Determination was
10 min into the heating/stirring block at 100 °C under performed using an enzymatic assay according to the
continuous stirring. Conversion of the fatty acids into description given by the Food Analysis Boehringer
methyl esters and separation of methyl esters was Mannheim Test No. 716251.
made as described for the direct method, except that
(because of the larger end volume) the content of the
15 mL culture tube must be transferred into a larger Statistical analysis
culture tube before separation. The measured values were evaluated by descriptive
For starch-rich food samples it is necessary to apply statistics (means and coefficients of variation). Differ-
an enzymatic degradation of gelatinized starch prior ences between the mean contents of each fatty acid in
to fat extraction: 1 mL of thermoresistent α-amylase all examined samples for both methods were analysed
preparation Termamyl 120L was mixed with 4 mL with the statistical program of Microsoft® Excel
pyrogallol in deionized water (1 g/100 mL) and this Version 5.0 (Microsoft Corporation, Redmond, WA,
solution added to the sample. The tube was capped U.S.A.). The paired two-sample t-test for means at the
and held for 30 min at 95 °C in a water-bath under 95% confidence level was used.
207
lwt/vol. 30 (1997) No. 2
Results and Discussion be omitted. This finding was expected because the
percentage of fatty acids in pasta is low (Table 2) and
In order to compare the direct method (Fig. 1) with the the share of pasta in the whole meal makes approx.
modified extraction method (Fig. 2) for the determina- 42%. The diminution of the total fatty acid content in
tion of fatty acid content and distribution in foodstuff, a the meal caused by the insufficiently extracted pasta
test meal (pasta with meat sauce), as well as some of its was 6.5% for the extraction method and 2.5% for the
ingredients, and mayonnaise were analysed. The lower direct method only. It was therefore within the meth-
detection limit for the quantitative determination of ods’ precision.
fatty acids in the meal was found to be approx. 10 mg Preliminary experiments proved the necessity to pre-
per 100 g fresh sample. The chromatogram of the vent the oxidation of unsaturated fatty acids in the
calibration solution of the standard mixture of FAME course of sample preparation. Addition of 3-tert-butyl-
(Fig. 3) and a representative chromatograph of the food 4-hydroxyanisole (BHA) solution (1 g/100 g) in ethanol
sample (Fig. 4) showed excellent separation of all fatty (direct method) or methanol (modified extraction
acids of interest. The selection of fatty acids can be method) led to two problems. Figure 4 reveals that
extended for example by C20:4 and C20:5 without any BHA coeluated with methyl ester of nonadecanoic acid
change of the operating conditions for GC (data not (C19:OMe) under the selected operating conditions.
shown). The data for minced beef, Parmesan cheese, This microbial fatty acid had been shown to be present
and mayonnaise are summarized in Table 1. The results in Parmesan cheese used in the meal (Table 1). Second,
indicated a high degree of precision for both methods the insolubility of BHA in water prevented its use
as shown by low coefficients of variation (CV) for the during the starch degradation step. Therefore pyro-
single fatty acids, ranging from 0.6 to 6.8. It was evident gallol at a concentration of 1 g/100 mL was used as
that both methods led to equivalent results. Application antioxidant. This substance is soluble in water, ethanol
of the paired two-sample t-test at the 95% confidence and methanol but insoluble in apolar solvents like
level revealed no significant differences between the hexane or in chloroform. The protective effect of
two methods for most of the fatty acids. A possible pyrogallol is therefore limited to the enzymatic starch
explanation for the only fatty acid, Eicosatrienoic acid degradation and to the saponification step. However,
(C 20:3), showing a different outcome of the statistical these two procedures would most likely cause oxidation
analysis could be its low concentration in some food because of the heating process involved.
samples making accurate integration of peaks more The accuracy of both methods was also evaluated by
difficult. analysis of glycerine mono-oleate. Using the direct
However, substantial differences resulted for cooked method an oleic acid recovery of 99.6% was obtained,
pasta. The extraction method gave much lower yields of whereas with the extraction method only 89.3% of the
fatty acids when compared to the direct method (Table theoretical value for oleic acid was recovered.
2). It was obvious that part of the lipids was ‘entrapped’
by the gelatinized starch and was thus not extracted.
Because of this problem, already described by the
method for determination of cholesterol (16), it was Conclusions
supposed that even by the direct method not all lipids
were liberated and analysed. The data from Table 2 The proposed direct method represents an accurate,
confirmed this assumption. The necessity to degrade the precise, simple and time saving method which enables
gelatinized starch in order to release total lipids has quantitative determination of cholesterol and fatty acid
already been recognized. The official method described contents in food on the same sample. It eliminates the
in the Swiss Manual for Food Analysis (17) uses Clara- use of hazardous chlorinated solvents. If the cholesterol
Diastase (an α-amylase preparation from Aspergillus content of a sample is not to be determined, the even
oryzae) for starch degradation. In our experiments the simpler modified extraction method could be used for
effectiveness of the Clara-Diastase was compared to the fatty acid analysis. The main advantage of this
thermoresistent α-amylase Termamyl 120L. Clara-Dia- modification is the possibility to introduce (like in the
stase at the recommended concentration level direct method) the internal standard to the sample at
improved lipid extraction from cooked pasta but the the very beginning of the analytical procedure which
amount of determined fatty acids was lower as com- improves its precision. Another advantage of the
pared to the amount obtained by the use of Termamyl procedure is a substantially reduced use of solvents.
120L (Table 2). Moreover, Clara-Diastase could not be
used in the direct method prior to the saponification
step because this enzyme needs buffered conditions for
its action. Another important advantage of using References
Termamyl 120L in the extraction method is that only
half the time (30 min instead of 60) is needed for 1 WILSON, P. W., CASTELLI, W. P. AND KANNEL, W. B. Coro-
sufficient starch degradation as compared to Clara- nary risk prediction in adults (The Framingham Study).
American Journal of Cardiology, 59, 91G–94G (1987)
Diastase. The need for a preliminary degradation of 2 TRUSWELL, A. S. Evolution of dietary recommendations,
starch for the test meal was also studied. The results are goals, and guidelines. American Journal of Clinical Nutri-
summarized in Table 3 and indicate that this step could tion, 45, 1060–1072 (1987)
208
lwt/vol. 30 (1997) No. 2
209