BIOMEDICAL CHROMATOGRAPHY, VOL. 6 , 120-123 (1992)
Simple and Highly Sensitive Determination of Free Fatty Acids in Human Serum by High Performance Liquid Chromatography with Fluorescence Detection Tetsuharu Iwata, Kouji Inoue, Masaru Nakamura and Masatoshi Yamaguchi* Faculty of Pharmaceutical Sciences, Fukuoka University, Nanakuma, Jonan-ku, Fukuoka 814-01, Japan
A highly sensitive and simple reversed phase high performance liquid chromatographic(HPLC) method for the quantitative determination of free fatty acids in human serum is presented. The method is based on the direct derivatization of serum fatty acids with 6,7-dimethoxy-l-methyl-2(1H)-quinoxalinone-3-propionylcarboxylic acid hydrazide. The derivatization reaction proceeds in aqueous solution in the presence of pyridine and l-ethyl3-(3-dimethylaminopropyl)carbodiimide at 37 "C. The resulting derivatives are separated within 75 min on a reversed phase column (YMC Pack Cs) with a gradient elution of aqueous acetonitrile and detected fluorimetrically. The detection limits are 2.5-5 fmol in a 10 pL injection volume. The sensitivity permits precise determination of free fatty acids in 5 pL serum. The method is simple and is without the conventional liquidliquid extraction steps of serum fatty acids.
INTRODUCTION Free fatty acids arise mainly from the hydrolysis of triacylglycerol in adipose tissues or from the action of lipoprotein lipase, and are released into the blood stream. The amounts of the acids in human serum/ plasma increase or decrease with the extent of diabetes, thyremphraxis and hepatic dysfunction. The determination of the acids is therefore very important in the diagnosis and therapy of these diseases. Some high performance liquid chromatographic (HPLC) methods with fluorescence detection have focussed on the sensitive determination of free fatty acids in human serum/plasma. The methods include the derivatization of fatty acids with fluorescence derivatization reagents-9-anthryldiazomethane (Shimomura et al., 1984, 1986; Ghiggeri et al., 1986; Hatsumi et a/., 1986; Kargas et al., 1990), 9-aminonaphthalene (Ikeda et al., 1984), 4-bromomethyl-7-acethoxycoumarin (Tsuchiya et af., 1982, 1984), 5-(dimethylamino)-lnaphthalenesulphonyl-semipiperazide(Yanagisawa et al., 1985) and 3-bromomethyl-6,7-dimethoxy-l-methyl2(IH)-quinoxalinone (Yamaguchi et al., 1986a, b ) . However, these reagents are not stable and generally require most dried aprotic solvents in the derivatization reaction. Therefore, the fatty acids should be extracted with organic solvents such as chloroform and nheptane, and the organic layer should be evaporated to dryness prior to the derivatization. Furthermore, the reagents need to react for a long time at high temperatures. Thus, the conventional fluorimetric HPLC methods are time-consuming and not ideal for routine clinical investigations. Recently, we have developed 6,7-dimethoxy1-methyl-2(IH) - quinoxalinone-3 -propionylcarboxylic acid hydrazide (DMEQ-hydrazide) as a highly sensitive Author to whom correspondence should be addresscd. 0269-3879/92/030120-04 $05 .oO @ 1992 by John Wiley & Sons. Ltd.
DMEQ-hydrazide CH3 I
CH,O*N+
CH~CH~CONHNHCOR
Scheme 1. Derivatization of fatty acids with DMEQ-hydrazide.
fluorescence derivatization reagent for fatty acids (Yamaguchi et af., 1990). The reagent readily reacts with fatty acids in aqueous solution in the presence of pyridine and l-ethyl-3-(3-dimethylaminopropyl)carbodiimide (EDC) at room temperature (0-37 "C) to produce intensely fluorescent DMEQ derivatives (Scheme 1). The purpose of the present research was aimed at establishing a sensitive and simple HPLC method utilizing DMEQ-hydrazide for the determination of free fatty acids in a minute amount of serum without cumbersome clean-up procedures.
EXPERIMENTAL Reagents and materials. All chemicals and solvents were of analytical reagent grade, unless otherwise stated. Deionized and distilled water was used. Lauric (C,,,,), myristoleic (C14:,),myristic (C13:0)r linolenie (Clx.n),eicosapentaenoic (C20:5 ) , palmitoleic (C16 :J, linoleic (CIS,*), arachidonic (C20.4)r docosahexaenoic (C22:6),palmitic (C16:o),dihomo-ylinolenic (Go:3), oleic (Clx,l),margaric (C17:o)and stearic acids were purchased from Sigma (St. Louis, MO, USA). DMEQ-hydrazide was prepared as described preReceived 4 July 1991 Accepted 30 August 1991
DETERMINATION OF FREE FA'ITY ACIDS I N HUMAN SERUM
viously (Yamaguchi et al., 1990). DMEQ-hydrazide (50 mM) solution was prepared in N,N-dimethylformamide. The DMEQ-hydrazide solution could be used for at least 1month even at room temperature. Pyridine (4%) solution was prepared in 20 mM hydrochloric acid ethanolic solution. Stock M) of fatty acids were prepared in solutions (2 x 10- 7-1 x ethanol. Serum specimens were obtained from fasting healthy volunteers in our laboratories.
HPLC apparatus and conditions. A Jasco Trirotor-V high performance liquid chromatograph (Tokyo, Japan), equipped with a Model VL-614 injector (100 pL loop), and a Jasco FP-210 spectrofluorometer, equipped with a 15 yL flow cell operated at an excitation wavelength of 360nm and an emission wavelength of 435 nm, were used. The column was a YMC-Pack Cx (250 x 4.6 mm i d . . particle size 10 pm, Yamamura Chemical Laboratories Kyoto, Japan). The column temperature was maintained at 30-+0.2"C with a Shimadzu CTO-6A column oven (Kyoto, Japan). For the separation of the DMEQ derivatives of the fatty acids on the column, a gradient elution with aqueous 5595% acetonitrile (Fig. 1) was carried out by using a Jasco GP-A40 Gradient Programmer. The flow rate was 1.O mL/min. Uncorrected fluorescence excitation and emission spectra of the eluate were measured with a Hitachi 650-60 fluorescence spectrophotometer fitted with a 20 yL flow cell; the spectral bandwidths were 5 n m in both the excitation and emission monochromators.
121
Procedure. A 5 pL aliquot of serum was mixed with 5 pL ethanol (or fatty acid standard solutions) and SO yL pyridine solution. The mixture was vortex mixed for ca. 10 s and then 25 pL DMEQ-hydrazide solution and 15 pL 2 M E D C in water were added. The resulting mixture was warmed at 37 "C for 10 min and was centrifuged at 1000 x g for ca. 5 min. The supernatant (10 pL) was injected into the chromatograph. The amounts of the free fatty acids were calibrated by means of the standard addition method: the 5 y L ethanol added to the serum in the procedure was replaced by 5 pL acid standard solution. The net peak heights of the individual fatty acids were plotted against the concentrations of the spiked fatty acids.
RESULTS AND DISCUSSION Derivatization reaction conditions
The optimal derivatization conditions for linear saturated fatty acids were previously examined using pyridine solution which was free from hydrochloric acid (Yamaguchi et al., 1990). We recently found that the yields of DMEQ derivatives of fatty acids increase in the presence of hydrochloric acid in the pyridine solution. Thus, the derivatization conditions were improved using pyridine solutions containing hydrochloric acid. l I
.-0
100
+J (u
I
w c W
4 80
V
co > W-.
a,?
60
_-_.
- w ..4
v
I
+J
.A
c
40
0
w
a,
W 4:
a, v)
l=
0
2
a v,
a, L
8
3
L 0
I
I
c,
u
12
a, 4-J a,
n
I
I
I
I
I
0
20
40
60
80
Retention tr?e ( m i n t Figure 1. Chromatogram of the DMEQ derivatives of fatty acids. A portion (5 pL) of a standard solution of fatty acids (2 nmol/rnL each) was treated as described in the text. Peaks: 1 =C,2:o; 2=C1,,1; 3 = C 1 4 : 0 ; 4 = C l S r 3 ; 5=C20:6; 6=C16.1; 7=c18:2; 8=c22:6 and C20:4; 9=c,6:,; I O = ~ , , : , ; 11 = C18:1; 12 = CI7:0; 1 3 = C18 o; 1 4 = DMEQ-hydrazide. Gradient elution with aqueous acetonitrile (0-56 min, 5 5 % ; 56-72 min, 55-95%; 72-76 min, 9 5 % 76-80 min, 55%).
TETSUHARU IWATA ET AL
122
Hydrochloric acid gave maximum and constant peak heights at the concentration of 1 0 - 3 0 m ~in the pyridine solution; the maximum peak height is 1.1 times that in the absence of hydrochloric acid in the pyridine solution. Hence, 20 mM hydrochloric acid was selected considering the following step of extracting the fatty acids from the serum. DMEQ-hydrazide gave the most intense peaks at concentrations greater than ca. 7 mM in the solution for all the fatty acids; 50mM was used as an optimum concentration. EDC and pyridine were used to facilitate the derivatization of fatty acids with DMEQ-hydrazide. Maximum and constant peak heights were attained at pyridine concentrations in the solution of 2-6010 ; 4% was selected as optimum. The peak heights for the acids were maximum and constant at concentrations of EDC higher than 1.8M; 2.0 M was used. The derivatization reaction of fatty acids with DMEQ-hydrazide apparently occurred even at 0 "C; higher temperatures allowed the fluorescence to develop more rapidly. However, at 60°C the peak heights decreased. At room temperature (15-37 "C), the peak heights for all the fatty acids were almost maximum after standing for 10 min. Hence, the solution was allowed to stand at 37°C for 10min. The DMEQ derivatives in the final mixture were stable for at least 24 h in daylight at room temperature. The detection limits for the fatty acids were 2-7 fino1 at a signal-to-noise ratio of 3, and became slightly better than those in a previous paper (Yamaguchi et al., 1990).
14
7 8
3
9
!
tl
20
0
13
60
40
80
Retention time (min) Figure 2. Chromatogram of the DMEQ derivatives of free fatty acids in normal human serum. A portion (5 pL) of serum was treated as in the detailed procedure. Peak number as in Fig. 1. Detector sensitivity: -, 1; ---, 16.
no effect on the fluorescence excitation (360nm) and emission maximum wavelengths (435 nm) and the fluorescence intensities of the DMEQ derivatives of all the fatty acids. The individual fatty acids gave single peaks in the chromatogram.
HPLC conditions
Determination of free fatty acids in human serum
The best separation of the DMEQ derivatives of 14 saturated and unsatured fatty acids was achieved on a reversed phase column, YMC Pack C,, by gradient elution with aqueous 55-95% ( v h ) acetonitrile (Fig. 1). However, the peaks for C22:6 and C20:4acids could not be resolved successfully under any HPLC conditions tested. A typical chromatogram obtained with a standard solution of the fatty acids is shown in Fig. 1. The change in acetonitrile concentration actually had
Chloroform has been commonly used for the liquidliquid extraction of fatty acids from serum. However, Hatsumi et ul. (1986) reported that some fatty acids were released from some phospholipids in biological materials during the chloroform extraction. To overcome the problems, Miwa and Yamamoto (1986) used ethanol both as a precipitating agent for serum proteins and an extraction solvent for serum fatty acids. Moreover, the serum sample was acidified with hyd-
Table 1. Concentration of free fatty acids in serum from healthy persons Concentration InmollmL) Age
Sex
C12.a
Ci4:i
C14.a
Cia 3
43 40 32 23 23 22 22 26 23 22 22 22 22 22 22
M M M M M M M F F
1.43 1.54 1.49 1.16 2.00 2.88 1.09 7.44 3.78 0.91 1.05 5.04 2.48 5.16 2.76
0.90 0.71 0.11 0.47 0.97 0.51 0.45 1.02 1.18 0.31 0.27 2.35 0.90 3.23 0.72
7.74 5.85 2.74 4.70 6.55 7.63 4.32 10.74 11.06 4.32 3.06 13.92 11.57 17.35 7.50
5.38 6.28 2.28 5.47 12.72 10.94 9.98 9.03 11.22 3.15 1.96 21.89 17.37 19.15 9.30
2.68 1.85
0.93 0.81
7.94 4.08
9.74 5.87
Mean S.D.
F F
F F F F
CZO:~
1.62 2.17 0.46 0.53 1.26 0.69 0.55 0.89 0.79 0.57 0.32 2.03 1.42 1.30 1.01
c16.1
C18:2
%:o
Cm:3
7.35 35.20 86.17 0.65 7.77 35.20 67.31 0.78 0.80 12.64 22.68 0.61 5.64 32.50 66.78 0.34 11.28 72.22 88.85 0.82 11.28 80.79 115.24 1.10 53.72 74.85 1.01 8.50 87.24 0.44 10.60 50.55 49.66 87.78 0.76 7.36 26.19 3.84 40.39 0.63 2.42 13.54 30.71 0.34 28.45 111.04 162.62 0.82 14.55 123.68 155.63 0.78 34.35 115.09 174.48 0.89 58.22 113.63 0.38 8.34
1.04 10.84 0.56 8.83
58.04 34.63
91.58 0.69 44.56 0.23
Cle I
C1s.a
87.93 62.90 17.59 58.85 112.96 133.27 104.84 100.84 98.10 40.59 27.06 253.00 179.26 224.60 109.59
27.98 22.07 10.88 27.56 30.93 34.20 31.55 27.82 35.75 12.90 14.62 55.45 44.1 5 55.1 1 32.80
107.42 30.92 65.65 12.91
DETERMINATION OF FREE FATTY AClDS IN HUMAN SERUM
rochloric acid to convert the serum fatty acids into the non-ionic forms from the dissociated forms. Thus, the acidic ethanol was used in the present work. The best recovery was made by adding more than 40vL of ethanolic hydrochloric acid (15-25 mM) solution. Ethanolic 20 r n M hydrochloric acid solution (50 pL) containing 4% pyridine was employed in the procedure. The recoveries (n= 8 each) of the fatty acids (0.2 nmol per 5 pL each) added to a pooled normal serum were 100.0-105.9%. A representative chromatogram obtained with normal human serum is shown in Fig. 2. All the peaks for the fatty acids were identified on the basis of their retention times and the fluorescence excitation and emission spectra of the eluates in comparison with the standard compounds, and also by co-chromatography of the standards and the serum sample with aqueous 50-100% acetronitrile or methanol as the mobile phase. Many substances such as alcohols, sugars, aldehydes, ketones, phenols, amines and a-keto acids gave no fluorescent derivatives under the described conditions. On the other hand, 17 different a-amino acids, hydroxycarboxylic acids (lactic and malic) , dicarboxylic acids (oxalic, malonic, succinic and adipic) and aromatic carboxylic acids (benzoic, salicylic and cinnamic) reacted with DMEQ-hydrazide to produce fluorescent derivatives. However, these compounds did not interfere with either the detection or the separation of the peaks for all the fatty acids even when they were spiked at unusually high concentrations in serum (10 pmol/mL of serum), because DMEQ derivatives of the compounds were co-eluted with DMEQ-hydrazide under the recommended HPLC conditions. Linear relationships were observed between the peak heights of the individual fatty acids and the amounts of the fatty acids added in the range 1pmol-5 nmol each
123
to 5 pL of serum. The precision was established by repeated determinations ( n = 7) using a normal serum. The coefficients of variation did not exceed 3.0% for all the fatty acids (n=7 in each case). C17 acid has been widely used as an internal standard for the analytical methods of free fatty acids in human plasma. However, Hatsumi et al. (1986) reported that C,, acid is present in human serum. Furthermore, the peak for C17 acid appeared in the chromatogram (Fig. 2) obtained with normal human serum. Hence, we did not use Ci7 acid as an internal standard in this study, since good precision was given even using the standard addition method. When cephalin and sphinogomyelin were treated as in the method, no peaks for the fatty acids were observed in the chromatogram. This supports the fact that no fatty acids were released from phospholipids during the present extraction and derivatization steps. The concentrations of free fatty acids in sera from healthy volunteers were determined by this method (Table 1). The mean values for the individual free fatty acids in normal serum were in good agreement with the published data. The present fluorimetric HPLC method using DMEQ-hydrazide gave a satisfactory sensitivity in quantitative analysis of fatty acids; the sensitivity permits the use of only 5 pL of normal human serum. DMEQ-hydrazide is very stable even at room temperature and reacts readily with fatty acids in aqueous solution under the mild derivatization conditions. Accordingly, the method is ideal for the derivatization of polyunsaturated fatty acids which are thermally labile. Furthermore, the present method does not require chloroform extraction of fatty acids and dehydration of the extracts, and permits the direct derivatization of serum fatty acids. This method is simple to perform and can therefore be used routinely.
REFERENCES Ghiggeri, G. M., Candiano, G., Delfino, G., Queirolo, C., Ginevri, F., Perfurno, F. and Gusmano, R. (1986). J. Chromatogr. 381, 411. Hatsumi. M., Kimata, S. and Hirosawa, K. (1986). J. Chromatogr. 380, 247. Ikeda, M., Shirnada, K., Sakaguchi, T. and Matsurnoto, U. (1984). J. Chromatogr. 305, 261. Kargas, G., Rudy, T.. Spennetta, T., Takayarna, K., Querishi, N. and Shrago, E. (1990). J. Chromatogr. 526, 331. Miwa, H. and Yamamoto, M. (1986). J. Chrornatogr. 351, 275. Shirnomura, Y., Taniguchi, K., Sugie, T., Murakami, M., Sugiyama, S. and Ozawa, T. (1984). Cfin. Chim. Acta 143, 361.
Shimomura, Y., Sugiyama, S., Takamura, T., Kondo, T. and Ozawa, T. (1986). J. Chromatogr. 383, 9. Tsuchiya, H., Hayashi, T., Naruse H. and Takagi, N. (1982). J. Chrornatogr. 234, 121. Tsuchiya, H., Hayashi, T., Sato, M., Tatsumi. M. and Takagi, N. (1984). J. Chrornatogr. 309, 43. Yarnaguchi, M., Matsunaga, R., Hara. S., Nakamura, M. and Ohkura, Y. (1986a). J. Chromatogr. 375, 27. Yamaguchi, M., Matsunaga, R., Fukuda, K., Nakarnura, M. and Ohkura, Y. (1986b). Anal. Biochem. 155, 256. Yarnaguchi, M., Iwata, T., Inoue, K., Hara, S. and Nakamura, M. (1990). Analyst. 115, 1363. Yanagisawa, I., Yamane. M. and Urayarna, T. (1985). J. Chromatogr. 345, 229.
Simple and Highly Sensitive Determination of Free Fatty Acids in Human Serum by High Performance Liquid Chromatography with Fluorescence Detection Tetsuharu Iwata, Kouji Inoue, Masaru Nakamura and Masatoshi Yamaguchi* Faculty of Pharmaceutical Sciences, Fukuoka University, Nanakuma, Jonan-ku, Fukuoka 814-01, Japan
A highly sensitive and simple reversed phase high performance liquid chromatographic(HPLC) method for the quantitative determination of free fatty acids in human serum is presented. The method is based on the direct derivatization of serum fatty acids with 6,7-dimethoxy-l-methyl-2(1H)-quinoxalinone-3-propionylcarboxylic acid hydrazide. The derivatization reaction proceeds in aqueous solution in the presence of pyridine and l-ethyl3-(3-dimethylaminopropyl)carbodiimide at 37 "C. The resulting derivatives are separated within 75 min on a reversed phase column (YMC Pack Cs) with a gradient elution of aqueous acetonitrile and detected fluorimetrically. The detection limits are 2.5-5 fmol in a 10 pL injection volume. The sensitivity permits precise determination of free fatty acids in 5 pL serum. The method is simple and is without the conventional liquidliquid extraction steps of serum fatty acids.
INTRODUCTION Free fatty acids arise mainly from the hydrolysis of triacylglycerol in adipose tissues or from the action of lipoprotein lipase, and are released into the blood stream. The amounts of the acids in human serum/ plasma increase or decrease with the extent of diabetes, thyremphraxis and hepatic dysfunction. The determination of the acids is therefore very important in the diagnosis and therapy of these diseases. Some high performance liquid chromatographic (HPLC) methods with fluorescence detection have focussed on the sensitive determination of free fatty acids in human serum/plasma. The methods include the derivatization of fatty acids with fluorescence derivatization reagents-9-anthryldiazomethane (Shimomura et al., 1984, 1986; Ghiggeri et al., 1986; Hatsumi et a/., 1986; Kargas et al., 1990), 9-aminonaphthalene (Ikeda et al., 1984), 4-bromomethyl-7-acethoxycoumarin (Tsuchiya et af., 1982, 1984), 5-(dimethylamino)-lnaphthalenesulphonyl-semipiperazide(Yanagisawa et al., 1985) and 3-bromomethyl-6,7-dimethoxy-l-methyl2(IH)-quinoxalinone (Yamaguchi et al., 1986a, b ) . However, these reagents are not stable and generally require most dried aprotic solvents in the derivatization reaction. Therefore, the fatty acids should be extracted with organic solvents such as chloroform and nheptane, and the organic layer should be evaporated to dryness prior to the derivatization. Furthermore, the reagents need to react for a long time at high temperatures. Thus, the conventional fluorimetric HPLC methods are time-consuming and not ideal for routine clinical investigations. Recently, we have developed 6,7-dimethoxy1-methyl-2(IH) - quinoxalinone-3 -propionylcarboxylic acid hydrazide (DMEQ-hydrazide) as a highly sensitive Author to whom correspondence should be addresscd. 0269-3879/92/030120-04 $05 .oO @ 1992 by John Wiley & Sons. Ltd.
DMEQ-hydrazide CH3 I
CH,O*N+
CH~CH~CONHNHCOR
Scheme 1. Derivatization of fatty acids with DMEQ-hydrazide.
fluorescence derivatization reagent for fatty acids (Yamaguchi et af., 1990). The reagent readily reacts with fatty acids in aqueous solution in the presence of pyridine and l-ethyl-3-(3-dimethylaminopropyl)carbodiimide (EDC) at room temperature (0-37 "C) to produce intensely fluorescent DMEQ derivatives (Scheme 1). The purpose of the present research was aimed at establishing a sensitive and simple HPLC method utilizing DMEQ-hydrazide for the determination of free fatty acids in a minute amount of serum without cumbersome clean-up procedures.
EXPERIMENTAL Reagents and materials. All chemicals and solvents were of analytical reagent grade, unless otherwise stated. Deionized and distilled water was used. Lauric (C,,,,), myristoleic (C14:,),myristic (C13:0)r linolenie (Clx.n),eicosapentaenoic (C20:5 ) , palmitoleic (C16 :J, linoleic (CIS,*), arachidonic (C20.4)r docosahexaenoic (C22:6),palmitic (C16:o),dihomo-ylinolenic (Go:3), oleic (Clx,l),margaric (C17:o)and stearic acids were purchased from Sigma (St. Louis, MO, USA). DMEQ-hydrazide was prepared as described preReceived 4 July 1991 Accepted 30 August 1991
DETERMINATION OF FREE FA'ITY ACIDS I N HUMAN SERUM
viously (Yamaguchi et al., 1990). DMEQ-hydrazide (50 mM) solution was prepared in N,N-dimethylformamide. The DMEQ-hydrazide solution could be used for at least 1month even at room temperature. Pyridine (4%) solution was prepared in 20 mM hydrochloric acid ethanolic solution. Stock M) of fatty acids were prepared in solutions (2 x 10- 7-1 x ethanol. Serum specimens were obtained from fasting healthy volunteers in our laboratories.
HPLC apparatus and conditions. A Jasco Trirotor-V high performance liquid chromatograph (Tokyo, Japan), equipped with a Model VL-614 injector (100 pL loop), and a Jasco FP-210 spectrofluorometer, equipped with a 15 yL flow cell operated at an excitation wavelength of 360nm and an emission wavelength of 435 nm, were used. The column was a YMC-Pack Cx (250 x 4.6 mm i d . . particle size 10 pm, Yamamura Chemical Laboratories Kyoto, Japan). The column temperature was maintained at 30-+0.2"C with a Shimadzu CTO-6A column oven (Kyoto, Japan). For the separation of the DMEQ derivatives of the fatty acids on the column, a gradient elution with aqueous 5595% acetonitrile (Fig. 1) was carried out by using a Jasco GP-A40 Gradient Programmer. The flow rate was 1.O mL/min. Uncorrected fluorescence excitation and emission spectra of the eluate were measured with a Hitachi 650-60 fluorescence spectrophotometer fitted with a 20 yL flow cell; the spectral bandwidths were 5 n m in both the excitation and emission monochromators.
121
Procedure. A 5 pL aliquot of serum was mixed with 5 pL ethanol (or fatty acid standard solutions) and SO yL pyridine solution. The mixture was vortex mixed for ca. 10 s and then 25 pL DMEQ-hydrazide solution and 15 pL 2 M E D C in water were added. The resulting mixture was warmed at 37 "C for 10 min and was centrifuged at 1000 x g for ca. 5 min. The supernatant (10 pL) was injected into the chromatograph. The amounts of the free fatty acids were calibrated by means of the standard addition method: the 5 y L ethanol added to the serum in the procedure was replaced by 5 pL acid standard solution. The net peak heights of the individual fatty acids were plotted against the concentrations of the spiked fatty acids.
RESULTS AND DISCUSSION Derivatization reaction conditions
The optimal derivatization conditions for linear saturated fatty acids were previously examined using pyridine solution which was free from hydrochloric acid (Yamaguchi et al., 1990). We recently found that the yields of DMEQ derivatives of fatty acids increase in the presence of hydrochloric acid in the pyridine solution. Thus, the derivatization conditions were improved using pyridine solutions containing hydrochloric acid. l I
.-0
100
+J (u
I
w c W
4 80
V
co > W-.
a,?
60
_-_.
- w ..4
v
I
+J
.A
c
40
0
w
a,
W 4:
a, v)
l=
0
2
a v,
a, L
8
3
L 0
I
I
c,
u
12
a, 4-J a,
n
I
I
I
I
I
0
20
40
60
80
Retention tr?e ( m i n t Figure 1. Chromatogram of the DMEQ derivatives of fatty acids. A portion (5 pL) of a standard solution of fatty acids (2 nmol/rnL each) was treated as described in the text. Peaks: 1 =C,2:o; 2=C1,,1; 3 = C 1 4 : 0 ; 4 = C l S r 3 ; 5=C20:6; 6=C16.1; 7=c18:2; 8=c22:6 and C20:4; 9=c,6:,; I O = ~ , , : , ; 11 = C18:1; 12 = CI7:0; 1 3 = C18 o; 1 4 = DMEQ-hydrazide. Gradient elution with aqueous acetonitrile (0-56 min, 5 5 % ; 56-72 min, 55-95%; 72-76 min, 9 5 % 76-80 min, 55%).
TETSUHARU IWATA ET AL
122
Hydrochloric acid gave maximum and constant peak heights at the concentration of 1 0 - 3 0 m ~in the pyridine solution; the maximum peak height is 1.1 times that in the absence of hydrochloric acid in the pyridine solution. Hence, 20 mM hydrochloric acid was selected considering the following step of extracting the fatty acids from the serum. DMEQ-hydrazide gave the most intense peaks at concentrations greater than ca. 7 mM in the solution for all the fatty acids; 50mM was used as an optimum concentration. EDC and pyridine were used to facilitate the derivatization of fatty acids with DMEQ-hydrazide. Maximum and constant peak heights were attained at pyridine concentrations in the solution of 2-6010 ; 4% was selected as optimum. The peak heights for the acids were maximum and constant at concentrations of EDC higher than 1.8M; 2.0 M was used. The derivatization reaction of fatty acids with DMEQ-hydrazide apparently occurred even at 0 "C; higher temperatures allowed the fluorescence to develop more rapidly. However, at 60°C the peak heights decreased. At room temperature (15-37 "C), the peak heights for all the fatty acids were almost maximum after standing for 10 min. Hence, the solution was allowed to stand at 37°C for 10min. The DMEQ derivatives in the final mixture were stable for at least 24 h in daylight at room temperature. The detection limits for the fatty acids were 2-7 fino1 at a signal-to-noise ratio of 3, and became slightly better than those in a previous paper (Yamaguchi et al., 1990).
14
7 8
3
9
!
tl
20
0
13
60
40
80
Retention time (min) Figure 2. Chromatogram of the DMEQ derivatives of free fatty acids in normal human serum. A portion (5 pL) of serum was treated as in the detailed procedure. Peak number as in Fig. 1. Detector sensitivity: -, 1; ---, 16.
no effect on the fluorescence excitation (360nm) and emission maximum wavelengths (435 nm) and the fluorescence intensities of the DMEQ derivatives of all the fatty acids. The individual fatty acids gave single peaks in the chromatogram.
HPLC conditions
Determination of free fatty acids in human serum
The best separation of the DMEQ derivatives of 14 saturated and unsatured fatty acids was achieved on a reversed phase column, YMC Pack C,, by gradient elution with aqueous 55-95% ( v h ) acetonitrile (Fig. 1). However, the peaks for C22:6 and C20:4acids could not be resolved successfully under any HPLC conditions tested. A typical chromatogram obtained with a standard solution of the fatty acids is shown in Fig. 1. The change in acetonitrile concentration actually had
Chloroform has been commonly used for the liquidliquid extraction of fatty acids from serum. However, Hatsumi et ul. (1986) reported that some fatty acids were released from some phospholipids in biological materials during the chloroform extraction. To overcome the problems, Miwa and Yamamoto (1986) used ethanol both as a precipitating agent for serum proteins and an extraction solvent for serum fatty acids. Moreover, the serum sample was acidified with hyd-
Table 1. Concentration of free fatty acids in serum from healthy persons Concentration InmollmL) Age
Sex
C12.a
Ci4:i
C14.a
Cia 3
43 40 32 23 23 22 22 26 23 22 22 22 22 22 22
M M M M M M M F F
1.43 1.54 1.49 1.16 2.00 2.88 1.09 7.44 3.78 0.91 1.05 5.04 2.48 5.16 2.76
0.90 0.71 0.11 0.47 0.97 0.51 0.45 1.02 1.18 0.31 0.27 2.35 0.90 3.23 0.72
7.74 5.85 2.74 4.70 6.55 7.63 4.32 10.74 11.06 4.32 3.06 13.92 11.57 17.35 7.50
5.38 6.28 2.28 5.47 12.72 10.94 9.98 9.03 11.22 3.15 1.96 21.89 17.37 19.15 9.30
2.68 1.85
0.93 0.81
7.94 4.08
9.74 5.87
Mean S.D.
F F
F F F F
CZO:~
1.62 2.17 0.46 0.53 1.26 0.69 0.55 0.89 0.79 0.57 0.32 2.03 1.42 1.30 1.01
c16.1
C18:2
%:o
Cm:3
7.35 35.20 86.17 0.65 7.77 35.20 67.31 0.78 0.80 12.64 22.68 0.61 5.64 32.50 66.78 0.34 11.28 72.22 88.85 0.82 11.28 80.79 115.24 1.10 53.72 74.85 1.01 8.50 87.24 0.44 10.60 50.55 49.66 87.78 0.76 7.36 26.19 3.84 40.39 0.63 2.42 13.54 30.71 0.34 28.45 111.04 162.62 0.82 14.55 123.68 155.63 0.78 34.35 115.09 174.48 0.89 58.22 113.63 0.38 8.34
1.04 10.84 0.56 8.83
58.04 34.63
91.58 0.69 44.56 0.23
Cle I
C1s.a
87.93 62.90 17.59 58.85 112.96 133.27 104.84 100.84 98.10 40.59 27.06 253.00 179.26 224.60 109.59
27.98 22.07 10.88 27.56 30.93 34.20 31.55 27.82 35.75 12.90 14.62 55.45 44.1 5 55.1 1 32.80
107.42 30.92 65.65 12.91
DETERMINATION OF FREE FATTY AClDS IN HUMAN SERUM
rochloric acid to convert the serum fatty acids into the non-ionic forms from the dissociated forms. Thus, the acidic ethanol was used in the present work. The best recovery was made by adding more than 40vL of ethanolic hydrochloric acid (15-25 mM) solution. Ethanolic 20 r n M hydrochloric acid solution (50 pL) containing 4% pyridine was employed in the procedure. The recoveries (n= 8 each) of the fatty acids (0.2 nmol per 5 pL each) added to a pooled normal serum were 100.0-105.9%. A representative chromatogram obtained with normal human serum is shown in Fig. 2. All the peaks for the fatty acids were identified on the basis of their retention times and the fluorescence excitation and emission spectra of the eluates in comparison with the standard compounds, and also by co-chromatography of the standards and the serum sample with aqueous 50-100% acetronitrile or methanol as the mobile phase. Many substances such as alcohols, sugars, aldehydes, ketones, phenols, amines and a-keto acids gave no fluorescent derivatives under the described conditions. On the other hand, 17 different a-amino acids, hydroxycarboxylic acids (lactic and malic) , dicarboxylic acids (oxalic, malonic, succinic and adipic) and aromatic carboxylic acids (benzoic, salicylic and cinnamic) reacted with DMEQ-hydrazide to produce fluorescent derivatives. However, these compounds did not interfere with either the detection or the separation of the peaks for all the fatty acids even when they were spiked at unusually high concentrations in serum (10 pmol/mL of serum), because DMEQ derivatives of the compounds were co-eluted with DMEQ-hydrazide under the recommended HPLC conditions. Linear relationships were observed between the peak heights of the individual fatty acids and the amounts of the fatty acids added in the range 1pmol-5 nmol each
123
to 5 pL of serum. The precision was established by repeated determinations ( n = 7) using a normal serum. The coefficients of variation did not exceed 3.0% for all the fatty acids (n=7 in each case). C17 acid has been widely used as an internal standard for the analytical methods of free fatty acids in human plasma. However, Hatsumi et al. (1986) reported that C,, acid is present in human serum. Furthermore, the peak for C17 acid appeared in the chromatogram (Fig. 2) obtained with normal human serum. Hence, we did not use Ci7 acid as an internal standard in this study, since good precision was given even using the standard addition method. When cephalin and sphinogomyelin were treated as in the method, no peaks for the fatty acids were observed in the chromatogram. This supports the fact that no fatty acids were released from phospholipids during the present extraction and derivatization steps. The concentrations of free fatty acids in sera from healthy volunteers were determined by this method (Table 1). The mean values for the individual free fatty acids in normal serum were in good agreement with the published data. The present fluorimetric HPLC method using DMEQ-hydrazide gave a satisfactory sensitivity in quantitative analysis of fatty acids; the sensitivity permits the use of only 5 pL of normal human serum. DMEQ-hydrazide is very stable even at room temperature and reacts readily with fatty acids in aqueous solution under the mild derivatization conditions. Accordingly, the method is ideal for the derivatization of polyunsaturated fatty acids which are thermally labile. Furthermore, the present method does not require chloroform extraction of fatty acids and dehydration of the extracts, and permits the direct derivatization of serum fatty acids. This method is simple to perform and can therefore be used routinely.
REFERENCES Ghiggeri, G. M., Candiano, G., Delfino, G., Queirolo, C., Ginevri, F., Perfurno, F. and Gusmano, R. (1986). J. Chromatogr. 381, 411. Hatsumi. M., Kimata, S. and Hirosawa, K. (1986). J. Chromatogr. 380, 247. Ikeda, M., Shirnada, K., Sakaguchi, T. and Matsurnoto, U. (1984). J. Chromatogr. 305, 261. Kargas, G., Rudy, T.. Spennetta, T., Takayarna, K., Querishi, N. and Shrago, E. (1990). J. Chromatogr. 526, 331. Miwa, H. and Yamamoto, M. (1986). J. Chrornatogr. 351, 275. Shirnomura, Y., Taniguchi, K., Sugie, T., Murakami, M., Sugiyama, S. and Ozawa, T. (1984). Cfin. Chim. Acta 143, 361.
Shimomura, Y., Sugiyama, S., Takamura, T., Kondo, T. and Ozawa, T. (1986). J. Chromatogr. 383, 9. Tsuchiya, H., Hayashi, T., Naruse H. and Takagi, N. (1982). J. Chrornatogr. 234, 121. Tsuchiya, H., Hayashi, T., Sato, M., Tatsumi. M. and Takagi, N. (1984). J. Chrornatogr. 309, 43. Yarnaguchi, M., Matsunaga, R., Hara. S., Nakamura, M. and Ohkura, Y. (1986a). J. Chromatogr. 375, 27. Yamaguchi, M., Matsunaga, R., Fukuda, K., Nakarnura, M. and Ohkura, Y. (1986b). Anal. Biochem. 155, 256. Yarnaguchi, M., Iwata, T., Inoue, K., Hara, S. and Nakamura, M. (1990). Analyst. 115, 1363. Yanagisawa, I., Yamane. M. and Urayarna, T. (1985). J. Chromatogr. 345, 229.
