Anal. Chem. 2001, 73, 698-702
Determination of Peptides and Proteins in Fats and Oils Francisco J. Hidalgo, Manuel Alaiz, and Rosario Zamora*
Consejo Superior de Investigaciones Cientı´ficas, Instituto de la Grasa, Avenida Padre Garcı´a Tejero 4, 41012 Sevilla, Spain
A method for the determination of proteins in fats and oils is described. Proteins were sequentially precipitated with acetone and hydrolyzed, and the produced amino acids were fractionated and quantificated. This analysis protocol afforded a method of high sensitivity and specificity which was fully evaluated and validated. The data obtained showed good accuracy and linearity with excellent reproducibility and recovery. When the method was applied to 40 olive oils, all of them contained proteins in the range 10-50 µg/100 g of oil, suggesting that proteins are nonpreviously described minor components of these oils. In addition, the proteins precipitated were almost exclusively composed by one polypeptide of apparent 4600 molecular weight, which was isolated from olive drupes and partially characterized by amino acid analysis. Similar polypeptides were also detected in other seeds, suggesting that they may constitute a new class of polypeptides in plants with oleosin-like characteristics. Furthermore, the method was also applied to different fats and oils, and all the samples analyzed contained proteins, suggesting that natural fats and oils always contain polypeptides and/or proteins as minor components. These results also suggest that some peptides are soluble in lipid matrixes, where they might be playing unknown functions. The developed procedure provides a methodology for the determination of these components. Olive oil is one of the oldest known vegetable oils. It is extracted from the fruit of the olive tree Olea europaea, and it is almost unique among vegetable oils because it can be consumed, without any refining treatment, in its crude form called virgin olive oil.1 It has also gained popularity in recent years, which can be attributed, in part, to reports of its potential health benefits, including a specific reduction in plasma non-high-density lipoprotein cholesterol levels and a protective effect in some types of cancer.2,3 As a consequence of this demand, the interest in the development of new methods of analysis,4-6 as well as for the * Corresponding author: (phone) +34 954 611550; (fax) +34 954 616790; (e-mail)
[email protected]. (1) Boskou, D. Olive Oil. Chemistry and Technology; AOCS Press: Champaign, IL, 1996. (2) Bruckner, G. In Fatty Acids in Foods and Their Health Implications; Chow, C. K., Ed.; Marcel Dekker: New York, 2000; pp 843-863. (3) Simonsen, N. R.; Ferna´ndez-Crehuet Navajas, J.; Martı´n-Moreno, J. M.; Strain, J. J.; Huttunen, J. K.; Martin, B. C.; Thamm, M.; Kardinaal, A. F.; van’t Veer, P.; Kok, F. J.; Kohlmeier, L. Am. J. Clin. Nutr. 1998, 68, 134141.
698 Analytical Chemistry, Vol. 73, No. 3, February 1, 2001
finding of new components that distinguish this oil from other oils,7 has been increased. In this context, the presence of different enzyme activities in virgin olive oils has been described.8,9 This implies the presence of proteins in these oils, although no previous reports have indicated that olive oils contain polypeptides or proteins. In addition, proteins have been detected in some other vegetable oils.10 However, this last study employed Bradford’s procedure,11 which was neither developed nor tested to be used with lipid matrixes. In fact, to our knowledge, none of the procedures developed for protein determination has yet been demonstrated to be useful in lipid matrixes. The present study was undertaken to develop a procedure that can be applied to the determination of proteins in fats and oils. Because olive oils can be consumed without refiningsa process where proteins may be lostsand proteins might contribute to some of the special characteristics of these oils, the procedure was primarily developed for its use in olive oils, and, lately, extended to other fats and oils. EXPERIMENTAL SECTION Materials. Virgin olive oils, refined olive oils, and crude soybean oils were obtained from our Institute’s experimental oil mill (Instituto de la Grasa, Sevilla, Spain) or Koipe S. A. (Andujar, Jae´n, Spain). Mature seeds of broad bean (Vicia faba L.), cabbage (Brassica oleracea L.), carrot (Daucus carota L.), corn (Zea mays L.), pine seed (Pinus pinea L.), rapeseed (Brassica napus L.), safflower (Carthamus tinctorius L.), soybean (Glycine max L.), and sunflower (Helianthus annuus L.) were obtained from Koipesol Semillas S. A. (Seville, Spain) or local groceries. Lard was bought in the supermarket. Olive drupes (O. europaea L. cv. Picual and cv. Arbequina) were harvested from our Institute’s field station. Reagents and solvents used were of analytical grade and were purchased from reliable commercial sources. Isolation of Peptides and Proteins from Fats and Oils. The oil (40 g) was maintained at 18 °C for at least 90 min prior to (4) Zhang, B.-L.; Trierweiler, M.; Jouitteau, C.; Martin, G. J. Anal. Chem. 1999, 71, 2301-2306. (5) Tian, K.; Dasgupta, P. K. Anal. Chem. 1999, 71, 1692-1698. (6) Tian, K.; Dasgupta, P. K. Anal. Chem. 1999, 71, 2053-2058. (7) Garcia-Mesa, J. A.; Luque de Castro, M. D.; Valcarcel, M. Anal. Chem. 1993, 65, 3540-3542. (8) Georgalaki, M. D.; Sotiroudis, T. G.; Xenakis, A. J. Am. Oil Chem. Soc. 1998, 75, 155-159. (9) Georgalaki, M. D.; Bachmann, A.; Sotiroudis, T. G.; Xenakis, A.; Porzel, A.; Feussner, I. Fett/Lipid 1998, 100, 554-560. (10) Klurfeld, D. M.; Kritchevsky, D. Lipids 1987, 22, 667-668. (11) Bradford, M. M. Anal. Biochem. 1976, 72, 248-254. 10.1021/ac000876o CCC: $20.00
© 2001 American Chemical Society Published on Web 12/30/2000
its treatment with 98 mL of acetone, which was previously cooled at 4 °C. The resulting mixture was maintained at 4 °C for 30 min and then filtered through Whatman no. 1 filter paper using a Buchner funnel. The paper was extracted by shaking it first in the presence of 5 mL of tetrahydrofuran and then with 5 mL of dioxane. The extracts were combined and taken to dryness with nitrogen. Quantification of Peptides and Proteins in the Isolated Fraction. Peptides and proteins were quantified by using amino acid analysis. The isolated extracts plus D,L-R-aminobutyric acid, which was added as internal standard, were dissolved in 1 mL of 6.0 M hydrochloric acid and hydrolyzed for 20 h at 110 °C. The hydrolyzed samples obtained were taken to dryness, dissolved in 3 mL of 1 M sodium borate buffer (pH 9.0), and derivatized with diethyl ethoxymethylenemalonate. Protected amino acids were, finally, fractionated by reverse-phase high-performance liquid chromatography (HPLC) with UV detection at 280 nm using a previously described gradient.12,13 Protein content was calculated from amino acid data. This method is the most accurate method for determining protein concentration.14 In addition, in the present study, it was much more exact than the Lowry15 or Bradford11 procedures, more likely as a consequence of the low solubility of the proteins studied and the interferences that residual lipids produced in those methods. Partial Characterization of the Polypeptides Obtained from Fats and Oils. Peptides and proteins obtained from fats and oils were characterized both by electrophoresis and amino acid analysis. Tricine-sodium dodecyl sulfate-polyacrylamide gel electrophoresis (tricine-SDS-PAGE) was performed according to Scha¨gger and Jagow16 with 16.5% total acrylamide gels containing 3% cross-linker. Gels were silver stained according to Morrissey.17 Briefly, gels were treated successively with glutaraldehyde, dithiothreitol, silver nitrate, formaldehyde/sodium carbonate, citric acid, water, Farmer’s reducer, and new silver staining. A typical calibration curve (r ) -0.987, p ) 0.018) was obtained with bovine serum albumin (66.0 kDa), chicken egg albumin (45.0 kDa), bovine erythrocytes carbonic anhydrase (29.0 kDa), chicken egg white lysozyme (14.3 kDa), and bovine insulin chain B (3.5 kDa). Amino acid analysis was carried out by both high-performance liquid chromatography, as described above, and gas chromatography (GC), according to Chen et al.18 For the GC method, amino acids were converted to tert-butyldimethylsilyl (TBDMS) derivatives prior to separation by capillary GC using a DB-1 fused-silica capillary column and detection with a flame ionization detector. Isolation from Olive Drupes of the Polypeptide Found in Olive Oils. The polypeptide found in olive oils was isolated from olive drupes to obtain an enriched fraction that could be employed to validate the developed procedure. This polypeptide was isolated from oil bodies of oil drupes. Oil bodies and microsomes were (12) Alaiz, M.; Navarro, J. L.; Giro´n, J.; Vioque, E. J. Chromatogr. 1992, 591, 181-186. (13) Hidalgo, F. J.; Zamora, R. Chem. Res. Toxicol. 2000, 13, 501-508. (14) Waterborg, J. H.; Mattheus, H. R. In The Protein Protocols Handbook; Walker, J. M., Ed.; Humana Press: Totowa, NJ, 1996; pp 7-9. (15) Lowry, O. H.; Rosebrough, N. J.; Farr, A. L.; Randall, R. J. J. Biol. Chem. 1951, 193, 265-275. (16) Scha¨gger, H.; von Jagow, G. Anal. Biochem. 1987, 166, 368-379. (17) Morrissey, J. H. Anal. Biochem. 1981, 117, 307-310. (18) Chen, Z.; Landman, P.; Colmer, T. D.; Adams, M. A. Anal. Biochem. 1998, 259, 203-211.
prepared at 0-4 °C following the procedures of Murphy et al.19 and Garce´s et al.,20 respectively, which were modified. Briefly, the mesocarp of olive drupes was homogenized in buffer with a Waring blender and the seeds were homogenized using a pestle and mortar. The homogenization medium (4 mL/g of tissue) contained 0.4 M sucrose, 100 mM Hepes/NaOH, pH 7.5, 10 mM KCl, 1 mM MgCl2, 1 mM EDTA, and 1% (w/v) ascorbic acid. The homogenate was then filtered through four layers of cheesecloth and centrifuged at 5000g for 15 min. The fat layer containing the crude oil bodies was recovered from the top of the 5000g supernatant, and the supernatant free of fat was used for the microsomal preparation (see below). The crude oil bodies were dispersed in 5 volumes of buffer and layered beneath a further 20 volumes of buffer containing 0.1 M sucrose. This was then centrifuged at 18000g for 15 min, after which the oil body fraction was again recovered from the top of the gradient. This dispersallayering-centrifugation procedure was repeated two more times. The purified oil bodies were dispersed in 3 volumes of water and extracted with 3 volumes of chloroform-methanol (2:1) in order to remove lipids. The layers were separated by centrifugation at 650g for 5 min and then removed. The defatted oil bodies were washed twice with chloroform-methanol (2:1), then dispersed in 25 mM sodium borate buffer, pH 8.35, containing 25 mM SDS, and heated for 1 h at 100 °C. The protein was recovered by centrifugation and dried under nitrogen. The whole procedure was repeated, and the resulting protein was dried under nitrogen and stored at -28 °C. This protein fraction was used to validate the procedure. The supernatant free of fat described above was centrifuged at 100000g for 60 min. The microsomal pellet was resuspended in phosphate buffer (0.1 M, pH 7.2) containing 0.33 M sucrose and was centrifuged at 100000g for 60 min. Finally, the microsomal pellet was resuspended in phosphate buffer (0.1 M, pH 7.2) and used as microsomal preparation. Method Validation. Linearity. Standard curves were prepared over a concentration range of 10-200 µg of protein/100 g of oil. For each curve, six different concentration levels were used. Protein determined for each oil was plotted against the protein added. The Y-intercept provided the amount of protein originally present in the oil. Recovery. Six different solutions of the protein obtained from olive drupes in an oil with a known concentration of protein were used in recovery determinations. The assay recovery was determined as follows: (protein determined)/(protein added + protein originally present in the oil) × 100. Precision. Precision was evaluated by analyzing different oils with different concentrations of proteins within the range 10200 µg/100 g of oil, on the same day (between three and seven replicates, within-day reproducibility) and over separate days (between 6 and 12 replicates, total reproducibility). Limits of Detection and of Quantification. The limit of detection (LOD) and limit of quantification (LOQ) were determined by analyzing decreasing concentrations of protein in refined olive oils. LOD was established as the lowest concentration that produces a response 3 times the noise level. The LOQ was defined as the lowest concentration that could be quantitated with a precision (19) Murphy, D. J.; Cummins, I.; Kang, A. S. Biochem. J. 1989, 258, 285-293. (20) Garce´s, R.; Sarmiento, C.; Mancha, M. Planta 1994, 193, 473-477.
Analytical Chemistry, Vol. 73, No. 3, February 1, 2001
699
Table 1. Effect of Temperature and Oil/Acetone Ratio on Protein Determined in a Virgin Olive Oil oil/acetone (°C)
oil/protein ratio
proteina
oil (°C)
acetone (°C)
(µg/100 g)
oil (g)
acetone (mL)
proteina (µg/100 g)
18 18 18 50
-28 4 18 18
26.0 44.5 34.8 31.0
30 40 50 60
110 98 85 70
39.9 44.5 38.2 37.8
a Values corresponded to the media of at least two independent experiments.
less than 15% and was established as the lowest point of the calibration curve. Extension of the Method to Other Fats and Oils. The method was applied as described above to the different vegetable oils. However, for the lard, the mixture of lard and acetone had to be heated at 40 °C for 5 min prior to the filtration to get the dissolution of the lard. RESULTS AND DISCUSSION Presence of Peptides and Proteins in Fats and Oils. The presence of peptides and proteins in vegetable oils may be easily detected, for example, by extraction of the oils with water and analysis of the aqueous fraction obtained by using Bradford’s reagent or amino acid analysis. However, to obtain quantitative results, the best procedure resulted from the precipitation of peptides and proteins with acetone and its determination by amino acid analysis. The extraction with water was not always complete, because peptides and proteins present in vegetable oils were not too soluble in water, and the presence of trace amounts of amino acids in the water was also an interference when a large amount of water was employed. In addition, residual amounts of organic solvents interfered with Bradford’s procedure, and the use of Lowry’s reagent always produced recoveries higher than the amount added to the oil (data not shown). In the precipitation with acetone, both the temperature of the oil and the solvent and the oil/acetone proportion were important. Table 1 shows the effect of the temperature and the oil/acetone ratio on the protein determined in a virgin olive oil. Two different temperatures for the oil and three for the acetone were assayed, and the resulting mixture was maintained at the temperature of the solvent for 30 min before filtration or centrifugation. The highest protein content was obtained when the oil was stored at 18 °C for at least 90 min, the acetone at 4 °C, and the resulting mixture at 4 °C for 30 min. Filtration and/or centrifugation of the protein always produced analogous results. Therefore, most of the experiments carried out in the present study were obtained by filtration of the protein, which resulted in the simplest way of recovering protein. Also few differences were observed for the oil/protein ratios assayed, and the use of 30-60 g of oil always produced similar results (Table 1). A proportion of 40 g of oil and 98 mL of acetone was selected, however, because it recovered a slight higher quantity of protein. Partial Characterization of the Proteins Obtained from Fats and Oils. The mixture of proteins precipitated with acetone 700 Analytical Chemistry, Vol. 73, No. 3, February 1, 2001
Figure 1. Tricine-SDS-PAGE of the following: 1, standard proteins; 2, proteins obtained from a virgin olive oil; 3, oil body proteins obtained from the mesocarp of olive drupes, cv. Picual; 4, oil body proteins obtained from the seeds of olive drupes, cv. Picual; 5, microsomal proteins obtained from the seeds of olive drupes, cv. Picual; and 6, standard proteins. Standard proteins included the following: bovine serum albumin (66.0 kDa), chicken egg albumin (45.0 kDa), bovine erythrocytes carbonic anhydrase (29.0 kDa), chicken egg white lysozyme (14.3 kDa), and bovine insulin chain B (3.5 kDa).
was studied by tricine-SDS-PAGE. Figure 1, lane 2, shows a typical gel obtained for the oils studied. Surprisingly, the proteins precipitated with acetone in the oils were almost exclusively composed by one polypeptide of apparent 4600 molecular weight. This fraction was studied by amino acid analysis. Table 2 collects the amino acid composition determined by both HPLC and GC. HPLC and GC procedures gave analogous results and indicated that the fraction precipitated with acetone had a small number of basic and sulfur-containing amino acids, a high number of amino acids with hydrophobic groups, and high content in asparagine/aspartic acid and glutamine/glutamic acid. This amino acid composition was similar to that described for oleosins.21-24 In addition, this fraction was soluble in buffers containing high concentrations of urea similar to that described for oleosins.25 Isolation from Olive Drupes of the Polypeptide Found in Olive Oils. To obtain the 4.6-kDa polypeptide in high quantity to validate the method developed, the presence of this polypeptide was investigated in the subcellular fractions of olive drupes associated to lipids: oil bodies and microsomes, both in the mesocarp and in the seeds of olive drupes. Figure 1 shows the (21) Vance, V. B.; Huang, A. H. C. J. Biol. Chem. 1987, 262, 11275-11279. (22) Lee, K.; Huang, A. H. C. Plant Physiol. 1991, 96, 1395-1397. (23) Hatzopoulos, P.; Franz, G.; Choy, L.; Sung, R. Z. Plant Cell 1990, 2, 457467. (24) Kalinski, A.; Loer, D. S.; Weisenann, J. M.; Matthews, B. J.; Herman, E. M. Plant Mol. Biol. 1991, 17, 1095-1098. (25) Millichip, M.; Tatham, A. S.; Jackson, F.; Griffiths, G.; Shewry, P. R.; Stobart, A. K. Biochem. J. 1996, 314, 333-337.
Table 2. Amino Acid Composition of the Protein Fraction Isolated from Virgin Olive Oils number of residues amino acid
HPLC
GC
mol %
Ala Arg Asxa Cys Glxa Gly His Ile Leu Lys Met Phe Pro Ser Thr Tyr Val
3.60 ( 0.27 2.09 ( 0.12 5.22 ( 0.29 0.28 ( 0.02 4.62 ( 0.31 4.11 ( 0.20 0.93 ( 0.10 2.84 ( 0.12 4.03 ( 0.15 1.52 ( 0.07 0.83 ( 0.03 2.09 ( 0.16 2.85 ( 0.11 2.65 ( 0.09 1.69 ( 0.09 3.61 ( 0.14
3.54 ( 0.16
8.0 4.7 10.7 0.8 10.0 8.8 2.2 6.0 9.0 3.1 2.0 4.5 4.8 8.5 5.6 3.9 7.6
a
4.36 ( 0.23 0.45 ( 0.05 4.30 ( 0.12 3.79 ( 0.10 1.00 ( 0.18 2.53 ( 0.04 1.24 ( 0.10 0.92 ( 0.03 1.94 ( 0.09 2.13 ( 0.11 2.74 ( 0.04 2.36 ( 0.06 1.81 ( 0.09 3.21 ( 0.12
Asx, aspartic acid + asparagine; Glx, glutamic acid + glutamine.
tricine-SDS-PAGE obtained for the oil body proteins obtained from the mesocarp (lane 3), the oil body proteins obtained from the seeds (lane 4), and the microsomal proteins obtained from the seeds (lane 5). Microsomal proteins obtained from the mesocarp were identical to oil body proteins obtained from the mesocarp (data not shown). When the gel was stained with Coomassie Blue, the only bands observed corresponded to the oleosins (two bands of ∼20 kDa in lanes 4 and 5), according to the literature.26 However, when it was silver stained, other bands could be observed. Thus, the lane corresponding to the oil bodies of the mesocarp of olive drupes (lane 3) only exhibited a single band, which corresponded to the polypeptide found in olive oils (lane 2). This band was also present in the oil bodies of the seeds (lane 4), but the seeds also exhibited the bands corresponding to the oleosins as described previously.26 When microsomes from olive seeds were studied by tricine-SDS-PAGE (lane 5), the band corresponding to the 4.6-kDa polypeptide was also present but it was very weak, suggesting that the polypeptide was a minor compound in this fraction. Although its presence in this fraction was always as a minor component, it was found in all the experiments carried out, suggesting that it was not a contamination produced during the isolation of this subcellular fraction. On the other hand, the 4.6-kDa polypeptide was absent when soluble proteins from olive drupes were studied by tricine-SDS-PAGE (data not shown). The oil body fraction of olive mesocarp was rich in the 4.6kDa polypetide, and this was the only peptidic component present in the fraction as deduced by tricine-SDS-PAGE. Therefore, this fraction was selected for its isolation in large scale. When the oil body fraction was defatted with chloroform-methanol (2:1), the solid obtained contained 8.9 ( 0.6% (n ) 4) of protein as deduced by amino acid analysis. Several procedures were assayed for increasing the protein content in this fraction, but most of them were unsuccessful. These included the treatment with several (26) Ross, J. H. E.; Sanchez, J.; Millan, F.; Murphy, D. J. Plant Sci. 1993, 93, 203-210.
Figure 2. Amino acid profile obtained by HPLC after acid hydrolysis and derivatization with diethyl ethoxymethylenemalonate of proteins obtained by acetone precipitation: A, virgin olive oil; B, crude soybean oil; and C, lard. Peaks are labeled with single-letter notations for amino acids; internal standard (IS), D,L-R-aminobutyric acid; *, reagent, diethyl ethoxymethylenemalonate.
buffers at different pHs and containing several concentrations of Triton X-100 and SDS. The best results were obtained with 0.20.4 N NaOH which dissolved almost completely the protein. However, the 4.6-kDa polypeptide was sensitive to this pH and the tricine-SDS-PAGE showed its decomposition (data not shown). To concentrate the 4.6-kDa polypeptide without appreciable changes as observed by both tricine-SDS-PAGE and amino acid analysis, the best results were attained by treating the defatted oil body fraction twice with sodium borate buffer, pH 8.35, containing 25 mM SDS and heating for 1 h at 100 °C. This treatment dissolved partially other nonproteinic components present in the defatted oil body fraction and the percentage of protein in the solid increased to 42.4 ( 3.3% (n ) 4). This solid was used to add to the oils for the validation of the method. Method Validation. Linearity. The calibration curves were linear over the specified range (10-200 µg/100 g of oil). A correlation coefficient of 0.993 or higher was obtained for the relationship between the protein determined and the protein added to the oil. The coefficients of variation (CV, %) for the regression slopes were low (CV < 5%, n ) 6), and the y-intercept for each calibration curve gave the protein originally contained in the oil before the dissolution of the added protein. The method showed good linearity over the concentration range found in the studied oils. Although we did not find any oil out of this range, it is possible to apply the method to oils with concentrations exceeding the calibration curve, by diluting the oil with a refined oil having a known concentration of protein. Recovery. The protein recovery was determined at six calibration points within the concentration range (10-200 µg/100 g of oil). The results obtained showed a high, reproducible, and fully Analytical Chemistry, Vol. 73, No. 3, February 1, 2001
701
Figure 3. Tricine-SDS-PAGE of the following: 1, standard proteins; 2, oil body proteins obtained from the mesocarp of olive drupes; 3, oil body proteins obtained from the seeds of olive drupes; 4, oil body proteins obtained from pine seeds; 5, oil body proteins obtained from cabbage seeds; 6, oil body proteins obtained from broad bean seeds; 7, oil body proteins obtained from corn seeds; 8, oil body proteins obtained from safflower seeds; 9, oil body proteins obtained from soybean seeds; and 10, standard proteins. Standard proteins were described in Figure 1.
concentration independent recovery within the range 30-200 µg/ 100 g of oil. Mean recovery was 93%. In the range 10-30 µg/100 g of oil, the recovery decreased slightly as a function of the concentration of the protein. Precision. The within-day and total reproducibility data for oils containing proteins in the range 30-200 µg/100 g of oil was high (CV < 4.2 and 5.7%, respectively). The CV increased in the range 10-30 µg/100 g of oil, but still was
