Extraction of Phospholipids from Salmon Roe with Supercritical Carbon Dioxide and an Entrainer

JOURNAL OF OLEO SCIENCE Copyright 2004 by Japan Oil Chemists Society JOS Extraction of Phospholipids from Salmon Roe with Supercritical Carbon Dioxide and an Entrainer Yukihisa TANAKA, Ikuko SAKAKI and Takeshi OHKUBO Tsukuba Research Laboratory, NOF CORPORATION (5-10 Tokodai, Tsukuba Ibaraki, 300-2635, JAPAN) Edited by K. Takahashi, Hokkaido Univ., and accepted April 19, 2004 (received for review March 29, 2004) Abstract: Supercritical carbon dioxide (SC-CO 2 ) is a suitable substance to extract nonpolar substances (triacylglycerols). However it has not proven effective in the extraction of polar substances. The efficient use of SC-CO 2 and ethanol mixture to extract and fractionate phospholipids from salmon fish roe was therefore investigated. Extraction was performed at low pressure and temperature (17.7 MPa and 33 ) to avoid oxidation of polyunsaturated fatty acids. Phospholipids were not found to be extracted with 0- and 5%-ethanol in SC-CO 2. However extractions with 10, 15 or 20%-ethanol in SC-CO 2 were effective in extracting phospholipids. The amount of extracted phospholipids increased with increased addition of ethanol. When the extraction was performed with SC-CO 2 and 20%-ethanol mixture, more than 80% of the phospholipids were recovered. Key words: supercritical carbon dioxide extraction, salmon roe, triacylglycerol, phospholipid, entrainer, ethanol 1 Introduction Supercritical carbon dioxide (SC-CO 2 ) fluid extraction is applied in the commercial production of flavoring cosmetics, pharmaceuticals and food products. Some examples are decaffeinating coffee (1), hop extraction (2), extraction of turmeric essential oils (3), and ginger flavoring (4). In the oleo-industry, a numerous of researchers have done oil-extraction from seeds and refined plant oils with SC-CO 2 (5-11). There are several advantages in using SC-CO 2 in industrial production. CO 2 has several desirable properties, such as, it is non-corrosive, non-toxic, non-flammable and nonexplosive. Because CO 2 is stable chemically, it does not react with other materials in treatment. Easy separation and removal of CO 2 from the products eliminates any problem related to toxic residual solvents. An added bonus is, it is inexpensive and readily available. A low critical temperature and pressure (Tc=31.1, Pc=7.4 MPa) can be utilized to establish an energy saving process. A great deal of research has been focused on the intake of polyunsaturated fatty acids (PUFA), especially n-3 PUFA, as they have been seen to showing them to play a beneficial role in the prevention of cardiovascular diseases (12), hypertriglyceridemia (13) and autoimmune diseases (14), etc. Some works refer to the application of SC-CO 2 extraction of marine materials to obtain PUFA. Yamaguchi et al. (15) reported on the extraction of lipids from Antarctic krill. According to their report, only non-polar components such as cholesterol, carotenoid triacylglycerols and their derivatives were extracted. Phospholipids did not appear in the extracted fractions. Cheung et al. (16) tried extracting lipids from brown seaweed. They reported that the extracting conditions affected the fatty acid profiles, that is, the concentration of total PUFAs increased reaching a higher value than Correspondence to: Yukihisa TANAKA, Tsukuba Research Laboratory NOF CORPORATION, 5-10 Tokodai, Tsukuba Ibaraki, 300-2635, JAPAN E-mail: yukihisa_tanaka@nof.co.jp Journal of Oleo Science ISSN 1345-8957 print / ISSN 1347-3352 online 417 http://jos.jstage.jst.go.jp/en/ RIMFROST EXHIBIT 1015 page 0001

Y. Tanaka, I. Sakaki and T. Ohkubo that obtained by solvent extraction. In our previous report (17) we extracted lipids from freeze-dried salmon roe to obtain an information on the effects of extracting conditions (pressure 9.8-31.4 MPa, temperature 40-80 ) and the behaviors of lipids in SC- CO 2. Triacylglycerols (TGs) were not extracted completely. Phospholipids (PLs) were not extracted with SC-CO 2 at all. SC-CO 2 does not provide a means to dissolve PLs effectively, but recovery of PLs can be achieved by the addition of a polar entrainer to SC-CO 2. The presence of an entrainer enhances the solubility of SC-CO 2. The choice of a suitable entrainer must be based on thermodynamic considerations and with regard to food safety, that is, it should be Generally recognized as safe (GRAS)(18). Prosise (19) noted that ethanol was an excellent solvent for isolating PLs for food use. Some researchers have already studied its role as an entrainer to extract PLs in SC-CO 2. Temelli (20) extracted PLs from canola flakes and presscake with SC-CO 2 and ethanol. Dunford et al. (21) reported a positive effect of ethanol on the extraction of PLs from canola meal. Montanari et al. (22) observed the extraction of PLs from soybean flakes with SC-CO 2 and 10 wt% ethanol. They reported phosphatidylcholine (PC) enrichment at low pressures although the total yields increased with increasing pressure. Teberikler et al. (23) used SC-CO 2 to produce a 95% purified PC from soybean lecithin containing a low percentage of PC. The authors report herein, the extraction of Docosahexaenoic acid (DHA)-rich PLs from freeze-dried salmon roe with SC-CO 2 and ethanol as the entrainer. 2 Experimental 2 1 Materials Frozen salmon roe was obtained from Nippon Kaken (Tokyo, Japan) and stored at -20 before use. It was thawed and freeze-dried. In this report freeze-dried salmon roe powder is referred to as the FD-sample. The lipid extracted from the FD-sample by Folch s method (24) is defined as the total lipid (TL). It contains TGs, PLs and their derivatives (diacylglycerols, monoacylglycerols and lysophospholipids, etc.). 2 2 SC-CO 2 Extraction The extraction vessel used in this work was of 10.0- mm interior diameter and 129 mm length (model EV-4, JASCO, Hachioji, Japan) with a volume of 10 ml. The equipment used for the work consisted of a high-pressure liquid chromatograph system (pump, JASCO PU- 1586, column oven, JASCO 865-CO) and back pressure regulator (JASCO 880-81). 4.0 g of the FD-sample was applied in the vessel. Extraction trials were performed at 33 and 17.7 MPa. The extracted lipid was collected several times during extraction. CO 2 and ethanol were delivered by two separate pumps, mixed and passed through a preheating coil. 2 3 Analysis In this work, the lipids extracted with SC-CO 2 are referred to as the extracted lipids (EL). After the SC- CO 2 extraction, the lipids retained in the spent FD-sample were extracted by Folch s method. This is referred to as the residual lipids (RL), in this work. The lipids content of EL, RL and TL were analyzed by means of silica gel thin layer chromatography (TLC, plate 5721, Merck, Darmstadt, Germany) with hexane-diethyl ether-acetic acid (80:20:1 v/v/v) or chloroformmethanol-water (65:25:4 v/v/v). The extracting yield is referred to as the ratio of the weight of EL to the weight of TL. TL and RL (150 mg) were run through the column (20 mm i.d. 200 mm height) with silica gel 60 (mesh 70-230, Merck) to fractionate the TGs and PLs. The TGs and PLs were eluted with 300 ml of chloroform and 200 ml of methanol, respectively. The TGs and PLs fractionated from TL are referred to as the original TGs and the original PLs. The fatty acid profiles were analyzed by gas chromatography of the methyl esters prepared by transmethylation with BF 3 /methanol. An Agilent 6890A series gas chromatograph (Yokogawa Analytical Systems, Musashino, Japan) equipped with a flame ionization detector (FID) and DB-WAX capillary column (30 M 0.25 mm i.d.) (J & W Scientific, Folsom, CA) was used. The column temperature was raised from 150 to 210 at 5 /min. Both the injector and detector temperatures were 250. The carrier gas was helium with hydrogen and air supplied to the FID. The fatty acids were identified by comparison of the retention times with lipid standards (Sigma, Saint Louis, MA). PLs analyses were performed by high pressured liquid chromatography (HPLC). A LC Model I HPLC system (Toso, Tokyo, Japan) equipped with DEVELOSIL 418 RIMFROST EXHIBIT 1015 page 0002

Extraction of Phospholipids from Salmon Roe with SC-CO 2 and Ethanol model 60-5 HPLC-column (259 mm 4.6 mmi.d.) (Nomura-chemicals, Tokyo, Japan) was used. The mobile phase was acetonitrile/methanol/phosphoric acid (780:50:9, v/v/v). All solvents were HPLC grade (Wako Pure Chemical Industries, Ltd., Osaka, Japan). The flow of mobile phase was 1.5 ml/min. The column oven temperature was 45. Detective absorbance was at 220 nm. Zephiramine (Wako Pure Chemical Industries, Ltd.,) was used as the inner standard to analyze quantitatively. 3 Results and Discussion 3 1 Effect of Extracting Temperature on the Lipid Yield at 17.7 MPa The authors investigated the most suitable conditions to extract lipids from salmon roe with SC-CO 2. The results in our previous report (17) suggested that the extraction at higher pressure gained a higher extraction yield. Extractions performed at 50 MPa and 40, 60 and 80 gave high extraction yields; 44.18 0.21, 46.71 0.83, and 51.03 0.71% (average SD), respectively. The extraction yields were significantly higher than those found at 31.4 MPa and 40-80, mentioned in the previous report. No significant amount of PLs was extracted at these conditions. The authors further investigated improved the extraction conditions from the following standpoints. The authors wanted to avoid PUFAs such as eicosapentaenoic acid (EPA; C20:5, n-3) and docosahexaenoic acid (DHA; C22:6, n-3) being oxidized through being exposed to high temperatures during extraction. Extraction at higher pressure also increases the risks of accidents in the handling of equipment. The authors furthermore tried to establish an energy saving protocol. Extraction at higher pressures and higher temperatures consume a great deal of energy to produce and maintain them. Since the authors had already shown that reducing the temperature lead to an increase in the extraction yield at 17.7 MPa (17), to extract lipids from salmon roe FD at lower temperatures than 40 suggested a higher extraction yield. The extraction yield at 33 was significantly higher than that at 40 (p<0.05, Fig. 1). This was comparable to that at 50 MPa and 60. No PLs were extracted under these conditions. The authors investigated the behavior of lipids in SC-CO 2 at 17.7 MPa and 33. Fig. 1 Effect of Extraction Temperature on Extraction Yield. SC-CO 2 flow was 3 ml/min, Extraction time was 10 hours. The extraction pressure was 17.7 MPa. Many researchers have already reported since a pure carbon dioxide does not dissolve PLs effectively, extraction of PLs might be achieved by the addition of a polar entrainer to SC-CO 2. An entrainer is a substance of medium volatility added to a mixture of compressed gas and a low volatility substance (20). As the solubility in SC-CO 2 at the same extracting conditions (temperature and pressure) is drastically enhanced, extraction can be conducted at a lower pressure (25). The logical choice for a co-solvent in the food industry would be ethanol. The authors used ethanol as the entrainer to extract PLs in SC-CO 2 because: (i) It is suitable for food use; and (ii) the phase behavior of CO 2 /ethanol mixes at high pressure is available (26, 27). CO 2 and ethanol were mixed and passed through the preheating coil, and delivered to the vessel in the oven to extract the lipids. Extractions of PLs from canola, soybean and cottonseed with SC-CO 2 /etanol mixture have been reported (21). In our study ethanol was used as the entrainer to extract PLs from salmon roe. 3 2 Effect of Ethanol on the Lipid Extraction The extraction was performed at 17.7 MPa and 33 with 5, 10, 15 or 20%-ethanol in SC-CO 2. The ethanol 419 RIMFROST EXHIBIT 1015 page 0003

Y. Tanaka, I. Sakaki and T. Ohkubo Fig. 2 Effect of Entrainer on Extraction Yield. flows were 5, 10, 15 or 20% of CO 2 flow by volume (CO 2 flow = 3.0 ml/min, Ethanol flow = 0.15, 0.30, 0.45 or 0.60 ml/min). The effect of the entrainer on the extraction yield is shown in Fig. 2. The extraction yield increased with increase in the ethanol percentage. 39% of the total lipids were extracted without the entrainer after 4 hours. When the extraction was performed with SC-CO 2 and 5%-ethanol mixture, the extraction yield rose above 50%, while an addition of 10-, 15-, or 20%- ethanol in SC-CO 2 achieved an extraction yield of more than 75%. The lipid compositions in all the fractions were observed using TLC. No PLs were observed in the EL fractions when extraction was performed with SC-CO 2 and 5%-ethanol mixture. When extraction was performed with SC-CO 2 and 10- or 15%-ethanol mixture, PLs were observed in the EL fractions. In all the extracted fractions both TGs and PLs were observed after a 4-hour extraction, and present in the RL fraction. When extraction was performed with SC-CO 2 and 20%- ethanol mixture within an initial 30 min extraction almost all the TGs were extracted. In the following fractions (1 h-4 h) and RL fraction, slight TGs and some amount of PLs were observed. The PL concentrations in all the fractions were quantified by HPLC. The results are shown in Table 1. Extraction with SC-CO 2 and 20%-ethanol mixture produced fractions containing PLs without TGs. The rate of extracted PLs to the original PLs are indicated as the PL recovery rate. The rate of extracted PLs with SC- CO 2 and ethanol mixtures increased with increase in the ethanol amount in SC-CO 2 (Fig. 3). When extraction was performed with SC-CO 2 and 10%-ethanol mixture 30% of the PLs were recovered in the EL fractions. When the ethanol percentage was increased up to 20% more than 80% of the PLs were recovered in the EL fractions. The rate of extracted TGs to the original TGs are indicated as the TG recovery rate. When extraction was performed with SC-CO 2 and 5%-ethanol mixture, nearly 80% of the TGs were recovered during the fourhour extraction. On the other hand, 20% of the TGs were not extracted with SC-CO 2 and remained present in the RL fraction. When the lipids were extracted with SC-CO 2 and 10-, 15-, or 20%-ethanol mixtures more than 90% of the TGs were recovered to the EL fraction (Fig. 4). 3 3 Effect of Ethanol on Fatty Acid Profiles of Lipids The fatty acid profiles of the TGs in all the fractions were analyzed. The concentrations of oleic acid (OA; C18:1, n-9) and DHA of the EL are shown in Figs. 5 and 6. The addition of the entrainer and extracting peri- Table 1 Phospholipid Content (wt%) in the Extracted Lipid Fractions. Ethanol (%) 10 15 20 Extraction time (h) 0.5 1 2 3 4 2.44 2.96 8.80 4.21 7.88 4.25 8.30 1.41 52.61 8.43 94.70 8.45 Phospholipid content was analyzed by HPLC. 12.55 3.10 67.37 6.71 84.90 8.41 10.65 5.61 64.13 7.70 92.99 4.02 14.26 7.36 39.31 4.25 97.20 1.33 420 RIMFROST EXHIBIT 1015 page 0004

Extraction of Phospholipids from Salmon Roe with SC-CO 2 and Ethanol Fig. 3 Effect of Entrainer on Phospholipids Recovery Rate. Fig. 5 Effect of Entrainer on Extracted Oleic Acid Concentration in TGs. Entrainer flows were 5 ( ), 10 ( ), 15 ( ) and 20 ( )%. ------- designates the oleic acid concentration in the original TG. Fig. 4 Effect of Entrainer on Triacylglycerols Recovery Rate. Fig. 6 Effect of Entrainer on Extracted Docosahexaenoic Acid Concentration in TGs. Entrainer flows were 5 ( ), 10 ( ), 15 ( ) and 20 ( )%. ------- designates the oleic acid concentration in the original TG. 421 RIMFROST EXHIBIT 1015 page 0005

Y. Tanaka, I. Sakaki and T. Ohkubo od did not affect the fatty acid profiles only the OA and DHA compositions. The OA concentrations in all the extracted fractions were lower than that in the original TGs. They increased with extracting time. The DHA concentrations in most of the extracted fractions were higher than that in the original TGs. After a four-hour extraction, TGs were observed in the RL of the extractions with SC-CO 2 and less than 15%-ethanol. The OA concentration in the RL was higher than in the original TGs. While the DHA concentration was lower than in the original TG. The results suggest a polarity of SC- CO 2 and ethanol mixture occurs. The polarity of SC- CO 2 added with ethanol accelerated the DHA extraction in the TGs. While the entrainer did not affect the fatty acid profiles of the extracted PLs, significantly, the fatty acid profiles of the PLs in the EL and RL fractions were the same as those in the original PLs. Temelli (20) in his report on the fatty acid profiles of canola oil extracted with ethanol, remarked that the relative concentrations of OA and linolenic acid (C18:3, n-3) decreased. 3 4 Separation of TGs and PLs with SC- CO 2 and Ethanol Some researches have reported a two-step process for the extraction of the PLs with SC-CO 2 to obtain high concentrated PLs. The first extraction is performed without ethanol to remove the oil. Subsequent extractions are performed with ethanol to extract the PLs. Temelli (20) and Montanari et al. (22) reported on the extraction of PLs from canola and soybean, respectively. According to their reports the oil was removed at high pressure (60-70 MPa) and high temperature (70-80 ). Since Montanari et al. (22) performed the extraction at high temperature and high pressure, almost of all soybean TGs were obtained. Both high temperature and pressure was necessary to extract the PLs from soybean. However when extraction temperature was high (80 ), the extracted amount of the PLs was reduced when extraction was performed at low pressure (23.9 MPa). Therefore a significant amount of the PLs was lost at 16.6 MPa. The extraction rate does not depend on SC-CO 2 and ethanol mixture density. In the present report, our choice of mild extraction conditions (33, 17.7 MPa) resulted in an incomplete extraction of TGs. According to the present results almost all lipids were extracted from the salmon roe FD with SC-CO 2 and 20%-ethanol mixture. During the initial 1h extraction almost all the TGs and 75% of the PLs were recovered at the same time. As a result both the TGs and PLs were observed in the same fraction. A further 22% of the PLs were extracted in the following 3 h. When lipids were extracted from the spent FD powder with an organic solvent. 3% of the PL was observed as the RL fraction. In the second the EL fraction and RL fraction, the PLs were isolated without TGs. With these extraction conditions, most of the PLs were obtained along with the TGs. The ratio of TGs and PLs in the first fraction was approximately the same as that for the TL such as 75:25 w/w. The authors tried to obtain fractions containing high concentrations of the PLs in the short time available 10min. taken for the TGs to be already extracted. The authors judged that it was difficult to separate TGs and PLs using their retention time lag. Instead, the authors investigated a three-step extraction. In the first step, a SC-CO 2 and 5%-ethanol mixture was used for 4 h to extract as much of the TGs as possible. In the second step, the amount of ethanol was increased from 5% to 20% in as short a time as possible, and the extraction progressed for a further 1 h. In the third step, the extraction was performed for 3 h with SC-CO 2 and 20%- ethanol. In the first step, approximately 80% of the TGs and none of the PLs were recovered. But thereafter, all of the TGs and part of PLs were recovered in the sec- Fig. 7 Extraction of Lipid from Salmon Roe FD by threestep Extraction. 422 RIMFROST EXHIBIT 1015 page 0006

Extraction of Phospholipids from Salmon Roe with SC-CO 2 and Ethanol ond step. In the third step, almost all of the PLs were recovered. The authors applied this three-step extraction to obtain high concentrated PLs from 4 g of salmon roe FD. The results are shown Fig. 7. The lipid delivery in the first, second, third and the RL fractions were 59.9 6.9, 17.1 5.6, 12.4 0.6, and 10.5 1.7% (average SD), respectively. In the third and the RL fractions the PLs but none of the TGs were observed. The ratio of the TGs and PLs was 2:1 (w/w) in the second fraction. Acknowledgments This work was performed for The Japanese Research and Development Association for New Functional Foods. References 1. K. ZOSEL, Separation with Supercritical Gases: Practical Application, Angew. Chem. Ind. Ed. Engl., Vol. 17, 702-709 (1978). 2. P. HUBERT and O.G. VITZTHUM, Fluid Extraction of Hops, Spices, and Tobacco with Supercritical Gases, Angew. Chem. Ind. Ed. Engl., Vol. 17, 710-715 (1978). 3. B. GOPALAN, M. GOTO, A. KODAMA and T. HIROSE, Supercritical Carbon Dioxide Extraction of Turmeric (Curcuma longa), J. Agric. Food Chem., Vol. 48, 2189-2192 (2000). 4. Y. YONEI, H. OHINATA, R. YOSHIDA, Y. SHIMIZU and C. YOKOYAMA, Extraction of Ginger Flavor with Liquid or Supercritical Carbon Dioxide, J. Supercritical Fluids., Vol. 8, 156-161 (1995). 5. P. BONDIOLI, C. MARIANI, A. LANZANI, E. FEDLI, A. MOSSA and A. MULLER, Lampante Olive Oil Refining with Supercritical Carbon Dioxide, J. Am. Oil Chem. Soc., Vol. 69, 477-480 (1992). 6. A. BIRTIGH, M. JOHANNSEN, G. BRUNNER and N. NAIR, Supercritical-fluid Extraction of Oil-palm Components, J. Supercritical Fluids, Vol. 8, 46-50 (1995). 7. G.R. LIST, J.W. KING, J.H. JOHNSON, K. WARNER and T.L. MOUNTG, Supercritical CO 2 Degumming and Physical Refining of Soybean Oil, J. Am. Oil Chem. Soc., Vol. 70, 473-476 (1993). 8. G.R. ZIEGLER and Y.-J. LIAW, Deodorization and Deacidification of Edibles Oils with Dense Dioxide, J. Am. Oil Chem. Soc., Vol. 70, 947-953 (1993). 9. S. LI and S. HARTLAND, A New Industrial Process for Extracting Cocoa Butter and Xanthines with Supercritical Carbon Dioxide, J. Am. Oil Chem. Soc., Vol. 73, 423-429 (1996). 10. M.S. KUK and M.K. DOWD, Supercritical CO 2 Extraction of Rice Bran, J. Am. Oil Chem. Soc., Vol. 75, 623-628 (1998). 11. I. PAPAMICHAIL, V. LOULI and K. MAGOULAS, Supercritical Fluid Extraction of Celery Seed Oil, J. Supercritical Fluids, Vol. 18, 213-226 (2000). 12. S.H. GOODNIGHT, The Effects of n-3 Fatty Acids on Arteriosclerosis and the Vascular Response to Injury, Arch. Pathol. Lab. Med., Vol. 117, 102-106 (1993). 13. W.C. HARRIS, D.W. ROTHROCK, A. FANNING, S.B. INKE- LES, S.H. GOODNIGHT, D.W. ILLINGWORTH and W.E. CONNOR, Fish Oil in Hypertriglyceridemia, A Dose Response Study, Am. J. Clin. Nutr., Vol. 51, 399-406 (1990). 14. J.M. KREMER, D.A. LAWRENCE, W. JUBIZ, R. Di GIACO- MA, K. RYNES, L.E. BARTHOLOMEW and M. SHERMAN, Dietary Fish Oil and Olive Oil Supplementation in Patients with Rheumatoid Arthritis, Arthritis Rheum., Vol. 33, 810-820 (1990). 15. K. YAMAGICHI, M. MURAKAMI, H. NAKANO, S. KONO- SU, T. KOKURA, H. YAMAMOTO, M. KOSAKA and K. HATA, Supercritical Carbon Dioxide Extraction of Oils from Antarctic Krill, J. Agric. Food Chem., Vol. 34, 904-907 (1986). 16. P.C.K. CHEUNG, A.Y.H. LEUNG and P.O. ANG, Jr., Comparison of Supercritical Carbon Dioxide and Soxhlet Extraction of Lipid from a Brown Seaweed, Sargassum hemiphyllum (Turn.) C. Ag., J. Agric. Food Chem., Vol. 46, 4228-4232 (1998). 17. Y. TANAKA and T. OHKUBO, Extraction of Lipids from Salmon Roe with Supercritical Carbon Dioxide, J. Oleo Sci., Vol. 52, 295-301 (2003). 18. L. MONTANARI, J.W. KING, G.R. LIST and K.A. RENNICK, Selective Extraction of Phospholipid Mixtures by Supercritical Carbon Dioxide and Cosolvents, J. Food Sci., Vol. 61, 1230-1233, 1253 (1996). 19. W.E. PROSISE, Commercial Lecithin Products: Food Use of Soybean Lecithin, in Lecithins, (B.F. SZUHAJ and G.R. LIST, ed), American Oil Chemists Society, Champaign IL, pp.163-182 (1985). 20. F. TEMELLI, Extraction of Triglycerides and Phospholipids from Canola with Supercritical Carbon Dioxide and Ethanol, J. Food Sci., Vol. 57, 440-442, 457 (1992). 21. N.T. DUNFORD and F. TEMELLI, Extraction of Phospholipids from Canola with Supercritical Carbon Dioxide and Ethanol, J. Am. Oil Chem. Soc., Vol. 72, 1009-1015 (1995). 22. L. MONTANARI, P. FANTOZZI, J.M. SNYDER and J.W. KING, Selective Extraction of Phospholipids from Soybeans with Supercritical Carbon Dioxide and Ethanol, J. Supercrit. Fluids, Vol. 14, 87-93 (1999). 23. L. TEBERIKLER, S. KOSEOGLU and A. AKGERMAN, Selective Extraction of Phosphatidylcholine from Lecithin by Supercritical Carbon Dioxide/Ethanol Mixture, J. Am. Oil Chem. Soc., Vol. 78, 115-119 (2001). 24. J. FOLCH, M. LEE and S.G.H. SLOANE, A Simple Method for the Isolation and Purification of Total Lipids from Animal Tissues, J. Biol. Chem., Vol. 226, 497-509 (1957). 25. G. BRUNNER and S. PETER, On the Solubility of Glycerides and Fatty Acids in Compressed Grases in the Presence of an Entrainer, Sep. Sci. Technol, Vol. 17, 199-214 (1982). 26. H. POHLER and E. KIRAN, Volumetric Properties of Carbon Dioxide + Ethanol at High Pressure, J. Chem. Eng. Data, Vol. RIMFROST EXHIBIT 1015 page 0007 423

Y. Tanaka, I. Sakaki and T. Ohkubo 42, 384-388 (1997). 27. C.Y. DAY, C.J. CHANG and C.Y. CHEN, Phase Equilibrium of Ethanol + CO 2 and Acetone + CO 2 at Elevated Pressures, J. Chem. Eng. Data, Vol. 41, 839-843 (1996). 424 RIMFROST EXHIBIT 1015 page 0008