ANALYST, FEBRUARY 1993, VOL. 118 131 Use of a Synthetic Detergent to Partition Protein Mixtures Wakako Tsuzuki, Hanae Kasumimoto and Shouichi Kobayashi National Food Research Institute, Ministry of Agriculture, Forestry and Fisheries, 2- 1-2 Kannondai, Tsukuba, lbaraki 305, Japan A previously reported method for the preparation of organic solvent soluble lipase in high yield has been applied to the partitioning of a protein mixture using a mixture of six known proteins as a model system. The organic solvent soluble complex of these proteins was obtained according to the previously reported method. In order to extract the proteins from the complex, the latter was dissolved in dichloromethane followed by the addition of buffer and triethylamine. By using this procedure, the proteins could be recovered from the complex formed with the detergent.It was found that the composition of the solvent used to prepare the complex influenced the ability of each protein to form a complex with the detergent. By making use of the differences in the efficiency of each protein to form a complex, several crude lipases could be successfully purified; in addition, their activities were retained during the purification procedure. The results suggest that the application of a synthetic detergent may be effective for the purification of proteins and enzymes. Keywords: Protein; purification; detergent; partition The amphiphilic properties of detergents have been utilized by a number of workers to purify proteins. In particular, insoluble proteins, such as membrane binding proteins, have been purified with detergents.1-6 The mechanism by which the detergent dissolves the proteins has also been s t ~ d i e d . ~ - ~ It was found that the properties of the detergent, including its co-operative binding to the proteins, its micellization with proteins and its dissolution of membranes, were useful for protein purification. Recently, methods in which a detergent is used to modify an enzyme so that it becomes soluble in organic solvents have been developed.I(b12 The aim of these studies was to conduct the enzymic reaction in an organic solvent and to modify the properties of the enzyme itself. In previous work,12 lipase was modified using a synthetic detergent (Fig. 1). The complex formed between lipase and the detergent was soluble in several organic solvents and retained its activity.The complex was designated as organic solvent soluble lipase. This paper describes the application of a synthetic detergent to the purification of proteins. A mixture of six proteins was used as the model system in order to determine the conditions for partitioning. Several crude lipases were successfully purified under the specified partition conditions. The par- titioning of proteins is based on the differences in the efficiency of each protein to form a complex with the detergent. In comparison with previous purification pro- cedures using detergents, the method described here is based on the formation of an insoluble complex between the proteins and the detergent and on the extraction of the proteins from the complex by organic solvents.Furthermore, partitioning using the detergent can be applied not only to insoluble proteins but also to soluble proteins, as it is thought that the surface of the proteins is bound to the hydrophilic region of the detergent in the complex12 (Fig. 2). Experimental Reagents A mixture of six proteins (phosphorylase b, bovine serum albumin, ovalbumin, carbonic anhydrase, soybean trypsin CH2-CH-CH-CH-CH-CO-NH-CH-CO-O-( CH2)l q-CH3 I I l l 1 I OH OH OH OH OH CH2 I CH2 --CO-O-(CH2)71-CH3 Molecular structure of the synthetic detergent, didodecyl- Fig. 1 glucosyl glutamate inhibitor and a-lactalbumin) was purchased from Pharmacia Fine Chemicals (Uppsala, Sweden) as an electrophoresis calibration kit, which was dialysed before use to remove sucrose. The detergent, didodecylglucosyl glutamate, used in the preparation of the complex was synthesized according to the method described previously.11 Several lipases were purchased commercially, viz., lipase P and lipase A from Amano Pharmaceutical (Nagoya, Japan), lipase PN from Wako Pure Chemicals (Osaka, Japan) and pancreatic lipase (porcine Type 11) €rom Sigma (St. Louis, MO, USA). Tetrahydrofuran (THF) without stabilizer, dichloromethane and triethylamine were purchased from Wako Pure Chemi- cals. All the other chemicals were of analytical-reagent grade and were purchased from Wako Pure Chemicals or Kanto Chemicals (Tokyo, Japan). Preparation of the Complex Between the Model Proteins and the Detergent The six dialysed proteins (the amount of phosphorylase b, bovine serum albumin, ovalbumin , carbonic anhydrase, soy- bean trypsin inhibitor and a-lactalbumin was fixed at 64, 83, 147, 83, 80 and 121 pg, respectively) in water were added to a solution of didodecylglucosyl glutamate (2.9 mg) in THF, followed by the addition of THF and water to give a final volume of 300 pl with a specific THF concentration.The complex between the protein mixture and the detergent was prepared as described previously. 12 Preparation of the Complex Between Lipase and the Detergent Lipase (100 mg) in water (2 ml) was added to the detergent (500 mg) in THF (4 ml) and the complex was prepared as described previously. 12 Lipase Detergent Organic solvent soluble lipase Fig. 2 Organic solvent soluble lipase132 ANALYST, FEBRUARY 1993, VOL.118 Extraction of Proteins From the Complex The complex was dissolved in the organic solvent, with the addition of the buffer [0.2 mol 1-1 Tris-HCI buffer (pH 7.5)] [Tris = tris(hydroxymethyl)methylamine]. After addition of triethylamine, the mixture was stirred vigorously using a vortex mixer and centrifuged at 15000 rev min-1 for 10 min. The aqueous phase was separated and analysed for the extracted proteins. Determination and Electrophoresis of Proteins The content of the proteins in the aqueous phase was determined using a protein assay kit (Bio-Rad Laboratories, Richmond, CA, USA). The extracted proteins were charac- terized by sodium dodecyl sulfate-polyacrylamide gel electro- phoresis (SDS-PAGE) using the Phast System (Pharmacia Fine Chemicals).Lipase Assay The lipase activity was determined according to the method of Shimura et al. 13 4-Methylumbelliferyl oleate was used as the substrate. Results Partition of Six Known Proteins A flow chart illustrating the application of the synthetic detergent to the partitioning of a protein mixture is shown in Fig. 3, This rapid and simple procedure involves two main steps: the preparation of the complex with the detergent and the extraction of the proteins from the complex. The conditions for the preparation of the organic solvent soluble proteins were studied in detail and the results are described below. The effect of the THF concentration on the yield of the complex was studied. As shown in Fig. 4, the proteins extracted from the complex, which was prepared at different Protein-water Detergent-THF I Stir at 4 "C Evaporate I PreciDitate 1- Wash with water I Precipitate Lyop h i I ize t I Complex 'powder Dissolve with organic solvent Buffer Triethylamine Mix vigorously Protein in i aqueous phase Fig.3 Flow chart of the protein purification procedure using the detergent THF concentrations, were characterized by SDS-PAGE. A marked difference was found in the characteristics of the extracted proteins depending on the THF concentration. When the THF concentration of the solution was 33%, five proteins were able to form a complex with the detergent. On the other hand, only three proteins, bovine serum albumin, soybean trypsin inhibitor and a-lactalbumin, formed a com- plex when the THF concentration was 67%.These findings demonstrated that the THF concentration of the solution affected the yield of the complex formed by each protein with the detergent. The yield of each protein purified in a solution of THF at a concentration of 67% was measured using the gel reader. The contents of bovine serum albumin, soybean trypsin inhibitor and a-lactalbumin were 28, 7 and 65%, respectively. Considering that the total recovery of protein was 42%, the yield of a-lactalbumin was close to 60%. The time course of the formation of the complex was studied. The THF concentration of the solution was fixed at 67%. As shown in Fig. 5 , the yield of the total proteins - B +C 4--D -E -F Fig. 4 SDS-PAGE of the proteins. Characteristics of the proteins extracted from the complex prepared at different concentrations of THF.A mixture of six proteins (d) (A. phosphorylase b; B, bovine serum albumin; C, ovalbumin; D, carbonic anhydrase; E. soybean trypsin inhibitor; and F, a-lactalbumin) was used to form a complex with the detergent. The com lex was prepared in a solution of THF at a concentration of ( a ) 33; 6) 50; and (c) 67%. The proteins were extracted from the complex according to the method described under Experimental - 0.4 E 2 0.3 E .- >. C 0.2 (u 0. .- +- 2 0.1 0 4 8 12 16 20 24 Time/h Fig. 5 Time course of the yield of the complex prepared in a solution of THF at a concentration of 67%. The total protein yield is expressed as a mass percentage based on the mass of the starting materials, which was taken as 100%ANALYST, FEBRUARY 1993, VOL.118 (a) ( b) (4 (d) 133 +B A- B- C- - E + F Fig. 6 SDS-PAGE of proteins extracted from the complex depend- ing on the time allowed for the formation of the complex between the detergent and the proteins. The mixture of six proteins (d) was added to the detergent for the formation of the complex in a solution with THF at a concentration of 67% for ( a ) 4; (b) 8; and (c) 24 h. For definitions of A-F see Fig. 4 Table 1 Effect of the solvent on the extraction efficiency of proteins. The yield of the extracted proteins is expressed as a percentage of the yield obtained by using dichloromethane containing 1 YO triethyl- amine. which was taken as 100% Solvent + 1% Solvent Solvent alone triethylamine Lauryl alcohol 1 69 Chloroform 0 18 Dichloromethane 3 100 Dichloroethane 0 42 Benzene 0 36 reached 42% compared with the original value after mixing with the detergent for 16 h.Further mixing did not increase the yield of the extracted proteins. Even for a mixing time of 8 h, 38% of the proteins had already formed a complex. The time course of the formation of the complex was also followed by SDS-PAGE (Fig. 6). As shown in Fig. 6, no differences could be detected in the characteristics of the proteins which formed a complex with the detergent for various periods of time. This observation suggests that the time required for the formation of the complex does not affect the efficiency of the proteins in forming the complex but does influence the yield of the complex. As a result, the time required for the formation of the complex was fixed at 16 h.The effect of the ratio of the proteins to the detergent on the yield of the complex was investigated. The proteins were treated with the detergent at protein : detergent ratios of 1 : 1, 1 : 2, 1 : 4 and 1 : 8. A slight difference was detected in the yield of the proteins extracted from the complex. The yield of the proteins extracted from the complex prepared with a protein : detergent ratio of 1 : 1 was 80% of that extracted from the complex preparcd with a protein : detergent ratio o€ 1 : 4. A 4-fold increase in the ratio of detergent to protein did not induce an increase in the yield of the proteins extracted from the complex. According to the analysis by SDS-PAGE, the characteristics of the extracted proteins did not change in the range of protein : detergent ratios studied (data not shown) and a protein:detergent ratio of 1 : 5 was used in the preparation of the complex.E- F - Fig. 7 Characteristics of the proteins extracted from the complex using different organic solvents. Six proteins (a) were used to form the complex. The complex was dissolved in chloroform ( b ) , dichloro- methane (c), dichloroethane (d), lauryl alcohol ( e ) and benzene cf). The proteins were extracted after the addition of triethylamine and buffer. For definitions of A-F see Fig. 4 I I I I I 0.5 0.4 - E g 0.3 -0 2 1 .- ). a 0, 0.2 .- .I- ? 0.1 0 1 2 3 4 5 Concentration of triethylamine (%) Fig. 8 Effect of the concentration of triethylamine on the extraction efficiency of proteins from the complex, which was prepared in a solution of THF at a concentration of 67%.The total protein yield is expressed as a mass percentage based on the mass of the starting materials, which was taken as 100% Hence, it was found that only the THF concentration influenced the efficiency of the proteins in forming a complex with the detergent. On the other hand, it was observed that the time required for the mixing of the proteins with the detergent and the ratio of the proteins to the detergent for the preparation of the complex were correlated with the yield of the complex but not with the ability of the detergent to become attached to the proteins. The complex formed by the proteins and the detergent could not be dissolved in an aqueous solution but was readily soluble in several organic solvents.The conditions for extraction of the proteins from the complex were investigated and the results are described below. The organic solvent in which the proteins could be efficiently extracted from the complex was selected. As shown in Table 1 , five organic solvents were used for the extraction of the proteins from the complex, which was prepared in solution at a THF concentration of 67%. The complex was dissolved in each organic solvent, followed by the addition of the buffer (0.2 moll-' Tris-HC1 buffer pH 7.5). After mixingvigorously, the proteins extracted into the buffer phase were analysed.134 ANALYST, FEBRUARY 1993, VOL. 118 Table 2 Purification of several lipases using the detergent Specific activity* Lipase (A) Lipase P 12.11 Lipase A I .98 Lipase PN 0.22 Pancreatic lipase 5.13 Yield of Specific extracted activity* Recovery of protein of extracted total (YO) protein (B) B/A activity (%) 0.8 778.23 64.3 51.2 1 .5 58.38 29.5 44.3 29.2 1.55 7.1 69.0 1.9 31.86 6.2 12.3 * The units of specific activity are expressed as the amount of 4-methylumbelliferonc released in nanomoles per milligram of protein per minute (nmol mg-1 min-1).- 94000 - 67000 c- 43000 - 30000 - 20000 f- 14400 Fig. 9 SDS-PAGE of lipase P. ( a ) Untreated lipase P available commercially. ( h ) Characteristics of the pattern of the proteins extracted from the complex between lipase P and the detergent. The values adjacent to the gels indicate the relative molecular masses of the calibration proteins Proteins were not released from the complex without the addition of triethylamine to each organic solvent, whereas the addition of triethylamine induced the release of the proteins from the complex.When the final concentration of triethyl- amine was fixed at 1%, the extraction efficiency depended on the organic solvent. The use of dichloromethane containing 1% triethylamine resulted in the most efficient extraction of the proteins (Table 1). After the extraction of the proteins from the complex with the dichloromethane containing 1% triethylamine, the absorbance of the complex was monitored at 250 nm and no absorption was detected. Dichloromethane including the complex exhibited some absorbance of proteins at 280 nm. The results suggested that the efficiency of the extraction of the proteins from the complex by using dichloro- methane containing 1% triethylamine was almost 100%.The characteristics of the proteins extracted from the complex with each of the organic solvents tested are presented in Fig. 7. Few differences were detected in the characteristics of the extrac- ted proteins, suggesting that the extraction efficiency, which depends on the type of protein, was not influenced by the organic solvent used. As a result, dichloromethane containing triethylamine was considered to be the most suitable solvent for extracting the proteins from the complex. The effect of the concentration of triethylamine in dichloro- methane was studied. The complex, which was prepared in a solution with a THF concentration of 67%, was dissolved in dichloromethane. After the addition of the buffer to the dichloromethane phase, triethylamine was added at various concentrations.As shown in Fig. 8, the efficiency of the extraction of the proteins from the complex into the aqueous phase was affected by the triethylamine concentration. The highest efficiency of protein extraction was achieved at a triethylamine concentration of 1-2%. Therefore: the concen- tration of triethylamine was fixed at 1% for the extraction of the proteins from the complex with dichloromethane. Application to the Purification of Crude Lipases The proposed method was applied to the purification of several commercially available lipases. The complex of each lipase with the detergent was prepared in a solution with a THF concentration of 67%. The proteins were extracted from the complex with the buffer and dichloromethane containing 1% triethylamine.The yields of the proteins extracted from the complex and their specific activities are listed in Table 2. A comparison of the specific activities of the extracted lipases with those of the untreated lipases indicates that all the lipases can be purified by about 6- to 64-fold. For lipase P, which appeared to be highly purified, the properties of the proteins obtained by SDS-PAGE revealed that the main protein extracted from the complex corresponded to lipase P itself (relative molecular mass 33000 Da) (Fig. 9). No marked inactivation of any of the lipases appeared to occur during the purification process using the detergent. These findings indicated that the lipases were not denatured by the organic solvent due to the formation of the complex with the detergent.These results suggest that the proposed partition method might also be effective for the purification of enzymes. Discussion In previous work, 12 a lipase could be modified in high yield to the organic solvent soluble lipase by adding THF to the solution used for preparing the complex. For lipase B, about a 4-fold increase in the organic solvent soluble lipase could be obtained at a THF concentration of 67% compared with a solution containing 100% buffer, suggesting that the addition of THF to the solution used to form the complex resulted in the efficient production of the organic solvent soluble lipases. The THF concentration appeared to be an important factor for the efficiency of the formation of the complex between the proteins and the detergent.The results suggested that the method used for the preparation of the organic solvent soluble lipase might be applicable to the partitioning of a protein mixture, if there was a difference in the efficiency of the formation of the complex among the proteins. This is the first report to demonstrate that each protein has a different efficiency in forming a complex with a detergent at a specified THF concentration depending on the surface property of the protein. This finding has allowed the develop- ment of a novel technique for partitioning proteins, although the phenomenon is not yet fully understood. It has been shown here that there are specific properties of the interaction between the interface of the protein and the detergent.The first is that the concentration of THF affects the efficiency of complex formation between the protein and the detergent. Based on this property, it should be possible to separate a specific protein from other proteins by adjusting the THF concentration to obtain the most efficient production of the complex of the protein. The second property of the interactionANALYST, FEBRUARY 1993, VOL. 118 135 between proteins and the detergent is the requirement for amines in the extraction of proteins from the complex. The addition of triethylamine also induces the release of the proteins from the complex. Little protein could be extracted from the complex using an organic solvent and buffer. These results suggest that amines might sever a bond between the surface of the proteins and the detergent.These findings are based on the establishment of a novel method for the purification of proteins using a synthetic detergent. The principle of the partitioning of the proteins is different from other current purification methods using detergents. Further investigations using various detergents and other proteins are required in order to elucidate the mechanism of the binding in the complex and the types of protein that could be purified by this method. It may then be possible to develop a very effective method for the purification of proteins. During the partitioning procedure used here, the denatura- tion of the proteins is considered to be minimal, because the proteins rcmain coated with the detergent throughout the process. In fact, the application of the method to the purification of several crude lipases did not result in their inactivation, suggesting that this method might also be suitable for the purification of enzymes. The purification of other enzymes and the use of a commercially available detergent are currently under investigation in this laboratory. 1 2 3 4 5 6 7 8 9 10 11 12 13 References Fowler, L. R., Richardson, S. H., and Hatefi, Y., Biochim. Biophys. Actu, 1962, 64, 170. Jacobs, E. E., Andrews, E. C., Cunningham, W.. and Crane, F. L., Biochem. Biophys. Res. Commun., 1966, 25, 87. Nakagawa, H., and Asano, A., J . Biochem., 1970, 68, 737. Tzagoloff, A., and Penefsky, H. S., Methods Enzymol., 1971, 22, 219. Marchesi, V. T., and Andrews, E. P., Science, 1971,174,1247. Gulik-Krzywicki, T., Biochim. BiophyJ. Actu, 1975, 415, 1. Nozaki, Y., Raynolds, J. A . , and Tanford, C., J. Biol. Chem., 1974, 249,4452. Helenius, A., and Simons, K . . Biochim. Biuphys. Actu, 1975, 415, 29. Takagi, T., Tsujii, K., and Shirahama, K., J. Biochem., 1975, 77, 939. Okahata, Y . , and Ijiro, K., J. Chem. Soc., Chem. Commun., 1988, 1392. Takahashi, K . , Saito, J . , and lnada, Y., J . Am. Oil Chem. Soc., 1988, 65, 911. Tsuzuki, W., Okahata, Y . , Katayama, O., and Suzuki, T., J. Chem. Soc., Perkin Trans. I , 1991, 1245. Shimura, S . , Tsuzuki, W., and Suzuki, T., Anal. Sci., 1991, 7, 15. Paper 2/04 7961 Received September 7, 1992 Accepted October 12, 1992