Anal. Chem. 1999, 71, 1323-1325
High Resolution of DL-Tryptophan at High Flow Rates Using a Bovine Serum Albumin-Multilayered Porous Hollow-Fiber Membrane M. Nakamura,† S. Kiyohara,‡ K. Saito,*,†,§ K. Sugita,† and T. Sugo⊥
Department of Materials Technology, Faculty of Engineering, Chiba University, Inage, Chiba 263, Japan, Fine Chemicals Research Laboratory, Nisshin Flour Milling Company, Ltd., Iruma, Saitama 356, Japan, “Form and Function”, PRESTO, Japan Science and Technology Corporation, Inage, Chiba 263, Japan, and Takasaki Radiation Chemistry Research Establishment, Japan Atomic Energy Research Institute, Takasaki, Gunma 370-12, Japan
We describe a bovine serum albumin (BSA)-multilayeradsorbed porous hollow-fiber membrane as a stationary phase that enables chiral separations at a high resolution and high rate. Epoxy-group-containing graft chains were uniformly immobilized on the surface of pores throughout a porous hollow-fiber membrane by radiation-induced graft polymerization. Subsequently, a diethylamino group as an anion-exchange moiety was introduced to the graft chains, which caused the chains to expand toward the interior of the pores due to mutual electrostatic repulsion. The expanding polymer chains provided multilayer binding sites for BSA as a chiral selector. BSA with a degree of multilayer binding of 4 specifically recognized Ltryptophan with a separation factor of 6.6 during permeation by a mobile phase (Tris-HCl buffer) injected with a racemic solution of DL-tryptophan through the BSAmultilayered porous membrane. In addition, the separation factor was constant irrespective of flow rates of the mobile phase because of negligible diffusional masstransfer resistance of tryptophan to BSA multilayered by the graft chains. Chiral separation, or optical resolution, is an essential technology for producing effective and safe pharmaceuticals. At present, selective crystallization and chiral chromatography are used to resolve various racemic chemicals;1,2 the latter is more versatile in that the chiral ligands can be immobilized on supporting matrixes. The chiral ligands can be classified into three categories: cellulose derivatives,3 synthetic chiral ligands, e.g., the ligands synthesized by Pirkle et al.,4 and proteins such as bovine serum albumin5 and ovoglycoprotein.6 * Address Correspondence to this author at Chiba University. TEL/FAX: +8143-290-3439. E-mail:
[email protected]. † Chiba University. ‡ Nisshin Flour Milling Co. § Japan Science and Technology Corp. ⊥ Japan Atomic Energy Research Institute. (1) Collet, A.; Brienne, M.-J.; Jacques, J. Chem. Rev. 1980, 80, 215-230. (2) Krstulovic, A. M. Chiral separations by HPLC; Ellis Horwood: New York, 1989. (3) Shibata, T.; Okamoto, I.; Ishii, K. J. Liq. Chromatogr. 1986, 9, 313-340. (4) Pirkle, W. H.; Finn, J. M.; Schreiner, J. L.; Hamper, B. C. J. Am. Chem. Soc. 1981, 103, 3964-3966. (5) Allenmark, S. J. Liq. Chromatogr. 1986, 9, 425-442. (6) Haginaka, J.; Seyama, C.; Kanasugi, N. Anal. Chem. 1995, 67, 2539-2547. 10.1021/ac9805596 CCC: $18.00 Published on Web 02/27/1999
© 1999 American Chemical Society
We have so far developed modified porous hollow-fiber membranes for membrane chromatography of biomolecules based on ion-exchange,7 affinity,8 and hydrophobic9 interactions. For example, polymer chains containing anion-exchange groups were appended uniformly across a porous hollow-fiber membrane, the thickness of which was about 1 mm. In addition, we demonstrated that, by permeating bovine serum albumin (BSA, Mr ) 66 300, pI ) 4.7-4.8) in a buffer solution (pH 8) through the pores, the membrane bound the BSA in the multilayers with a capacity as much as 10-fold greater than the monolayer binding capacity.10 This phenomenon can be explained by considering that the anionexchange-group-containing polymer chains expand from the pore surface toward the pore interior due to mutual electrostatic repulsion and provide three-dimensional binding sites for proteins with negative charges. This BSA-multilayered structure on the surface of the pores is evaluated as a novel stationary phase for the chiral separation. We describe the chiral separation of DL-tryptophan with high resolution and at a high flow rate using a BSA-multilayer-adsorbed porous hollow-fiber membrane. The chiral separation of amino acids and their derivatives using the hollow-fiber membrane module based on liquid extraction has been previously reported;11,12 therefore, the membrane works as a support of the chiral selector-containing liquid, while the system suggested here enables high-rate separation aided by permeation of the solution through the porous hollow-fiber membranes based on specific adsorption. EXPERIMENTAL SECTION Preparation of a BSA-Multilayered Porous Membrane. A BSA-multilayer-adsorbed porous membrane was prepared via three steps, as shown in Figure 1. (A) Graft Polymerization of an Epoxy-Group-Containing Monomer onto a Porous Membrane. A porous hollow-fiber (7) Matoba, S.; Tsuneda, S.; Saito, K.; Sugo, T. Bio/Technology 1995, 13, 795797. (8) Kim, M.; Saito, K.; Furusaki, S.; Sugo, T.; Ishigaki, I. J. Chromatogr. 1991, 586, 27-33. (9) Kubota, N.; Kounosu, M.; Saito, K.; Sugita, K.; Watanabe, K.; Sugo, T. Biotechnol. Prog. 1997, 13, 89-95. (10) Tsuneda, S.; Saito, K.; Furusaki, S.; Sugo, T. J. Chromatogr. A 1995, 689, 211-218. (11) Ding, H. B.; Carr, P. W.; Cussler, E. L. AIChE J. 1992, 38, 1493-1498. (12) Pirkle, W. H.; Bowen, W. E. Tetrahedron 1994, 5, 773-776.
Analytical Chemistry, Vol. 71, No. 7, April 1, 1999 1323
Figure 1. Preparation scheme of a BSA-multilayered porous hollow-fiber membrane.
as follows:
q)
∫
Vs
0
(C0 - C) dV/W
(1)
where V and W are the volume of the effluent penetrating the outside surface of the DEA-EA fiber and the weight of the fiber, respectively. Vs is the effluent volume when C reaches C0. In addition, the degree of BSA multilayer binding was evaluated by dividing the value of q by a theoretical monolayer binding capacity, qt, defined as
qt ) Mrav/(aNA) Figure 2. Experimental apparatus for chromatography using a single hollow-fiber module.
membrane made of polyethylene, which was supplied by Asahi Chemical Industry Co., Ltd., Japan, was used as the trunk polymer for grafting. The hollow fiber had inner and outer diameters of 1.8 and 3.1 mm, respectively, with a pore size of 0.4 µm and porosity of 70%. The hollow fiber, 10 cm long, previously irradiated with an electron beam at a total dose of 200 kGy, was immersed in a 10% (v/v) solution of glycidyl methacrylate (GMA, CH2dCCH3COOCH2CHOCH2) in methanol at 313 K. The amount of grafted GMA was set to 1.9 times the mass of the trunk polymer. (B) Introduction of an Anion-Exchange Group. 60% of the epoxy groups of the poly-GMA chain were converted to diethylamino (DEA) groups (-N(C2H5)2) to selectively bind BSA, and the remaining epoxy groups were reacted with ethanolamine (NH2CH2CH2OH) to reduce nonselective adsorption of BSA. The resultant anion-exchange porous hollow-fiber membrane is referred to as a DEA-EA fiber. (C) Multilayer Binding of BSA onto the Anion-Exchange Porous Hollow-Fiber Membrane. The DEA-EA fiber (i.d. ) 2.4 mm, o.d. ) 4.4 mm, length ) 10 cm) was set in an I configuration with one end connected to a syringe infusion pump and the other end sealed. A 2 mg of BSA/mL of Tris-HCl buffer solution (pH 8) was forced to permeate radially outward through the pores, the surfaces of which were covered with the DEA-group-containing polymer chains, at a transmembrane pressure of 0.1 MPa. The amount of BSA bound, q, was calculated by integrating the concentration difference between the feed and effluent, C0 - C, 1324 Analytical Chemistry, Vol. 71, No. 7, April 1, 1999
(2)
where av is the pore surface area per gram of the fiber (6.0 m2/ g). The term a is the area occupied by a molecule of BSA with an end-on orientation10 (1.4 × 10-17 m2); Mr and NA are the molecular weight of BSA (66 300) and Avogadro’s number, respectively. The amount of the DEA groups was determined from the weight gain of the hollow fiber, which is reported to be equivalent to that determined from the measurement of the total ionexchange capacity by titration.13 To determine the distribution of the DEA groups introduced and the BSA adsorbed across the hollow-fiber thickness, the intensity of the X-ray characteristic of the chloride adsorbed onto the DEA groups and sulfur contained in the BSA molecule was measured with an electron probe X-ray microanalyzer (JEOL JXA-733). The permeation rate of the fiber was determined by the deadend type apparatus described previously.9 The hollow fiber was positioned in a U configuration. Liquid was forced to permeate radially outward at a constant transmembrane pressure of 0.01 MPa. The flow rate of the liquid penetrating the outside of the hollow fiber was measured. Chiral Chromatography. The BSA-multilayer-adsorbed porous hollow-fiber membrane was mounted on a single hollow-fiber module, as shown in Figure 2. The module size was 5 mm in inner diameter and 40 mm in effective length. The module was incorporated into a liquid chromatography system (Shimadzu LC6A). A sample of 20 µL of a racemic tryptophan solution, whose concentration was 0.6 mM, was injected to the flow of the mobile (13) Tsuneda, S.; Saito, K.; Furusaki, S.; Sugo, T.; Ishigaki, I. J. Membr. Sci. 1992, 71, 1-12.
Figure 3. Chiral separation of DL-tryptophan during the permeation of a racemic solution through a BSA-multilayered porous hollow-fiber membrane. (a) Membrane containing 2-hydroxyethylamino group (degree of multilayer binding: monolayer). (b) Membrane containing both diethylamino and 2-hydroxyethylamino groups (degree of multilayer binding, four layers).
phase, a Tris-HCl buffer (pH 8) at ambient temperature. The flow rate of the buffer ranged from 0.3 to 3 mL/min. The tryptophan concentration of the effluent was continuously monitored by measuring the UV absorbance at 280 nm. The separation factor is defined as
separation factor ) (tL - t0)/(tD - t0)
(3)
where tD and tL are the retention times of D- and L-tryptophan in the chromatogram, respectively. t0 is the void time of the membrane module. For comparison, a similar experiment was performed using a module consisting of an exclusively 2-hydroxyethylamino-groupcontaining porous hollow-fiber membrane onto which a monolayer of BSA was adsorbed. RESULTS AND DISCUSSION Properties of BSA-Multilayered Porous Hollow-Fiber Membrane. A DEA group was appended onto a polyethylene (14) Erlandsson, P.; Hansson, L.; Isaksson, R. J. Chromatogr. 1986, 370, 475483. (15) Erlandsson, P.; Nilsson, S. J. Chromatogr. 1989, 482, 35-51. (16) Gilpin, R. K.; Ehtesham, S. E.; Gregory, R. B. Anal. Chem. 1991, 63, 28252828. (17) Andersson, S.; Thompson, R. A.; Allenmark, S. G. J. Chromatogr. 1992, 591, 65-73. (18) Simek, Z.; Vespalec, R. J. Chromatogr. A 1994, 685, 7-14.
porous hollow-fiber membrane with a density of 2.2 mmol/g of the resultant DEA-EA fiber. The inner and outer diameters of the hollow fiber were 2.4 and 4.4 mm, respectively. The liquid permeability, i.e., the permeation rate per unit of inside surface area, of the DEA-EA fiber for the buffer was maintained at 50% of that of the original hollow fiber. Volume swelling of the porous hollow fiber, accompanied by graft polymerization, prevented the graft chains from filling the pores. During permeation of BSA in a Tris-HCl buffer, 190 mg of BSA/g of the membrane was bound to the DEA-EA fiber due to the anion-exchange interaction of BSA with the graft chains. This protein binding capacity was equivalent to 4 times the theoretical monolayer binding capacity (48 mg/g) calculated according to eq 2. A uniform distribution of the DEA groups introduced and the BSA adsorbed across the membrane was ascertained by observing the uniform distribution of chloride and sulfur, respectively, by X-ray microanalysis. The BSA-multilayer-adsorbed hollow fiber exhibited 50% of the flux of the DEA-EA fiber for the buffer. This decrease in the liquid permeability reflects the multilayer adsorption of BSA into the graft chains extending from the pore surface toward the pore interior. Resolution of DL-Tryptophan. A sample of 20 µL of 0.6 mM DL-tryptophan solution was loaded into the BSA-four-layer-adsorbed porous hollow-fiber module at a flow rate of buffer as the mobile phase of 0.5 mL/min. DL-Tryptophan was resolved with a separation factor of 6.6, as shown in Figure 3. In contrast, a BSAmonolayer-adsorbed porous hollow-fiber module exhibited a lower separation factor of 2.9. BSA adsorbed in the multilayers by the polymer chains grafted onto the pore surface worked well at prolonging the retention time of L-tryptophan. The separation factor of DL-tryptophan for BSA immobilized on various supports under different mobile-phase conditions reported previously ranged from 1.2 to 14.14-18 The separation factor obtained with the BSA-multilayered porous hollow-fiber membrane ensures a satisfactory resolution. The separation factor was determined at various flow rates of mobile phase ranging, from 0.3 to 3 mL/min; i.e., the residence times ranged from 4.6 to 46 s in the BSA-multilayered porous hollow fiber. As a result, the separation factor was constant irrespective of the flow rate because diffusional mass-transfer resistance of the target molecule (tryptophan) to the recognition site of the chiral ligand (BSA) was negligible. This is a favorable attribute for a scale-up for the chiral separation of racemic chemicals. Therefore, a separation at high resolution and high flow rate was realized by injecting the racemic tryptophan to the BSA-multilayered porous hollow-fiber membrane. Leakage of BSA was not detected after 20 repeated injections. Multilayer binding of the proteins by the graft chain was observed for various proteins, such as lactoglobulin,10 bovine γ globulin,10 and urease;7 immobilization of other proteins will enable chiral separations of other chiral molecules at a high resolution and high flow rate. ACKNOWLEDGMENT The authors thank Kohei Watanabe and Noboru Kubota of the Industrial Membranes Development Department of Asahi Chemical Industry Co., Ltd., Japan, for providing the starting membrane. Received for review May 21, 1998. Accepted December 22, 1998. AC9805596 Analytical Chemistry, Vol. 71, No. 7, April 1, 1999
1325
