- Email: [email protected]
Modification of poly(glycidyl methacrylate–divinylbenzene) porous microspheres with polyethylene glycol and their adsorption property of protein
Colloids and Surfaces B: Biointerfaces 51 (2006) 93–99
Modification of poly(glycidyl methacrylate–divinylbenzene) porous microspheres with polyethylene glycol and their adsorption property of protein Renwei Wang, Ying Zhang, Guanghui Ma, Zhiguo Su ∗ National Key Laboratory of Biochemical Engineering, Institute of Process Engineering, Chinese Academy of Sciences, P.O. Box 353, Beijing 100080, China Received 21 December 2005; received in revised form 19 April 2006; accepted 21 May 2006 Available online 26 May 2006
Abstract Rigid porous poly(glycidyl methacrylate–divinylbenzene) (P(GMA–DVB)) microspheres were synthesized through suspension polymerization with a mixture of isooctane and 4-methyl-2-pentonal as the porogen. The microspheres were intended to use as column packing materials for protein separation. However, irreversible adsorption of protein was found on the polymer microsphere. To circumvent the problem, polyethylene glycol (PEG) was coupled to the microspheres. The coupling reaction took place between the hydroxyl group of PEG and the epoxy group of the P(GMA–DVB) solid medium in the presence of boron trifluoride. The density of PEG immobilized onto the P(GMA–DVB) can be determined easily by saponification of modified microsphere firstly and then titration of glycerol–PEG. The effect of the cross-linker content of microsphere on the density of PEG immobilization was investigated. Molecular weight of PEG was found to influence the PEG-immobilization density, which subsequently affects the hydrophilicity of the modified P(GMA–DVB). Bovine serum albumin (BSA) and trypsin were used as model proteins to examine the adsorption and desorption properties of the modified P(GMA–DVB) microspheres. The results demonstrated that P(GMA–DVB) porous microsphere with 20% DVB and modified with PEG4000 showed excellent adsorption and desorption properties. Adsorption capacity of BSA on the modified microsphere attained to 51.6 mg/g microsphere, and BSA mass recovery and trypsin activity recovery was up to 97.6% and 98.7%, respectively. The modified microsphere was demonstrated to be a promising hydrophobic interaction chromatography material for purification of protein. © 2006 Elsevier B.V. All rights reserved. Keywords: Poly(glycidyl methacrylate–divinylbenzene); Polyethylene glycol; Modification; Adsorption; Protein
1. Introduction The screening of a suitable matrix is very important in protein adsorption and separation. Agarose beads have been a main matrix in this field due to its large network structure, high hydrophilicity and low irreversible adsorption of protein. However, agarose beads have their own drawbacks, including low rigidity and high cost. Rigid porous polymer microsphere is a more promising alternative candidate for adsorption and separation of protein. However, in many cases, proteins are found to bind irreversibly to the polymer particles or to undergo denaturation due to their strongly hydrophobic interaction with the
∗
Corresponding author. Tel.: +86 10 62561817. E-mail address: [email protected] (Z. Su).
0927-7765/$ – see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.colsurfb.2006.05.015
polymer particles and hard elution conditions. For these reasons, the modification of polymer particles to make them more hydrophilic is desirable. Polyethylene glycol (PEG) is a hydrophilic macromolecular, which only shows mild hydrophobic interaction with proteins at higher salt concentration, and hardly affect the bioactivity of protein under adequate protein purification condition. The supports consisting of covalently bound PEG on polysaccharide gels have been used for the fractionation of protein mixtures by hydrophobic chromatography [1–3]. Poly(glycidyl methacrylate) (PGMA) consists of epoxy side groups, which offer the advantage that the ligand can be covalently immobilized on it by a single chemical reaction step. We synthesized rigid poly(glycidyl methacrylate–divinylbenzene) (P(GMA–DVB)) porous microspheres with porous structure by suspension polymerization in the presence of a mixture of isooctane and 4-
94
R. Wang et al. / Colloids and Surfaces B: Biointerfaces 51 (2006) 93–99
methyl-2-pentonal as the porogen. PEG of various molecular weight was coupled onto P(GMA–DVB) microspheres to increase the hydrophilicity of the polymer microspheres. It was difficult to determine the density of PEG on the agarose or other matrix by easily chemical methods. In this study, it was invented a new method that modified P(GMA–DVB) microspheres in sodium hydroxide solution was saponified to free glycol–PEG from microsphere, and the density of PEG on the P(GMA–DVB) was finally determined by titrating glycol–PEG. The effect of the cross-linker content of microsphere to the density of PEG coupled to P(GMA–DVB), and the effect of the molecular weight of PEG and PEG-immobilized density to the hydrophilicity of the modified P(GMA–DVB) were investigated. In order to find suitable microspheres for purification of proteins, bovine serum albumin (BSA) and trypsin were used as model proteins to examine the adsorption and desorption properties of modified P(GMA–DVB) microspheres. 2. Materials and methods 2.1. Materials Glycidyl methacrylate (GMA, Fluka) was distilled under reduced pressure. Divinylbenzene (DVB, 55 wt.%, Sigma) was washed with 5 wt.% aqueous sodium hydroxide solution and deionized water, and then dried by anhydrous sodium sulfate. Benzoyl peroxide (BPO) (with 25 wt.% moisture content), 4methyl-2-pentanol (analytical grade) and isooctane (analytical grade) were purchased from Beijing Chemical Reagents Company. Poly(vinyl alcohol) (PVA GH20, 87.7 mol% of hydrolysis degree) was provided by Nippon Synthetic Chemical Industry Co. Polyethylene glycols (PEG) (MW 1000, 2000, 4000, 10,000), sodium hydroxide, ammonium sulfate, phosphate, dioxane, glycol and boron trifluoride–etherate (47.5 wt.%) was purchased from Sinopharm Chemical Reagent Co. Ltd. Bovine serum albumin (BSA, MW 66,300) and trypsin (MW 23,800) were from Beijing Xin Jing Ke Biotechnology Co. Ltd., China. 2.2. Preparation of P(GMA–DVB) microspheres P(GMA–DVB) porous microspheres were prepared by suspension copolymerization of GMA and DVB in the presence of a mixture of 4-methyl-2-pentanol and isooctane as porogen. Benzoyl peroxide was used as the initiator. Phase separation between polymer and porogen leaved pores in the structure of the final microspheres after the porogen was washed away. The standard recipe of preparation of P(GMA–DVB) is shown in Table 1. The polymerization was conducted at 80 ◦ C under N2
Table 1 Standard recipe of P(GMA–DVB) microspheres Oil phase Monomers: GMA/DVB = 4/1, 3/2, 1/1, 2/3 (w/w) Porogen: 4-methyl-2-pentonal/isooctane = 3/2 (w/w) Monomers/porogen = 1/1 (w/w) Initiator/monomers = 1 wt.% Aqueous solution PVA/water = 1 wt.% NaNO2 /water = 0.5 wt.% Oil phase/aqueous solution = 1/7 (w/w)
Table 2 Effects of DVB content in monomer to pore size and specific surface area of P(GMA–DVB) microspheres
DVB content in monomer (wt.%) Average pore size (nm) Specific surface area (m2 /g) Total microsphere yield (%) Microsphere yield between 30 and 50 m (%)
Run101
Run102
20
40
36.59 29.97 83.5 62.6
25.80 61.88 85.3 64.3
Run103 50 22.56 168.9 87.1 66.4
Run104 60 13.79 312.7 91.2 67.7
atmosphere using an anchor-type stirrer (500 rpm) for 12 h. The obtained microspheres were firstly washed with water, and then the porogen was removed by extracting with acetone in a Soxhlet apparatus. The microspheres were finally dried in vacuum at 40 ◦ C for 12 h and sieved. The microspheres with the diameter of 30–50 m were used in the following modification experiment. The yield of microspheres after polymerization was form 83.5% (w/w) to 91.2% (w/w), and the fraction of the microspheres between 30 and 50 m was around 65% (w/w), which were shown in Table 2. 2.3. Modification of P(GMA–DVB) microspheres The modification of P(GMA–DVB) microspheres is based on the reaction between the epoxy groups of the microspheres and the hydroxyl groups of PEG, and the reaction was catalyzed by boron trifluoride, showing in Fig. 1. P(GMA–DVB) microspheres were mixed with 5% PEG, (w/v) in dioxane and the amount of PEG was equal to the theoretical epoxy groups of microspheres. The mixture was stirred and kept at 80 ◦ C for 6 h, and then the unreacted epoxy groups were deactivated by incubating in 1% (v/v) glycol for 3 h at 80 ◦ C. The resulted product was washed by deionized water and ethanol, and then dried at room temperature in vacuum for 24 h.
Fig. 1. Reaction of Microspheres with PEG.
R. Wang et al. / Colloids and Surfaces B: Biointerfaces 51 (2006) 93–99
95
Fig. 2. Saponification of modified P(GMA–DVB) in sodium hydroxide solution.
2.4. Characterization of microsphere JSM-6700F (JEOL) of scanning electron microscopy (SEM) was used to observe the surface features of polymer microspheres. The specific surface area and pore size were analyzed by BET nitrogen adsorption method with Autosorb 1 analyzer (Quantachrome Corp., USA). The amount of accessible epoxy groups was determined by titration [4,5]. Briefly, P(GMA–DVB) microspheres were dispersed in tetraethylammonium bromide/acetic acid solution and titrated with perchloric acid/acetic acid to the blue green end point of crystal violet indicator. The method to determine the amount of PEG coupled on P(GMA–DVB) was based on the saponification reaction shown in Fig. 2 . The modified microspheres were suspended in 5.0 M sodium hydroxide solution with agitation at 80 ◦ C for 12 h. Then the microspheres were separated from solution by filtration and rinsed by deionized water and filtrated for three times. The filtrates were gathered together. The amounts of glycerol–PEG in the filtrates were determined by titration on spectrophotometric measurement, a barium–iodide complex of the glycol formed by barium and iodine gives a band at 535 nm [6,7]. So, the density of PEG immobilized on the P(GMA–DVB) microspheres can be determined. The hydrophilicity of the modified microspheres was evaluated gravimetrically. The microspheres were suspended in deionized water with agitation at room temperature for 6 h. Then, the microspheres were filtrated to remove water on surface and weighted to get the weight of wet microspheres. Finally, the wet microspheres were dried at 40 ◦ C in vacuum for 24 h and weighed to get the weight of dry microspheres. The hydrophilicity of the microspheres was described with water content and defined as the formulae (1), in which H was hydrophophility: H =1−
weight of dry microspheres weight of wet microspheres
(1)
2.5. Adsorption and desorption of proteins on PEG-modified porous microspheres The intention of the modification of P(GMA–DVB) microsphere was to prepare suitable media for protein adsorption and separation. Here bovine serum albumin (BSA) and trypsin were used as model proteins to perform the adsorption and desorption experiments. BSA adsorption was conducted by mixing 20 mg of modified P(GMA–DVB) microsphere with 10 ml of BSA in 0.1 M phosphate buffer (pH 7.0). The microspheres were washed three times with the buffers before use, by the repeated operation of centrifugation (RCF: 8497 × g, 10 min, 4 ◦ C), decantation
and redispersion. BSA solutions were prepared beforehand and the concentration was varied from 0.2 to 2.0 mg/ml. Ammonium sulfate was added to solution to increase hydrophobic interaction between proteins and microspheres. The suspensions were incubated on a shaking incubator at 25 ◦ C for 6 h. Then, the re-suspension and centrifugation (RCF: 8497 × g, 10 min, 4 ◦ C) were repeated for three times. All of the suspensions were gathered together, and BSA concentration was determined. Desorption experiments were performed by replacing the ammonium sulfate solution with 0.1 M phosphate buffer (pH 7.0) in the eluant. The mass of the eluted BSA was also determined. The trypsin adsorption experiment was conducted in 0.05 M Tris–HCl buffer (pH 7.6), plus 1 M ammonium sulfate, and at 25 ◦ C. The 20 mg modified microspheres suspended in 10 ml trypsin solution were added into a series of tubes. The concentration range of trypsin in these tubes is from 0 to 2 mg/ml. The suspensions were incubated on a shaking incubator at 25 ◦ C for 12 h to promote trypsin adsorption onto the microsphere. Then, the microsphere was separated by centrifugation (RCF: 8497 × g, 10 min, 4 ◦ C) and the residual trypsin (namely free trypsin) in supernatant is measured. The re-suspension and centrifugation were repeated for three times. The absorbed trypsin on the microspheres was eluted with 0.05 M Tris–HCl buffer (pH 7.6) and the activity and mass of eluted trypsin were also determined. The protein concentrations in solution were determined by Bradford method [8]. The activity of trypsin was determined according to the protocols of reference by hydrolyzing casein [9]. The principle of this method was based on the shift adsorption of dye from 365 to 595 nm after the binding of dye coomassie brilliant blue G-250 to the protein. In this method, when casein is hydrolyzed by trypsin the casein–dye complex cannot be formed. Therefore, the decline of the adsorption in 595 nm accurately reflects the activity of trypsin. In this study, the mass and activity recoveries of protein were defined as the percentage of that of free protein and eluted protein in that of total protein to investigate the extent of irreversible adsorption of protein on microspheres, that is, high recovery means low irreversible adsorption. 3. Results and discussion 3.1. Characters of P(GMA–DVB) porous microspheres P(GMA–DVB) porous microspheres were successfully obtained by suspension polymerization with a mixture of 4methyl-2-pentanol and isooctane used as a porogen. The mor-
96
R. Wang et al. / Colloids and Surfaces B: Biointerfaces 51 (2006) 93–99
Fig. 4. Content of theoretical () and measured () epoxy group in copolymer microsphere.
3.2. Density of PEG immobilized onto the P(GMA–DVB) microspheres
Fig. 3. SEM photo of P(GMA–DVB) microspheres (Run101): (a) total image; (b) surface image.
phology and surface feature are shown in Fig. 3. The effect the DVB content from 20 to 60 wt.% in monomer to the pore size and specific surface of the microsphere P(GMA–DVB) was investigated and the results was shown in Table 2. The specific surface area of the porous microspheres increased and the average pore size decreased with the increase of DVB content in total monomer mixture. Higher specific surface area was obtained by adding higher fraction of DVB, however, higher cross-linker content could result in smaller pores which would be unfavorable for the immobilization of PEG, therefore, and the balance should be made in this study. The theoretical epoxy group in copolymer microsphere was dependent on GMA content in monomer. Here, the theoretical epoxy was defined as the proportion of epoxy in total polymer by assuming all of the monomers were converted to the polymer. Fig. 4 shows the measured and the theoretical epoxy group contents in the copolymer microsphere. Because a part of epoxy groups was buried in the inner copolymer, they were not chemically reactive under titration condition [10]. With the increase of DVB content, or with the decrease of GMA content, more surfaces of microspheres were titrated and therefore the measured amount of epoxy group was close to the theoretical value.
P(GMA–DVB) microspheres were modified with PEG by using boron trifluoride as a catalyst. Fig. 5 shows the effect of DVB content in monomer to the immobilized density of PEG of various molecular weights. In Run101 microsphere, GMA content was 80 wt.% (where DVB content is 20 wt.%), and immobilized-PEG density was higher than that of the other microsphere, which implied that the immobilized-PEG density was mainly depended on GMA content. The available epoxy groups were almost linearly decreased with the increase of DVB content in monomer, which has been shown in Fig. 4. However, immobilized-PEG density did not linearly decrease with the decrease of GMA content, i.e. the increase of DVB content as shown in Fig. 5, and the decrease trend became slower with the increase of DVB content. This result showed that the immobilized-PEG density was also affected by specific surface area of microsphere, which was increased with the increment of
Fig. 5. Density of PEG on microspheres of various cross-linking degrees. Molecular weight of PEG was () 1000; (䊉) 2000; () 4000; () 6000; () 10,000.
R. Wang et al. / Colloids and Surfaces B: Biointerfaces 51 (2006) 93–99
DVB content in monomer. Fig. 5 also demonstrates that lower density of PEG molecules coupled onto the P(GMA–DVB) microsphere was obtained at the higher molecular weight of PEG. It is hypothesized that the immobilized-PEG in the micropore may prevent other PEGs to enter into micropores and react with epoxy group, so the density of immobilized-PEG was affected apparently by the steric action of PEG molecule, and this action became more apparent in the case of PEG molecular weight over 4000. The investigation demonstrate that the immobilized density of PEG onto the P(GMA–DVB) microsphere was affected by GMA content in monomer, PEG molecular weight, pore size, and specific surface area of microsphere. Further more pore size and specific surface also related to each other. 3.3. Hydrophilicity of modified P(GMA–DVB) porous microspheres The hydrophilicity of P(GMA–DVB) microspheres was increased after being modified by PEG. The hydrophilicity was affected by both the molecular weight of PEG and cross-linking degree controlled by the DVB contents in monomer (shown in Fig. 6). The hydrophilicity of modified microspheres increased with the increase of molecular weight of PEG. However, the hydrophilicity of modified microspheres decreased gradually if the molecular weight was larger than 4000. It was also observed that Run101 containing the least DVB (20%) was the highest hydrophilicity. Even though the specific surface area of Run104 microsphere was the largest, its smaller pore size and lower content epoxy groups limited the density of PEG with larger molecular weight on it. This showed that the hydrophilicity of modified microsphere depended on both the density of PEG on microsphere and molecular weight of PEG. 3.4. Adsorption of protein on modified P(GMA–DVB) microsphere
97
tion, and also little denaturation to the protein during the adsorption and desorption. Although the increase of DVB content can enhance the rigidity of the P(GMA–DVB) microsphere and specific surface area for protein adsorption, it might cause irreversible adsorption of protein. Therefore, the adsorption and desorption behavior of proteins to modified P(GMA–DVB) microsphere were investigated. The adsorption of BSA on microspheres can be described by Langmuir equation shown as the formulae: q=
q mc kd + c
1 kd 1 1 = × + q qm c qm
(2) (3)
where c (mg/ml) is the equilibrium concentration of BSA in bulk solution, q (mg/g microsphere) the adsorbed density of protein to the adsorbent, qm (mg/g microsphere) the saturated adsorption, and kd is the dissociation constant. Parameters of kd and qm in the Langmuir equation were obtained with Eq. (3) using the linear regression method with the experimental data. The effect of structure of microspheres, the molecular weight of PEG and salt concentration to the adsorption capacity of the microsphere was investigated. 3.4.1. Effect of DVB content to adsorption capacity of modified microspheres The effects of DVB content of microspheres modified with PEG4000 to BSA adsorption property was shown in Fig. 7, demonstrated that higher adsorption capacity qm and lower dissociation constant kd were obtained by increasing DVB content in microspheres, in which the qm of Run101 (DVB content = 20 wt.%) was 51.6 mg/g microsphere while Run104 ((DVB content = 60 wt.%) was 185.2 mg/g microsphere. The increase of DVB content in microspheres led to lower immobilized-PEG density, which resulted in higher
An ideal microsphere media for purification of protein should have higher adsorption capacity and lower irreversible adsorp-
Fig. 6. The effect of molecular weight of PEG and cross-linking degree on the hydrophilicity of modified microsphere. () Run101; (䊉) Run102; () Run103; () Run104.
Fig. 7. The equilibration adsorption curve of microspheres modified with PEG4000. (NH4 )2 SO4 concentration: 1.0 M. () Run101, qm = 51.6 mg/g microsphere, kd = 0.50; (䊉) Run102, qm = 83.3 mg/g microsphere, kd = 0.13; () Run103, qm = 103.1 mg/g microsphere, kd = 0.06; () Run104, qm = 185.2 mg/g microsphere, kd = 0.02.
98
R. Wang et al. / Colloids and Surfaces B: Biointerfaces 51 (2006) 93–99
Fig. 8. Effect of ammonium sulfate concentration of solution to the BSA adsorption on Run101 microspheres modified with PEG4000. () (NH4 )2 SO4 concentration: 0.5 M, qm = 22.6 mg/g microsphere, kd = 0.73; () (NH4 )2 SO4 concentration: 1.0 M, qm = 51.55 mg/g microsphere, kd = 0.49; () (NH4 )2 SO4 concentration: 1.5 M, qm = 52.51 mg/g microsphere, kd = 0.40; (䊉) (NH4 )2 SO4 concentration: 2.0 M, qm = 54.05 mg/g microsphere, kd = 0.14 × 10−3 .
protein adsorption on surface of the microsphere by strong hydrophobic interaction between DVB and protein. 3.4.2. Effect of ammonium sulfate concentration to adsorption capacity of microspheres The effects of salt concentration of ammonium sulfate to the BSA adsorption on Run101 microspheres modified with PEG4000 are shown in Fig. 8. BSA was not adsorbed onto the modified microspheres without ammonium sulfate in solution. The adsorption capacity of microspheres for BSA was increased from qm 22.57 to 51.55 with the increase of ammonium sulfate concentration from 0.5 to 1.0 M, but no further obvious increase was observed if ammonium sulfate concentration was over 1.0 M. 3.4.3. Effect of molecular weight of PEG to the adsorption capacity of microspheres Protein adsorption capacity of microsphere was depended on both of the microsphere structure and PEG on microsphere, including PEG molecular weight and PEG density. In Fig. 9, the microsphere of 40 wt.% DVB (Run102) showed higher protein adsorption capacity than that of the microsphere of 20 wt.% DVB (Run101), which was explained by stronger hydrophobic interaction between DVB and protein and higher specific surface area of the microsphere. The microsphere modified by PEG with larger molecular weight (<4000) showed higher adsorption capacity. However, when PEG molecular weight was higher than 4000, the density of PEG coupled onto microsphere was limited. As a result, the adsorption amount was decreased due to the decrease of hydrophobic adsorption of PEG to protein. Microspheres modified with PEG4000 shown the highest adsorption capacity due to the both effects of larger PEG molecular weight and higher PEG density on this microsphere.
Fig. 9. Adsorption capacity of microspheres modified with different molecular weight of PEG: () Run101; (䊉) Run102.
3.4.4. Recovery of protein adsorbed on microspheres It is necessary to investigate the mass recovery and the activity recovery of protein on the modified microsphere, in order to guarantee the performance of P(GMA–DVB) microsphere used as protein adsorption media. The mass recovery of BSA and the activity recovery of trypsin were shown in Figs. 10 and 11, respectively. These figures demonstrated that lower mass recovery and activity recovery of protein was obtained in the unmodified microsphere than that in the modified microsphere. However, the recovery of protein was greatly decreased with the increase of DVB content in microsphere. It may be hypothesized that the hydrophobic group of DVB was not covered enough, and the strong interaction between DVB and protein led to the irreversible adsorption of protein. The maximal mass recovery of BSA was 97.6% (Fig. 10) and activity recovery of trypsin was 98.7% (Fig. 11) in the case of modified microsphere of 20% DVB (Run101) with PEG4000, showing that the protein adsorption on this microsphere was reversible. This implied
Fig. 10. Mass recovery of BSA on Run101: () unmodified microsphere; () modified microsphere with PEG4000.
R. Wang et al. / Colloids and Surfaces B: Biointerfaces 51 (2006) 93–99
99
microsphere and protein. A new method was developed to evaluate PEG density coupled on microspheres. By investigating the cross-linker content of microspheres and PEG molecular weight, it was found that the modified microspheres of low cross-linking degree (20 wt.%, DVB) with PEG4000 showed the highest hydrophilicity, and had reversible hydrophobic interaction with the model protein. Adsorption capacity of the modified microspheres attained to 51.6 mg/g microsphere. BSA mass recovery and trypsin activity recovery attained to 97.6% and 98.7%, respectively. Therefore, this microsphere is expected to become a promising protein purification media. Acknowledgements
Fig. 11. Activity recovery of trypsin on Run101: () unmodified microsphere; () modified microsphere with PEG4000.
We would like to thank the financial support from National Natural Science Foundation of China under 20136020 and 20125616. The support from Chinese Academy of Sciences is also gratefully acknowledged. References
that the adsorption capacity of this P(GMA–DVB) microsphere is acceptable. And, the higher mechanical strength of polymer porous microsphere also makes it promising a potential media for protein purification. 4. Conclusion In order to develop an alternative of conventional hydrogel for protein purification such as agarose, a rigid porous microsphere P(GMA–DVB) was synthesized and modified further by PEG to decrease the irreversible interaction between
[1] U. Matsumoto, Y. Shibusawa, J. Chromatogr. 187 (1980) 351–362. [2] T.G.I. Ling, B. Mattiasson, J. Chromatogr. 254 (1983) 83–89. [3] R. Mathis, P. Hubert, E. Dellacherie, J. Chromatogr. 474 (1989) 396–399. [4] R.R. Jay, Anal. Chem. 36 (1964) 667–668. ˇ ˇ [5] V. Smigol, F. Svec, J. Appl. Polym. Sci. 46 (1992) 1439–1448. [6] C.E. Childs, Microchem. J. 20 (1975) 190–192. [7] G.E.C. Sims, T.J. Snape, Anal. Biochem. 107 (1980) 60–63. [8] M.M. Bradford, Anal. Biochem. 72 (1976) 248–254. [9] H. Zhou, Z.B. Lu, J. Qi, W. Li, G.F. Bickerstaff, Chinese Biochem. J. 10 (1994) 630–632. [10] H. Kuroda, Z. Osawa, Eur. Polym. J. 31 (1995) 57–62.
Modification of poly(glycidyl methacrylate–divinylbenzene) porous microspheres with polyethylene glycol and their adsorption property of protein Renwei Wang, Ying Zhang, Guanghui Ma, Zhiguo Su ∗ National Key Laboratory of Biochemical Engineering, Institute of Process Engineering, Chinese Academy of Sciences, P.O. Box 353, Beijing 100080, China Received 21 December 2005; received in revised form 19 April 2006; accepted 21 May 2006 Available online 26 May 2006
Abstract Rigid porous poly(glycidyl methacrylate–divinylbenzene) (P(GMA–DVB)) microspheres were synthesized through suspension polymerization with a mixture of isooctane and 4-methyl-2-pentonal as the porogen. The microspheres were intended to use as column packing materials for protein separation. However, irreversible adsorption of protein was found on the polymer microsphere. To circumvent the problem, polyethylene glycol (PEG) was coupled to the microspheres. The coupling reaction took place between the hydroxyl group of PEG and the epoxy group of the P(GMA–DVB) solid medium in the presence of boron trifluoride. The density of PEG immobilized onto the P(GMA–DVB) can be determined easily by saponification of modified microsphere firstly and then titration of glycerol–PEG. The effect of the cross-linker content of microsphere on the density of PEG immobilization was investigated. Molecular weight of PEG was found to influence the PEG-immobilization density, which subsequently affects the hydrophilicity of the modified P(GMA–DVB). Bovine serum albumin (BSA) and trypsin were used as model proteins to examine the adsorption and desorption properties of the modified P(GMA–DVB) microspheres. The results demonstrated that P(GMA–DVB) porous microsphere with 20% DVB and modified with PEG4000 showed excellent adsorption and desorption properties. Adsorption capacity of BSA on the modified microsphere attained to 51.6 mg/g microsphere, and BSA mass recovery and trypsin activity recovery was up to 97.6% and 98.7%, respectively. The modified microsphere was demonstrated to be a promising hydrophobic interaction chromatography material for purification of protein. © 2006 Elsevier B.V. All rights reserved. Keywords: Poly(glycidyl methacrylate–divinylbenzene); Polyethylene glycol; Modification; Adsorption; Protein
1. Introduction The screening of a suitable matrix is very important in protein adsorption and separation. Agarose beads have been a main matrix in this field due to its large network structure, high hydrophilicity and low irreversible adsorption of protein. However, agarose beads have their own drawbacks, including low rigidity and high cost. Rigid porous polymer microsphere is a more promising alternative candidate for adsorption and separation of protein. However, in many cases, proteins are found to bind irreversibly to the polymer particles or to undergo denaturation due to their strongly hydrophobic interaction with the
∗
Corresponding author. Tel.: +86 10 62561817. E-mail address: [email protected] (Z. Su).
0927-7765/$ – see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.colsurfb.2006.05.015
polymer particles and hard elution conditions. For these reasons, the modification of polymer particles to make them more hydrophilic is desirable. Polyethylene glycol (PEG) is a hydrophilic macromolecular, which only shows mild hydrophobic interaction with proteins at higher salt concentration, and hardly affect the bioactivity of protein under adequate protein purification condition. The supports consisting of covalently bound PEG on polysaccharide gels have been used for the fractionation of protein mixtures by hydrophobic chromatography [1–3]. Poly(glycidyl methacrylate) (PGMA) consists of epoxy side groups, which offer the advantage that the ligand can be covalently immobilized on it by a single chemical reaction step. We synthesized rigid poly(glycidyl methacrylate–divinylbenzene) (P(GMA–DVB)) porous microspheres with porous structure by suspension polymerization in the presence of a mixture of isooctane and 4-
94
R. Wang et al. / Colloids and Surfaces B: Biointerfaces 51 (2006) 93–99
methyl-2-pentonal as the porogen. PEG of various molecular weight was coupled onto P(GMA–DVB) microspheres to increase the hydrophilicity of the polymer microspheres. It was difficult to determine the density of PEG on the agarose or other matrix by easily chemical methods. In this study, it was invented a new method that modified P(GMA–DVB) microspheres in sodium hydroxide solution was saponified to free glycol–PEG from microsphere, and the density of PEG on the P(GMA–DVB) was finally determined by titrating glycol–PEG. The effect of the cross-linker content of microsphere to the density of PEG coupled to P(GMA–DVB), and the effect of the molecular weight of PEG and PEG-immobilized density to the hydrophilicity of the modified P(GMA–DVB) were investigated. In order to find suitable microspheres for purification of proteins, bovine serum albumin (BSA) and trypsin were used as model proteins to examine the adsorption and desorption properties of modified P(GMA–DVB) microspheres. 2. Materials and methods 2.1. Materials Glycidyl methacrylate (GMA, Fluka) was distilled under reduced pressure. Divinylbenzene (DVB, 55 wt.%, Sigma) was washed with 5 wt.% aqueous sodium hydroxide solution and deionized water, and then dried by anhydrous sodium sulfate. Benzoyl peroxide (BPO) (with 25 wt.% moisture content), 4methyl-2-pentanol (analytical grade) and isooctane (analytical grade) were purchased from Beijing Chemical Reagents Company. Poly(vinyl alcohol) (PVA GH20, 87.7 mol% of hydrolysis degree) was provided by Nippon Synthetic Chemical Industry Co. Polyethylene glycols (PEG) (MW 1000, 2000, 4000, 10,000), sodium hydroxide, ammonium sulfate, phosphate, dioxane, glycol and boron trifluoride–etherate (47.5 wt.%) was purchased from Sinopharm Chemical Reagent Co. Ltd. Bovine serum albumin (BSA, MW 66,300) and trypsin (MW 23,800) were from Beijing Xin Jing Ke Biotechnology Co. Ltd., China. 2.2. Preparation of P(GMA–DVB) microspheres P(GMA–DVB) porous microspheres were prepared by suspension copolymerization of GMA and DVB in the presence of a mixture of 4-methyl-2-pentanol and isooctane as porogen. Benzoyl peroxide was used as the initiator. Phase separation between polymer and porogen leaved pores in the structure of the final microspheres after the porogen was washed away. The standard recipe of preparation of P(GMA–DVB) is shown in Table 1. The polymerization was conducted at 80 ◦ C under N2
Table 1 Standard recipe of P(GMA–DVB) microspheres Oil phase Monomers: GMA/DVB = 4/1, 3/2, 1/1, 2/3 (w/w) Porogen: 4-methyl-2-pentonal/isooctane = 3/2 (w/w) Monomers/porogen = 1/1 (w/w) Initiator/monomers = 1 wt.% Aqueous solution PVA/water = 1 wt.% NaNO2 /water = 0.5 wt.% Oil phase/aqueous solution = 1/7 (w/w)
Table 2 Effects of DVB content in monomer to pore size and specific surface area of P(GMA–DVB) microspheres
DVB content in monomer (wt.%) Average pore size (nm) Specific surface area (m2 /g) Total microsphere yield (%) Microsphere yield between 30 and 50 m (%)
Run101
Run102
20
40
36.59 29.97 83.5 62.6
25.80 61.88 85.3 64.3
Run103 50 22.56 168.9 87.1 66.4
Run104 60 13.79 312.7 91.2 67.7
atmosphere using an anchor-type stirrer (500 rpm) for 12 h. The obtained microspheres were firstly washed with water, and then the porogen was removed by extracting with acetone in a Soxhlet apparatus. The microspheres were finally dried in vacuum at 40 ◦ C for 12 h and sieved. The microspheres with the diameter of 30–50 m were used in the following modification experiment. The yield of microspheres after polymerization was form 83.5% (w/w) to 91.2% (w/w), and the fraction of the microspheres between 30 and 50 m was around 65% (w/w), which were shown in Table 2. 2.3. Modification of P(GMA–DVB) microspheres The modification of P(GMA–DVB) microspheres is based on the reaction between the epoxy groups of the microspheres and the hydroxyl groups of PEG, and the reaction was catalyzed by boron trifluoride, showing in Fig. 1. P(GMA–DVB) microspheres were mixed with 5% PEG, (w/v) in dioxane and the amount of PEG was equal to the theoretical epoxy groups of microspheres. The mixture was stirred and kept at 80 ◦ C for 6 h, and then the unreacted epoxy groups were deactivated by incubating in 1% (v/v) glycol for 3 h at 80 ◦ C. The resulted product was washed by deionized water and ethanol, and then dried at room temperature in vacuum for 24 h.
Fig. 1. Reaction of Microspheres with PEG.
R. Wang et al. / Colloids and Surfaces B: Biointerfaces 51 (2006) 93–99
95
Fig. 2. Saponification of modified P(GMA–DVB) in sodium hydroxide solution.
2.4. Characterization of microsphere JSM-6700F (JEOL) of scanning electron microscopy (SEM) was used to observe the surface features of polymer microspheres. The specific surface area and pore size were analyzed by BET nitrogen adsorption method with Autosorb 1 analyzer (Quantachrome Corp., USA). The amount of accessible epoxy groups was determined by titration [4,5]. Briefly, P(GMA–DVB) microspheres were dispersed in tetraethylammonium bromide/acetic acid solution and titrated with perchloric acid/acetic acid to the blue green end point of crystal violet indicator. The method to determine the amount of PEG coupled on P(GMA–DVB) was based on the saponification reaction shown in Fig. 2 . The modified microspheres were suspended in 5.0 M sodium hydroxide solution with agitation at 80 ◦ C for 12 h. Then the microspheres were separated from solution by filtration and rinsed by deionized water and filtrated for three times. The filtrates were gathered together. The amounts of glycerol–PEG in the filtrates were determined by titration on spectrophotometric measurement, a barium–iodide complex of the glycol formed by barium and iodine gives a band at 535 nm [6,7]. So, the density of PEG immobilized on the P(GMA–DVB) microspheres can be determined. The hydrophilicity of the modified microspheres was evaluated gravimetrically. The microspheres were suspended in deionized water with agitation at room temperature for 6 h. Then, the microspheres were filtrated to remove water on surface and weighted to get the weight of wet microspheres. Finally, the wet microspheres were dried at 40 ◦ C in vacuum for 24 h and weighed to get the weight of dry microspheres. The hydrophilicity of the microspheres was described with water content and defined as the formulae (1), in which H was hydrophophility: H =1−
weight of dry microspheres weight of wet microspheres
(1)
2.5. Adsorption and desorption of proteins on PEG-modified porous microspheres The intention of the modification of P(GMA–DVB) microsphere was to prepare suitable media for protein adsorption and separation. Here bovine serum albumin (BSA) and trypsin were used as model proteins to perform the adsorption and desorption experiments. BSA adsorption was conducted by mixing 20 mg of modified P(GMA–DVB) microsphere with 10 ml of BSA in 0.1 M phosphate buffer (pH 7.0). The microspheres were washed three times with the buffers before use, by the repeated operation of centrifugation (RCF: 8497 × g, 10 min, 4 ◦ C), decantation
and redispersion. BSA solutions were prepared beforehand and the concentration was varied from 0.2 to 2.0 mg/ml. Ammonium sulfate was added to solution to increase hydrophobic interaction between proteins and microspheres. The suspensions were incubated on a shaking incubator at 25 ◦ C for 6 h. Then, the re-suspension and centrifugation (RCF: 8497 × g, 10 min, 4 ◦ C) were repeated for three times. All of the suspensions were gathered together, and BSA concentration was determined. Desorption experiments were performed by replacing the ammonium sulfate solution with 0.1 M phosphate buffer (pH 7.0) in the eluant. The mass of the eluted BSA was also determined. The trypsin adsorption experiment was conducted in 0.05 M Tris–HCl buffer (pH 7.6), plus 1 M ammonium sulfate, and at 25 ◦ C. The 20 mg modified microspheres suspended in 10 ml trypsin solution were added into a series of tubes. The concentration range of trypsin in these tubes is from 0 to 2 mg/ml. The suspensions were incubated on a shaking incubator at 25 ◦ C for 12 h to promote trypsin adsorption onto the microsphere. Then, the microsphere was separated by centrifugation (RCF: 8497 × g, 10 min, 4 ◦ C) and the residual trypsin (namely free trypsin) in supernatant is measured. The re-suspension and centrifugation were repeated for three times. The absorbed trypsin on the microspheres was eluted with 0.05 M Tris–HCl buffer (pH 7.6) and the activity and mass of eluted trypsin were also determined. The protein concentrations in solution were determined by Bradford method [8]. The activity of trypsin was determined according to the protocols of reference by hydrolyzing casein [9]. The principle of this method was based on the shift adsorption of dye from 365 to 595 nm after the binding of dye coomassie brilliant blue G-250 to the protein. In this method, when casein is hydrolyzed by trypsin the casein–dye complex cannot be formed. Therefore, the decline of the adsorption in 595 nm accurately reflects the activity of trypsin. In this study, the mass and activity recoveries of protein were defined as the percentage of that of free protein and eluted protein in that of total protein to investigate the extent of irreversible adsorption of protein on microspheres, that is, high recovery means low irreversible adsorption. 3. Results and discussion 3.1. Characters of P(GMA–DVB) porous microspheres P(GMA–DVB) porous microspheres were successfully obtained by suspension polymerization with a mixture of 4methyl-2-pentanol and isooctane used as a porogen. The mor-
96
R. Wang et al. / Colloids and Surfaces B: Biointerfaces 51 (2006) 93–99
Fig. 4. Content of theoretical () and measured () epoxy group in copolymer microsphere.
3.2. Density of PEG immobilized onto the P(GMA–DVB) microspheres
Fig. 3. SEM photo of P(GMA–DVB) microspheres (Run101): (a) total image; (b) surface image.
phology and surface feature are shown in Fig. 3. The effect the DVB content from 20 to 60 wt.% in monomer to the pore size and specific surface of the microsphere P(GMA–DVB) was investigated and the results was shown in Table 2. The specific surface area of the porous microspheres increased and the average pore size decreased with the increase of DVB content in total monomer mixture. Higher specific surface area was obtained by adding higher fraction of DVB, however, higher cross-linker content could result in smaller pores which would be unfavorable for the immobilization of PEG, therefore, and the balance should be made in this study. The theoretical epoxy group in copolymer microsphere was dependent on GMA content in monomer. Here, the theoretical epoxy was defined as the proportion of epoxy in total polymer by assuming all of the monomers were converted to the polymer. Fig. 4 shows the measured and the theoretical epoxy group contents in the copolymer microsphere. Because a part of epoxy groups was buried in the inner copolymer, they were not chemically reactive under titration condition [10]. With the increase of DVB content, or with the decrease of GMA content, more surfaces of microspheres were titrated and therefore the measured amount of epoxy group was close to the theoretical value.
P(GMA–DVB) microspheres were modified with PEG by using boron trifluoride as a catalyst. Fig. 5 shows the effect of DVB content in monomer to the immobilized density of PEG of various molecular weights. In Run101 microsphere, GMA content was 80 wt.% (where DVB content is 20 wt.%), and immobilized-PEG density was higher than that of the other microsphere, which implied that the immobilized-PEG density was mainly depended on GMA content. The available epoxy groups were almost linearly decreased with the increase of DVB content in monomer, which has been shown in Fig. 4. However, immobilized-PEG density did not linearly decrease with the decrease of GMA content, i.e. the increase of DVB content as shown in Fig. 5, and the decrease trend became slower with the increase of DVB content. This result showed that the immobilized-PEG density was also affected by specific surface area of microsphere, which was increased with the increment of
Fig. 5. Density of PEG on microspheres of various cross-linking degrees. Molecular weight of PEG was () 1000; (䊉) 2000; () 4000; () 6000; () 10,000.
R. Wang et al. / Colloids and Surfaces B: Biointerfaces 51 (2006) 93–99
DVB content in monomer. Fig. 5 also demonstrates that lower density of PEG molecules coupled onto the P(GMA–DVB) microsphere was obtained at the higher molecular weight of PEG. It is hypothesized that the immobilized-PEG in the micropore may prevent other PEGs to enter into micropores and react with epoxy group, so the density of immobilized-PEG was affected apparently by the steric action of PEG molecule, and this action became more apparent in the case of PEG molecular weight over 4000. The investigation demonstrate that the immobilized density of PEG onto the P(GMA–DVB) microsphere was affected by GMA content in monomer, PEG molecular weight, pore size, and specific surface area of microsphere. Further more pore size and specific surface also related to each other. 3.3. Hydrophilicity of modified P(GMA–DVB) porous microspheres The hydrophilicity of P(GMA–DVB) microspheres was increased after being modified by PEG. The hydrophilicity was affected by both the molecular weight of PEG and cross-linking degree controlled by the DVB contents in monomer (shown in Fig. 6). The hydrophilicity of modified microspheres increased with the increase of molecular weight of PEG. However, the hydrophilicity of modified microspheres decreased gradually if the molecular weight was larger than 4000. It was also observed that Run101 containing the least DVB (20%) was the highest hydrophilicity. Even though the specific surface area of Run104 microsphere was the largest, its smaller pore size and lower content epoxy groups limited the density of PEG with larger molecular weight on it. This showed that the hydrophilicity of modified microsphere depended on both the density of PEG on microsphere and molecular weight of PEG. 3.4. Adsorption of protein on modified P(GMA–DVB) microsphere
97
tion, and also little denaturation to the protein during the adsorption and desorption. Although the increase of DVB content can enhance the rigidity of the P(GMA–DVB) microsphere and specific surface area for protein adsorption, it might cause irreversible adsorption of protein. Therefore, the adsorption and desorption behavior of proteins to modified P(GMA–DVB) microsphere were investigated. The adsorption of BSA on microspheres can be described by Langmuir equation shown as the formulae: q=
q mc kd + c
1 kd 1 1 = × + q qm c qm
(2) (3)
where c (mg/ml) is the equilibrium concentration of BSA in bulk solution, q (mg/g microsphere) the adsorbed density of protein to the adsorbent, qm (mg/g microsphere) the saturated adsorption, and kd is the dissociation constant. Parameters of kd and qm in the Langmuir equation were obtained with Eq. (3) using the linear regression method with the experimental data. The effect of structure of microspheres, the molecular weight of PEG and salt concentration to the adsorption capacity of the microsphere was investigated. 3.4.1. Effect of DVB content to adsorption capacity of modified microspheres The effects of DVB content of microspheres modified with PEG4000 to BSA adsorption property was shown in Fig. 7, demonstrated that higher adsorption capacity qm and lower dissociation constant kd were obtained by increasing DVB content in microspheres, in which the qm of Run101 (DVB content = 20 wt.%) was 51.6 mg/g microsphere while Run104 ((DVB content = 60 wt.%) was 185.2 mg/g microsphere. The increase of DVB content in microspheres led to lower immobilized-PEG density, which resulted in higher
An ideal microsphere media for purification of protein should have higher adsorption capacity and lower irreversible adsorp-
Fig. 6. The effect of molecular weight of PEG and cross-linking degree on the hydrophilicity of modified microsphere. () Run101; (䊉) Run102; () Run103; () Run104.
Fig. 7. The equilibration adsorption curve of microspheres modified with PEG4000. (NH4 )2 SO4 concentration: 1.0 M. () Run101, qm = 51.6 mg/g microsphere, kd = 0.50; (䊉) Run102, qm = 83.3 mg/g microsphere, kd = 0.13; () Run103, qm = 103.1 mg/g microsphere, kd = 0.06; () Run104, qm = 185.2 mg/g microsphere, kd = 0.02.
98
R. Wang et al. / Colloids and Surfaces B: Biointerfaces 51 (2006) 93–99
Fig. 8. Effect of ammonium sulfate concentration of solution to the BSA adsorption on Run101 microspheres modified with PEG4000. () (NH4 )2 SO4 concentration: 0.5 M, qm = 22.6 mg/g microsphere, kd = 0.73; () (NH4 )2 SO4 concentration: 1.0 M, qm = 51.55 mg/g microsphere, kd = 0.49; () (NH4 )2 SO4 concentration: 1.5 M, qm = 52.51 mg/g microsphere, kd = 0.40; (䊉) (NH4 )2 SO4 concentration: 2.0 M, qm = 54.05 mg/g microsphere, kd = 0.14 × 10−3 .
protein adsorption on surface of the microsphere by strong hydrophobic interaction between DVB and protein. 3.4.2. Effect of ammonium sulfate concentration to adsorption capacity of microspheres The effects of salt concentration of ammonium sulfate to the BSA adsorption on Run101 microspheres modified with PEG4000 are shown in Fig. 8. BSA was not adsorbed onto the modified microspheres without ammonium sulfate in solution. The adsorption capacity of microspheres for BSA was increased from qm 22.57 to 51.55 with the increase of ammonium sulfate concentration from 0.5 to 1.0 M, but no further obvious increase was observed if ammonium sulfate concentration was over 1.0 M. 3.4.3. Effect of molecular weight of PEG to the adsorption capacity of microspheres Protein adsorption capacity of microsphere was depended on both of the microsphere structure and PEG on microsphere, including PEG molecular weight and PEG density. In Fig. 9, the microsphere of 40 wt.% DVB (Run102) showed higher protein adsorption capacity than that of the microsphere of 20 wt.% DVB (Run101), which was explained by stronger hydrophobic interaction between DVB and protein and higher specific surface area of the microsphere. The microsphere modified by PEG with larger molecular weight (<4000) showed higher adsorption capacity. However, when PEG molecular weight was higher than 4000, the density of PEG coupled onto microsphere was limited. As a result, the adsorption amount was decreased due to the decrease of hydrophobic adsorption of PEG to protein. Microspheres modified with PEG4000 shown the highest adsorption capacity due to the both effects of larger PEG molecular weight and higher PEG density on this microsphere.
Fig. 9. Adsorption capacity of microspheres modified with different molecular weight of PEG: () Run101; (䊉) Run102.
3.4.4. Recovery of protein adsorbed on microspheres It is necessary to investigate the mass recovery and the activity recovery of protein on the modified microsphere, in order to guarantee the performance of P(GMA–DVB) microsphere used as protein adsorption media. The mass recovery of BSA and the activity recovery of trypsin were shown in Figs. 10 and 11, respectively. These figures demonstrated that lower mass recovery and activity recovery of protein was obtained in the unmodified microsphere than that in the modified microsphere. However, the recovery of protein was greatly decreased with the increase of DVB content in microsphere. It may be hypothesized that the hydrophobic group of DVB was not covered enough, and the strong interaction between DVB and protein led to the irreversible adsorption of protein. The maximal mass recovery of BSA was 97.6% (Fig. 10) and activity recovery of trypsin was 98.7% (Fig. 11) in the case of modified microsphere of 20% DVB (Run101) with PEG4000, showing that the protein adsorption on this microsphere was reversible. This implied
Fig. 10. Mass recovery of BSA on Run101: () unmodified microsphere; () modified microsphere with PEG4000.
R. Wang et al. / Colloids and Surfaces B: Biointerfaces 51 (2006) 93–99
99
microsphere and protein. A new method was developed to evaluate PEG density coupled on microspheres. By investigating the cross-linker content of microspheres and PEG molecular weight, it was found that the modified microspheres of low cross-linking degree (20 wt.%, DVB) with PEG4000 showed the highest hydrophilicity, and had reversible hydrophobic interaction with the model protein. Adsorption capacity of the modified microspheres attained to 51.6 mg/g microsphere. BSA mass recovery and trypsin activity recovery attained to 97.6% and 98.7%, respectively. Therefore, this microsphere is expected to become a promising protein purification media. Acknowledgements
Fig. 11. Activity recovery of trypsin on Run101: () unmodified microsphere; () modified microsphere with PEG4000.
We would like to thank the financial support from National Natural Science Foundation of China under 20136020 and 20125616. The support from Chinese Academy of Sciences is also gratefully acknowledged. References
that the adsorption capacity of this P(GMA–DVB) microsphere is acceptable. And, the higher mechanical strength of polymer porous microsphere also makes it promising a potential media for protein purification. 4. Conclusion In order to develop an alternative of conventional hydrogel for protein purification such as agarose, a rigid porous microsphere P(GMA–DVB) was synthesized and modified further by PEG to decrease the irreversible interaction between
[1] U. Matsumoto, Y. Shibusawa, J. Chromatogr. 187 (1980) 351–362. [2] T.G.I. Ling, B. Mattiasson, J. Chromatogr. 254 (1983) 83–89. [3] R. Mathis, P. Hubert, E. Dellacherie, J. Chromatogr. 474 (1989) 396–399. [4] R.R. Jay, Anal. Chem. 36 (1964) 667–668. ˇ ˇ [5] V. Smigol, F. Svec, J. Appl. Polym. Sci. 46 (1992) 1439–1448. [6] C.E. Childs, Microchem. J. 20 (1975) 190–192. [7] G.E.C. Sims, T.J. Snape, Anal. Biochem. 107 (1980) 60–63. [8] M.M. Bradford, Anal. Biochem. 72 (1976) 248–254. [9] H. Zhou, Z.B. Lu, J. Qi, W. Li, G.F. Bickerstaff, Chinese Biochem. J. 10 (1994) 630–632. [10] H. Kuroda, Z. Osawa, Eur. Polym. J. 31 (1995) 57–62.