Macroporous Copolymer Matrix: 1. Effectiveness of Different

2,4-Tolylene diisocyanate was established as the best spacer arm on the basis of high ligand binding and low nonspecific interactions. The characteris...
1 downloads 0 Views 135KB Size
Download PDF
Biomacromolecules 2001, 2, 1116-1123

1116

Macroporous Copolymer Matrix: 1. Effectiveness of Different Diisocyanate Spacer Arms to Bind Cyclodextrins Deba P. Nayak, Arika M. Kotha, Omprakash S. Yemul, Surendra Ponrathnam,* and Rajan C. Raman Chemical Engineering Division, National Chemical Laboratory, Pune 411 008, Maharashtra, India Received March 21, 2001; Revised Manuscript Received July 4, 2001

Cross-linked macroporous beaded polymer matrices, with pendant hydroxyl groups, were synthesized by the copolymerization of 2-hydroxyethyl methacrylate and ethylene glycol dimethacrylate using suspension polymerization methodology. Novel affinity chromatography matrices were synthesized using various diisocyanates as bifunctional reagents to couple the macroporous polymeric supports, of controlled particle size distribution, with R and β-cyclodextrins. The optimal conditions to couple the hydroxyl groups of cyclodextrin (ligand) and the polymeric supports through urethane linkages were established iteratively using various diisocyanates. Efficacy of ligand binding on the matrix and nonspecific interactions of the synthesized affinity matrices were evaluated to establish the best support and spacer arm. 2,4-Tolylene diisocyanate was established as the best spacer arm on the basis of high ligand binding and low nonspecific interactions. The characteristics of the synthesized affinity matrices toward the adsorption of R and β-cyclodextrin glycosyltransferase (CGTase) were investigated. The binding of β-CGTase was the highest on affinity matrices with the polymeric methylene diisocyanate spacer. The optimal conditions to regenerate the matrices were also established. 1. Introduction Total synthesis of new affinity matrices entails the sequential optimization of procedures aimed at (i) synthesis of suitable support material with a functional group, (ii) modification of the functional group, and (iii) attachment of the ligand to the modified support material. With an established matrix, for the above steps to be successful, the derivatization and coupling reactions should be optimized. The derivatization and coupling chemistry depend on the nature of the functional group present on the support matrix as well as on the ligand. Matrices and the ligands with hydroxyl groups are conventionally activated by epoxides, cyanogen bromide, and divinyl sulfone. The ether bond generated by the transformation of the epoxy group, formed by reaction of matrices with epoxides, is very stable. This epoxidation is carried out in the pH range of 9.0-13.0 (alkaline conditions). Thus, this procedure is suitable for ligands stable under alkaline conditions.1 The hydroxyl can be transformed into the oxirane group, by the reaction with bisoxirane, such as 1,4-bis(2,3-epoxypropoxy) butane.2 Simultaneously, the desired long hydrophilic spacer arm to delink the ligand from the bulk of the support is also introduced. The simultaneous cross-linking increases stability but at the same time decreases flexibility and permeability. Cyanogen bromide offers facile activation of the matrices. However, it is hazardous and the process is a multistep one.3,4 The matrix shrinks very noticeably under extremely high activation as a result of cross-linking.3 The other method of * To whom correspondence should be addressed. Fax: 91-20-5893041. E-mail: [email protected].

activation of the hydroxyl matrices is with divinyl sulfone (DVS). DVS is useful with alkali-resistant nonshrinking polymers. This activation increases the rigidity of the matrices.5 There are a number of reports on binding cyclodextrins (CDs) to polymeric supports. Most of the reported affinity matrices are Sepharose 4B, Sepharose 6B, and Sepharose 6FF coupled to R, β, and γ-CDs through oxiranes.6-10 R-CD has also been coupled to Sepharose 6B via dioxirane.11 Villette et al. prepared polymeric β-CD, by cross-linking with epichlorohydrin, to purify β-cyclodextrin glycosyltransferase (CGTase).12 The type of CD used as ligand is dictated by the affinity for a particular type of CGTase; that is, R-CGTase will bind more strongly to a Sepharose-R-CD matrix than on a Sepharose-β-CD or Sepharose-γ-CD matrix. The purification of R-CGTase has been demonstrated with a Sepharose 6FF-R-CD matrix,6 β-CGTase has been purified using Sepharose 6B-β-CD,7-9 and Sepharose 4B-γ-CD has been used to purify γ-CGTase.10 Rapid and selective adsorption of β-CGTase was established on β-CD cross-linked with epichlorohydrin.12 The adsorbed CGTase was eluted with the corresponding cyclodextrin solution.10,12 Earlier on, there have been reports of adsorbing β-CGTase to starch by suspending the enzyme in starch slurry either with or without ammonium sulfate.13-16 A packed bed of the starch column was successfully evaluated for affinity adsorption of β-CGTase.17 The adsorbed enzyme could be eluted with a solution of β-CD. There have been no literature reports on using diisocyanates as binding reagents in the development of affinity matrices. In this communication, we report our work on the

10.1021/bm010059d CCC: $20.00 © 2001 American Chemical Society Published on Web 08/23/2001

Macroporous Copolymer Matrix

use of reactive diisocyanates in the preparation of affinity chromatography matrices. We synthesized macroporous, beaded, cross-linked copolymers of 2-hydroxyethyl methacrylate and ethylene glycol dimethacrylate as model matrices, R- and β-CDs as ligands, and R- and β-CGTase as proteins. The hydroxyl groups of the matrices were coupled to the hydroxyl groups of the CDs using a variety of diisocyanates to obtain affinity matrices differing in the spacer arm. The nonspecific interactions, extent of ligand binding, and stability toward clean-in-place reagents were assessed to evaluate the suitability of these affinity matrices. 2. Materials and Methods 2.1. Chemicals. 2-Hydroxyethyl methacrylate (HEMA) and ethylene glycol dimethacrylate (EGDM) were procured from Fluka AG (Germany). R and β-Cyclodextrins were gifted by Cerestar (USA) and SA Chemicals (Mumbai, India), respectively. 1,6-Hexamethylene diisocyanate (HMDI), isophorone diisocyanate (IPDI), 2,4-tolylene diisocyanate (TDI), and dibutyltin dilaurate were obtained from Sigma (USA). Phenylene methylene diisocyanate (PMDI) was from BASF (Germany). 1,4-Dioxane, dimethyl sulfoxide (DMSO), and cyclohexanol were procured from SD Fine Chemicals (Boisar, India). Azobis(isobutyronitrile) (AIBN) was obtained from SISCO (Mumbai, India). All fermentation media components were purchased from Himedia (Mumbai, India). Soluble starch was from E-Merck (Mumbai, India). 2.2. Instrumental. Infrared (IR) spectra were recorded on a model IR 470 spectrophotometer (Shimadzu, Japan). Samples of HEMA-EGDM copolymer beads, R and β-CD, and affinity polymer were mixed in KBr pellets and analyzed. 2.3. Preparation of the Polymeric Matrices. The beaded macroporous polymeric matrices were synthesized by suspension copolymerization18 of HEMA and EGDM. The mole ratio of EGDM relative to HEMA was in the range of 25200%. The density of the HEMA-EGDM copolymer matrix with varying cross-link density was measured under the swollen state by the flotation method.19 The solvent mixture used in the experiment was methylene chloride and methyl salicylate. 2.4. Coupling of Ligands. 1,4-Dioxane was made anhydrous by keeping it in sodium hydroxide pellets overnight and was distilled and stored over dried molecular sieves (3 Å). Similarly, DMSO was kept over chromatographic grade alumina overnight, distilled, and stored over dried molecular sieves (4 Å). Diisocyanates are efficiently coupled to macroporous HEMA-EGDM copolymer matrices in the ambient in an aprotic solvent.20 The macroporous copolymer matrices with isocyanate linkages so obtained were coupled to R/βcyclodextrin by a sequential procedure.20 2.5. Ligand Concentration Measurement. The R- and β-CD affinity matrices synthesized were assayed by methyl orange21 and phenolphthalein methods,22 respectively. The affinity matrices were added to the specified solution of the dyes till the decrease in absorbance was nearly 5-6%. The respective underivatized HEMA-EGDM matrices were used

Biomacromolecules, Vol. 2, No. 4, 2001 1117

as controls. The amount of CD bound was calculated from the difference in test and control values. 2.6. Estimation of Bound Diisocyanates. The bound diisocyanates were estimated as follows. Samples (1 g) of the affinity matrices were delivered to a round-bottom flask, and 1 mL of 0.001 N solution of di-n-butylamine in 25 mL of toluene solution was added. The mixture was heated under reflux for 15 min. On cooling, 50 mL of methanol was added through the condenser and the content of the flask was titrated against 0.001 N HCl solution using bromophenol blue as the indicator. A blank experiment was run in parallel. The isocyanate content (expressed in µmol NCO/g of the matrix) was calculated from the difference of the blank and sample readings. 2.7. Porosity Measurement. The pore size distribution, pore volume, and pore surface area were measured by a mercury intrusion porosimetry instrument in the pressure range of 0-4000 kg/cm2 (Quantachrome, USA). The mercury contact angle was 140°. The polymers were also sieved and were in the range of 10-250 µm. 2.8. Nonspecific Interactions. In a 50 mL conical flask, affinity matrix (1 g, 2 µmoles of R-CD/g of dried matrix) was equilibrated with 0.05 M Tris-HCl buffer of pH 7.0 to check for nonspecific ionic interactions. A portion (10 mL) of 0.1 mg/mL bovine serum albumin (BSA) (made in 50 mM Tris-HCl buffer, pH 7.0) was added to each flask, and the mixture was shaken well at 12 °C and 200 rpm for 8 h. The plain matrices with BSA solution served as the control. BSA solution was also shaken under similar conditions, to estimate the change in its concentration in native form under identical conditions. The matrices were decanted and washed with 0.05 M Tris-HCl buffer (pH 7.0) at 12 °C. The adsorbed protein was eluted with 5 mL of 6 M urea solution. The desorbed protein was estimated as described by Lowry.23 Similarly, the hydrophobic interaction was studied at room temperature with protein solution containing 1 M NaCl, under conditions identical to that used for nonspecific ionic interactions. The matrices were washed with 1 M NaCl solution at 25 °C and eluted with 0.05 M Tris-HCl buffer (pH 7) at 12 °C. 2.9. Production of r- and β-CGTase. R and β-CGTase were produced by fermentation, using Klebsiella pneumoniae pneumoniae NCIM 512124 and Bacillus firmus NCIM 5119,25 respectively, in a 1 L magnetically driven top suspended fermenter (Gallenkamp, U.K.). The optimized fermentation medium used was as described earlier.25,26 The fermented broth was centrifuged at 7200g for 20 min. The supernatant was assayed for adsorption of the enzyme in the synthesized affinity matrices. 2.10. Adsorption Studies. Adsorption of both R- and β-CGTase was studied in shake flasks. The matrices (1 g, 2 µmol of R/β-CDs) were equilibrated with 10 mL of 0.05 M Tris-HCl buffer, pH 7.0. The fermented broth (50 mL), containing 3.42 U/mL of CGTase, was added, and the flasks were shaken at 12 °C for 16 h. The plain matrices were also shaken with the fermented broth under similar conditions, to check for nonspecific binding. Similarly, fermented broth was also shaken alone to establish the stability of the enzyme under these conditions. The adsorbed R- and β-CGTase were

1118

Biomacromolecules, Vol. 2, No. 4, 2001

Nayak et al.

Table 1. Composition of Various Cross-Link Densities of HEMA-EGDM Matricesa HEMA sample no.

volume, mL

1 2 3 4 5 6 7 8

21.60 16.88 13.85 11.74 10.19 9.00 8.06 7.30

EGDM

mole

volume, mL

mole

HEMA-EGDM copolymer cross-link density (%)

0.178 0.139 0.114 0.097 0.084 0.074 0.066 0.060

8.40 13.12 16.15 18.26 19.81 21.00 21.94 22.70

0.044 0.070 0.086 0.097 0.105 0.111 0.116 0.120

25 50 75 100 125 150 175 200

a The continuous phase is composed of 1% PVP in water. The discontinuous phase consists of 30 mL of HEMA and EGDM with 0.6 g of AIBN. The polymerization was carried out at 70 °C.

eluted with 10 mM CaCl2 solution in 50 mM Tris-HCl buffer, pH 7.0, at 25 and 55 °C, respectively. Protein was assayed by the Lowry method.23 R- and β-CGTase were assayed by methyl orange21 and phenolphthalein methods,27 respectively. 2.11. Ligand Stability Measurement. To 1 g of affinity matrix, 10 mL portions of various regenerating solutions were added and shaken for 2 h. The regenerating solutions were decanted, and the treated affinity matrices were washed 5 times with 10 mL of distilled water till free from regenerating solutions. The matrices were dried under vacuum at 60 °C and then analyzed for R/β-CD content, as described in section 2.5. 3. Results and Discussion 3.1. Synthesis of Polymeric Beads. Macroporous HEMAEGDM beaded copolymers were synthesized by suspension polymerization in a jacketed cylindrical polymerization reactor at 70 °C using AIBN as a free radical thermal initiator. The continuous phase comprised an aqueous solution of poly(vinyl pyrrolidone) (PVP). The discontinuous organic phase consisted of HEMA and EGDM (Table 1). A series of copolymers, differing in cross-link density (CLD), were synthesized. CLD is defined as the mole percent of cross-linking comonomer relative to the moles of reactive functional monomer. Since all copolymerizations were taken to 100% conversion, the average cross-link density in all copolymers is effectively the relative mole percentage of the cross-linking comonomer in the feed. The morphology of the pores of the HEMA-co-EGDM (CLD, 25 mol %) on the scanning electron microscope (SEM) is presented in Figure 1, which indicates that these beads are built up of aggregates of compact microspheres. 3.2. Development of Affinity Matrices. The present activation and coupling chemistry employs very facile bifunctional reagents such as diisocyanates. Isocyanates are carbonyl compounds with double bonds. The urethane bond formation starts with the reaction at ambient temperature between the carbonyl carbon of the isocyanates and the alcohol oxygen. Ozawa employed hexamethylene diisocyanate as a bridging reagent to stabilize lysine and enzyme proteins.28 However, this reagent has not been applied for the development of affinity matrices through urethane linkages. Furthermore, a detailed study is desirable to

Figure 1. Internal porous structure of HEG beads (CLD 25 mol %) on SEM.

establish the suitability of this highly reactive reagent as linking groups in affinity matrices. A number of diisocyanates were evaluated for their ability to couple the ligands to the matrices. HEMA-EGDM copolymer beads (CLD 25%) were coupled in two sequential steps to R- and β-CDs through urethane linkages using different diisocyanates. The first activation step was optimized using different molar quantities of -NCO group (present in diisocyanates) with respect to a specific mole of -OH group in the matrix. This step was conducted in a stoppered 250 mL conical flask in a suitable aprotic solvent under stirring. The other parameters varied were temperature, addition of catalyst, and the nature of the gaseous overlay. After completion of the activation step, the matrices were washed with anhydrous 1,4-dioxane. The second step was the coupling of the ligand (CD) to the activated matrices. Cyclodextrin solution was prepared in anhydrous DMSO and added to the activated matrices. The coupling reaction was conducted at 200 rpm. The other parameters established were the effects of temperature and addition of catalyst20 (Scheme 1). Figure 2 shows the IR spectra of HEMA-EGDM copolymer, R-CD, and a typical affinity matrix comprising HEMAEGDM copolymer coupled to R-CD. The spectra show strong absorption peaks at 3500 (-OH), 2920 (-CH3), 1715 (-Cd O), and 1040 cm-1 for HEMA-EGDM copolymer, at 1027 cm-1 (-OH) for R-CD, and at 3300, 2950, 1700, and 1020 cm-1 for HEMA-EGDM copolymer coupled with R-CD through TDI. Strong peaks at 3300, 2950, and 1700 cm-1 confirm urethane bond formation in the derivatized copolymer. The typical spectra of R-CD are the strong peak concentrated around 1020 cm-1. The spectrum of R-CD coupled to HEMA-EGDM copolymer by 2,4-TDI has all the features of the spectra of HEMA-EGDM copolymer and R-CD. This indicates the formation of urethane linkages between HEMA-EGDM copolymer and R-CD. With HMDI in anhydrous 1,4-dioxane in a sealed conical flask under nitrogen overlay, a maximum of 21.6 µmol of R-CD could be bound to 1 g of matrix (Table 2). The maximum binding of R-CD with reactive IPDI, TDI, and

Biomacromolecules, Vol. 2, No. 4, 2001 1119

Macroporous Copolymer Matrix Scheme 1. Activation and Coupling of HEMA-EGDM Copolymer Beads (4:1)a

Table 2. Binding of Cyclodextrina to the HEMA-EGDM Matrixb with Diisocyanates diisocyanates HMDI IPDI PMDI TDI

µmol of R-CD bound/g of matrices 21.6 15.0 72.0 87.0

a In an inert atmosphere of nitrogen, the derivatization was carried out. The molar ratio of -CNO to -OH groups was kept at 1:1. b Mole percent of the cross-link monomer relative to the moles of the reactive comonomer of the polymeric matrix was 25%.

a Step 1: Activation of the matrix by HMDI, IPDI, PMDI, and TDI. Step 2: Coupling of R-CD (aCD) and/or β-CD (bCD) to the activated matrix. The -R group in the affinity matrix is represented as Ha/Hb (HMDI), Ta/ Tb (TDI), Pa/Pb (PMDI), and Ia/Ib (IPDI) coupled with R- and β-CD, respectively.

Figure 2. IR spectra of (a) HEMA-EGDM copolymer beads, (b) R-CD, and (c) HEMA-EGDM copolymer beads activated with 2,4TDI and coupled to R-CD.

PMDI spacers was 15.0, 72.0, and 87.0 µmol per g of matrix, respectively (Table 2). The maximum cyclodextrin binding was noted on copolymer with TDI, while the minimum was with copolymer coupled to IPDI (Table 2) as the spacer arm. From these observations, the reactivity of different diisocy-

anates in an inert atmosphere (nitrogen) is as follows: TDI > PMDI > HMDI > IPDI. Therefore, in our subsequent studies of binding of cyclodextrins to the HEMA-EGDM polymeric support the reactions were carried out with TDI as the spacer under a dry nitrogen atmosphere. 3.2.1. Effect of Cross-Link Density of the Polymeric Beads. As the maximum binding of R-CD was noted with TDI, copolymers of differing CLD were next evaluated for binding of R-CD using TDI as the spacer. With an increase in CLD, the pore surface area is known to increase while pore size distribution tends to broaden.29 In addition to this, we observed an increase in density of the polymeric beads from 1.21 to 1.26 g/mL with an increase in CLD, that is, an increase in mole fraction of cross-linking monomer, EGDM. The higher density of the polymeric beads is useful in expanded bed adsorption (EBA) of proteins from the fermentation broth. EBA is an increasingly popular unit operation for singlestep recovery of proteins from the feedstocks.30 CLD, the fraction of cross-linking monomer, was increased to synthesize a series of HEMA-EGDM copolymer beads. A maximum of 1495 µmol of R-CD was bound to the HEMAEGDM polymer of 200% cross-linking density, which is an increase of 17-fold over that on HEMA-EGDM copolymer with 25 mol % CLD (Table 4). Reactive porous synthetic copolymer beaded matrices ought to embody specific pore size, pore size distribution, pore volume and pore surface area, and optimum concentration of reactive groups to be effective for the covalent binding of the ligands. Pore size distribution should be definitive to anchor the larger ligands. The pore surface should comprise functional groups, which can react efficiently with the ligand (irrespective of the size of the ligand). It is observed that the pore surface area is not in an increasing trend with an increase in CLD as described by Kotha et al.29 This anomaly in pore surface area with an increase in CLD is probably due to the randomness in precipitation polymerization. However, the pore surface area has the same trend (Table 3) as the coupled R-CD to the matrices (Table 5). The pore volume was increased and a broader distribution of pores was observed with an increase in %CLD value. Furthermore, since HEMA contains one -OH group, whereas EGDM does not contain an -OH group, an increase in %CLD decreases the number of reactive functional groups (-OH) per gram of the macroporous copolymer. This indicates that a combined effect of the irregular profile of pore surface area, which is an important parameter of accessibility for coupling of the ligand to the reactive functional group of the copolymeric matrix, and the decreasing trend of -OH groups with an

1120

Biomacromolecules, Vol. 2, No. 4, 2001

Nayak et al.

Table 3. Effect of Cross-Link Densitya of HEMA-EGDM Matrices on Binding of R-Cyclodextrinb % CLD

µmol of R-CD bound/ g of matrices

% CLD

µmol of R-CD bound/ g of matrices

25 50 75 100

86 598 822 608

125 150 175 200

910 506 372 1495

a CLD is defined as the mole percent of cross-linking monomer relative to the moles of reactive functional comonomer. b TDI was used as the coupling reagent.

Table 4. Effect of Cross-Link Density of HEMA-EGDM Matrices on Swelling in DMSO and Bound Diisocyanates mol % CLD

swelling (mL/g)a

bound diisocyanates (µmol/g)

25 50 75 100 125 150 175 200

4.10 4.05 4.00 3.94 3.90 3.80 3.70 3.70

95 605 855 615 920 485 408 1502

a Packed volume of the swollen matrix in DMSO per gram dry weight of the matrix.

increase in %CLD are not sufficient clues to explain the abnormality in cyclodextrin coupling to the copolymeric matrices (Table 3). The degree of swelling within the series of matrices (differing in CLD) was evaluated in DMSO to determine whether the permeation of CD into the beads causes the higher binding of R-CD. This means that increased swelling would lead to higher permeation of CD into the beads resulting in greater binding of CD. Since DMSO is the solvent used in the coupling reaction of cyclodextrins, the swelling study was conducted with this solvent. It was observed that the swelling of the matrices decreased with an increase in CLD of HEMA-EGDM copolymer matrices (Table 4). This indicates that the increasing binding of R-CD to higher cross-linked polymeric matrices is not due to swelling of the polymer matrix. It is well-known that the pores present in highly cross-linked polymers are permanent ones while those in less cross-linked polymers (25 mol % CLD) are due to solvent-induced swelling. Since both degree of swelling of the matrices in DMSO and the pore surface area data could not explain the abnormality in the trend of cyclodextrin binding with the matrices of varying CLD, the bound diisocyanates were analyzed. The bound diisocyanates are in the same trend as the trend of binding of cyclodextrin to the matrices of varying CLD (Table 4). Therefore, it was concluded that the surface area data have the dominating role rather than the concentration of -OH groups in coupling and the degree of swelling of β-CD of the HEMA-EGDM copolymeric matrices of varying %CLDs. 3.2.2. Effect of Catalyst. Dibutyltin dilaurate (1 wt %) was added as a catalyst to facilitate the rate of the reaction between -NCO groups of diisocyanates with -OH groups of the HEMA-EGDM copolymer.31 This reaction was carried out with HEMA-EGDM copolymer beads of 25 mol % CLD under a nitrogen atmosphere. Each isocyanate showed differing reactivities; in all the cases, coupling reactions (second reaction) were much slower than the

Figure 3. Coupling of β-CD to HEMA-EGDM copolymer beads: (9) Activation of the matrix and coupling of β-CD to the activated matrix were at 25 °C in the absence of catalyst. (b) Activation of the matrix and coupling of β-CD to the activated matrix were at 25 and 45 °C, respectively, in the absence of catalyst. (2) Activation of the matrix and coupling of β-CD to the activated matrix were at 25 °C in the presence of catalyst. (1) Activation of the matrix and coupling of β-CD to the activated matrix were at 25 and 45 °C, respectively, in the presence of catalyst. -NCO/-OH is the molar ratio of -NCO (of TDI) to -OH groups (polymeric matrix).

activation step (first reaction). This is evident from the activation energies and rate constants for the first and second reactions cited in the literature.32 Therefore, the catalyst was tested at two temperatures to see its effects. The matrices were activated at room temperature. The coupling step with β-CD was studied at ambient (25 °C) and at 45 °C. In a parallel control set, without catalyst, both reactions were conducted at 25 °C. In a third set, again devoid of catalyst, the first step was at 25 °C while the second step was carried out at 45 °C. The profile indicates (Figure 3) that a higher reaction temperature is preferable for the coupling step with CD, while the activation step between matrix and isocyanate should be performed at a lower temperature. The first (activation) step is slower in the absence of catalyst and at 25 °C, whereas the second (coupling) step is moderate when carried out either without catalyst at a higher temperature or at 25 °C in the presence of catalyst. However, the exact reason for the kinetics of binding is not clear, as the reactive group to one end of the isocyanate (first reaction) is a crosslinked polymer whereas the other end (second reaction) comprised cyclodextrin. The groups differ in structure and thus might show different reactivities. This kinetics shows a differing trend from that observed for various diisocyanates with n-butanol in toluene.32 At a higher temperature (45 °C), in the presence of catalyst, the coupling step does not increase the binding of β-CD with an increase in concentration of diisocyanates. However, at a lower temperature (25 °C) there was a greater increase in binding of β-CD with the diisocyanate concentration. 3.3. Nonspecific Interactions. Nonspecific interactions of proteins with solid supports, such as the affinity matrix, arise as a result of retention of the proteins by some general factors such as ion exchange, hydrophobic interactions, charge-

Biomacromolecules, Vol. 2, No. 4, 2001 1121

Macroporous Copolymer Matrix Table 5. Pore Characteristics with Varying CLD pore size distribution (vol %), radius in Å % CLD

3000

pore volume (cc/g)

surface area (m2/g)

25 50 75 100 125 150 175 200

58.8 50.6 68.6 48.5 14.0 15.8 17.4 10.0

32.4 30.6 14.0 21.6 14.3 8.8 8.0 8.0

1.7 6.9 4.2 12.5 12.1 15.5 10.0 6.5

1.1 2.8 0 5.4 10.4 29.3 19.0 4.0

1.1 2.6 0 4.8 18.5 21.3 36.4 7.6

1.1 3.5 8.0 3.2 18.6 5.8 5.7 11.8

1.7 3.0 2.9 2.0 8.6 2.7 2.4 17.0

1.3 0.0 1.0 2.0 3.4 0.8 1.3 27.1

0.8 0.0 1.3 0.0 0.1 0.0 0.0 7.9

0.1 0.13 0.61 0.55 0.92 0.63 0.53 1.87

1.7 9.1 93 85 162 145 127 283

transfer complexion of phenolic groups, and so forth. The effectiveness and selectivity of an affinity chromatography matrix is depressed by these nonspecific sorptions. Even a small percentage of nonspecific adsorption on the column destroys its usefulness. Here, it is essential to have a hydrophilic nonionic matrix with little or no capacity for nonspecific adsorption. Nonspecific adsorption should not occur in the derivatized matrix, though the unreacted matrix may provide considerable adsorption or hydrophobicity. The adsorption of BSA was tested on the affinity matrix at pH 7.0 and 12 °C in a shake flask. BSA adsorbed to the matrices was eluted with 6 M urea at room temperature. It was presumed that the adsorption of protein to the matrices was nonspecific, if the protein could be eluted with the desorbing buffer. The nonspecific interaction (ionic and hydrophobic) studies were tested on the 25 mol % CLD HEMA-EGDM copolymer beads coupled to R-CD using different diisocyanates. The results presented in Table 6 reveal low nonspecific ionic interaction for matrices with HMDI and PMDI spacers. The absence of nonspecific ionic interaction of HMDI and PMDI cannot be explained on the basis of resonance and inductive effects, which create charged pockets and bind BSA, since the system is structurally complex. However, nonspecific interaction of BSA increased with an increase in CLD of the polymeric matrix (Table 7). Therefore, the polymeric matrix of lower CLD (25 mol %) was used in the adsorption and desorption studies. This increase in nonspecific interaction may be explained on the basis of greater surface area and pore volume with increased CLD (Table 5). Hydrophobicity on the matrices denatures the protein and binds at higher salt concentration and temperature. The protein is desorbed by low concentrations of buffer at low temperatures. BSA was used as a model protein to study the adsorption to the matrices with 1% NaCl at 25 °C. BSA was desorbed at 4 °C and with 0.05 M Tris-HCl buffer, pH 7.0. The result indicates that the very marginal hydrophobicity of the plain HEMA-EGDM matrices disappears on activation with HMDI, IPDI, and TDI (Table 6). With PMDI modification, the hydrophobicity of the matrix is depressed. Long-chain aliphatic compounds and aromatic rings have hydrophobic pockets. However, the combined effect of the polymer coupled to neutral cyclodextrin by a urethane bond on hydrophobicity is not known. The increase in hydrophobic interaction with cross-link density arises from the higher hydrophobicity of the cross-linking monomer (EGDM), which is more hydrophobic relative to HEMA (Table 7). Though the density of the matrix and the pore surface area

Table 6. Nonspecific (Ionic and Hydrophobic) Interactions of BSA with the Matrices diisocyanate bound matrices plain matrices HMDI IPDI PMDI TDI

ionica

hydrophobicb

nil 6.0 nil 5.0 nil

95.0 nil nil 30.0 nil

a Amount of BSA adsorbed per gram of dried matrix at 12 °C in 50 mM Tris-HCl buffer, pH 7.0. b Amount of BSA adsorbed per gram of dried matrix at 25 °C in the presence of 1% NaCl solution in 50 mM Tris-HCl buffer, pH 7.0.

Table 7. Nonspecific Interaction (Ionic and Hydrophobic) of Protein on HEMA-EGDM Matrices with Varying Cross-Link Densitya % CLD 25 50 75 100

ionicb nil 59 65 71

hydrophobicc

% CLD

ionicb

hydrophobicc

95 153 284 297

150 175 200

77 123 183

478 487 547

a CLD is defined as the mole percent of cross-linking monomer relative to the moles of reactive functional comonomer. b Amount of BSA adsorbed per gram of dried matrix at 12 °C in 50 mM Tris-HCl buffer, pH 7.0. c Amount of BSA adsorbed per gram of dried matrix at 25 °C in the presence of 1% NaCl solution in 50 mM Tris-HCl buffer, pH 7.0.

increase with an increase in cross-linker content, the higher nonspecific adsorption of protein to the matrix with increasing CLD of copolymer limits the use of copolymers with CLD in excess of 25 mol % in chromatographic operations. 3.4. Adsorption in Shake Flasks. The adsorptions of CGTases on the optimally synthesized affinity matrices were evaluated next. Since the HEMA-EGDM matrix showed the lowest nonspecific interaction at 25% CLD, affinity matrices prepared with different diisocyanate spacers and R/β-CD as ligand were taken for adsorption studies in shake flasks. One gram samples of affinity matrices prepared with different diisocyanates spacers and having 2 µmol of ligand/g of dried matrices were used in this study. The affinity matrices were shaken with 15 mL (6.95 U/mL) of β-CGTase and 10 mL (31.45 U/mL) of R-CGTase from the clarified fermented broth at 100 rpm and 12 °C for 16 h. The matrices were washed with 10 mL of 1 mM CaCl2 solution in 50 mM Tris-HCl buffer (pH 7.0). The adsorbed enzymes were desorbed with 5 mL of 10 mM CaCl2 solution in 50 mM Tris-HCl buffer shaken for 4 h at pH 7.0 at 25 and 55 °C for R-CGTase and β-CGTase, respectively. The data are presented in Table 8. An approximately identical amount of β-CGTase was observed to be adsorbed to the affinity HEMA-EGDM copolymer matrices coupled to β-CD modi-

1122

Biomacromolecules, Vol. 2, No. 4, 2001

Nayak et al.

Table 8. Adsorption and Elution of R- and β-CGTase on Affinity Matricesa matrices

U of CGTase adsorbed/ g of matrices

% eluted

HEG-HMDI-β-CDb HEG 25-IPDI-β-CDc HEG-PMDI-β-CDd HEG-PMDI-R-CDe HEG-TDI-β-CDf

70.00 69.00 76.00 65.50 69.00

89 91 79 92 95

a Affinity matrices consist of beaded copolymeric HEMA-EGDM coupled to R- and β-CDs by the respective diisocyanates. The ligand concentrations in each of the shake flask experiments were kept at 2 µmol of R/β-CD per g of dry matrix. b Beaded copolymeric HEMA-EGDM (CLD 25%) coupled to β-CD by HMDI. c Beaded copolymeric HEMA-EGDM (CLD 25%) coupled to β-CD by IPDI. d Beaded copolymeric HEMAEGDM (CLD 25%) coupled to β-CD by PMDI. e Beaded copolymeric HEMA-EGDM (CLD 25%) coupled to R-CD by PMDI. f Beaded copolymeric HEMA-EGDM (CLD 25%) coupled to β-CD by TDI.

fied with HMDI, IPDI, and TDI spacers. On the contrary, the affinity matrix with the PMDI spacer with β-CD adsorbed a higher amount of β-CGTase as compared to matrices with HMDI, PMDI, and TDI spacers and β-CD-based affinity matrices. Elution of the enzymes from the respective affinity matrices resulted in a lower recovery from the PMDImodified β-CD-based affinity matrix. The higher amount of adsorption and lower recovery of β-CGTase for the matrix modified with PMDI can be explained on the basis of hydrophobicity. The fermentation medium of β-CGTase from Bacillus firmus contained 1% Na2CO3, and pH was 8.9. The adsorption studies with BSA also confirm the increased adsorption at a higher salt concentration due to the hydrophobic interaction (Table 6). However, a nearly identical amount of β-CGTase was adsorbed to the matrices modified with HMDI, IPDI, and TDI whereas a differential behavior was noticed with the PMDI matrix, which showed an enhanced adsorption. Thus, the increased binding of β-CGTase relative to R-CGTase on HEMA-EGDM copolymer beads modified with PMDI is due to increased hydrophobic interactions, which were prominent at a higher salt concentration. This also results in a lower desorption of β-CGTase (Table 8). 3.5. Stability of the Coupled Ligand. The operational stability of the matrices requires stable ligands, coupling groups, and matrices. The affinity matrices were tested with various regenerating reagents employed in clean-in-place procedures. The regenerating reagents used in our experiments were 1 and 2 M NaCl, 0.5 and 1 M NaOH, 30% (v/v) 2-propanol/water, 6 M urea, 10% ethanol amine, sodium dodecyl sulfate (SDS), and water at 121 °C for 30 min. The matrices (CLD 25%) coupled with R-CD (2 µmol/g of dried matrices) by TDI were shaken with the regenerating reagents for 5 h in a shaker at 400 rpm. The regenerating solution was decanted, and the matrices were washed five times with 10 mL of water. The matrices were dried under a vacuum at 60 °C. The dried matrices were assayed for R-CD. Drastic conditions such as 0.5 and 1 N NaOH cleave the bond. Milder reagents (NaCl up to 2 M concentration, 30% 2-propanol, 6 M urea, 10% ethanolamine, and SDS) and higher temperatures (up to 121 °C) were found to be suitable for the clean-in-place procedure.

4. Conclusions The cyclodextrin-based beaded macroporous affinity matrices based on HEMA-EGDM using diisocyanates as spacers were optimized with various physical and chemical variables. IR spectra confirm the formation of urethane linkages with the CDs. 2,4-TDI was found to be the best coupling reagent because of higher coupling of the ligand and lower nonspecific and hydrophobic interactions. Adsorption of R- and β-CGTase and desorption from the affinity supports were studied in shake flasks. The affinity copolymer beads prepared with HMDI, IPDI, and TDI show nearly identical adsorption and desorption of β-CGTase. Higher binding and lower elution of β-CGTase were observed on HEMA-EGDM copolymer beads modified with PMDI. This procedure of coupling CD to the matrices with -OH groups can be extended to other matrices and ligands having -OH groups. The matrices could be regenerated with mild reagents and water at higher temperatures. Acknowledgment. A Senior Research Fellowship to Deba Prasad Nayak from the Council of Scientific Industrial Research, New Delhi, India, is duly acknowledged. Financial support from the Department of Biotechnology, Government of India, is also acknowledged. Our sincere acknowledgment to Dr. Mary McNamara, School of Chemistry, Dublin Institute of Technology, Dublin, Ireland, for various technical discussions. References and Notes (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15) (16) (17) (18) (19) (20) (21) (22)

Porath, J.; Fornstedt, N. J. Chromatogr. 1970, 51, 479. Sundberg, L.; Porath, J. J. Chromatogr. 1974, 90, 87. Axen, R.; Porath, J.; Earnback, S. Nature 1967, 214, 1302. Porath, J.; Aspberg, K.; Drevin, H.; Axen, R. J. Chromatogr. 1973, 86, 53. Porath, J.; Sundberg, L. Nature (London), New Biol. 1972, 238, 261. Wind, R. D.; Liebel, W.; Buitelaar, R. M.; Penninga, D.; Spreinat, A.; Dijkhuijen, L.; Bahl, H. Appl. EnViron. Microbiol. 1995, 61, 1527. Ahn, J.-H.; Hwang, J.-B.; Kim, S.-H. Korean J. Appl. Microbiol. Biotechnol. 1990, 18, 585. Ahn, J.-H.; Hwang, J.-B.; Kim, S.-H. Korean J. Appl. Microbiol. Biotechnol. 1991, 19, 113. Larsen, K. L.; Duedahl-Olesen, L.; Christensen, H. J. S.; Mathiesen, F.; Pedersen, L. H.; Zimmerman, W. Carbohydr. Res. 1998, 310, 211. Mori, S.; Hirose, S.; Oya, T.; Kitahata, S. Biosci. Biotechnol. Biochem. 1994, 58, 1968. La´szlo´, E.; Ba´nky, B.; Seres, G.; Szejtli, J. Starch/Staerke 1981, 33, 281. Villette, J. R.; Looten, P. J.; Bouquelet, S. J.-L. Chromatographia 1991, 32, 341. Kitahata, S.; Okada S. J. Jpn. Soc. Starch Sci. 1982, 29, 7. Makela, M. J.; Mattson P.; Schinina, M. E.; Korpela, T. K. Biotechnol. Appl. Biochem. 1988, 10, 414. Nakamura, N.; Horikoshi, K. Agric. Biol. Chem. 1976, 40, 935. Wind, R. D.; Liebel, W.; Buitelaar, R. M.; Penninga, D.; Spreinat, A.; Dijkhuizen, L.; Bahl, H. Appl. EnViron. Microbiol. 1995, 61, 1257. Nakamura, N.; Horikoshi, K. Agric. Biol. Chem. 1976, 40, 1785. Pongsawasdi, P.; Yagisawa, M. Agric. Biol. Chem. 1988, 52, 1099. Kotha, A.; Selvaraj, L.; Rajan, C. R.; Ponrathnam, S.; Kumar, K. K.; Ambekar, G. R.; Shewale, J. G. Appl. Biochem. Biotechnol. 1991, 30, 297. Nayak, D. P.; Rajan, C. R.; Patkar, A. Y.; Yemul, O. S.; Ponrathnam, S. A Indian Patent Application No. 1034/DEL/2000. McCaffery, E. L. In Laboratory Preparation for Macromolecular Chemistry; McGraw-Hill: New York, 1970; p 13. Lejuene, A.; Sakaguchi, K.; Imanaka, T. Anal. Biochem. 1989, 181, 6. Kaneko, T.; Kato, T.; Nakamura, N.; Horikoshi, K. J. Jpn. Soc. Starch Sci. 1987, 34, 45.

Macroporous Copolymer Matrix (23) Lowry, O. H.; Rosenbrough, N. J.; Farr, A. L.; Randall, R. J. J. Biol. Chem. 1951, 193, 265. (24) Gawande, B. N.; Patkar, A. Y. Indian Patent Application No. 1104/ DEL/98, 1998. (25) Gawande, B. N.; Singh, R. K.; Chauhan, A. K.; Goel, A.; Patkar, A. Y. Enzyme Microb. Technol. 1998, 22, 288. (26) Gawande, B. N.; Patkar, A. Y. Biotechnol. Bioeng. 1999, 64, 168. (27) Goel, A.; Nene, S. Starch 1995, 47, 399. (28) Ozawa, H. J. Biochem. (Tokyo) 1967, 62, 419.

Biomacromolecules, Vol. 2, No. 4, 2001 1123 (29) Kotha, A.; Raman, C. R.; Ponrathnam, S.; Shewale, J. React. Funct. Polym. 1996, 28, 227. (30) Chase, H. A. Trends Biotechnol. 1994, 12, 296. (31) Hostettler, F.; Cox, E. F. Ind. Eng. Chem. 1960, 52, 609. (32) Saunders, J. H.; Frisch, K. C. Polyurethanes: Chemistry and Technology Part I. Chemistry; Interscience Publishers: New York, 1962; p 157.

BM010059D