Effect of ionic capacity on dynamic adsorption behavior of protein in ion-exchange electrochromatography

Separation and Purification Technology 68 (2009) 109–113

Contents lists available at ScienceDirect

Separation and Purification Technology journal homepage: www.elsevier.com/locate/seppur

Short communication

Effect of ionic capacity on dynamic adsorption behavior of protein in ion-exchange electrochromatography Wei Yuan, Yan Sun ∗ Department of Biochemical Engineering, School of Chemical Engineering and Technology, Tianjin University, Tianjin 300072, China

a r t i c l e

i n f o

Article history: Received 9 March 2009 Accepted 2 April 2009 Keywords: Electrochromatography Protein adsorption Breakthrough curve Pore diffusion Electroosmosis

a b s t r a c t Four anion-exchangers of different ion-exchange capacities (21–129 ␮mol/mL) were prepared by coupling diethylaminoethyl to Sepharose 6FF. The adsorbents were used to study the dynamic protein adsorption in ion-exchange electrochromatography with an oscillatory transverse electric field perpendicular to the mobile-phase flow (IEEC). The static adsorption capacity of bovine serum albumin (BSA) increased from 67 to 186 mg/mL in the ionic capacity range, but the effective pore diffusion coefficient of the protein decreased from 10.9 × 10−12 to 1.85 × 10−12 m2 /s with increasing the ionic capacity due to the hindrance effect of the bound protein molecules at the pore entrance. So, the dynamic binding capacity (DBC) of protein in ion-exchange chromatography decreased with increasing the ionic capacity. By applying an electric field of 30 mA, the DBC in IEEC packed with ion-exchangers of high ionic capacities (53 and 129 ␮mol/mL) increased significantly (over 30% and 100%, respectively). This was because the high surface charge density led to high electroosmotic flow that enhanced intraparticle mass transfer in IEEC. In comparison, the DBC in IEEC packed with ion-exchangers of low ionic capacities (21 and 35 ␮mol/mL) increased only slightly (ca. 10%) under the same condition. The results indicated the minor effect of electrophoretic mobility on the intraparticle mass transfer. Hence, it is beneficial to use ion-exchangers of high ionic capacity for high-capacity purification of proteins by IEEC. © 2009 Elsevier B.V. All rights reserved.

1. Introduction Preparative electrochromatography is a kind of separation technology with an external electric field applied to the chromatographic column [1–3]. In the presence of electric field, electrophoresis of charged solutes and electroosmotic flow (EOF) at a charge surface occur, so the separation performance can be enhanced with electrochromatography. For example, the resolution of proteins or nucleic acids in size exclusion electrochromatography is improved because the electrophoresis of charged solutes provides additional retention mechanisms [4–7]. In the case of adsorption chromatography with charged adsorbents, EOF occurs at the pore surface, which would, together with the electrophoresis of charged solutes, intensify intraparticle mass transfer, leading to the increase of dynamic binding capacity (DBC) [8–11]. However, it is still not clear to what extent the electroosmosis and electrophoresis contribute to the intraparticle mass transfer in addition to diffusive transport phenomena [9,10]. Hence, to investigate the relative importance of the two kinds of electric-kinetic phenomena, we have herein fabricated anion-exchange adsorbents with

∗ Corresponding author. Tel.: +86 22 27404981; fax: +86 22 27406590. E-mail address: [email protected] (Y. Sun). 1383-5866/$ – see front matter © 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.seppur.2009.04.002

four different coupling densities of ion-exchange groups (diethylaminoethyl). This is expected to create different magnitudes of intraparticle EOFs because the rate of EOF is proportional to surface charge density [12]. The research was carried out in preparative electrochromatography with an oscillatory transverse electric field perpendicular to the mobile-phase flow proposed by Sun and coworkers [6,10]. Protein adsorption isotherms and kinetics to the porous ion-exchangers were determined for discussion of electric field effect on the dynamic binding behavior of protein in the ionexchange electrochromatography (IEEC).

2. Materials and methods 2.1. Materials Tris-(hydroxymethyl) aminomethane (Tris) and Sepharose 6 Fast Flow were purchased from GE Healthcare (Uppsala, Sweden). The Sepharose gel has an average diameter of 90 ␮m and a mean pore size of about 35 nm [13]. Diethylaminoethyl chloride (DEAECl), glycine (Gly) and bovine serum albumin (BSA) were obtained from Sigma (St. Louis, MO, USA). Other reagents of analytical grade were all from local sources. Deionized water was used to prepare aqueous solutions.

110

W. Yuan, Y. Sun / Separation and Purification Technology 68 (2009) 109–113

2.2. Preparation of DEAE–Sepharose gels

pore diffusion model [15]

Four anion-exchange beads with deferent surface charge densities were prepared by coupling DEAE-Cl to Sepharose 6FF with the method described by Wang et al. [14]. Briefly, 10 g of Sepharose 6FF was suspended in 20 mL DEAE-Cl (0.6–4.0 mol/L), and the suspension was incubated at 60 ◦ C by shaking at 170 rpm. After 10 min, 20 mL of pre-heated NaOH solution (3.5 mol/L) at 60 ◦ C was added. The alkaline suspension was kept shaking for 60 min before it was cooled in an ice-water bath. The beads were collected on a G3 filter and washed thoroughly with deionized water to remove the residual reactants. In the coupling reaction, DEAE-Cl concentration was changed at the range of 0.6–4.0 mol/L to obtained different DEAE densities. The ion-exchangers were stored in 20% (v/v) ethanol solution. Each of the resins prepared above was packed in an HR 5/10 column, and the column was connected to ÄKTA Explorer 100 system (GE Healthcare). The total ionic capacity of the resins was measured by determination of adsorbed chloride ions [14]. In the measurements, the packed column was first equilibrated with 10 column volumes (CVs) of 1 mol/L NaCl, and then washed with deionized water until the conductivity signal decreased to the baseline. Then, 10 CVs of 0.5 mol/L sodium sulfate solution were applied to the column and the effluent solution was collected. In this process, Cl− ions were replaced by SO4 2− ions. The chloride ions in the eluate were determined by titration with 0.01 mol/L AgNO3 solution and the total ionic capacity was calculated by mass balance. A fresh sodium sulfate solution was used as a reference for the titration.

∂cp ∂q De ∂ + = 2 εp ∂t ∂t r ∂r



r

2 ∂cp



∂r

(2)

where De is the effective pore diffusivity, εp the effective porosity for BSA (0.55 [16]), cp the protein concentration in pores, t the time, and r the coordinate in radial direction. Neglecting the external mass transfer resistance, the mass transfer of protein from the bulk liquid phase to the solid phase in the finite batch adsorption system is expressed by



dcp  dc 3F = − De R dt dr 

(3) r=R

where c is protein concentration in the bulk liquid phase, R the particle radius, and F the volume ratio of solid to liquid phase. The initial condition (IC) and boundary condition (BC) for Eqs. (2) and (3) are given by IC : t = 0, q = 0, cp = 0, c = c0

(4a)

BC1 : r = R, cp = cb

(4b)

BC2 : r = 0,

∂cp =0 ∂r

(4c)

where c0 is the initial protein concentration in the bulk liquid phase (2.0 mg/mL). Matching the model with the c vs. t curves determined by the dynamic adsorption experiments led to the estimation of De . 2.4. Electrochromatography

2.3. Equilibrium and dynamic adsorption experiments Adsorption isotherms of BSA to the ionic-exchangers were determined by static batch adsorption experiments. All experiments were performed at 25 ◦ C in 3.9 mmol/L Tris-47 mmol/L Gly-5 mmol/L NaCl buffer (TG buffer), pH 8.3. The beads were first equilibrated with the buffer for 30 min and drained on a G3 sintered glass filter. Then, 0.1 g of the drained adsorbent was weighed into 25-mL flasks and 10 mL of the buffer solution containing BSA of 0.1–2.0 mg/mL was mixed with it. The sealed flasks were shaken at 25 ◦ C in an incubator for 24 h to achieve adsorption equilibrium. After centrifugation at 1000 rpm for 5 min, protein concentration in the supernatant was measured at 280 nm with a Lamda 35 UV–VIS spectrophotometer (PerkinElmer Instrument, USA). The bound amount of protein per milliliter of drained beads (q) was calculated by mass balance [14]. The Langmuir equation was used to fit the equilibrium data: q=

qm c Kd + c

(1)

where c and q are the free and adsorbed protein concentration in equilibrium, respectively, qm the adsorption capacity (mg/mL) and Kd the dissociation constant (mg/mL). The dynamic uptake of BSA to DEAE–Sepharose was carried out in a 250-mL three-neck flask equipped with a half-moon paddle for agitation. In the flask, 100 mL of 2.0 mg/mL BSA solution in the TG buffer was stirred at 200 rpm in an incubator of 25 ◦ C. The online protein concentration was determined by a UV monitor at 280 nm, with the circulation flow rate of 20 mL/min by a P-50 peristaltic pump (GE Healthcare) through a 2-␮m stainless filter. Then, 1.0 g drained adsorbent was added into the flask. The UV signal of the bulk liquid phase began to decrease due to protein adsorption into the adsorbent. The UV signal was recorded by the data acquisition software, and then converted into dynamic protein concentration vs. time curves. The dynamic data was analyzed with the effective

The experimental system for the IEEC was set up as described earlier [6,10]. In this work, however, the dimensions (length × width × depth) of the column central compartment were modified to 40 mm × 7 mm × 7 mm, corresponding to a packed-bed volume of 1.96 mL. As shown in Fig. 1a, electric field is applied across the direction of column width while cooled electrode solution (TG buffer, 8 ◦ C in inlet) flows upward in the neighboring electrode compartments. Each DEAE–Sepharose gel is packed into the central compartment by gravity sedimentation. The central compartment is separated from electrode compartments by dialysis membranes (molecular weight cutoff, 3–8 kDa) supported by porous ceramic plates of 5 mm thickness. These combinatorial isolators could not only prevent proteins from penetrating into the electrode compartments, but also prevent solvent osmosis between the central part and the electrode part. In the electrode compartments, two platinum plates are respectively mounted inside, and the large area avoids the electrode gases enwrapping the electrodes. The column was connected to ÄKTA Explorer 100 system controlled by Unicorn 4.11 for data acquisition and processing. An oscillatory direct current with an equal duration of positive and negative polarities (10 s each) was applied by a DDY-8C electrophoretic power supply (LiuYi Analytical Instrument, Beijing, China). In other words, the time of current cycle was 20 s, or the frequency was 1/20 Hz. The TG buffer was used as mobile phase. As shown in Fig. 1a, the pressure-driven flow is in the axial direction of the column, while the EOF on pore surface of the anion-exchanger is towards the anode, vertical to the mobile-phase flow. The electrophoretic migration of negatively charged protein (BSA) is also vertical to the mobile-phase flow, towards anode. So, the electro-kinetic flow in this case is composed of electroosmosis and electrophoresis in the same direction (Fig. 1b). 2.5. Dynamic protein adsorption in IEEC Frontal analysis in the IEEC was conducted to determine the breakthrough curves of BSA in the columns packed with

W. Yuan, Y. Sun / Separation and Purification Technology 68 (2009) 109–113

111

Fig. 2. Adsorption isotherms of BSA to DEAE–Sepharose of different ionic capacities. The solid lines are calculated from the Langmuir isotherms model.

The dynamic binding capacity (DBC) of BSA to DEAE–Sepharose in the IEEC was evaluated by the ratio of bound protein at 10% breakthrough (Q10 ) to the static adsorption amount (Qs ). Qs is the equilibrium adsorption amount calculated by Eq. (1) with the feeding protein concentration, while Q10 is obtained from breakthrough curves with the following equation: Q10 =

c0 (V10 − V0 ) (1 − εb )Vb

(5)

where c0 is feeding protein concentration (2 mg/mL), Vb is the bed volume; εb is the bed voidage, V10 is the effluent volume at 10% breakthrough, and V0 is the breakthrough volume under non-retained condition. V0 was determined with a packed bed of Sepharose 6FF gel. A bed voidage of 0.4 was assumed [17]. Fig. 1. Schematic view of the pressure-driven flow and the electro-kinetic transport phenomena in IEEC. (a) Structure and operation of the IEEC, redrawn after that in [10]. (b) Electrophoresis of negatively charged BSA and EOF in IEEC.

3. Results and discussion 3.1. Adsorption equilibrium and kinetics

DEAE–Sepharose of different coupling densities. BSA solution in the TG buffer at 2.0 mg/mL was used in the experiments. The flow rate was accurately controlled at 1.3 mL/min (superficial velocity, 160 cm/h). Prior to each BSA feeding, the column was equilibrated with the TG buffer until the absorbance at 280 nm reached the baseline. Then, an oscillatory electric field was applied in the direction perpendicular to the mobile-phase flow. The electric current was set at 20 or 30 mA, 1/20 Hz. A few minutes later, BSA solution was continuously loaded and protein concentration in outlet stream was detected with UV monitor at 280 nm. After the outlet BSA concentration exceeded 30% of the feed concentration, the sample loading was stopped and so was the electric current. Then, adsorbed protein was eluted with 20 CVs of the TG buffer containing 1.0 mol/L NaCl. Both of the equilibration and elution buffers were filtered through 0.45-␮m filter and degassed by sonication for 15 min prior to use.

We have produced DEAE–Sepharose gels of four different coupling densities, as listed in Table 1. The resins were then named with their DEAE densities (in ␮mol/mL), that is, DEAE-21, DEAE-35, DEAE-53 and DEAE-129 (Table 1). It is notable that the coupling densities are in the range of commercial ion-exchangers [18], and DEAE-129 has a similar ionic capacity with that of DEAE–Sepharose FF produced by GE Healthcare (110–160 ␮mol/mL). The adsorption isotherms of BSA to the anion-exchangers are shown in Fig. 2. The Langmuir equation (Eq. (1)) was used to fit the equilibrium adsorption data and the Langmuir parameters are given in Table 1. As can be seen from Fig. 2 and Table 1, the isotherms are nearly rectangular and the dissociation constants (Kd ) are very small, indicating the strong binding of BSA to the anion-exchanger. This is because of the very low ionic strength of the TG buffer used in the adsorption. Moreover, in the ionic capacity range, the adsorp-

Table 1 Equilibrium and kinetic parameters for protein adsorption to DEAE–Sepharose of different ionic capacities. DEAE–Sepharose

Ionic capacity (␮mol/mL)

DEAE-21 DEAE-35 DEAE-53 DEAE-130

21.0 34.8 52.6 129

± ± ± ±

3.0 5.0 2.0 1.2

qmax (mg/mL)

Kd (mg/mL)

De (10−12 m2 /s)

66.9 97.3 136 186

0.0015 0.0012 0.0010 0.0076

10.9 7.59 3.57 1.85

112

W. Yuan, Y. Sun / Separation and Purification Technology 68 (2009) 109–113

Fig. 3. Experimental and simulated dynamic uptake curves of BSA DEAE–Sepharose. The solid lines are calculated from the pore diffusion model.

to

tion capacity of BSA increased with increasing the density of the anion-exchange groups. Fig. 3 shows the dynamic adsorption data of BSA to the four different DEAE–Sepharose resins. Because the increase of ionic capacity led to the increase of protein adsorption amount, the dynamic adsorption time to approach equilibrium also increased significantly with the ionic capacity. The pore diffusion model described above was used to fit the dynamic adsorption data of BSA, as given by the solid lines in Fig. 3. The data of effective pore diffusion coefficient (De ) thus obtained are summarized in Table 1. As can be seen, the effective pore diffusion coefficients of BSA in the DEAE–Sepharose decreased distinctly with increasing the ionic capacity of the resins. It implies that the increasing amount of adsorbed protein on the adsorbent led to the decrease of protein diffusion rate inside the particles. These results are similar to Chen et al. [19], who reported that pore diffusivity of ␥-globulin into a porous anion-exchanger, DEAE–Spherodex M, decreased exponentially with increasing protein adsorption density. In the present work, the strong adsorption of BSA (Stokes radius = 3.59 nm [20]) at the pore entrance decreased the pore channel diameter of Sepharose gel (35 nm in average [13]), resulting in an increase of the hindrance effect of the solid phase on the diffusive mass transfer [21]. Therefore, although a resin of high ionic capacity can offer a high static adsorption capacity, it may not be kinetically favorable for protein adsorption in a dynamic process. 3.2. Dynamic protein adsorption in IEEC Fig. 4 shows the effect of electric field strength on the breakthrough curves of BSA in IEEC packed with different ion-exchangers. It is obvious that the breakthrough curves in DEAE-21 and DEAE35 were less affect by the electric field, while those in DEAE-53 and DEAE-130 were clearly delayed by increasing electric current.

Fig. 4. Breakthrough curves of BSA in chromatography of DEAE–Sepharose gels at different current densities.

This is also obvious by examining the dynamic adsorption capacity (Q10 ) and its ratio to the static adsorption amount (Q10 /Qs ), as listed in Table 2. Due to the unfavorable diffusive mass transfer of protein to the resins of high ionic capacities (Table 1), Q10 for the ion-exchange chromatography (IEC) decreased with increasing the ionic capacity. Upon the application of electric field, there was only slight increase (∼10%) of Q10 or Q10 /Qs for the resins of DEAE-21 Table 2 Effect of electric field strength on the DBC of the DEAE–Sepharose beads of different ionic capacities. DEAE–Sepharose

Qs a , b

Electric current density (mA) 0 Q10

DEAE-21 DEAE-35 DEAE-53 DEAE-130 a

66.82 97.25 136.22 184.91

20 a

44.2 49.1 38.6 30.4

30

Q10 /Qs

Q10

Q10 /Qs

Q10

Q10 /Qs

0.661 0.505 0.283 0.164

44.2 48.3 49.5 48.9

0.661 0.497 0.363 0.265

48.5 53.4 60.0 64.5

0.726 0.549 0.440 0.349

The unit for Qs and Q10 is mg/mL. Qs was calculated with Eq. (1) with c = 2 mg/mL and the Langmuir parameters provided in Table 1. b

W. Yuan, Y. Sun / Separation and Purification Technology 68 (2009) 109–113

and DEAE-35, but over 30% increase for DEAE-53 and over 100% increase for DEAE-130 at a current density of 30 mA. This implies that the enhancement effect of electric field on the DBC of protein is more effective for the high ionic-capacity resins. In other words, EOF played the most important role in this enhancement because the magnitude of EOF (vEOF ) is directly proportional to the surface charge density, as expressed by the following equation [12]:

vEOF =

−1 E 

(6)

where  is the surface charge density, −1 the thickness of electrical double layer,  the buffer viscosity and E the electric field strength. Thus, high EOF could occur in the high ionic-capacity resins. In comparison to EOF, the electrophoretic mobility of BSA in the pores contributed less to the enhancement in DBC. This is clear by examining the inappreciable effect of electric field for the low ionic-capacity resins (DEAE-21 and DEAE-35). This indicates that the electrophoretic rate of BSA was much lower than its diffusive transport in the ion-exchange resins. In contrast, the EOF rate was comparable or even higher than the diffusive mass transport in the high ionic-capacity resins, so the DBC increased significantly in IEEC with increasing the current density. 4. Conclusions Anion-exchangers of four different ionic capacities were prepared with Sepharose gel. The static protein capacities increased significantly with increasing the ionic capacity, but the effective pore diffusion coefficient decreased distinctly due to the hindrance effect of the bound protein molecules at the pore entrance. As a result, the DBC of protein in IEC decreased with increasing the ionic capacity. In the presence of an oscillatory transverse electric field, significant increase of the DBC was observed for the ion-exchangers with ionic capacities higher than 50 ␮mol/mL. This was because high surface charge density led to high surface zeta potential, and subsequently high EOF that enhanced intraparticle mass transfer in IEEC. In comparison, the DBC in IEEC packed with ion-exchangers of low ionic capacities was not significantly affected by electric field strength, indicating the minor effect of electrophoretic mobility of protein on the intraparticle mass transfer. Hence, ion-exchangers of high ionic capacity should be used in IEEC to achieve high DBC in protein purification. Acknowledgment This work was supported by the Natural Science Foundation of China (Nos. 20636040 and 20776105).

113

References [1] K.D. Cole, H. Cabezas Jr., Improved preparative electrochromatography column design, J. Chromatogr. A 760 (1997) 259–263. [2] S.K. Basak, M.R. Ladisch, Mechanistic description and experimental studies of electrochromatography of proteins, AIChE J. 41 (1995) 2499–2507. [3] S.R. Rudge, S.K. Basak, M.R. Ladisch, Solute retention in electrochromatography by electrically induced adsorption, AIChE J. 39 (1993) 797–808. [4] C.M. Tellez, K.D. Cole, Preparative electrochromatography of proteins in various types of porous media, Electrophoresis 21 (2000) 1001–1009. [5] K.D. Cole, Preparative concentration and size fractionation of DNA by porous media using a combination of flow and low electric field strength, Biotechnol. Prog. 13 (1991) 289–295. [6] G.M. Tan, Q.H. Shi, Y. Sun, Retention behavior of proteins in size-exclusion electrochromatography with a low-voltage electric field perpendicular to the liquid phase streamline, Electrophoresis 26 (2005) 3084–3093. [7] G.M. Tan, X.Y. Dong, Y. Sun, Oscillatory transverse electric field enhances protein resolution and capacity of size-exclusion chromatography, J. Sep. Sci. 29 (2006) 684–690. [8] Z. Liu, G. Yin, S.H. Feng, D.H. Wang, F.X. Ding, N.J. Yuan, Oscillatory electroosmosis-enhanced intra/inter-particle liquid transport and its primary applications in the preparative electrochromatography of proteins, J. Chromatogr. A 921 (2001) 93–98. [9] G.M. Tan, Q.H. Shi, Y. Sun, Oscillatory transverse electric field enhances mass transfer and protein capacity in ion-exchange electrochromatography, J. Chromatogr. A 1098 (2005) 131–137. [10] W. Yuan, G.F. Zhao, X.Y. Dong, Y. Sun, High-capacity purification of hen-white proteins by ion-exchange electrochromatography with an oscillatory transverse electric field, J. Sep. Sci. 29 (2006) 2383–2389. [11] G.-D. Jia, X.-Y. Dong, Y. Sun, Dye–ligand affinity electrochromatography with transverse and/or longitudinal electric field, Sep. Purif. Technol. 59 (2008) 277–285. [12] M.G. Cikalo, K.D. Bartle, P. Myers, Influence of the electrical double-layer on electroosmotic flow in capillary electrochromatography, J. Chromatogr. A 836 (1999) 35–51. [13] J.C. Bosma, J.A. Wesselingh, pH dependence of ion-exchange equilibrium of proteins, AIChE J. 44 (1998) 2399–2409. [14] D.M. Wang, G. Hao, Q.H. Shi, Y. Sun, Fabrication and characterization of superporous cellulose bead for high-speed protein chromatography, J. Chromatogr. A 1146 (2007) 32–40. [15] K. Yang, Q.H. Shi, Y. Sun, Modeling and simulation of protein uptake in cation exchanger visualized by confocal laser scanning microscopy, J. Chromatogr. A 1136 (2006) 19–28. [16] B.J. Horstmann, C.N. Kenney, H.A. Chase, Adsorption of proteins on Sepharose affinity adsorbents of varying particle size, J. Chromatogr. 361 (1986) 179–190. [17] P.M. Boyer, J.T. Hsu, Experimental studies of restricted protein diffusion in an agarose matrix, AIChE J. 38 (1992) 259–271. [18] E. Karlsson, L. Rydén, J. Brewer, Ion exchange chromatography, in: J.-C. Janson, L. Rydén (Eds.), Protein Purification, second ed., John Wiley & Sons, Inc., New York, 1998, pp. 145–206. [19] W.D. Chen, X.-Y. Dong, S. Bai, Y. Sun, Dependence of pore diffusivity of protein on adsorption density in anion-exchange adsorbent, Biochem. Eng. J. 14 (2003) 45–50. [20] S.P. Zhang, Y. Sun, Steric mass-action model for dye–ligand affinity adsorption of protein, J. Chromatogr. A 957 (2002) 89–97. [21] C. Harinarayan, J. Mueller, A. Ljunglof, R. Fahrner, J. van Alstine, R. van Reis, An exclusion mechanism in ion exchange chromatography, Biotechnol. Bioeng. 95 (2006) 775–787.