Isoelectric Separation of Proteins using Charged Ultrafilter

780

Ind. Eng. Chem. Res. 2010, 49, 780–789

Isoelectric Separation of Proteins using Charged Ultrafilter Membranes with Different Functionality under Coupled Driving Forces Arunima Saxena and Vinod K. Shahi* Electro-Membrane Processes DiVision, Central Salt & Marine Chemicals Research Institute, Council of Scientific and Industrial Research (CSIR), G. B. Marg, BhaVnagar-3640021, (Gujarat) INDIA

A simple membrane process for the protein fractionation under coupled driving forces (pressure and potential difference) has been developed using acidic functionalized (sulphonated, carboxyleted, and phosphorylated) ultrafilter membranes, based on the interpolymer of poly(vinyl chloride) (PVC) and styrene-divinyl benzene (DVB) copolymer. Introduction of the different functional groups was confirmed by Fourier transform infrared (FTIR), CHNS analysis, and ion-exchange capacity measurements. Molecular weight cut off (MWCO) determination of these membranes suggested their ultrafilter nature, while their contact angle values showed hydrophilic characteristics. The apparent pore radius of these membranes was estimated by water permeation studies, while electro-osmotic permeation data was used for the determination of zeta potential under the operating environment. Systematic studies on the effects of pH, or nature of the charge on the casein (CAS) and lysozyme (LYS), on their adsorption characteristic using these charged ultrafilter membranes were carried out. Protein transmission (selectivity) and membrane throughput across both membranes were studied using binary mixture of protein under different gradients at pH points: 2.0, 5.0, 10.7, and 13.0. It was concluded that separation from the binary mixture of CAS-LYS of LYS at pH 5.0 (pI of CAS) using charged ultrafilter membranes was possible with high selectivity and throughput. It was observed that transmission of protein can be governed by varying the nature and extent of charge on the protein (pH) and membrane matrix, polarity of applied potential gradient with an ultrafilter membrane of given pore dimensions. In these novel processes, charge on the protein, nature and extent of the charge on the membrane interfaces, and polarity of the potential gradient all are governing the transport of a given protein across the membrane, which resulted high selectivity and membrane throughput under coupled driving forces. 1. Introduction In chemical and biochemical sciences, molecular size or charge based separation of a component from the mixture is often required. An increase in complexity of chemical, biochemical, or biological systems have required more sophisticated, selective, and efficient technique to resolve/separate individual component. Recently, membrane-based processes gained importance in biotechnology due to their ability for size and/or charge based protein separation with high purity and throughput.1-7 Ultrafiltration (UF) of protein solution is gradually emerging as a powerful bioseparation process for diversified fields such as: biotechnology, biomedicine, and the dairy and food industry.8-12 Although protein concentration using UF has become a routine and successful operation in biotechnology, but fractionation of proteins using UF is still a technological challenge. Its effectiveness and efficiency are strongly dependent on operating parameters such as pH, salt concentration, permeate flux, and system hydrodynamics.13-21 Additionally, UF is a size based separation process, and it is difficult to achieve high selectivity and throughput. To solve these problems, charged ultrafilter membranes were used for protein separation and electrostatic interactions between charged proteins and membranes were also studied.19,22,23 Many processes with an electric gradient as the driving force such as: electrophoresis, gel electrophoresis, capillary electrophoresis, etc. were also developed for commercial practice of protein separation.24,25 In all these aforementioned processes, only one property of a given molecule either its molecular size or charge/isoelectric * To whom correspondence should be addressed. Tel.: +91-2782569445. Fax: +91-278-2567562 /2566970. E-mail: [email protected]; [email protected].

point was used to achieve separation under a single driving force (either a pressure or electrical gradient). All processes involve one selectivity parameter (either molecular size or charge) and one driving force, which results in less selectivity and membrane throughput. No report is available about the nature of the charge on the protein; its molecular size; nature, and extent on the membrane matrix were used to increase the selectivity of separation with high membrane throughput under coupled driving forces (pressure and potential gradient). In the literature, electro-ultrafiltration (UF under an electric field) is also reported to achieve fractionation/separation of a protein with enhanced throughput using uncharged UF membranes.26,27 Also, the amino acid transport through the membrane follows an electro-diffusional mechanism rather than a pure diffusional mechanism, and the effect of the electric potential gradient on the amino acid flux should be considered. There is clear experimental evidence that nature of charge plays a key role in the amino acid transport through hydrophilic charged pores.28,29 Earlier, researchers studied electrostatic interactions between charges on proteins and membranes and demonstrated that pH values and ionic strengths have profound effects on protein separation.28,30,31 No report is available in which the nature of charge on the protein and its molecular size and extent on the membrane matrix were used to increase the selectivity of separation with high membrane throughput under coupled driving forces (pressure and potential gradient). In this investigation, we prepared three types of acidic functionalized (sulphonated, carboxyleted, and phosphorylated) ultrafilter membranes, based on the interpolymer of poly(vinyl chloride) (PVC) and styrene-divinyl benzene (DVB) copolymer. Transmission of proteins was studied in different environments under simultaneous action of coupled driving forces. Because

10.1021/ie900258d  2010 American Chemical Society Published on Web 12/04/2009

Ind. Eng. Chem. Res., Vol. 49, No. 2, 2010

of three selectivity parameters (i.e., charge on the membrane matrix, nature of charge, and molecular size of the protein) and dual driving forces, high selectivity and maximum membrane throughput were achieved. These charged ultrafilter membranes were characterized to investigate their physicochemical and electrochemical properties and protein binding capacity. Casein and lysozyme were used as model proteins. 2. Experimental Section 2.1. Materials and Membrane Preparations. Acryl amide, phosphorous acid, and maleic anhydride were received from Sigma-Aldrich Chemicals. PVC, hexanone, formaldehyde, benzoyl peroxide (BP), styrene, DVB, sulphuric acid, sodium hydroxide, hydrochloric acid, sodium hydrogen phosphate, disodium hydrogen phosphate, etc. of AR grade were obtained from SD. Fine Chemicals, India, and used without any further purification. CAS (Mw: 20 000 Da) and LYS (Mw: 14 600 Da) were received from HiMedia Laboratories Pvt. Ltd. India. Double-distilled water was used in all experiments. Three types of acidic functionalized (sulphonated, carboxyleted, and phosphorylated) ultrafilter membranes, based on the interpolymer of PVC and styrene-DVB copolymer, were prepared. For the preparation of sulfonic acid functionalized ultrafilter membrane (MSO3H), 5 g PVC was dissolved in 70.0 mL of hexanone. To the clear solution, styrene (5.0 g) and 50% DVB (1 g) were added. The temperature of the solution was raised up to 70 °C, and then, initiator BP (0.02 g) was added. Afterward the temperature was maintained for 6 h. The obtained viscous solution was transformed into thin film on a cleaned glass plate. Partially dried film (2 h at 30 °C) was gelated by immersing in deionized water for 45 min at 5 °C. The film obtained thusly was immersed in 85% sulphuric acid for sulfonation. Sulfonated membrane was washed with distilled water and then successively conditioned in 1 M HCl and 1 M NaOH and stored. Carboxylic acid functionalized ultrafilter membrane (MCOOH) was prepared by dissolving PVC (5.0 g) in hexanone (70.0 mL). To a clear solution, maleic anhydride (5.0 g), styrene (4.0 g), and 50% DVB (1.0 g) were added at a temperature of 70 °C in the presence of BP (0.02 g). The temperature was maintained for 6 h. The resultant viscous solution was cast in the form of a thin film of desired thickness on a cleaned glass plate and partially dried for 2 h at ambient temperature (30 °C). Then, it was gelated in the deionized water for 45 min at 5 °C. The membrane obtained thusly was kept in 1 M NaOH to hydrolyze the anhydride group to a carboxylic acid group. The carboxylated membrane was washed with distilled water and then successively conditioned in 1 M HCl and 1 M NaOH and stored. For the preparation of phosphonic acid functionalized ultrafilter membrane (MPO3H2), five grams of PVC was dissolved in 70.0 mL of hexanone. To the clear solution, acryl amide (5.0 g), styrene (4 g), and 50% DVB (0.5 mL) were added at a temperature of 70 °C in the presence of BP (0.02 g). The solution was stirred for 6 h at a temperature of 70 °C and then for a further 6 h at room temperature. The resultant viscous solution was transformed into a thin film on a cleaned glass plate of desired thickness and partially dried for 2 h at ambient temperature (30 °C). Then dried film was gelated in the deionized water for 45 min at 5 °C. The membrane obtained thusly was treated with the mixture of 37% formaldehyde (100 mL) and phosphorous acid (37 g) at 80 °C for 8 h, to introduce the phosphonic acid group. Phosphorylated membrane was washed with distilled water and then successively conditioned in 1 M HCl and 1 M NaOH and stored. A schematic presentation

781

for the preparations of MSO3H, MCOOH, and MPO3H2 membranes and their structures is given in Figure 1. A membrane was also prepared in a similar manner as described above without acidification and hydrolysis as a pristine membrane and named as Mo. Thicknesses of all prepared membranes were kept constant (200 µm) for comparison purposes. 2.2. FTIR, TGA, DSC, DMA, and SEM Studies and Elemental Analysis. The Fourier transform infrared (FTIR) spectra of completely dried membranes were recorded using an attenuated total reflectence (ATR) technique with a spectrum GX series 49387 spectrometer in the range of 4000-600 cm-1. Thermal degradation behaviors of membranes were investigated using thermogravimetric analysis (TGA) (Mettler Toledo TGA/SDTA851e with stare software), under nitrogen atmosphere with heating rate of 10 °C/min from 50 to 600 °C. Differential scanning calorimetry (DSC) measurements were carried out using a Mettler Toledo DSC822e thermal analyzer with stare software. The mechanical strength of the membranes was analyzed by a Mettler Toledo dynamic mechanical analysis (DMA) 861c instrument with starc software under isothermal conditions. For scanning electron microscopy (SEM), gold sputter coatings were carried out on the dried membrane samples at pressures ranging in between 0.1 and 1 Pa. A sample was loaded in the machine, which was operated at 10-3-10-2 Pa with EHT 15.00 kV with 300 V collector bias using a Leo microscope, and SEM images were recorded. Elemental analysis (CHNS) of each membrane was carried out using Perkin-Elmer-2400, CHNS Analyzer. 2.3. Contact Angle, Ion-Exchange Capacity (IEC), Water Content (λ), and Molecular Weight Cut Off Measurements. For assessing the hydrophilic nature of membranes, the contact angle of membranes in water was measured with a tensiometer (DCAT 21, Dataphysics, Filderstadt, Germany) based on an energy balance approach to the three-phase equilibrium. The thickness of the wet membrane was determined by a digital micrometer (Mitutoyo, Japan, resolution 0.001 mm, six digit liquid crystal display (LCD)). IEC can be defined as the ratio between numbers of exchangeable groups (equivalents) and weight of the dry membrane. IEC was measured by a classical titration method as reported previously.32 Membrane samples with known dry weight were equilibrated in 1.0 M HCl for converting all charge sites into H+ form. The membranes were then thoroughly washed with double distilled water to remove the last trace of acid. Then, it was equilibrated in 0.1 M NaCl for 24 h and ionexchange capacity was determined from the increase in acidity or basicity by acid or base titration. For the measurement of water content (λ), the membrane samples (2 × 2 cm) were immersed in the distilled water for 24 h, their surfaces were wiped with filter paper, and then the wet membrane was weighed. This wet membrane dried at a fixed temperature of 60 °C until a constant weight, dry membrane, was obtained. The value of λ (%) was estimated by the following equation. λ(%) )

Wwet - Wdry × 100 Wdry

(1)

where Wwet is the weight of wet membrane sample and Wdry is the weight of dry membrane sample. MWCO of different membranes were determined by polyethylene glycol (PEG) rejection analysis in UF experiments by plotting the rejection against the molecular weight. The MWCO curve was obtained by plotting the rejection of each solute

782


Figure 1. Schematic presentation of membrane preparations and chemical structures for (A) MSO3H; (B) MCOOH; and (C) MPO3H2 membranes.

against the molecular weight. MWCO was considered as thr molecular weight of PEG at which a rejection of 90% is achieved. The rejection (Ri) of each test compound was determined by the following equation:

(

Ri(%) ) 1 -

)

(CP)i × 100 (CF)i

(2)

(CP)i and (CF)i are the concentration in the permeate and feed solution, respectively. The concentration of PEG in the feed and permeate was analyzed by following reported procedure.33 2.4. Membrane Conductivity and Counterion Transport Numbers. Membrane conductivity measurements were performed by potentiostatic, two-electrode mode with alternating current (ac). Both of the electrodes were not in direct contact with the membrane. Membrane resistance (Rm) was estimated by subtracting electrolyte solution resistance (Rsol) from total cell resistance (Rcell) [Rm ) Rcell - Rsol]. The membrane resistance was measured with the help of a digital conductivity

meter (century, model CC601). The process was repeated until reproducible values (within (0.01 mS) were obtained. Counterion transport numbers across the membranes (tim) were estimated from membrane potential (E m) measurements in equilibration with NaCl solutions of 0.01 and 0.10 M concentrations, according to a previously reported methodology34 using following equation: Em ) (2tm i - 1)

RT a1 ln nF a2

(3)

where R is the universal gas constant, F is the Faraday constant, T is the temperature, n is the electrovalence (1 in this case), and a1 and a2 are activities of the NaCl solutions used, respectively. 2.5. Electro-osmotic Permeability Measurements. The electro-osmotic permeability of different charged ultrafilter membranes were measured in a two-compartment membrane cell35 with an effective membrane area of 24.0 cm2, in


783

measurement, the membrane was soaked in distilled water for 2 h then equilibrated in the experimental solution. The pressure was varied in between 2 and 4 × 10-11 N cm-2, while the applied electrical gradient was varied between 0.67 and 2.00 V cm-1. To prevent gas evolution at the electrodes by electrolysis and heat generation, the ionic strength of the buffer solution and applied electrical gradient were kept relatively low. Also microfilters separated two electrodes and restricted the approach of proteins on the electrodes. Permeate volume was measured with respect to time at each applied pressure with an accuracy of 0.01 cm3. Permeate flux aroused in the ultrafiltration was described as volume permeated per unit area per unit time using the following equation: Jv )

Figure 2. Experimental cell for protein separation under coupled driving forces.

equilibration with 0.001 M NaCl solutions. Both compartments were kept in a state of constant agitation by means of a magnetic and mechanical stirrer. A known potential difference was imposed across the membrane with the help of an electronically operated power supply using Ag/AgCl electrodes. Volumetric flux was measured by observing movement of liquid in a horizontally fixed capillary tube of known radius. The current flowing through the system was also measured using a digital ammeter connected in series. Several measurements were performed in order to obtain reproducible values. 2.6. Protein Analysis and Equilibrium Adsorption Studies. Fresh solutions of CAS and LYS were separately prepared for each experiment in phosphate buffer (Na2HPO4 and NaH2PO4) of desired pH and composition. Sodium azide (0.1% w/v) was added to prevent microbial growth. Single and mixed protein concentration was determined by high performance liquid chromatography (HPLC) (Prominence, Shimadzu HPLC) using a BioSep-SEP-S2000 (Phenomenex) column with geometry of 300 mm × 7.80 mm. The mobile phase was a phosphate buffer (NaH2PO4/Na2HPO4; 5.0 mM, pH 7.0). The flow rate of the mobile phase was adjusted to 1.0 mL/min, and the oven temperature was set to 30.0 °C. The retention time for CAS was observed at 10.23 min while for LYS it was at 12.58 min. Determination of the proteins concentrations was achieved up to (0.001 mg/mL. CAS was employed for the adsorption study. Membranes were cleaned and conditioned with the deionized water before the experiment. Different membrane samples were cut into small pieces (25 cm2) and introduced in a small beaker containing a 25 mL CAS solution of different initial concentrations and pH values. These suspensions were kept for 6 h at 25 °C under shaking. The concentrations of the supernatants were determined by HPLC for estimation of the CAS binding capacity. 2.7. Ultrafiltration Experiments under the Applied Potential Gradient. Ultrafiltration (UF) experiments, with or without applied potential gradient, using desired solution of CAS or LYS and their mixture, were conducted in dead-end permeation modes using plate and frame model. The experimental cell, shown in Figure 2, was made of acrylic with compressed air as the driving force having volume of 200 cm3. The effective membrane area was 65 cm2. The experiments were conducted in batch mode of operation under stirred conditions using magnetic stirrer to minimize the effects of concentration polarization. There was provision for applying desired potential across the membrane by using two platinum electrodes using a regulated digital power supply (Aplab, India, model 1285). Effective pressure in the UF cell was measured by a pressure meter fitted on the top of cell with a safety valve. Before each

dV A dt

(4)

Where V is the volume (cm3), A is the effective membrane area (cm2), and t is the time (s). The concentration of CAS or LYS in the permeate side was determined by high performance liquid chromatography (HPLC) as described above. The observed protein transmission (ζobs) and selectivity parameter (SP) of the UF also can be defined as ζobs )

(CP)i (CF)i

(5)

SP )

(ζobs)i (ζobs)j

(6)

and

At least three protein samples were collected at regular intervals for subsequent analysis. After completion of the experiments, membrane was washed thoroughly with double distilled water and original membrane permeability was reproduced to ensure the absence of any adsorption or fouling. Denaturation or degradation of CAS and LYS was studied by recording the retention time in the chromatogram before and after the experiments, and no change was observed. No damage or denaturation of the proteins under electric field strengths was also reported in the literature.9,36 3. Results and Discussion 3.1. Membranes Preparation and Their Physicochemical and Electrochemical Properties. Interpolymer of PVC and styrene-DVB was prepared and conventional methods were used to introduce sulfonic, carboxylic, and phosphonic acid functionality in the polymer matrix. Introduction of different functional groups was confirmed by FTIR spectra for three membranes (MSO3H, MCOOH, and MPO3H2) are presented in Figure 3. For MPO3H2, peaks at 1300-1140 cm-1 was assigned to PdO stretching, while peaks around 3200, 2358, and 1040 cm-1 rose due to OH stretching. For MCOOH, peaks around 3200 and 2300 cm-1 give information about the acidic proton, while strong absorption at 1700 cm-1 confirms the CdO stretching of carboxylic acid. Strong characteristic peaks at 1030 and 1096 cm-1 were observed for MSO3H, due to symmetric and asymmetric stretching of sulfonate group. The presence of different functional groups were also supported by the elemental (CHNS) analysis data (Table 1). On the basis of the membrane forming materials and functional groups, chemical structures of different charged ultrafilter membranes are presented in Figure 1A-C. Thermal strengths for different membranes were illustrated by TGA analysis, representative curves for MSO3H, MCOOH,

784


Figure 3. FTIR spectra for MSO3H, MCOOH, and MPO3H2 membranes. Table 1. Physicochemical Properties for Different Membranes CHNS analysis membranes

Λ (%)a

IEC (mequiv/g of dry membrane)b

Xm (m mol/dm3)

C (%)

H (%)

N (%)

S (%)

MSO3H MCOOH MPO3H2

48.3 59.4 41.3

0.59 0.29 0.86

0.32 0.34 0.28

38.25 42.46 20.31

3.95 4.13 4.89

0.01 0.01 0.01

2.95 0.01 0.01

a

0.1% uncertainty. b 0.01 mequiv/g of dry membrane uncertainty.

MPO3H2, and Mo membranes under flowing nitrogen are presented as Figure S1 in the Supporting Information. Membranes (MSO3H, MCOOH, MPO3H2 membranes) showed twostage thermal degradation (loss in bound water and polymer) curve. The first weight loss below 100 °C was attributed to the loss of absorbed water in the membrane matrix. The second weight loss started at about 350 °C for MPO3H2, while it started at 250 °C for the MSO3H and MCOOH membranes. The second weight loss was attributed to defunctionalization of the membrane, and it was delayed for the MPO3H2 membrane because of the comparatively thermal stable nature of -PO3H2 groups. All membranes attained complete weight losses in the final stage at around 400 °C. The DSC analysis of these membranes was carried out under nitrogen with a 5 °C/min heating rate and is presented as Figure S2 in the Supporting Information. Membranes exhibited Tg values around 150, 120, and 117 °C for MSO3H, MCOOH, and MPO3H2, respectively. Tg values for MCOOH and MPO3H2 membranes are comparable, while it was relatively higher for MSO3H membrane. Difference in Tg values may be attributed to nature of functional groups. Furthermore Mo also showed similar thermal strength. DMA curves for charged ultrafilter membranes, at 50 N applied force and 1 Hz frequency at room temperature, are presented as Figure S3 in the Supporting Information. All membranes exhibited good mechanical stability without breaking of the polymer film under testing conditions. Charged ultrafilter membranes with different functionality were also characterized by measuring their physicochemical properties such as water content, ion exchange capacity, contact angle, and MWCO. The water content (λ) for different types of membranes is summarized in Table 1, which follows the trend: MCOOH > MSO3H > MPO3H2. Water absorption involves two types of water molecule population: (i) those “bound” to the hydrophilic groups and (ii) and those physically absorbed as “free water”. Water content mainly represents the latter type (in this case). Membranes with same

Table 2. Contact Angle, Membrane Conductivity, Counterion Transport Number, Membrane Porosity, and Surface Charge Concentration for Different Membranes (200 µm Thickness) membranes

MWCO (kDa)

contact angle (deg)

κm (mS/cm)a

MSO3H MCOOH MPO3H2

25 28 25

59.31 80.70 81.61

2.89 2.68 3.01

tim

b

0.75 0.64 0.69

a Measured in the equilibration with 0.001 mol/dm3 NaCl solution. The uncertainty of measurement is 0.01 mS/cm. b Estimated from membrane potential measurements using NaCl solutions of 0.01/0.001 mol/dm3 concentration across the membrane. The uncertainty of measurement is 0.01 mV.

degree of cross-linking or void volume absorb the same amount of water if the density of ionizable groups or hydrophilic nature is same.37,38 Thus, water content depends on (i) hydrophilic nature of the membrane matrix and (ii) void volume available for water sorption in the membrane matrix. Contact angle values (Table 2) revealed the comparatively more hydrophilic nature of the MSO3H membrane, while MCOOH membrane exhibited the highest (λ) value. It seems that factor ii predominates and water content depends on void volume in the membrane matrix. The swelling nature of developed membranes was recorded in an acidic and basic medium (pH: 2.0-13.0), and it was found to be negligible. Furthermore, we have an idea about the density of the exchangeable functional groups from their IEC values presented in Table 1. The IEC values for MSO3H were higher than those for MCOOH, because of the high dissociation of the sulfonic acid group at neutral pH in comparison to a carboxylic acid group. The IEC value of MPO3H2 was very high due to the bifunctional nature of phosphonic acid groups, which indicates that phosphonic acid groups were completely introduced in the membrane matrix and no side reactions such as cross-linking of the functional groups had occurred. Furthermore, λ values in conjunction with IEC values can also be used for the determination of fixed ion concentration


785

Figure 4. SEM images for different membranes.

(Xm) of the membrane in units of moles of sites per unit volume of wet membrane by the following equation Xm ) τ(IEC)Fd /∆V

(7)

Where, Fd is the density of dry membrane. The void porosity (τ, volume of free water within membrane per unit volume of wet membrane) was obtained by35,39

Figure 5. Binding capacity for (A) CAS and (B) LYS solutions (5 mg/mL each separately) with (1) MPO3H2; (2) MCOOH; (3) MSO3H; and (4) Mo membranes at different solution pH for 6 h adsorption.

∆V (8) 1 + ∆V ∆V denotes the volume increase of the membrane after absorption of the water per unit of dry membrane volume, which may be estimated using following equation τ)

∆V )

(Wh - Wd)Fd FwWd

(9)

Fw is the density of water. Xm values for different membranes follow the trend: MCOOH > MSO3H > MPO3H2 (Table 1). Contact angle values (Table 2) also revealed hydrophilic nature of these membranes. SEM images of these membranes (Figure 4) show smooth and homogeneous surfaces without any phase separation. The MWCO of these membranes by PEG rejection analysis was found to be between 28 and 25 kDa (Table 2). This information demonstrates the mildly charged, hydrophilic nature of membranes with ultrafilter characteristics. The mildly charged nature of these ultrafilter membranes was also revealed by their electrochemical properties such as membrane conductivity (κ m) and counterion transport number in the membrane phase(tim), values presented in Table 2. The reasonably good conductivity (2.68-3.01 mS/cm) of these membranes indicates their suitability for electrochemical devices. Also, higher (tim) values than in the solution phase under similar experimental conditions (Na+ transport number is 0.392) revealed that cation’s selective nature due the negatively charged acidic functional groups on the membrane matrix. On the basis of these characterizations, we can conclude the cation selective charged nature of developed ultrafilter membranes. 3.2. Protein Adsorption on Charged Ultrafilter Membranes. Charged ultrafilter membranes had higher dynamic, i.e., practical binding capacity for proteins, and exhibited a higher throughput for protein recovery. Introduction of a high density of functional groups in the membrane was required to maintain high selectivity for the protein adsorption/separation.40 Several interactions, namely hydrophobic interactions, hydrogen bonding, and electrostatic interactions simultaneously controlled membrane behavior for protein adsorption. Binding capacities of different membranes and protein (CAS and LYS) were determined in batch mode at different pH values (5 mg/mL) for 6 h adsorption, and relevant data are presented in Figure 5A and B. For each type of membrane, protein binding capacity was highly dependent on pH of the solution as well as nature of functional groups on the membrane matrix. pKa values for MSO3H, MCOOH, and MPO3H2 are 3.00, 4.76, and 2.00, respectively. It seems that

Figure 6. Variation of volumetric flux for (1) MSO3H; (2) MCOOH; (3) MPO3H2; and (4) Mo membranes under (A) applied current and (B) applied pressure gradient.

dissociation of functional groups and protein binding capacity depends on membrane nature and pKa values. At constant pH, CAS and LYS solutions exhibited protein binding capacity: MSO3H > MCOOH > MPO3H2. Thus protein binding capacity depends on the hydrophilic nature of the membrane. The MSO3H membrane was most hydrophilic in nature (Table 2) and showed comparatively high protein binding capacity, while the MPO3H2 membrane was comparatively less hydrophilic in nature. Further, the binding capacity of CAS was almost constant in lower pH and decreased with an increase in the pH. At pH < 5 (pI of CAS), CAS carried positive charges (CAS+) and its adsorption on a negatively charged matrix was favored due to mutual electrostatic attraction. Meanwhile, LYS carried a positive charge at pH < 10.7 (pI of LYS) and adsorption of LYS+ on the negatively charged ultrafilter was favored. At a higher pH, both proteins carried a negative charge and their binding capacity on the negatively charged interface was very low due to electrostatic repulsive forces. Furthermore, MSO3H and MCOOH membranes showed higher protein binding capacity in comparison with MPO3H2, while the Mo membrane showed very low binding capacity under similar experimental conditions. Thus, the pH of the protein solution is an important parameter for achieving high adsorption and transport/separation across charged ultrafilter membrane. 3.3. Membrane Permeation Studies. The impact of the membrane forming materials, nature of functional groups, and optimum membrane casting and gelation conditions was investigated by permeation characteristics of charged ultrafilter membranes. Volumetric flux (Jv) data under applied pressure and electric gradient are presented in Figure 6A and B. In all cases, straight lines were obtained. Hydrodynamic and electro-

786


Table 3. Hydrodynamic Permeability (Lp), Equivalent Pore Radius (rp), Electro-osmotic Permeability (Jv/I), and Membrane Zeta Potential (ξ) Values for Different Membranes membrane

L11 × 10-6 (cm3/(N s))

rpa (nm)

L12 × 10-3 (cm/(A s))

ξb (mV)

MSO3H MCOOH MPO3H2

1.62 1.13 1.23

11.31 15.27 10.21

0.33 0.30 0.28

-5.22 -4.25 -4.63 from

Figure 7. Schematic presentation of (A) conventional UF and (B) UF under coupled driving forces using a charged membrane.

osmotic permeability may be defined by phenomenological coefficients L11 and L12, respectively, by following equations.

values for membranes were also presented in Table 3. Usually, measurement of Eeff at the membrane interface is difficult, thus it is convenient to use current (I) for estimating electro-osmotic flux (L12 ) Jv/I). In general, the zeta potential of the membranes is supposed to represent the charge characteristics of membranesolution interfacial zone due to the presence of charges at membrane surfaces. MSO3H showed relatively high ξ values (-5.22 mV) because of the presence of strong acidic functional groups. The mildly charged nature of the ultrafilter membranes is responsible for the observed zeta potential as revealed by Xm values (Table 1). Furthermore, the negative values of zeta potentials of these membranes suggest acidic functional groups and their capability for charge based selection/rejection. 3.4. Transmission of a Single Protein under Coupled Driving Forces. Protein transmissions across the charged ultrafilter membranes and their effect on the pH of the feed protein solution were studied under coupled driving forces (pressure and electric gradient) to investigate the effect of the nature of the charge on the protein molecule (CAS or LYS), membrane matrix, and direction of electric polarity applied. Schematic diagrams for conventional UF and UF under coupled driving forces (pressure and electric gradient) using a charged ultrafilter membrane are presented in Figure 7A and B, respectively. The polarity of the electric gradient (Figure 7B) was fixed in such a way that a positively charged protein (for example CAS+, pH < pI) was facilitated to migrate through negatively charged membrane from anode side compartment toward cathode side compartment. Under this mode of operation, three screening parameters, pore size, nature of the charge on the membrane, and electric gradient, enhance the protein transmission and membrane throughput will also be increased due to the electrophoretic migration of the solvent along with the protein. In this case, convective transport will be approaching zero and total volumetric flux obtained under the simultaneously active forces (pressure and electric gradient in this case) can be written as follows in the form of a phenomenological equation based on the nonequilibrium thermodynamic principle.8,38

a Estimated from water permeability data. b Estimated electro-osmotic data in equilibration with 0.001 mM NaCl solution.

( )

) L11

(10)

( )

) L12

(11)

Jv ∆P

∆E)0

and Jv ∆E

∆P)0

Where, ∆P and ∆E are applied pressure difference and potential gradient, respectively. Phenomenological coefficients (L11 and L12) were estimated from the slope of straight lines (Figure 6) and are presented Table 3. Among three types of functionalized membranes, MSO3H showed the highest permeabilities (hydrodynamic and electro-osmotic), which may be due to the highly charged nature and pore dimensions. Further, the L11 values were used with advantage for the estimation of apparent pore radius (rp) using the Hagen-Poiseuille equation: L11

rp2 ) 8η(δ/λ)

(12)

Here, δ denotes membrane thickness, and η, the coefficients of viscosity of permeate. Estimated rp values are presented in Table 3, for different membranes. In spite of the pressure driven permeability value (L11) for the MSO3H membrane, its relatively low apparent pore radius may be explained due to its low water content (λ) values in compare with MCOOH membrane. Interestingly, the Mo membrane showed relatively low Jv values (Figure 6) and thus lower L11 and L12 values. Results revealed relatively high permeability values of charged ultrafilter membranes compare with uncharged ultrafilter (Mo), prepared under similar conditions. The estimated apparent pore radius of these membranes also suggested the ultrafilter nature of these membranes as also observed by MWCO studies. The nature of the functional groups had a pronounced effect on membrane pore dimensions under similar membrane forming material and casting conditions. Also, the nature and extent of surface charge in the membrane matrix influenced their flux and separation properties. Knowledge of the electrokinetic properties such as zeta potential of a particular membrane provides useful information to asses its suitability for a specific separation process. Surface charge on synthetic membrane has significant effect on its separation properties41,42 and fouling tendency.43,44 The zeta potential (ξ) can be obtained from electro-osmotic permeability (L12) values presented in Table 3, using the Smoluchowski equation.8,45 L12 )

ε0εrξ ηκ

(13)

Where ε0 is the permittivity of vacuum, εr is the relative dielectric constant of the electrolyte, η is the viscosity of electrolyte, and κ is the conductivity of electrolyte. Zeta potential

(Jv)coup ) L11∆P + L12∆E

(14)

In the case of coupling of driving forces, (Jv)∆E)0 < (Jv)coup > (Jv)∆P)0. Figure 8 presents the volumetric flux of CAS solutions at different pH values (between 2.0 and 10.7) across the MSO3H membrane, as a representative case, under constant applied pressure (3 × 10-11 N/cm2) with or without an applied electric gradient. It can be seen that without any electric gradient (at 0 V/cm) and under constant pressure, flux values were lowest, which were progressively increased with the increase in electric gradient, especially at pH 2. Also at pH 5 (pI of CAS) in Jv with an increase in the applied electric gradient, was very low. While at pH 10.7 (CAS-; pH > pI), no clear alteration in Jv was observed with the increase in applied electric gradient. The filtrate flux is an important parameter of amelioration when evaluating the separation/fractionation process. After bioproduction, the targeted product should be quickly separated


787

Figure 10. Variation of observed protein transmission (ζobs) for (A) CAS solution and (B) LYS solution (5 mg/mL each) at different pH values under 3 × 10-11 N/cm2 applied pressure and 2.0 V/cm applied electric gradient for (1) MSO3H; (2) MCOOH; (3) MPO3H2; and (4) Mo. membranes. Figure 8. Variation of Jv at different pH values of CAS solution (5 mg/ mL) under 3 × 10-11 N/cm2 applied pressure and varied applied electric gradient for MSO3H membrane.

Figure 9. Variation of volumetric flux (Jv) for (A) CAS solution and (B) LYS solution (5 mg/mL each) at different pH under 3 × 10-11 N/cm2 applied pressure and 2.0 V/cm applied electric gradient for (1) MSO3H; (2) MCOOH; (3) MPO3H2; and (4) Mo membranes.

because enzyme activity reduced with time.16 Here, it can be seen that filtration velocity was accelerated by coupling of driving forces, pressure, and electric gradient, for both type of membranes. The pH of the protein solution has a marked effect on the permeation properties of protein during UF.7,46,47 Volumetric flux data are also presented, as a function of pH of the feed solution, in Figure 9, for charged ultrafilter membranes under 3 × 10-11 N/cm2 applied pressure and 2.0 V/cm applied potential gradient. Filtrate flux was highly dependent on pH of the feed CAS or LYS solution. All charged ultrafilter membranes showed high flux at pKa values of their functional groups. Interestingly, both protein showed highest flux below their pI, while at pH > pI, fluxes were almost constant. All three membranes showed similar type behavior, while the MSO3H membrane showed the comparatively highest filtrate flux, in spite of its low MWCO and apparent pore radius. These observations also verify the coupling of both the driving forces. This may be because of its comparatively high charge concentration (Xm) values. Also alteration in the filtrate flux under different experimental conditions may be attributed to the combined electrophoretic and electroosmotic effects. For the protein separation/fractionation, the transmission or selectivity of the targeted protein molecules is also an important parameter in addition to the throughput. The observed protein transmission (ζobs) was estimated from the ratio of protein concentration in the feed side to the concentrate side (eq 5). Representative curves for CAS or LYS transmission (ζobs) against the pH of feed solution under 3 × 10-11 N/cm2 applied

pressure and 2.0 V/cm applied potential gradient for three membranes are presented in Figure 10A and B. It was observed that under suitable experimental conditions, coupled driving forces not only influence the filtrate flux but also enhance the protein transmission for charged ultrafilter membrane with different functionalities. Similar to the filtrate flux values, the ζobs value for CAS+ or LYS+ (pH below their respective pI values) was very high, which reduced significantly their pI (CAS or LYS) or above due to alteration in the charged nature of the protein with respect to pH. Among the three types of membranes, the MSO3H membrane showed a comparatively high ζobs value, similar to filtrate flux data. It was interesting to note that membranes showed high ζobs values at pKa values of their functional groups. Also, these observations revealed that, at a pI of CAS (pH: 5.0), a relatively high difference in ζobs value for both proteins can be used with advantages for their separation in the mixture. This difference in their transmission was aroused due to their different charged nature. At pH 5, CAS was in zwitterionic form, whereas LYS was LYS+ form. Thus, their transmission of proteins was highly dependent on their charged nature along with applied electric gradient. Thus in this study, three parameters such as applied pressure, electric gradient, and the nature of the charge on the membrane matrix govern the filtrate flux and protein transmission across the membranes with similar pore characteristics. Here, it is important to record that fouling was noticeably absent in all cases with less than a 1% reduction in the membrane permeability after repeated use, due to the applied electric gradient. 3.5. Protein Transmission and Selectivity in the Mixture of CAS and LYS. The selective separation of CAS and LYS from their equi-gram mixed solution was carried out at different pH values ranging between 2.0 and 13.0. Representative ζobs values for CAS and LYS separately, and for a mixed protein solution, are presented against pH in Figure 11. For each case, separation was achieved under 3 × 10-11 N/cm2 applied pressure and 2.0 V/cm1 applied electric gradient. For both proteins, BSA, and LYS, transmission was strongly dependent on the nature of the charge on the transmitting species (pH) and the membranes and the electric gradient. At pH 2.0, CAS and LYS both existed in positively charged species (BSA+ or LYS+) and thus showed very high ζobs values due to the negatively charged nature of membranes and applied electrical polarity. At pH 5.0, the transmission of CAS across charged ultrafilter membranes was relatively lower because of the uncharged nature of permeating species (pI 5.0). Transmission of LYS+ at pH 5.0, across the membranes, was high due to electrostatic attraction between migrating species and the membrane interface. Thus, at the isoelectric point of CAS, both proteins can be separated efficiently. Negatively charged membranes allowed high trans-

788


membranes follow the trend: MPO3H2 > MCOOH > MSO3H. Thus, it is possible to develop a rapid and highly selective protein separation method using an MPO3H2 membrane under coupled driving forces (pressure and electric gradient). In this process, no appreciable membrane fouling or protein denaturation was observed. By isoelectric focusing of one component, the other component in the charged state (depending on pH) may be effectively separated. In this case, the difference in the pI values of LYS and CAS was quite high. But in general cases where pI values were close, this process also can be efficiently used by varying the pore size of the membrane and pH of the permeating protein. Two screening parameters such as pI and molecular size were taken into account for achieving a high degree of selectivity. Figure 11. ζobs values for CAS (solid line) and LYS (dotted line) in equilibration with CAS-LYS mixed solution (1 mg/mL each) at different pH values under 3 × 10-11 N/cm2 applied pressure and 2.0 V/cm applied electric gradient for (1) MSO3H; (2) MCOOH; and (3) MPO3H2 membranes.

Figure 12. Separation factor (SF) as function of pH for LYS in its mixed solution with CAS (1 mg/mL each) under constant 3 × 10-11 N/cm2 applied pressure: (A) 2.0 V/cm applied potential for (1) MPO3H2, (2) MCOOH, and (3) MSO3H membranes; (B) for MPO3H2 membrane and varied applied potential (1) 1.0, (2) 2.0, and (3) 3.0 V/cm.

mission of BSA+ or LYS+, while the electrical gradient enhanced their flux in spite of similar pore dimensions. At pH 10.7, transmission of LYS0 (pI 10.7) and CAS- were reduced significantly. Similar trends were also observed at higher pH. Thus, selection of charged membranes with suitable pore dimensions and electrical polarity is very much necessary for electrostatic repulsion/attraction of permeating protein. Also protein transmission depends on charge concentration and nature of functional groups on the membrane matrix. Thus, it is possible to separate LYS from the mixture of CAS-LYS at pH 5.0 (pI of CAS) using negatively charged ultrafilter membranes with high selectivity. Also in all cases due to coupling of driving forces, filtrate flux and protein transmission were enhanced. A further idea about protein separation and selectivity can be obtained from the separation factor (SF), defined as the ratio of transmission of individual proteins from their mixture (eq 6). Estimated SF values for different charged ultrafilter membranes are presented in Figure 12A as a function of the pH of the mixed protein feed solution. SF values for the MPO3H2 membrane were higher than those for MSO3H and MCOOH membranes, in spite of its low filtrate flux. Also from the data, it is clearly evident that separation of CAS and LYS can be efficiently achieved by using a charged ultrafilter membrane. An increase in SF values with applied electric gradient (Figure 12B) suggested a highly selective separation of proteins under coupled driving forces. However, the applied electric gradient was so small that no denaturation of the protein is possible under given experimental conditions.9,36 SF values for different

4. Conclusions An interpolymer of PVC and styrene-DVB was prepared, and thin films of desired thickness were cast. Partially dried films were gelated by in the deionized water under optimized conditions for tailoring the pore structure of the membrane matrix. Different functional groups such as sulfonic, carboxylic, or phosphonic acid were introduced in the membrane matrix for the preparation of charged ultrafilter membrane. Introduction of functional groups was also identified by FTIR and ionexchange capacity studies. TGA, DSC, and DMA testing of these membranes revealed their thermal and mechanical stabilities. Electrochemical characteristics such as membrane conductivity, surface charge density, and counterion transport number of these membranes also revealed the charged nature of the membrane matrix. It was observed that the nature of the functional groups (hydrophilic nature) has a pronounced effect on membrane pore dimensions, in spite of same membrane forming material and casting conditions. These membranes were employed for the separation of proteins from their mixture under coupled driving forces. Coupling of forces were also confirmed by volumetric filtrate flux under different charged states of permeating protein. A single protein transmission study revealed that, at a pI of CAS (pH 5.0), the difference in ζobs values for both proteins can be used with the advantage for their separation in the mixture. It was concluded that separation of LYS from the mixture of CAS-LYS at pH 5.0 (pI of CAS) using negatively charged ultrafilter membranes, especially with MPO3H2 membrane, was possible with high selectivity. Also in all cases due to coupling of driving forces, filtrate flux, and protein transmission was enhanced. Furthermore, the applied electric gradient further progressively enhanced the SF values and suggested a highly selective separation of protein under coupled driving forces. In the present work, product yield defined as transmission through the membrane could be highly raised due to coupling of driving forces. Meanwhile, three screening parameters, charge on the protein; charge on the membrane matrix, and superimposed electric gradient, were responsible for high selectivity. Thus this is a novel process using a charged ultrafilter membrane for improving product purity and filtration velocity (throughput). Both the electrophoretic migration of molecules and the positive effect of electro-osmosis lead to an increase in product purity. Also, an additional driving force (electric field) could raise the filtration velocity even when the increase in applied pressure had reached to its limit. The reported process is a promising tool to make downstream processing faster and product purer in certain fields of application.


Acknowledgment The authors are extremely thankful to the Department of Atomic Energy, Government of India, for providing financial assistance by sanctioning project no. 2007/35/35/BRNS/102. We also acknowledge the services of the analytical science division, CSMCRI, Bhavnagar, for instrumental support. Supporting Information Available: Figures S1-S3. This information is available free of charge via the Internet at http:// pubs.acs.org. Literature Cited (1) Ulbricht, M. Advanced functional polymer membranes. Polymer 2006, 47, 2217. (2) Baker, R. W. Membrane Technology and Applications; Wiley: Chichester, 2004. (3) Blanch, H. W.; Clark, D. S. Biochemical Engineering; Marcel Dekker: New York, 1996. (4) Van Reis, R.; Zydney, A. L. Bioprocess membrane technology. J. Membr. Sci. 2007, 297, 16. (5) Nagarale, R. K.; Gohil, G. S.; Shahi, V. K. Recent developments on ion-exchange membranes and electro-membrane processes. AdV. Colloid Interface Sci. 2006, 119, 97. (6) Saxena, A.; Gohil, G. S.; Shahi, V. K. Electrochemical Membrane Reactor: single-step separation and ion substitution for the recovery of lactic acid from lactate salts. Ind. Eng. Chem. Res. 2007, 46, 1270. (7) Saxena, A.; Tripathi, B. P.; Kumar, M.; Shahi, V. K. Membranebased techniques for the separation and purification of proteins: An overview. AdV. Colloid Interface Sci. 2009, 145, 1. (8) Saxena, A.; Shahi, V. K. pH controlled selective transport of proteins through charged ultrafilter membranes under coupled driving forces: An efficient process for protein separation. J. Membr. Sci. 2007, 299, 211. (9) Kappler, T.; Posten, C. Fractionation of proteins with two-sided electro-ultrafiltration, J. Biotechnol. J. Biotechnol. 2007, 128, 895. (10) Ghosh, R.; Cui, Z. F. Fractionation of BSA and lysozyme using ultrafiltration: effect of pH and membrane pretreatment. J. Membr. Sci. 1998, 139, 17. (11) Feins, M.; Sirkar, K. K. Highly selective membranes in protein ultrafiltration. Biotechnol. Bioeng. 2004, 86, 603. (12) Maruyama, T.; Katoh, S.; Nakajima, M.; Nabetani, H. Mechanism of bovine serum albumin aggregation during ultrafiltration. Biotechnol. Bioeng. 2001, 75, 233. (13) Saksena, S.; Zydney, A. L. Effect of solution pH and ionic strength on the separation of albumin from immunoglobulins (IgG) by selective filtration. Biotechnol. Bioeng. 1994, 43, 960. (14) Wan, Y.; Ghosh, R.; Cui, Z. Separation of human serum albumin and human immunoglobulins using carrier phase ultrafiltration. Biotechnol. Prog. 2004, 20, 1103. (15) Muller, C. H.; Agarwal, G. P. Melin, Th.; Wintgens, Th. Study of ultrafiltration of a single and binary protein solution in a thin spiral channel module. J. Membr. Sci. 2003, 227, 51. (16) Shukla, R.; Balakrishnan, M.; Agarwal, G. P. Bovine serum albuminhemoglobin fractionation: significance of ultrafiltration system and feed solution characteristics. Bioseparation 2000, 9, 7. (17) Iritani, E.; Mukai, Y.; Kiyotomo, Y. Effects of electric field on dynamic behaviors of dead-end inclined and downward ultrafiltration of protein solutions. J. Membr. Sci. 2000, 164, 51. (18) Nystrom, M.; Aimar, P.; Luque, S.; Kulovaara, M.; Metsamuuronen, S. Fractionation of model proteins using their physiochemical properties. Colloids Surf. A: Physicochem. Eng. Aspects 1998, 138, 185. (19) Ghosh, R. Fractionation of biological macromolecules using carrier phase ultrafiltration. Biotechnol. Bioeng. 2001, 74, 1. (20) Ehsani, N.; Parkkinen, S.; Nystrom, M. Fractionation of natural and model egg-white protein solutions with modified and unmodified polysulfone UF membranes. J. Membr. Sci. 1997, 123, 105. (21) Fane, A. G.; Fell, C. J. D.; Suki, A. The effect of pH and ionic environment on the ultrafiltration of protein solutions with retentive membranes. J. Membr. Sci. 1983, 16, 195. (22) Burns, D. B.; Zydney, A. L. Buffer effects on the zeta potential of ultrafiltration membranes. J. Membr. Sci. 2000, 172, 39. (23) Cheang, B.; Zydney, A. L. Separation of R-lactalbumin and β-lactoglobulin using membrane ultrafiltration. Biotechnol. Bioeng. 2003, 83, 201.

789

(24) Carlsson, N.; Sanandaji, N.; Voinova, M.; Akerman, B. Bicontinuous Cubic Phase of Monoolein and Water as Medium for Electrophoresis of Both Membrane-Bound Probes and DNA. Langmuir 2006, 22, 4408. (25) Lilley, K. S.; Razzaq, A.; Dupree, P. Two-dimensional gel electrophoresis: recent advances in sample preparation, detection and quantitation. Curr. Opin. Chem. Biol. 2001, 6, 46. (26) Hulteen, J. C.; Jirage, K. B.; Martin, C. R. Introducing chemical transport selectivity into gold nanotubule membranes. J. Am. Chem. Soc. 1998, 120, 6603. (27) Chun, K. Y.; Stroeve, P. Protein transport in nanoporous membranes modified with self-assembled monolayers of functionalized thiols. Langmuir 2002, 18, 4653. (28) Ku, J. R.; Lai, S. M.; Ileri, N.; Ramır´ez, P.; Mafe, S.; Stroeve, P. pH and ionic strength effects on amino acid transport through Au-nanotube membranes charged with self-assembled monolayers. J. Phys. Chem. C 2007, 111, 2965. (29) Chun, K. Y.; Mafe, S.; Ramır´ez, P.; Stroeve, P. Protein transport through gold-coated, charged nanopores: Effects of applied voltage. Chem. Phy. Lett. 2006, 418, 561. (30) Pujar, N. S.; Zydney, A. L. Electrostatic effects on protein partitioning in size-exclusion chromatography and membrane ultrafiltration. J. Chromatogr., Sect. A 1998, 796, 229. (31) Ku, J. R.; Stroeve, P. Protein diffusion in charged nanotubes: “OnOff” behavior of molecular transport. Langmuir 2004, 20, 2030–2032. (32) Miyatake, K.; Chikashige, Y.; Higuchi, E.; Watanabe, M. Tuned polymer electrolyte membranes based on aromatic polyethers for fuel cell applications. J. Am. Chem. Soc. 2007, 129, 3879. (33) Sims, G.; Snape, T. A method for the estimation of polyethylene glycol in plasmaprotein fractions. Anal. Biochem. 1980, 107, 60. (34) Gohil, G. S.; Shahi, V. K.; Rangarajan, R. Comparative studies on electrochemical characterization of homogeneous and heterogeneous type of ion-exchange membranes. J. Membr. Sci. 2004, 240, 211. (35) Nagarale, R. K.; Gohil, G. S.; Shahi, V. K.; Rangarajan, R. Preparation of organic-inorganic composite anion-exchange membranes via aqueous dispersion polymerization and their characterization. J. Colloid Interface Sci. 2005, 287, 198. (36) Saxena, A.; Tripathi, B. P.; Shahi, V. K. An improved process for separation of proteins using modified chitosan-silica cross-linked charged ultrafilter membranes under coupled driving forces: Isoelectric separation of proteins. J. Colloid Interface Sci. 2008, 319, 252. (37) Nagarale, R. K.; Gohil, G. S.; Shahi, V. K.; Rangarajan, R. Organicinorganic hybrid membranes: Thermally stable cation-exchange membrane by sol-gel method. Macromolecules 2004, 37, 10023. (38) Koter, S.; Piotrowskia, P.; Kerres, J. Comparative investigations of ion-exchange membranes. J. Membr. Sci. 1999, 153, 83. (39) Capeci, S. W.; Pintauro, P. N.; Bennion, D. N. The molecular level interpretation of salt uptake and anion transport in Nafion membrane. J. Electrochem. Soc. 1989, 136, 10. (40) Kubota, N.; Miura, S.; Saito, K.; Sugita, Watanabe, K.; Sugo, T. Comparison of protein adsorption by anion-exchange interaction onto porous hollow-fiber membrane and gel bead-packed bed. J. Membr. Sci. 1996, 117, 135. (41) Wang, D. I. C.; Sinskey, A. J.; Sonoyama, T. Recovery of biological materials through ultrafiltration. Biotechnol. Bioeng. 1969, 11, 987. (42) Dukhin, S. S.; Derjaguin, B. W. Surface and Colloid Science; Matijevic, E., Ed.; Wiley: New York, 1974. (43) Larsson, K. Interfacial phenamena-bioadhesion and biocompatibility. Desalination 1980, 35, 105. (44) Breslau, B. R.; Testa, A. J.; Milnes, B. A.; Medjanis, G. Ultrafiltration Membranes and Applications; Cooper, A. R., Ed.; Plenum Press: New York, 1980. (45) Kim, K. J.; Fane, A. G.; Nystrom, M.; Pihlajamaki, A.; Bowen, W. R.; Mukhtar, H. Evaluation of electroosmosis and streaming potential for measurement of electric charges of polymeric membranes. J. Membr. Sci. 1996, 116, 149. (46) Reis, R. V.; Brake, J. M.; Charkoudian, J.; Burns, D. B.; Zydney, A. L. High performance tangential flow filtration using charged membranes. J. Membr. Sci. 1999, 159, 133. (47) Ricq, L.; Narcon, S.; Reggiani, J. C.; Pagetti, J. Streaming potential and protein transmission ultrafiltration of single proteins and proteins in mixture: lactoglobulin and lysozyme. J. Membr. Sci. 1999, 156, 81.

ReceiVed for reView February 17, 2009 ReVised manuscript receiVed October 3, 2009 Accepted November 16, 2009 IE900258D

Isoelectric Separation of Proteins using Charged Ultrafilter

Recommend Documents