Controlling protein transport in ultrafiltration using small charged ligands

Previous studies have demonstrated that protein transport during ultrafiltration can be strongly influenced by solution pH and ionic strength. The objective of this study was to examine the possibility of controlling protein transmission using a small, highly charged ligand that selectively binds to the protein of interest. Experiments were performed using bovine serum albumin and the dye Cibacron Blue. Protein sieving data were obtained with essentially neutral and negatively charged versions of a composite regenerated cellulose membrane to examine the effects of electrostatic interactions. The addition of only 1 g/L of Cibacron Blue to an 8 g/L BSA solution reduced the BSA sieving coefficient through the negatively-charged membrane by more than two orders of magnitude, with this effect being largely eliminated at high salt and with the neutral membrane. Protein sieving data were in good agreement with model calculations based on the partitioning of a charged sphere in a charged pore accounting for the change in net protein charge due to ligand binding and the increase in solution ionic strength due to the free ligand in solution.     

Use of protein charge ladders to study electrostatic interactions during protein ultrafiltration

Recent studies have demonstrated the importance of electrostatic interactions in membrane systems, but there is still controversy about the underlying phenomena. Protein charge ladders, consisting of a set of chemical derivatives of a given protein that differ by single charge groups, were used to quantify the electrostatic interactions during protein ultrafiltration. Myoglobin charge ladders were generated by acylation, with the different derivatives analyzed simultaneously by capillary electrophoresis. Filtration experiments were performed using polyethersulfone and composite regenerated cellulose membranes, with the membrane charge determined from the streaming potential. As expected, the rejection increased as the protein became more heavily charged due to the increase in electrostatic repulsion. However, the transmission of the weakly charged myoglobin species increased dramatically at very low ionic strength. This increase in transmission was attributed to a shift in pH within the pore caused by hydrogen ion partitioning into the charged membrane. The sieving data were in good agreement with theoretical calculations accounting for the effects of this pH shift on the electrostatic interactions.     

Effect of solution pH and ionic strength on the separation of albumin from immunoglobulins (IgG) by selective filtration

Although protein fractionation by selective membrane filtration has numerous potential applications in both the downstream processing of fermentation broths and the purification of plasma proteins, the selectivity for proteins with only moderately different molecular weights has generally been quite poor. We have obtained experimental data for the transport of bovine serum albumin (BSA) and immunoglobulins (IgG) through 100,000 and 300,000 molecular weight cutoff polyethersulfone membranes in a stirred ultrafiltration device at different solution pH and ionic strength. The selectivity was a complex function of the flux due to the simultaneous convective and diffusive solute transport through the membrane and the bulk mass transfer limitations in the stirred cell. Under phsioligical conditions (pH 7.0 and 0.15 M NaCI) the maximum selectivity for the BSA-IgG separation was only about 2.0 due primarily to the effects of protein adsorption. In contrast, BSA-IgG selectivities as high as 50 were obtained with the same membranes when the protein solution was at pH 4.8 and 0.0015 M NaCl. This enhanced selectivity was a direct result of the electrosatatic contributions to both bulk and membrane transport. The membrane selectivity could actually be reversed, with higher passage of the larger IgG molecules, by using a 300,000 molecular weight cutoff membrane at pH 7.4 and an ionic strength of 0.0015 M NaCl. These results clearly demonstrate that the effectiveness of selective protein filtration can be dramatically altered by appropriately controlling electrostatic interactions through changes in pH and/or ionic strength. (c) 1994 John Wiley & Sons, Inc.     

Retention of small charged impurities during ultrafiltration

Ultrafiltration is used to remove small impurities from a variety of processing streams. However, the clearance of small charged impurities may be inadequate due to electrostatic exclusion by the charged ultrafiltration membranes, an effect that has been largely unappreciated. Ultrafiltration experiments were performed to evaluate the transmission of several model impurities with different electrical charge through ultrafiltration membranes having different surface charge characteristics. Highly charged impurities are strongly rejected by charged cellulose and polyethersulfone membranes even though these solutes are much smaller than the membrane pore size. These effects could be eliminated by using high ionic strength solutions to shield the electrostatic interactions. The sieving data are in good agreement with model calculations based on the partitioning of charged spheres into charged cylindrical pores. Guidelines are developed for estimating conditions needed to obtain effective removal of small charged impurities through charged ultrafiltration membranes.     

Effect of solution pH on protein transport through ultrafiltration membranes

Although a number of previous studies have demonstrated that solution pH can have a dramatic effect on protein transport through ultrafiltration membranes, the exact origin of this behavior has been unclear. Experimental data were obtained for the transport of a broad range of proteins with different surface charge and molecular weight. The effective hydrodynamic size of the proteins was evaluated using size‐exclusion chromatography. The membrane charge, both before and after exposure to a given protein, was evaluated using streaming potential measurements. In most cases, the electrostatic interactions were dominated by the distortion of the electrical double layer surrounding the protein, leading to a distinct maximum in protein transmission at the protein isoelectric point. Attractive electrostatic interactions did occur when the protein and membrane had a large opposite charge, causing a second maximum in transmission at a pH between the isoelectric points of the protein and membrane. The sieving data were in good agreement with theoretical calculations based on available models for the partitioning of charged solutes in cylindrical pores. © 1999 John Wiley & Sons, Inc. Biotechnol Bioeng 64: 27–37, 1999.     

Effect of membrane charge on flow and protein transport during ultrafiltration

Although several recent studies have demonstrated the importance of electrostatic interactions in ultrafiltration, there have been few quantitative studies of the effects of membrane charge density on protein transport and membrane hydraulic permeability. Data were obtained using a series of charge-modified cellulose membranes, with the surface charge density controlled by varying the extent of addition of a quaternary amine functionality. The membrane charge was evaluated from streaming potential measurements. Protein transmission decreased by a factor of 100 as the membrane zeta potential increased from 0.3 to 6.6 mV. The protein sieving data were in good agreement with a partitioning model accounting for electrostatic effects, while the hydraulic permeability data were consistent with a flow model accounting for the effects of counter-electroosmosis. The results provide the first quantitative analysis of the effects of membrane charge density on the performance of ultrafiltration membranes.     

Fractionation of bovine serum albumin and monoclonal antibody alemtuzumab using carrier phase ultrafiltration

Protein transmission and hence selectivity of separation can be significantly affected by solution pH and ionic strength in protein fractionation using ultrafiltration. Using parameter scanning ultrafiltration, the transmission of bovine serum albumin (BSA) and monoclonal antibody alemtuzumab (Campath-1H) through 300 kDa polyethersulfone (PES) ultrafiltration membranes were studied over a range of pH and salt concentrations, with focus on the likely conditions for achieving "reverse selectivity," i.e., obtaining purified alemtuzumab (approximately 155 kDa) in the permeate. Experimental results demonstrate that the reverse selectivity could be obtained by manipulating the operating conditions such as the solution pH, ionic strength, permeate flux, and system hydrodynamics. With a two-stage batch ultrafiltration process under suitable conditions, the monoclonal antibody alemtuzumab with a purity of > 98% was obtained in the permeate from a feed solution initially containing 0.50 g/l each of BSA and alemtuzumab. Further purity can be expected by selecting more suitable membranes and optimizing operating conditions.     

Further studies on the contribution of electrostatic and hydrophobic interactions to protein adsorption on dye-ligand adsorbents.

The adsorption equilibria of bovine serum albumin (BSA), gamma-globulin, and lysozyme to three kinds of Cibacron blue 3GA (CB)-modified agarose gels, 6% agarose gel-coated steel heads (6AS), Sepharose CL-6B, and a home-made 4% agarose gel (4AB), were studied. We show that ionic strength has irregular effects on BSA adsorption to the CB-modified affinity gels by affecting the interactions between the negatively charged protein and CB as well as CB and the support matrix. At low salt concentrations, the increase in ionic strength decreases the electrostatic repulsion between negatively charged BSA and the negatively charged gel surfaces, thus resulting in the increase of BSA adsorption. This tendency depends on the pore size of the solid matrix, CB coupling density, and the net negative charges of proteins (or aqueous - phase pH value). Sepharose gel has larger average pore size, so the electrostatic repulsion-effected protein exclusion from the small gel pores is observed only for the affinity adsorbent with high CB coupling density (15.4 micromol/mL) at very low ionic strength (NaCl concentration below 0.05 M in 10 mM Tris-HCl buffer, pH 7.5). However, because CB-6AS and CB-4AB have a smaller pore size, the electrostatic exclusion effect can be found at NaCl concentrations of up to 0.2 M. The electrostatic exclusion effect is even found for CB-6AS with a CB density as low as 2.38 micromol/mL. Moreover, the electrostatic exclusion effect decreases with decreasing aqueous-phase pH due to the decrease of the net negative charges of the protein. For gamma-globulin and lysozyme with higher isoelectric points than BSA, the electrostatic exclusion effect is not observed. At higher ionic strength, protein adsorption to the CB-modified adsorbents decreases with increasing ionic strength. It is concluded that the hydrophobic interaction between CB molecules and the support matrix increases with increasing ionic strength, leading to the decrease of ligand density accessible to proteins, and then the decrease of protein adsorption. Thus, due to the hybrid effect of electrostatic and hydrophobic interactions, in most cases studied there exists a salt concentration to maximize BSA adsorption.     

Salt-induced changes in plasmid DNA transmission through ultrafiltration membranes

Recent studies have demonstrated the feasibility of using ultrafiltration for the purification of plasmid DNA, but there is still little understanding of the factors governing DNA transmission. Experimental data were obtained for the transmission of a 3.0 kbp supercoiled plasmid DNA through composite regenerated cellulose ultrafiltration membranes as a function of solution ionic environment in a stirred ultrafiltration cell. The dependence on salt concentration was quite dramatic, with the sieving coefficient increasing by more than 80-fold as the NaCl concentration increased from 1 to 150 mM at a fixed filtrate flux. At the same total ionic strength, the sieving coefficient in an MgCl2 solution was significantly larger than that evaluated in NaCl. The sieving results are consistent with independent studies showing a reduction in the effective plasmid size due to salt specific shielding of intramolecular electrostatic interactions. DNA transmission was also a strong function of the filtrate flux, with negligible transmission below a critical value of the flux. The predicted values of the critical filtrate flux determined using a modified elongational flow model were in excellent agreement with the experimental data. These results clearly demonstrate that salt-induced changes in plasmid DNA structure have a significant effect on plasmid DNA transmission through ultrafiltration membranes.     

Diffusion and partitioning of proteins in charged agarose gels.

The effects of electrostatic interactions on the diffusion and equilibrium partitioning of fluorescein-labeled proteins in charged gels were examined using fluorescence recovery after photobleaching and gel chromatography, respectively. Measurements were made with BSA, ovalbumin, and lactalbumin in SP-Sepharose (6% sulfated agarose), in phosphate buffers at pH 7 and ionic strengths ranging from 0.01 to 1.0 M. Diffusivities in individual gel beads (D) and in the adjacent bulk solution (D infinity) were determined from the spatial Fourier transform of the digitized two-dimensional fluorescence recovery images. Equilibrium partition coefficients (phi) were measured by recirculating protein solutions through a gel chromatography column until equilibrium was reached, and using a mass balance. Diffusion in the gel beads was hindered noticeably, with D/D infinity = 0.4-0.5 in each case. There were no effects of ionic strength on BSA diffusivities, but with the smaller proteins (ovalbumin and lactalbumin) D infinity increased slightly and D decreased at the lowest ionic strength. In contrast to the modest changes in diffusivity, there were marked effects of ionic strength on the partition coefficients of these proteins. We conclude that for diffusion of globular proteins through gel membranes of like charge, electrostatic effects on the effective diffusivity (Deff = phi D) are likely to result primarily from variations in phi with only small contributions from the intramembrane diffusivity.     

Separation of beta-casein peptides through UF inorganic membranes.

Ultrafiltration of peptide mixtures is studied under various operating conditions (transmembrane pressure, tangential flow-rate) using two ultrafiltration inorganic membranes M5 and M1 with molecular weight cut-offs, MWCO 10 and 70 kD, respectively. It is shown that the separation of peptides is controlled by a dual mechanism: size exclusion and electrostatic repulsion. When the ionic strength is high enough to screen out the electrostatic interactions, experimental data are in good agreement with a sieving model developed to estimate the intrinsic transmission from the molecular weight of a component and from the MWCO of the membranes. Although the transmission so found is altered by concentration polarisation and pore blocking mechanisms, the results explain the apparent low transmission of peptides by ultrafiltration membranes. If the ionic strength of the fluid is low, electrostatic interactions can influence the transport phenomena, provided that the molecules are highly charged (at pHs away from the pI). For attractive interactions, an apparent partition coefficient larger than 1 is observed. Otherwise, the transmission is lower than predicted by the sieving theoretical equation, as if the partition coefficient were smaller than 1.     

Intrinsic electric fields and proton diffusion in immobilized protein membranes. Effects of electrolytes and buffers.

We present a theory for proton diffusion through an immobilized protein membrane perfused with an electrolyte and a buffer. Using a Nernst-Planck equation for each species and assuming local charge neutrality, we obtain two coupled nonlinear diffusion equations with new diffusion coefficients dependent on the concentration of all species, the diffusion constants or mobilities of the buffers and salts, the pH-derivative of the titration curves of the mobile buffer and the immobilized protein, and the derivative with respect to ionic strength of the protein titration curve. Transient time scales are locally pH-dependent because of protonation-deprotonation reactions with the fixed protein and are ionic strength-dependent because salts provide charge carriers to shield internal electric fields. Intrinsic electric fields arise proportional to the gradient of an "effective" charge concentration. The field may reverse locally if buffer concentrations are large (greater to or equal to 0.1 M) and if the diffusivity of the electrolyte species is sufficiently small. The "ideal" electrolyte case (where each species has the same diffusivity) reduces to a simple form. We apply these theoretical considerations to membranes composed of papain and bovine serum albumin (BSA) and show that intrinsic electric fields greatly enhance the mobility of protons when the ionic strength of the salts is smaller than 0.1 M. These results are consistent with experiments where pH changes are observed to depend strongly on buffer, salt, and proton concentrations in baths adjacent to the membranes.     

Influence of the NaCl or CaCl2 concentration on the structure of heat-set bovine serum albumin gels at pH 7

The structure of heat-set systems of the globular protein bovine serum albumin (BSA) was investigated at pH 7 in different salt conditions (NaCl or CaCl(2)) using light scattering. Cross-correlation dynamic light scattering was used to correct for multiple scattering from turbid samples. After heat treatment, aggregates are formed whose size increases as the protein concentration increases. Beyond a critical concentration that decreases with increasing salt concentration, gels are formed. The heterogeneity and the reduced turbidity of the gels were found to increase with increasing salt concentration and to decrease with increasing protein concentration. The structure of the gels is determined by the strength of the repulsive electrostatic interactions between the aggregated proteins. The results obtained in NaCl are similar to those reported in previous studies for other globular proteins. CaCl(2) was found to be much more efficient in reducing electrostatic interactions than NaCl at the same ionic strength.     

Brownian dynamics simulations of probe and self-diffusion in concentrated protein and DNA solutions.

We have developed a Brownian dynamics algorithm for simulating probe and self-diffusion in concentrated solutions of DNA and protein. In these simulations, proteins are represented as spheres with radii given by their hydrodynamic radii, while DNA is modeled as a wormlike chain of hydrodynamically equivalent spherical frictional elements. The molecular interaction potentials employed by the program allow for intramolecular stretching and bending motions of the DNA chains, short-range Lennard-Jones interactions, and long-range electrostatic interactions. To test the program, we have carried out simulations of bovine serum albumin (BSA) probe diffusion and DNA self-diffusion in solutions of short-chain DNA as a function of both DNA concentration and solution ionic strength. In addition, we report on simulations of BSA self-diffusion as a function of BSA concentration and ionic strength. Based on a comparison to available experimental data, we find that our simulations accurately predict these transport properties under conditions of physiological salt concentration and predict the stronger concentration dependence observed at lower salt concentrations. These results are discussed in light of the nature of the intermolecular interactions in such systems and the approximations and limitations of the simulation algorithm.     

No salting-in of lysozyme chloride observed at low ionic strength over a large range of pH.

Solubility of lysozyme chloride was determined in the absence of added salt and in the presence of 0.05-1.2 M NaCl, starting from isoionic lysozyme, which was then brought to pH values from 9 to 3 by addition of HCl. The main observation is the absence of a salting-in region whatever the pH studied. This is explained by a predominant electrostatic screening of the positively charged protein and/or by adsorption of chloride ions by the protein. The solubility increases with the protein net charge at low ionic strength, but the reverse is observed at high ionic strength. The solubility of lysozyme chloride seems to become independent of ionic strength at pH approximately 9.5, which is interpreted as a shift of the isoionic pH (10.8) to an isoelectric pH due to chloride binding. The crystallization at very low ionic strength, where lysozyme crystallizes at supersaturation values as low as 1.1, amplifies the effect of pH on protein solubility. Understanding the effect of the net charge and of ionic strength on protein-protein interactions is valuable not only for protein crystal growth but more generally also for the formation of protein-protein or protein-ligand complexes.     

Studies on Box-Behnken design experiments: cellulose acetate-polyurethane ultrafiltration membranes for BSA separation

Ultrafiltration membranes were prepared from mixtures of cellulose acetate-polyurethane blend membranes. During the last 1 or 2 decades, the concentration purification and separation of Albumin by ultrafiltration through semipermeable membranes have been put into practice and hence membrane separation is considered as the unit operation. The blend solution was prepared from cellulose acetate and polyurethane in polar solvent in presence of polyvinylpyrrolidone as additive. The performance of modified blend membranes applied for Bovine Serum Albumin (BSA) separation by ultrafiltration technique using Box-Behnken design with three variables: additive, time and pressure. Three different levels was studied to identify a significant correlation between the effect of these variables on the amount of separation of BSA. The methodology identifies the principal experimental variables, which have the greatest effect on the separation process. The experimental values are in good agreement with predicted values, the correlation coefficient was found to be 0.9871.     

Membrane interactions in nerve myelin. I. Determination of surface charge from effects of pH and ionic strength on period.

We have used x-ray diffraction to study the interactions between myelin membranes in the sciatic nerve (PNS) and optic nerve (CNS) as a function of pH (2-10) and ionic strength (0-0.18). The period of myelin was found to change in a systematic manner with pH and ionic strength. PNS periods ranged from 165 to 250 A or more, while CNS periods ranged from 150 to 230 A. The native periods were observed only near physiological ionic strength at neutral or alkaline pH. The smallest periods were observed in the pH range 2.5-4 for PNS myelin and pH 2.5-5 for CNS myelin. The minimum period was also observed for PNS myelin after prolonged incubation in distilled water. At pH 4, within these acidic pH ranges, myelin period increased slightly with ionic strength; however, above these ranges, the period increased with pH and decreased with ionic strength. Electron density profiles calculated at different pH and ionic strength showed that the major structural alteration underlying the changes in period was in the width of the aqueous space at the extracellular apposition of membranes; the width of the cytoplasmic space was virtually constant. Assuming that the equilibrium myelin periods are determined by a balance of nonspecific forces/i.e., the electrostatic repulsion force and the van der Walls attractive force, as well as the short-range repulsion force (hydration force, or steric stabilization), then values in the period-dependency curve can be used to define the isoelectric pH and exclusion length of the membrane. The exclusion length, which is related to the minimum period at isoelectric pH, was used to calculate the electrostatic repulsion force given the other forces. The electrostatic repulsion was then used to calculate the surface potential, which in turn was used to calculate the surface charge density (at different pH and ionic strength). We found the negative surface charge increases with pH at constant ionic strength and with ionic strength at constant pH. We suggest that the former is due to deprotonation of the ionizable groups on the surface while the latter is due to ion binding. Interpretation of our data in terms of the chemical composition of myelin is given in the accompanying paper (Inouye and Kirschner, 1988). We also calculated the total potential energy functions for the different equilibrium periods and found that the energy minima became shallower and broader with increasing membrane separation. Finally, it was difficult to account directly for certain structural transitions from a balance of nonspecific forces.(ABSTRACT TRUNCATED AT 400 WORDS)     

Effect of electrostatic interactions on the ultrafiltration behavior of charged bacterial capsular polysaccharides

Charged polysaccharides are used in the food industry, as cosmetics, and as vaccines. The viscosity, thermodynamics, and hydrodynamic properties of these charged polysaccharides are known to be strongly dependent on the solution ionic strength because of both inter‐ and intramolecular electrostatic interactions. The goal of this work was to quantitatively describe the effect of these electrostatic interactions on the ultrafiltration behavior of several charged capsular polysaccharides obtained from Streptococcus pneumoniae and used in the production of Pneumococcus vaccines. Ultrafiltration data were obtained using various Biomax? polyethersulfone membranes with different nominal molecular weight cutoffs. Polysaccharide transmission decreased with decreasing ionic strength primarily because of the expansion of the charged polysaccharide associated with intramolecular electrostatic repulsion. Data were in good agreement with a simple theoretical model based on solute partitioning in porous membranes, with the effective size of the different polysaccharide serotypes evaluated using size exclusion chromatography at the same ionic conditions. These results provide fundamental insights and practical guidelines for exploiting the effects of electrostatic interactions during the ultrafiltration of charged polysaccharides. © 2016 American Institute of Chemical Engineers Biotechnol. Prog., 32:1531–1538, 2016     

The interaction of trichosanthin with supported phospholipid membranes studied by surface plasmon resonance

Trichosanthin (TCS) is a toxic protein isolated from a Chinese herbal medicine, the root tuber of Trichosanthes kirilowii Maximowicz of the Curcurbitaceae family. It is now used in China to terminate early and mid-trimester pregnancies. The ribosome inactivating property is thought to be account for its toxicity; it can inactivate the eukaryotic ribosome through its RNA N-glycosidase activity. The interactions of TCS with biological membrane is thought to be essential for its physiological effect, for it must get across the membrane before it can enter the cytoplasm and exert its RIP function. In the present work, the interaction of TCS with supported phospholipid monolayers is studied by surface plasmon resonance. The results show that electrostatic forces dominate the interaction between TCS and negatively charged phospholipid containing membranes under acid condition and that both the pH value and the ionic strength can influence its binding. It is proposed that, besides electrostatic forces, hydrophobic interaction may also be involved in the binding process.     

Separation of hemoglobin and myoglobin from yellowfin tuna red muscle by ultrafiltration: Effect of pH and ionic strength

Efficiency and selectivity of 30 and 150 kd inorganic ultrafiltration membranes (Techsep) toward tuna hemoglobin and myoglobin were studied. The influence of pH and ionic strength was investigated. Mass flow of myoglobin was higher at its isoelectric pH (8.6) and for low ionic strength (1.5 mM). This result was related to the absence of electrostatic repulsion between myoglobin and the surface of the dynamic membrane. The use of high ionic strength 0.15 M NaCl involved an apparent dimerisation of myoglobin and consequently a lower permeation through the membrane due to the molecular weight increase. The permeation and retention of hemoglobin did not agree with the effect of pH observed with myoglobin (best permeation at isoelectric pH) but followed the behavior of myoglobin. This was explained by a myoglobin concentration 10 times higher than hemoglobin concentration. The yield of retention selectivity was investigated. Selectivity of the membrane at pH 8.6 and 1.5 mM was favorable to myoglobin (increase of 40%) whereas a reversed selectivity was observed at pH 7.3, 0.15 M. (c) 1996 John Wiley & Sons, Inc.