Effect of phenyl sepharose ligand density on protein monomer/aggregate purification and separation using hydrophobic interaction chromatography

In the large-scale manufacturing and purification of protein therapeutics, multiple chromatography adsorbent lots are often required due to limited absorbent batch sizes or during replacement at the end of the useful column lifetime. Variability in the adsorbent performance from lot to lot should be minimal in order to ensure that consistent product purity and product quality attributes are achieved when a different lot or lot mixture is implemented in the process. Vendors of chromatographic adsorbents will often provide release specifications, which may possess a narrow range of acceptable values. Despite relatively narrow release specifications, the performance of the adsorbent in a given purification process could still vary from lot to lot. In this case, an alternative use test (one that properly captures the lot to lot variability) may be required to determine an acceptable range of variability for a specific process. In this work, we describe the separation of therapeutic protein monomer and aggregate species using hydrophobic interaction chromatography, which is potentially sensitive to adsorbent lot variability. An alternative use test is formulated, which can be used to rapidly screen different adsorbent lots prior to implementation in a large-scale manufacturing process. In addition, the underlying mechanism responsible for the adsorbent lot variability, which was based upon differences in protein adsorption characteristics, was also investigated using both experimental and modeling approaches.     

High-throughput protein precipitation and hydrophobic interaction chromatography: salt effects and thermodynamic interrelation

Salt-induced protein precipitation and hydrophobic interaction chromatography (HIC) are two widely used methods for protein purification. In this study, salt effects in protein precipitation and HIC were investigated for a broad combination of proteins, salts and HIC resins. Interrelation between the critical thermodynamic salting out parameters in both techniques was equally investigated. Protein precipitation data were obtained by a high-throughput technique employing 96-well microtitre plates and robotic liquid handling technology. For the same protein-salt combinations, isocratic HIC experiments were performed using two or three different commercially available stationary phases-Phenyl Sepharose low sub, Butyl Sepharose and Resource Phenyl. In general, similar salt effects and deviations from the lyotropic series were observed in both separation methods, for example, the reverse Hofmeister effect reported for lysozyme below its isoelectric point and at low salt concentrations. The salting out constant could be expressed in terms of the preferential interaction parameter in protein precipitation, showing that the former is, in effect, the net result of preferential interaction of a protein with water molecules and salt ions in its vicinity. However, no general quantitative interrelation was found between salting out parameters or the number of released water molecules in protein precipitation and HIC. In other words, protein solubility and HIC retention factor could not be quantitatively interrelated, although for some proteins, regular trends were observed across the different resins and salt types.     

Retention model of protein for mixedmode interaction mechanism in ion exchange and hydrophobic interaction chromatography

A unified retention equation of proteins was proved to be valid for a mixed-mode interaction mechanism in ion exchange chromatography (IEC) and hydrophobia interaction chro-matography (HIC). The reason to form a "U" shape retention curve of proteins hi both HIC and IEC was explained and the concentration range of the strongest elution ability for the mobile phase was determined with this equation. The parameters in this equation could be used to characterize the difference for either HIC or IEC adsorbents and the changes in the molecular conformation of proteins. With the parameters in this equation, the contributions of salt and water in the mobile phase to the protein retention in HIC and IEC were discussed, respectively. In addition, the comparison between the unified equation and Melander' s three-parameter equation for mixed-mode interaction chromatography was also investigated and better results were obtained in former equation.     

Changes in solvent exposure reveal the kinetics and equilibria of adsorbed protein unfolding in hydrophobic interaction chromatography

Hydrogen exchange has been a useful technique for studying the conformational state of proteins, both in bulk solution and at interfaces, for several decades. Here, we propose a physically based model of simultaneous protein adsorption, unfolding and hydrogen exchange in HIC. An accompanying experimental protocol, utilizing mass spectrometry to quantify deuterium labeling, enables the determination of both the equilibrium partitioning between conformational states and pseudo-first order rate constants for folding and unfolding of adsorbed protein. Unlike chromatographic techniques, which rely on the interpretation of bulk phase behavior, this methodology utilizes the measurement of a molecular property (solvent exposure) and provides insight into the nature of the unfolded conformation in the adsorbed phase. Three model proteins of varying conformational stability, α-chymotrypsinogen A, β-lactoglobulin B, and holo α-lactalbumin, are studied on Sepharose™ HIC resins possessing assorted ligand chemistries and densities. α-Chymotrypsinogen, conformationally the most stable protein in the set, exhibits no change in solvent exposure at all the conditions studied, even when isocratic pulse-response chromatography suggests nearly irreversible adsorption. Apparent unfolding energies of adsorbed β-lactoglobulin B and holo α-lactalbumin range from −4 to 3 kJ/mol and are dependent on resin properties and salt concentration. Characteristic pseudo-first order rate constants for surface-induced unfolding are 0.2–0.9 min⁻¹. While poor protein recovery in HIC is often associated with irreversible unfolding, this study documents that non-eluting behavior can occur when surface unfolding is reversible or does not occur at all. Further, this hydrogen exchange technique can be used to assess the conformation of adsorbed protein under conditions where the protein is non-eluting and chromatographic methods are not applicable.     

Solute retention in reversed-phase chromatography as a function of stationary phase properties: Effect of n-alkyl chain length and ligand density

Summary Two series of bonded phases were synthesized employing LiChrospher Si 100, 10 m and n-alkyldimethylmonochlorosilanes as silanizing reagents. In series A the n-alkyl chain length, n, of the bonded phase was varied between 1 and 20 at a constant ligand density of 3.5±0.2mol·m^–2. In series B the ligand density, d, was gradually changed from 0 to 4.1mol·m^–2 on the C₁, C₄, C₆, C₈ and C₁₈ bonded phases, respectively.The capacity factors of benzoic acid esters and anilines as solutes were found to increase linearly with the n-alkyl chain length of packings at constant eluent composition (series A) up to a so-called critical chain length, n_crit, where the capacity factor remained constant. n_crit was in the range from 11 to 14. The same pattern was observed when plotting the solute capacity factor against the ligand density of the packing at constant n and constant eluent composition (series B). The critical ligand density, _crit, varied between 2.3 and 3.2 mol·m^–2 depending on n and the solute. Furthermore, solute retention was slightly higher on RP packings with an even number of carbon atoms in the alkyl chain than on those with an odd number.These findings are consistent with the results of Berendsen and de Galan (J. Chromatogr., 196, 21 (1980)), Dill (J. Phys. Chem., 91, 1980 (1987)) and Simpson and Lau (to be published). The observed phenomena reflect the dynamic structure of RP silicas which are related to the mobility of solvated n-alkyl chains. Due to the lack of a precise conformational analysis of the surface of RP silicas, a semiquantitative model was applied to interprete the described dependencies.     

Retention and selectivity effects caused by bonding of a polar urea-type ligand to silica: a study on mixed-mode retention mechanisms and the pivotal role of solute-silanol interactions in the hydrophilic interaction chromatography elution mode

The separation properties of five silica packings bonded with 1-[3-(trimethoxysilyl)propyl]urea in the range of 0–3.67 μmol m⁻² were investigated in the hydrophilic interaction chromatography (HILIC) elution mode. An increase of the ligand surface density promoted retention of non-charged polar compounds and even more so for acids. An opposite trend was observed for bases, while the amphoteric compound tyrosine exhibited a U-shaped response profile. An overall partitioning retention mechanism was incompatible with these observations; rather, the substantial involvement of adsorptive interactions was implicated. Support for the latter was provided by column-specific changes in analyte retention and concomitant selectivity effects due to variations of salt concentration, type of salt, pH value, organic modifier content, and column temperature. Silica was more selective for separating compounds differing in charge state (e.g. tyramine vs. 4-hydroxybenzoic acid), while in cases where structural differences of solutes resided in non-charged polar groups (e.g. tyramine vs. 5-hydroxydopamine, nucleoside vs. nucleobase) more selective separations were obtained on bonded phases. Hierarchical cluster analysis of the home-made urea-type and three commercial amide-type bonded packings evinced considerable differences in separation properties. The present data emphasise that the role of the packing material under HILIC elution conditions is hardly just the polar support for a dynamic coating with a water-enriched layer. Three major retention mechanisms are claimed to be relevant on bare silica and the urea-type bonded packings: (i) HILIC-type partitioning, (ii) HILIC-type weak adsorption such as hydrogen bonding between solutes and ligands or solutes and silanols (potentially influenced by individual degrees of solvation, salt bridging, etc.), (iii) strong electrostatic (ionic) solute–silanol interactions (attractive/repulsive). Even when non-charged polar bonded phases are used, solute–silanol interactions should not be discounted, which makes them a prime parameter to be characterised by HILIC column tests. Multi/mixed-mode type separations seem to be common under HILIC elution conditions, associated with a great deal of selectivity increments. They are accessible and controllable by a careful choice of the type of packing, the mobile phase composition, and the temperature.     

Microcalorimetric studies of interactions between proteins and hydrophobic ligands in hydrophobic interaction chromatography: effects of ligand chain length, density and the amount of bound protein

Using isothermal titration calorimetry (ITC), this investigation directly measured the adsorption enthalpies of proteins on various hydrophobic adsorbents. Various amounts of butyl and octyl groups were attached onto CM-Sepharose to form C4 and C8, two types of hydrophobic adsorbents. The adsorption enthalpies of both trypsinogen and alpha-chymotrypsinogen A were measured at 4.0 M NaCl and pH 10.0, in which most ionic interaction was suppressed. The adsorption isotherms of both proteins on various adsorbents were also measured, thus allowing us to calculate the Gibbs free energy and entropy of adsorption. Experimental results indicated that the adsorption of both proteins on butyl-containing adsorbents was exothermic, while their adsorption on octyl ones was endothermic. In addition, binding of both proteins with the butyl ligand is basically an adsorption process, while binding with the octyl ligand is adsorption and partition processes. Moreover, on both butyl or octyl, the adsorption enthalpy became increasingly positive as the ligand density increased, while the adsorption entropy became more positive as the alkyl chain length or density of the adsorbent increased. In addition, ITC was used to measure protein-protein interaction. The adsorption enthalpy of both proteins increased as the amount of bound protein increased, and the enthalpy increase of trypsinogen appeared to be higher than that of alpha-chymotrypsinogen A. This observation implies that protein-protein repulsion was stronger among trypsinogen molecules in the experiments.     

Generalizing a two-conformation model for describing salt and temperature effects on protein retention and stability in hydrophobic interaction chromatography

A two-conformation adsorption model that includes the effects of salt concentration and temperature on both stability and adsorption has been developed to describe the effects of secondary protein unfolding on hydrophobic interaction chromatography (HIC). The model has been applied to a biotech protein and to beta-lactoglobulin on Phenyl Sepharose 6FF low sub HIC media. Thermodynamic property models for adsorption and protein stability with parameters obtained from experimental chromatographic data successfully describe observed chromatographic behavior over ranges of temperature and salt concentration, provide predictions of distribution among different conformers, and give a basis for calculating trends in retention strength and stability with changing conditions, that might prove useful in HIC process development.     

Modeling the effects of type and concentration of organic modifiers, column type and chemical structure of analytes on the retention in reversed phase liquid chromatography using a single model

A previously proposed model for representing the retention factor (k) of an analyte in mixed solvent mobile phases was extended to calculate the k of different analytes with respect to the nature of analyte, organic modifier, its concentration and type of the stationary phase. The accuracy of the proposed method was evaluated by calculating mean percentage deviation (MPD) as accuracy criterion. The predicted vs. observed plots were also provided as goodness of fit criteria. The developed model prediction capability compared with a number of previous models (i.e. LSER, general LSER and Oscik equation) through MPD and fitting plots. The proposed method provided acceptable predictions with the advantage of modeling the effects of organic modifiers, mobile phase compositions, columns and analytes using a single equation. The accuracy of developed model was checked using the one column and one analyte out cross validation analyses and the results showed that the developed model was able to predict the unknown analyte retention and the analytes retentions on unknown column accurately.     

Protein adsorption in porous adsorbent particles: A multiscale modeling study on inner radial humps in the concentration profiles of adsorbed protein induced by nonuniform ligand density distributions

Recently published results determined from molecular dynamics (MD) modeling and simulation studies have shown that the spatial distribution of the density of immobilized charged ligands in ion‐exchange porous adsorbent particles is most likely nonuniform and the adsorbent particles also exhibit local nonelectroneutrality. In this work, the functional forms of the nonuniform spatial distributions of the density of the immobilized ligands in four different porous adsorbent media that were determined by MD studies were employed in a macroscopic continuum model describing the transport and adsorption of a single protein in the porous particles of the four different adsorbent media. The results clearly show that inner radial humps in the concentration profiles of the adsorbed protein can occur when the spatial distribution of the density of the immobilized ligands in the porous adsorbent particles is nonuniform and also has local maxima or minima along the radial direction in the particle. The results also indicate that the rate at which the equilibrium condition is approached depends significantly on the functional form of the spatial distribution of the density of the immobilized ligands. When adsorption equilibrium has been reached, the concentration profile of the adsorbed protein exhibits the shape of the spatial distribution of the density of the immobilized ligands. The results suggest that the technique of confocal scanning laser microscopy could be used to measure the concentration profile of an adsorbed protein at equilibrium and this measurement could provide the spatial distribution of the density of the immobilized ligands, and such measurements could also be used for quality control of the adsorbent medium. The results in this work have also implications in the modeling, design, analysis, and quality control of systems involving biocatalysis. Furthermore, the results clearly indicate that it is very important to study the dynamic behavior of an adsorption system having a nonuniform spatial distribution in the density of the immobilized charged ligands and where (i) both monovalent and multivalent interactions between the single charged adsorbate and the immobilized charged ligands occur and (ii) the values of the pH and ionic strength are such that the electrophoretic effects are active.     

The effects of ligand chain length, salt concentration and temperature on the adsorption of bovine serum albumin onto polypropyleneglycol-Sepharose

The interaction thermodynamics associated with bovine serum albumin adsorption on polypropylene glycol (n=3)-Sepharose CL-6B and polypropylene glycol (n=7)-Sepharose CL-6B, using ammonium sulfate as the modulator was studied. Analysis of data under linear conditions was accomplished with the stoichiometric displacement retention model, preferential interaction approach and van't Hoff plots applied to HIC systems. Preferential interaction analysis indicated a strong entropic driving force under linear conditions, due to the release of a large amount of solvent on adsorption. In contrast, flow microcalorimetry under overloaded conditions showed that the adsorption of bovine serum albumin may be entropically or enthalpically driven. It is postulated that adsorption in the nonlinear region is influenced by the degree of water release, protein-protein interactions on the surface, reorientation of ligand, and conformational changes in the protein.     

A new approach for the removal of protein impurities from purified biologicals using combinatorial solid-phase ligand libraries

The removal of last impurity traces from a purified protein is generally called polishing. It is an important step in downstream processing since protein impurities may generate undesirable side effects when the preparation is intended for research, diagnostic and more importantly therapeutic applications. Polishing is generally achieved by using orthogonal separation methods to previous steps, the most common being gel permeation chromatography. In spite of its polishing effectiveness, this technique suffers from a poor separation capacity and modest productivity as a result of low speed. Other approaches, for instance, based on anion exchange or on hydrophobic chromatography, that may be optimized for a given process cannot be used as generic methods. This document reports for the first time the use of a combinatorial solid-phase peptide library as a general method for the removal of impurity traces. Several examples of impurity trace removal are reported; starting material is either a pure protein spiked with serum proteins or with Escherichia coli extracts or current purified proteins still containing a small percentage of impurities. Among polished proteins are recombinant human albumin expressed in Pichia pastoris and human transferrin purified from whole plasma. This new method is used in neutral or even physiological pH and ionic strength conditions, with a remarkable capability to remove impurities. The process is as rapid as current adsorption chromatography procedures usable for the removal of a large number of protein impurities, with each one present in small amounts, such as host cell proteins.     

Effects of ligand density and pore size on the adsorption of bovine IgG with DEAE ion-exchange resins

Immunoglobulin G is an important plasma protein with many applications in therapeutics and diagnostics, which can be purified effectively by ion exchange chromatography. The ligand densities and pore properties of ion-exchange resins have significant effects on the separation behaviors of protein, however, the understandings are quite limited. In this work, with bovine immunoglobulin as the model IgG, the adsorption isotherms and adsorption kinetics were investigated systematically with series of diethylaminoethyl ion-exchange resins with different ligand densities and pore sizes. The Langmuir equation and pore diffusion model were used to fit the experimental data. The influences of ligand density and pore size on the saturated adsorption capacity, the dissociation constant and the effective diffusivity were discussed. The adsorption capacities increased with the increase of ligand density and the decrease of pore size, and an integrative parameter was proposed to describe the combined effects of ligand density and pore size. It was also found that the effective pore diffusion coefficient of the adsorption kinetics was influenced by pore sizes of resins, but was relatively independent on the ligand densities of resins. For a given protein, the ligand density and pore size should be optimized for improving the protein adsorption.     

Influence of the sample-solvent on protein retention,mass transfer and unfolding kinetics in hydrophobic interaction chromatography

Typical mobile phase employed in hydrophobic interaction chromatography contains cosmotropic salts, which promote retention and simultaneously reduce the protein solubility in the mobile phase. To increase mass overloading in the separation process the protein can be dissolved in a sample-solvent with concentration of salt lower than that in the mobile phase or in salt free solutions. However, this methodology may cause band splitting and band deformation, which results in yield losses. In this study, these phenomena were analyzed based on the retention behavior of two model proteins, i.e., lysozyme and bovine serum albumin. Retention of these proteins was accompanied by strong band broadening originated from slow rates of mass transfer and/or of adsorption–desorption process involving the protein conformational changes. The mass transport resistances and unfolding kinetics were found to contribute to the sample-solvent effects. To avoid band deformations the process variables such as the salt concentration and temperature were adjusted in such a way that complete resolution between band profile of the sample-solvent and the protein was achieved. For the process simulation a dynamic model, which accounted for underlying kinetics was used. General guidelines of the process design were developed.     

Electrostatic model for protein adsorption in ion-exchange chromatography and application to monoclonal antibodies,lysozyme and chymotrypsinogen A

A model for the adsorption equilibrium of proteins in ion-exchange chromatography explicitly accounting for the effect of pH and salt concentration in the limit of highly diluted systems was developed. It is based on the use of DLVO theory to estimate the electrostatic interactions between the charged surface of the ion-exchanger and the proteins. The corresponding charge distributions were evaluated as a function of pH and salt concentration using a molecular approach. The model was verified for the adsorption equilibrium of lysozyme, chymotrypsinogen A and four industrial monoclonal antibodies on two strong cation-exchangers. The adsorption equilibrium constants of these proteins were determined experimentally at various pH values and salt concentrations and the model was fitted with a good agreement using three adjustable parameters for each protein in the whole range of experimental conditions. Despite the simplifications of the model regarding the geometry of the protein–ion-exchanger system, the physical meaning of the parameters was retained.     

蛋白质在疏水作用色谱上的保留模型及其模型参数间关系的研究

在优先水化作用和相关作用原理的基础上,从热力学出发推导出蛋白质在高效疏水作用色谱(HIC)上的保留模型,建立了模型参数间的关系式,使过去观察到的经验规律有了理论依据,并利用上述模型解释了一些蛋白质在HIC上的重要色谱特性。     

Immobilized metal ion affinity chromatography. Effect of solute structure, ligand density and salt concentration on the retention of peptides

The adsorption characteristics of a variety of synthetic peptide hormones and di-, tri- and tetrapeptides on Cu(II) immobilized on two commercially available high-performance chelating gels run under various experimental conditions are described. Methods for determining the concentration of immobilized Cu(II) in situ are also described. The Cu(II)-charged columns exhibit a net negative charge as judged from the significantly higher retention of some basic peptides in the absence of NaCl in the equilibration and elution buffers. At higher NaCl concentrations (2-4 M), aromatic interactions seem to be superimposed on the metal ion affinity characteristics of the peptides. The relationship between resolution of peptides and the concentration of immobilized Cu(II) ions has also been established for the Chelating Superose gel where 40 mumol Cu(II) ml-1 gel apparently gives the optimum resolution. The nature of the gel matrix also plays a role in the resolution of some peptides, the extent of which is difficult to predict. The results obtained also suggest that peptides containing aromatic and hydroxy amino acids are retarded more than those which lack them. Moreover, these same amino acids apparently strengthen the existing strong binding of peptides containing His, Trp or Cys to a Chelating Superose-Cu(II) column. Dipeptides with C-terminal His (i.e., X-His) are neither bound nor retarded on a column of Chelating Superose-Cu(II) whereas those having the structure His-X are strongly bound. Some tri- and tetrapeptides containing His were also found not to bind to the column. The underlying cause of this anomalous adsorption behaviour is discussed and is ascribed to "metal ion transfer" arising from the relatively higher affinity of such peptides towards immobilized Cu(II) ions than the chelator groups (iminodiacetate) which are covalently bound to the gel matrix.     

Role of the ligand density in cation exchange materials for the purification of proteins

The performance of functionalized materials, such as cation exchange resins, is dependent not only on the ligand type and ligand density, but also on the pore accessibility of the target molecule. In the case of large molecules such as antibodies this latter parameter becomes crucial, because the size of such molecules falls somewhere inside the pore size distribution of the resin. The influence of the ligand density and accessibility on the overall performance of the material is explored systematically. Five different materials, having the same chemistry as the strong cation exchange resin Fractogel EMD SO₃⁻ (M) , have been analyzed. These materials only differ in the ligand density. It is shown that the ligand density directly influences the porosity of the materials as well as the pore diffusivity and the dynamic binding capacity. For a given purification problem an optimal ligand density can be found. Based on the above results a new material is proposed, showing superior properties in terms of dynamic binding capacity. This is achieved by an optimization of the ligand density and by a decrease of the particle size of the stationary phase. The material properties are modeled with a general rate model. Further simulations were conducted to evaluate the performance of the new material in comparison with a conventional resin.     

Comparison of the adsorption mechanisms of pyridine in hydrophilic interaction chromatography and in reversed-phase aqueous liquid chromatography

The adsorption isotherms of pyridine were measured by frontal analysis (FA) on a column packed with shell particles of neat porous silica (Halo), using water–acetonitrile mixtures as the mobile phase at 295 K. The isotherm data were measured for pyridine concentrations covering a dynamic range of four millions. The degree of heterogeneity of the surface was characterized by the adsorption energy distribution (AED) function calculated from the raw adsorption data, using the expectation-maximization (EM) procedure. The results showed that two different retention mechanisms dominate in Per aqueous liquid chromatography (PALC) at low acetonitrile concentrations and in hydrophilic interaction chromatography (HILIC) at high acetonitrile concentrations. In the PALC mode, the adsorption mechanism of pyridine on the silica surface is controlled by hydrophobic interactions that take place on very few and ultra-active adsorption sites, which might be pores on the irregular and rugose surface of the porous silica particles. The surface is seriously heterogeneous, with up to five distinct adsorption sites and five different energy peaks on the AED of the packing material. In contrast, in the HILIC mode, the adsorption behavior is quasi-homogeneous and pyridine retention is governed by its adsorption onto free silanol groups. For intermediate mobile phase compositions, the siloxane and the silanol groups are both significantly saturated with acetonitrile and water, respectively, causing a minimum of the retention factor of pyridine on the Halo column.