C H A P T E R
T W E N T Y- S I X
Affinity Chromatography: General Methods Marjeta Urh, Dan Simpson, and Kate Zhao Contents 1. Introduction 2. Selection of Affinity Matrix 2.1. General features of the support material 2.2. Selectivity 2.3. Stability 2.4. Magnetic affinity beads 3. Selection of Ligands 3.1. General considerations for ligands design and selection 3.2. Characterization of immobilized ligand 3.3. Affinity matrices carrying specific ligands 3.4. Immunoglobulin binding proteins 3.5. Lectins 3.6. Biomimetic ligands 3.7. Covalent affinity chromatography 4. Attachment Chemistry 4.1. Activation of surface 4.2. Ligand attachment 5. Purification Method 5.1. Sample preparation 5.2. Binding and wash 5.3. Elution References
418 419 419 420 422 422 423 423 424 425 425 426 427 428 429 430 430 433 433 433 434 435
Abstract Affinity chromatography is one of the most diverse and powerful chromatographic methods for purification of a specific molecule or a group of molecules from complex mixtures. It is based on highly specific biological interactions between two molecules, such as interactions between enzyme and substrate, receptor and ligand, or antibody and antigen. These interactions, which are typically reversible, are used for purification by placing one of the Promega Corporation, Madison, Wisconsin, USA Methods in Enzymology, Volume 463 ISSN 0076-6879, DOI: 10.1016/S0076-6879(09)63026-3
2009 Elsevier Inc. All rights reserved.
Marjeta Urh et al.
interacting molecules, referred to as affinity ligand, onto a solid matrix to create a stationary phase while the target molecule is in the mobile phase. Successful affinity purification requires a certain degree of knowledge and understanding of the nature of interactions between the target molecule and the ligand to help determine the selection of an appropriate affinity ligand and purification procedure. With the growing popularity of affinity purification, many of the commonly used ligands coupled to affinity matrices are now commercially available and are ready to use. However, in some cases new affinity chromatographic material may need to be developed by coupling the ligand onto the matrix such that the ligand retains specific binding affinity for the molecule of interest. In this chapter, we discuss factors which are important to consider when selecting the ligand, proper attachment chemistry, and the matrix. In recent years, matrices with unique features which overcome some of the limitations of more traditional materials have been developed and these are also described. Affinity purification can provide significant time savings and several hundred-fold or higher purification, but the success depends on the method used. Thus, it is important to optimize the purification protocol to achieve efficient capture and maximum recovery of the target.
1. Introduction Affinity chromatography is a method for selective purification of a molecule or group of molecules from complex mixtures based on highly specific biological interaction between the two molecules. The interaction is typically reversible and purification is achieved through a biphasic interaction with one of the molecules (the ligand) immobilized to a surface while its partner (the target) is in a mobile phase as part of a complex mixture. The capture step is generally followed by washing and elution, resulting in recovery of highly purified protein. Highly selective interactions allow for a fast, often single step, process, with potential for purification in the order of several hundred to thousand-fold. Additional uses of affinity chromatography include the ability to concentrate substances present at low concentration and the ability to separate proteins based on their biological function where an active form can be separated from the inactive form or a form with different biological function. Recent decades have seen tremendous advancements in the utility of affinity chromatography, with developments in support materials such as flow-through beads, magnetic beads and monolithic materials as well as new ligands with a variety of interesting biological properties. In addition, this approach is no longer used only for purification of specific biomolecules. It is also quickly becoming a method of choice to study biological interactions and can be used for preparation of samples for mass spectrometry or for specific removal of contaminants.
Overview of Affinity Chromatography Methods
While the first examples of affinity chromatography exploited binding of enzyme amylase to insoluble starch (Starkenstein, 1910), it was the development of solid support materials and the chemistry used to attach ligands to these solid supports that truly enabled affinity chromatography (Campbell et al., 1951; Cuatrecasas et al., 1968; Lerman, 1953a,b). Since these seminal advances, there has been an explosion in the development of more support materials, affinity ligands, and improvements in the methods that make affinity purification an attractive approach to purify biomolecules from complex mixtures. While affinity chromatography can be used to purify any biological molecule with a specific interacting ligand, this chapter will focus on purification of proteins. Several excellent reviews and books have been written on this subject (Hage et al., 2006; Hirabayashi, 2008; Labrou, 2003; Ostrove, 1990; Zachariou, 2007); thus, we will not review all the proteins and methods that have been developed up to date, but will rather give general guidelines and considerations important for the selection of different support materials, ligands, attachment chemistry, and optimization of the purification and discuss some of the latest developments in affinity chromatography.
2. Selection of Affinity Matrix Successful affinity purification depends on the selection of a suitable solid support and a suitable immobilized ligand. The affinity matrix (i.e., a solid support onto which ligand is immobilized) should ideally be macroporous with high chemical and physical stability, and should selectively capture target of interest while at the same time exhibit low nonspecific adsorption and maintain good flow properties throughout processing. They are preferably inexpensive, readily available, and simple to use. The affinity matrix may be commercially available or can be made by attaching a suitable ligand to a solid support via the appropriate chemistry (see Section 4). There are many detailed reviews on how to select a suitable affinity matrix (GE Healthcare, 2007; Gustavsson and Larsson, 2006; Zachariou, 2007), a summary of some of the key consideration points are given in the following sections.
2.1. General features of the support material The matrix should be macroporous with uniform particle and pore size and with good flow properties. Pore size of a matrix is inversely correlated to its surface area, which in turn, directly affects the amount of immobilized ligand and thus the capacity. The pore size correlates to exclusion limit, which is the size (molecular weight) or the size range of proteins that cannot enter the pore. Large pores do not suffer from size exclusion effect and allow
Marjeta Urh et al.
unhindered access of large molecules to the immobilized ligands, but have reduced surface area and lower ligand density that can result in lower capacity. According to the Renkin equation (Renkin, 1954), the size of the pores should be at least five times larger than the average size of a biomolecule for its easy access to immobilized ligands. The size of the pores ˚ or greater if assuming an average protein size of should be at least 300 A ˚ 60 A. Most commonly available supports satisfy this requirement (Table 26.1). Some soft gel-based matrices are cross-linked to increase their mechanical stability, but this process can reduce porosity (pore volume), thus reducing the amount of attached ligand and binding capacity. In most cases, a compromise is made between pore size and surface area to fit a particular application. Other important considerations are diameter of particles and particle size distribution. In theory, smaller particles are better because they allow for faster mass transfer between outer flow and interior of the particle, making higher flow rates possible while maintaining efficient affinity capture. However, they lead to higher flow resistance, greater potential for particle collapse, and increased sensitivity to contaminants such as particles and denatured proteins in the sample which can lead to high back pressure. The selection of the particle size will largely depend on the purpose and method of the separation where particles of 10 mm are best suited for HPLC separation and 400 mm for preparative purposes. The particle size distribution should be as uniform as possible so that smaller particles do not fill the void volume and restrict the flow.
2.2. Selectivity One key feature of an affinity matrix is its selectivity. It should be specific for a protein of interest as determined by the specific ligand coupled to the matrix and inert to all other compounds present in the complex sample. Since most applications are performed in aqueous solutions, often with low ionic strength, the support should be hydrophilic and contain limited charge that may lead to undesirable ionic interaction. Many commercially available supports fulfill these requirements, either with their native structure or by coating with suitable materials. Common supports include beaded agarose and cellulose, available commercially from a number of vendors including Sepharose from GE Healthcare and Affigel from Bio-Rad (Table 26.1). Nonspecificity can come from the support itself, such as hydrophobicity associated with polystyrene beads and negative charge on the surface of silica. It can also be introduced when modifying a matrix to accept a particular ligand. In these cases, the attachment chemistry, the ligand, and the spacer between the ligand and the matrix should be carefully designed, screened and optimized for selectivity, capacity for target of interest, and low nonspecific binding.
Table 26.1 Examples of commercially available matrices Name
Particle size (mm)
Exclusion limit (Da)
SepharoseTM CL6B SepharoseTM CL4B Bio-gel A-5m medium Perloza MT100 medium Perloza MT50 medium AllTech Macrosphere Bio-gel P-100 medium SephacrylTM Poros 50
75–300 (50–200 mesh)
Pore size 60–300 A˚ 5000–100,000
GE Healthcare Applied Biosystems
Cross-linked ally dextrose Cross-linked poly(styrenedivinylbenzene)
1000–100,000 Pore size 50–100 nm
Marjeta Urh et al.
2.3. Stability The affinity matrix must also be chemically and physically stable during the process, such that the support material itself as well as the attached ligand should not react to the solvents used in the process, nor should they be degraded or damaged by enzymes and microbes that might be present in the sample. The chemical compatibilities of commercially available affinity matrices are usually supplied by the manufacturer, which should be used as guidance for developing a successful purification protocol. Cross-linked agarose can usually withstand a wide pH range (e.g., pH 3–12), most aqueous solvents (including denaturants), many organic solvents or modifiers, and enzymatic treatments. Materials such as glass and silica are not stable at alkaline conditions due to hydrolysis; therefore, coating of these surfaces is often needed before attaching ligands. The matrix should also withstand physical stress, such as pressure, especially when packed into a column, and remain intact during the purification process. High pressure can compress the matrix, causing it to collapse. Agarose beads and other soft gel matrices are more susceptible to pressure, relative to stronger supports, such as silica, polystyrene and other highly cross-linked materials. Monoliths are macroporous, nonbeaded single matrix that can be made from different materials such as agarose, silica, GMA/EDMA, and cryogel (Mallik and Hage, 2006; Plievaa et al., 2009). It possesses many of the desired properties of an affinity support and has become popular due to the presence of large flow-through pores with no void volume permitting convective flow instead of diffusion and high flow rate for shortened run time. In addition, the matrix will not compress and has less pressure drop during column chromatography and can be made to withstand larger pH ranges and harsh chemicals. There are several limitations of monoliths as compared to traditional matrix, such as lower capacity, special processes required for making each type of monolith affinity supports and the current limited range of available affinity types (Mallik and Hage, 2006).
2.4. Magnetic affinity beads Magnetic separation can significantly shorten the purification process by quick retrieval of affinity beads at each step (e.g., binding, wash, and elution), and reduce sample dilution usually associated with traditional column-based elution. The method can be used on viscous materials that will otherwise clog traditional columns and can therefore simplify the purification process by eliminating sample pretreatment, such as centrifugation or filtration to remove insoluble materials and particulates. The capability of miniaturization and parallel screening of multiple conditions, such as growth conditions for optimal protein expression and buffer conditions for purification, makes magnetic separation amenable
Overview of Affinity Chromatography Methods
to high-throughput analysis which can significantly shorten the purification process (Saiyed et al., 2003). Paramagnetic particles are available as unmodified, modified with common affinity ligands (e.g., streptavidin, GSH, Protein A, etc.), and conjugated particles with specific recognition groups such as monoclonal and polyclonal antibodies (Koneracka et al., 2006). In addition to target protein purification, they can also be used to immobilize a target protein which then acts as a bait to pull down its interaction partner(s) from a complex biological mixture. See Chapter 16.
3. Selection of Ligands Selection of the appropriate ligand requires a certain degree of knowledge and understanding of the nature of interactions between the ligand and the target molecule; where the ligand must specifically bind the target molecule and should be stable in different binding and elution conditions. Additionally, when developing affinity purification scheme it is important to consider whether the ligand is commercially available or de novo development of the ligand and affinity matrix will be required. The success of the second scenario will largely depend on existing knowledge of protein structure and the nature of interaction, requiring the use of molecular modeling and combinatorial organic synthesis coupled with immobilization chemistry and selective binding analysis. The time and effort to design a novel ligand and to develop appropriate coupling chemistry and matrix may prove to be too lengthy and costly, whereas the use of nonaffinity-based purification technique such as ion exchange and hydrophobic interaction schemes may be a better choice.
3.1. General considerations for ligands design and selection Affinity between ligand and target molecule is one of the most important considerations when developing a new affinity purification material; low affinity can reduce the binding efficiency resulting in poor yield while high affinity can lead to inefficient elution or inactivation of the target protein by harsh elution conditions also giving low yield. Generally, an affinity constant in the range of 106–108 M 1 can be used for affinity-based purification. It is preferred that binding affinity of ligand and target protein be evaluated prior to the creation of an affinity matrix, but one has to be aware that affinity in solution may differ from affinity after immobilization and in some cases immobilization may even result in a change in specificity. When attaching ligand to a matrix, covalent coupling is preferred due to the reduced risk of ligand leaching during purification. However,
Marjeta Urh et al.
noncovalent attachment strategies such as nonspecific adsorption, biospecific interaction (biotin–streptavidin), entrapment, and most recently developed, molecular imprinting (Alexander et al., 2006) may also be utilized in the absence of reactive groups or to reduce the risk of ligand denaturation. See Section 4 for more details. Besides weak affinity, reduced binding between ligand and target can also be a result of steric hindrance caused either by the matrix or by other ligands. Introduction of a spacer between the matrix and the ligand can reduce the steric hindrance caused by the matrix (Hage et al., 2006). When designing a spacer arm, consideration should be given to both its length and the hydrophobic character, as a more hydrophilic nature is desirable to reduce the nonspecific interactions with molecules other than the target. Optimization of spacer length is also important, since shorter spacers may not relieve the steric hindrance effect of the matrix and longer ones may promote nonspecific interaction or may fold back onto themselves thereby limiting specific interactions. The density of ligand on the surface should also be optimized. Very high density of a ligand may have an adverse effect and lead to loss of binding capacity either because of close proximity of binding sites causing steric hindrance or strong binding that may prevent effective elution (Hage et al., 2006). Other factors that may influence selection of the best ligand are ability to be sterilized, stability of the ligand, proper storage conditions and cost.
3.2. Characterization of immobilized ligand An important factor determining the utility of the affinity resin is the ability to reproducibly make the resin with similar performance characteristics. Thus, it is important to determine optimal reaction conditions including the amount of the ligand needed for the synthesis to assure consistency in performance. Determining the optimal amount of ligand in the synthesis may also reduce the cost by preventing wasteful use of excess ligand. For covalent attachment it is often desirable to determine how much ligand is attached on the matrix and to determine how coupling conditions affect the amount of immobilized ligand. A simple approach to determine ligand density is to analyze the amount of the ligand left in the reaction after coupling is complete and subtract it from the total starting amount. Depending on the nature of the ligand, different detection methods can be applied, such as spectrophotometric detection, BCA assay for proteins, Ellman’s reagent can be used to detect sulfhydryl groups, fluorescent detection and others (Guilbault, 1988; Langone, 1982). In some instances, such as amino acid or elemental analysis, the ligand can be measured directly on the support, but this leads to destruction of the material and should therefore be done on a small fraction. An indirect estimation of immobilized ligand can be achieved by estimating the binding
Overview of Affinity Chromatography Methods
capacity by determining the amount of target molecule retained by the affinity matrix under desired conditions. This may be the most relevant measurement of the ligand immobilization since it will directly correlate to the performance of the affinity matrix. Maximum binding capacity should be performed at equilibrium, usually at a slow flow rate or in batch binding mode. Note that dynamic binding capacity may differ from optimal binding capacity as it will be affected by the flow rate, mass transfer through matrix pores and affinity constant.
3.3. Affinity matrices carrying specific ligands In recent years, we have witnessed an explosion in the development of affinity ligands and ready-to-use affinity matrices specific for different target molecules. An overview of commonly used, commercially available ligands is given in Table 26.2, which is by no means a comprehensive list of all the available ligands. Most available affinity matrices carry what are often called ‘‘group specific ligands’’ that exhibit binding affinity for a group of structurally or functionally related proteins and thus can be used for purification of different target proteins with similar functionality. In the following sections, we describe some commonly used ligands for affinity matrices.
3.4. Immunoglobulin binding proteins Purification of antibodies, one of the most effective and frequently used application of affinity purification methods is based on binding between the constant region (Fc) of many types of immunoglobulins and protein A or protein G. Protein A from Staphylococcus aureus and protein G made by Streptococcus are related bacterial proteins that bind the IgG class of antibodies, with differences in subclass specificity and source organism. Detailed lists of different binding affinities of protein A and G can be found in many sources including several commercial suppliers (Guss et al., 1986; Hage et al., 2006). Other immunoglobulin binding proteins that have lately become popular are protein B, a surface protein of group A Streptococci bacteria, which binds several subclasses of human IgA antibodies, and protein L (Peptostreptococcus magnus), which has ability to interact with kappa light chains without affecting the antigen-binding site of antibodies (Faulmann et al., 1991; Hermanson, 1992). This makes protein L uniquely suited for purification of antibodies lacking Fc regions. In most cases, good binding to antibodies can be achieved at or near neutral pH, but the optimal pH for sample application can differ depending on the protein used. Protein A binds antibodies strongest at pH 8.2, protein L at pH 7.5 and protein G at pH 5 but protein G can also be used at pH 7–7.5. For elution, solutions at acidic pH (2.5–3.0) are often applied
Marjeta Urh et al.
Table 26.2 Commonly used ligands and specificity Ligand
Cibacron Blue F3G-A Blue B Orange A Green A Polymixin Benzamidine Biotin Gelatin Heparin
Albumin, kinases, dehydrogenases, enzymes requiring adenylyl-containing cofactors, NADþ Kinases, dehydrogenases, nucleic acid binding proteins Lactate dehydrogenase HAS, dehydrogenases Endoproteins Serine proteases (thrombin, trypsin, kallikrein) Streptavidin, avidin Fibronectin DNA binding proteins, serine protease inhibitors (antithrombin III), growth factors, lipoproteins, hormone receptors, coagulation factors, DNA, RNA Plasminogen, rRNA, dsDNA Serine proteases with affinity for arg, fibronectin, prothrombin Enzymes with affinity for NADPþ NAD-dependent dehydrogenases and ATP-dependent kinases Dehydrogenases Glycoproteins, polysaccharides, glycolipids Calmodulin binding proteins, ATPase, adenylate cyclase, kinases, phosphodiesterase, Fc regions of many IgG subtypes, species dependent, weak interactions with IgA, IgM, IgD Fc region of many IgG subtypes, species dependent Kappa light chains of antibodies (Fab, single chain variable region scFv)
Lysine Arginine ADT AMP NAD, NADP Lectins Calmodulin Protein A Protein G Protein L
and samples are collected into buffers with neutral or slightly basic pH to avoid denaturation and loss of activity. In cases where biological stability is lost during elution at low pH, elution with a pH gradient in combination with salt can be explored.
3.5. Lectins Lectins are a diverse group of proteins which bind carbohydrates with high degree of specificity where each lectin has its own specificity profile. They are often used in affinity purification or enrichment of carbohydrate moieties of complex glycoconjugates, such as polysaccharides, glycolipids,
Overview of Affinity Chromatography Methods
and glycoproteins. They also allow for a specific isolation of different glycolforms of a specific protein depending on the nature of glycosylation. Recently, lectins have been used not only for the purification of sugar-containing molecules but also for the enrichment of subgroups of glycoproteins for analysis in mass spectrometry (Hirabayashi, 2008). See Chapter 34. Most commercially available lectins are of plant origin, and there are over 100 different lectins available either in conjugated or free form. Lectin from Canavalia ensiformis, known as Conacanavalin A (ConA) has the affinity for a-D-mannose, a-D-glucose, and N-acetylglucosamine and is probably the most frequently used lectin (Hermanson, 1992). Two other popular lectins are wheat germ agglutinin (WGA) and Jacalin; WGA binds to sialic acid and molecules containing N-acetyl-D-glucosamine residue and Jacalin binds to a-D-galactosyl groups. Coupling of lectin onto resin is often performed at neutral pH and in the presence of sugar to preserve sugar binding site. Binding to the target molecule is also usually carried out in neutral pH; note that some lectins require the presence of divalent metal ions, Ca2þ and Mn2þ in the case of ConA, for optimal binding. Elution is accomplished by adding an access of the specific sugar molecule to the elution buffer and this can be done as stepwise or gradient elution. After elution, the free sugar should be removed by dialysis or size exclusion chromatography.
3.6. Biomimetic ligands Most of affinity chromatography ligands are naturally occurring and the appeal of these ligands is high selectivity and capacity of binding, but they often suffer from several drawbacks such as low stability, need for purification in their own right, variability in performance from lot to lot, contamination with other biomolecules, sensitivity to sterilization methods and high cost. To overcome some of these limitations and accelerate the use of affinity chromatography in purification of therapeutic proteins, there is an increased focus on the development of synthetic or altered affinity ligands, biomimetics, which mimic the structure of and binding of natural biological ligands. One of the most popular group of biomimetic ligands are reactive textile dyes which gained popularity because of their flexibility and ability to assume polarity and geometry of the surface of a variety of competitive biomolecules and can function in solution as a competitive inhibitor, coenzyme or effector of many proteins (Madoery and Minchiotti, 2006; Stellwagen, 1990). The majority of these ligands, including the most popular dye, Cibacron blue F3G-A, contain a triazine scaffold which can be modified for improved specificity, forming the basis for the biomimetic dye–ligands concept. Over the past decades, a plethora of dyes were developed and used in the purification of a broad spectrum of proteins including
Marjeta Urh et al.
blood proteins such as albumin, oxidoreductase, decarboxylases, glycolitic enzymes, nucleases, hydrolases, lyases, synthetases, and transferases, and many dye–ligand-based affinity chromatographic materials are now commercially available materials (see Table 26.2) (Labrou and Clonis, 1994, 2002). In addition to the expansion of commercially available materials, there is a growing popularity in creating tailor-made biomimetics for better performance. These ligands are designed to target specific proteins by mimicking peptide templates, natural biological recognition motifs, or complementary surface-exposed residues and have been generated by a combination of rational design, combinatorial library synthesis, and subsequent screening (Cecı´lia et al., 2005; Labrou, 2003; Lowe et al., 2001). The immobilization of triazine containing dye onto affinity matrices such as agarose, dextran, and cellulose can be achieved under alkaline conditions by using nucleophilic displacement of the dye’s chlorine by hydroxyl groups on the support surface (Labrou, 2000; Labrou and Clonis, 1994; Labrou et al., 1995).
3.7. Covalent affinity chromatography Affinity chromatography typically relies on reversible interaction between the target molecule and the ligand; however, covalent interactions have also been utilized for isolation of specific target molecules (Blumberg and Strominger, 1972; GE Healthcare, 2007; Hage, 2006). One such approach is based on the covalent binding between a thiolcontaining target molecule and activated thiol immobilized onto purification matrix. Bound protein can be eluted by reducing the cysteine disulfide with 2-mercaptoethanol, TCEP, or dithiothreitol. Recently, another covalent binding-based purification has been developed on the basis of specific and covalent interaction between a chloroalkane ligand and a protein called HaloTag, Figure 26.1 (Los et al., 2008; Ohana et al., 2009). HaloTag is a 34-kDa monomeric protein fusion tag which can be genetically fused to any protein of interest either at its C- or N-terminus. HaloTag protein was created by first modifying the active site of bacterial haloalkane dehydrolase so that a permanent covalent bond can be formed with the specific chloroalkane ligand. This modification was followed by additional mutagenesis to increase stability of the protein and the binding rate, which is similar to that of the biotin–streptavidin. To make use of HaloTag for purification, a chromatographic matrix, the HaloLink resin carrying the chloroalkane ligand, was created. The covalent nature of the HaloTag-fusion protein capture, combined with rapid binding kinetics, overcome the equilibrium-based limitations associated with traditional affinity purification and enables efficient capture of target proteins even at very low abundance. Furthermore, it allows for extensive and stringent washing without losing the bound protein. While the covalent association
Overview of Affinity Chromatography Methods
B N Solid support Cl
Chloroalkane ligand HaloTag protein
Figure 26.1 HaloTag technology comprises two components: (A) The HaloTag protein shown on the left with covalently bound HaloTag-TMR ligand. N- and C-termini are indicated. (B) The chloroalkane ligand. Different functional groups including, but not limited to, fluorescent dyes or solid support surfaces can be attached to the chloroalkane ligand, which covalently binds to HaloTag protein and imparts different functionalities including protein immobilization and fluorescent labeling.
clearly has its advantages, it also creates a challenge in eluting the protein of interest. Because the covalent bond between HaloTag and chloroalkane cannot be reversed, traditional approaches to elute proteins from the resin cannot be utilized. Instead the protein of interest can be released from HaloTag by specific protease (TEV) as the TEV recognition site is present between the two moieties (the HaloTag and fusion target protein). Upon cleavage, HaloTag stays bound to the resin while the fusion partner is released yielding highly pure protein free of tag (Urh et al., 2008). Besides using HaloTag for purification of fusion proteins described earlier, the immobilized HaloTag fusions can also be considered as affinity ligands in their own right, and can, similarly to protein G, be used to capture antibodies or other proteins which specifically bind to the proteins fused to HaloTag. The advantage of this system is that unlike other covalent immobilization techniques, where binding of protein is random and may lead to multiple attachment sites and improper orientation, immobilization using HaloTag is a single point attachment through active site and therefore oriented. Single point, oriented attachment increases capacity, effectiveness and reproducibility of the system, and HaloTag fusions covalently bound to the matrix can therefore improve purification of specific antibodies or isolation of binding partners.
4. Attachment Chemistry This section will briefly review some of the more common chemistry for the covalent attachment of affinity ligands to conventional surfaces such as agarose, cellulose, silica, glass, and synthetic polymeric supports. Strategically, the process can be divided into three components: (1) the surface or
Marjeta Urh et al.
resin (matrix); (2) the linkage or resin-activation component (spacer); and (3) the affinity ligand. The constraint that most often dictates how the ligand is attached to a surface is the type and availability of reactive moieties on the ligand. Common moieties used are primary amine, thiol, and carboxyl groups. Alternatively, modification of a ligand to create a reactive group, such as the oxidation of a carbohydrate or the covalent attachment of an orthogonal reactive group, provides additional options for ligand immobilization. Attachment strategies for small synthetic molecules tend to follow the same chemical pathway as biomolecules but often with somewhat different design considerations.
4.1. Activation of surface The process of preparing a covalent affinity-based surface generally begins with activation of the surface to either directly accept the ligand or to accept an activated linker to physically ‘‘space’’ the ligand away from the surface. Representative attachment chemistries are summarized in Table 26.3, listed with established surfaces and shown in contrast with the ligand entity used to attach to the activated surface. A variety of preactivated surfaces are commercially available for most of the chemistries listed in Table 26.3. If the appropriate chemical entity is readily displayed on the ligand, the remaining chemistry is reduced to defining the optimal conditions for the coupling reaction to occur weighed against the stability constraints of the ligand. Most ligand coupling reactions are done in aqueous pH-controlled buffers for periods between 1 and 20 h at temperatures between 4 C and room temperature. Since the reaction is heterogenous, both concentration and mixing are important to ensure an even distribution of reactants. Whenever possible keep the solution phase of the reaction to no more than twice the resin volume. Ligand concentration will vary somewhat by objective but for proteins a 2.5–10 mg/ml is a reasonable starting point. Depending on the activation chemistry and ligand reactivity, protocols exist for a fairly broad range of pH and buffer conditions (Hermanson, 1992). Efficiencies will vary from system to system and as a consequence it is essential to cap left over surface reactive groups to neutralize or reduce nonspecific interactions.
4.2. Ligand attachment For a peptide, protein, or nucleic acid, the covalent attachment is most commonly done through an amine group on the ligand to an amine reactive functionality on the resin surface. The most common amine-reactive attachment is an imidocarbonate that results from the reaction of cyanogen bromide activated surface with a primary amine, most typically a lysine side chain (Hermanson, 1992; Porath et al., 1967) (see Chapter 28). Advantages
Overview of Affinity Chromatography Methods
Table 26.3 Commonly used attachment chemistries
Bead or surface
Activation and/or linkage
Soft gels: agarose, cellulose Synthetic supports: Polyacrylamide beads, Trisacryl, Sephacryl, Ultragel, Azlactone beads, Methacrylate (TSK gel), Eupergit, Polystyrene (Poros supports)
Cyanogen bromide Aldehyde (reductive amination) Activated carboxyl ester (succinimidyl ester) Carbonyldiimidazole Carboxyl (activation concurrent with coupling) FMP activation Divinyl sulfone Azlactone Epoxy (bisoxirane, epichlorohydrin) Tresyl chloride Haloacetyl (iodo or bromo) Maleimide Pyridyl disulfide Amine
Inorganics: controlled pore glass, silica, alumina, zeolites, etc.
3-(Glycidyloxypropyl) trimethoxy-silane 3-(Aminopropyl) trimethoxysilane
Affinity ligand reactive group
Amine Amine Amine Amine Amine
Amine, thiol Amine, thiol Amine, thiol Amine, thiol Amine, thiol Thiol Thiol Thiol Carboxyl (activation concurrent with coupling) Carbohydrate (periodate reduced) Amine Carboxyl (after activation), aldehyde
of this activation include commercially available preactivated resin and very efficient protein coupling, near 100%. While this is historically a popular method, it suffers a number of drawbacks including: (1) linkage leading to leaching of the ligand off of the resin, (2) attachment directly to the surface with no spacer, and (3) required additional safety precautions due to the toxicity of cyanogen bromide. In addition, cross-linking may be needed to help limit the leakiness of this immobilization (Korpela and Hinkkanen, 1976; Kowal and Parsons, 1980).
Marjeta Urh et al.
Covalent attachment through the formation of an amide bond is an alternative to the cyanogen bromide chemistry but requires either an activated carboxyl surface (such as a N-hydroxysuccinimidyl ester) or in situ activation of the carboxyl group with a coupling agent such as N-ethyl-N0 -(3-dimethylaminopropyl)carbodiimide (EDC) (Wilchek et al., 1984, 1994). Another stable alternative is the formation of a secondary amine linkage resulting from a reductive alkylation of a Schiff-base intermediate formed between a primary amine (lysine or N-terminus) to an aldehyde activated surface. This reductive alkylation attachment mechanism is a popular method to immobilize enzymes to carbohydrate surfaces (agarose or cellulose) as the conditions for coupling are mild and the immobilized enzyme has been reported to retain more activity than observed with other methods of attachment (Hermanson, 1992). Amine reactive linkages are also a common approach to attachment of small molecule ligands to surfaces. The primary requirement being that the ligand has the predesigned amine group for attachment. A number of the chemistries that are applied to amine immobilization also apply to sulfhydryl (or thiol)-reactive immobilization which is in concept the same as aminereactive immobilization but relies on the reaction of a cysteine with the reactive surface. For thiol specific immobilization, the two most prominent strategies are through haloacetyl (iodo or bromo) and maleimide activated surfaces (Mallik et al., 2007). Sulfhydryl linkage may be advantageous in that it can also be made reversible through the use of a disulfide linkage that can be removed from the surface by treatment with a reducing agent such as DTT or TCEP (Brena et al., 1993). Immobilization of antibodies or glycosylated protein to a surface can be done by the methods described earlier but an additional option that offers a potentially more oriented attachment is also available through modification of the carbohydrate (Oates et al., 1998; Vijayendran and Leckband, 2001). Immobilization through carbohydrate requires a mild oxidation (i.e., periodate) of the carbohydrate to form reactive aldehydes with the sugar residues. The aldehydes produced by oxidation are then used to immobilize the protein or antibody to a hydrazide reactive surface. This approach is commonly used for antibodies, glycoproteins, glycopolymers, and ribonucleic acids (O’Shannessy and Wilchek, 1990). In most cases, it is sufficient to rely on the attachment of a ligand through one of the reactive groups discussed earlier, but in some cases, it may be desirable to space the linker further from the surface. One option for doing so is to orthogonally modify the ligand of interest with a spacer and reactive group selective for a second group that will preferentially recognize the orthogonal label attached to a surface. This requires modification of both the surface and the ligand with the appropriate reactive groups. A current popular embodiment of this approach is the copper(I)-catalyzed 1,3-dipolar cycloaddition of azides to terminal alkynes to form 1,2,3-triazoles known
Overview of Affinity Chromatography Methods
commonly now as ‘‘click chemistry’’ (Chandran et al., 2009; Gauchet et al., 2006). The advantage of this approach is that the chemistry is mild, the selectivity is unique and reactive groups can be easily switched between the ligand and the surface. Because the reaction itself requires catalysis, the reactive groups on their own are much more stable than reactive groups commonly used with thiol and amine labeling.
5. Purification Method Purification by affinity starts with proper handling of the sample and the matrix, followed by selective binding (capture) of the target, washing to remove nonspecific background, and, finally, elution of the bound target. Successful affinity purification depends on a number of notable factors including, the amount and accessibility of the ligand on the resin, the strength of the interaction, and the integrity of protein to be immobilized. Usually, conditions are optimized to maximize the interaction between a target and immobilized ligand during the binding and wash process, then, switched to substantially weaken the interaction thus allowing for release of the target. It is recommended to perform small scale trials to select for the best purification conditions. A short summary of some common practical issues and considerations affecting affinity purification is given in the following sections.
5.1. Sample preparation When preparing sample for purification, conditions should be selected to retain the proper fold and functionality of the target of interest. It is also highly recommended to remove insoluble materials and to reduce viscosity because both of these factors could clog the column, reduce the flow rate, and increase back pressure. Some proteins tend to aggregate at high concentrations, which results in increased apparent molecular weight, decreased diffusion rate, and reduced capture by the affinity matrix. Dilution of the sample or cell lysate in a larger volume may be needed to reduce aggregation and increase protein capture and recovery under this circumstance. Conversely, if the sample is too dilute, binding rate and capture efficiency can be reduced especially for low-affinity binders.
5.2. Binding and wash Efficiency of binding is related to the strength and the kinetics of protein– ligand interaction which can be affected by the nature of the interaction, the concentration of applied target, the amount of immobilized ligand, and
Marjeta Urh et al.
the flow rate used for binding. The binding process can be simplified as Eq. (26.1), assuming a 1:1 molar ratio where Ka is the association equilibrium constant, [L] is ligand concentration, [T ] is the concentration of target protein, and [LT ] is the concentration of the complex. Ka equals [LT ]/([L] [T ]), which can also be expressed as ka/kd, where ka is second order association rate constant that depends on the concentrations of both L and T, and kd is the first order dissociation constant that does not depend on ligand concentration. Higher Ka usually leads to a higher adsorption ratio, defined as the ratio of bound to total applied target, thus, better binding. Normally, the ligand concentration on the matrix is around 10 2–10 4 M and to achieve efficient binding, the Ka value should be in the range of 104–106 M 1. ka
L þ T ! LT
Affinity immobilization onto solid support can be achieved by passing the sample through a column packed with the affinity matrix, usually under ambient pressure and a slow flow rate. Generally, a higher flow rate will reduce the binding efficiency, especially, when the interaction between the ligand and protein is weak or the mass-transfer rate in the column is slow. The binding process can also be performed in batch, where the resin and sample are constantly mixed. Batch binding promotes effective contact between target and immobilized ligand and often saves time, especially when dealing with large sample volumes; however, nonspecific binding can also increase. It is a good practice to optimize the amount of resin used during purification, where saturation of the resin with target during binding is preferred since excess resin can result in an increase in nonspecific binding as well as reduced target recovery due to readsorption, unless the latter is required under special circumstances. Following binding, protein bound by nonspecific interactions can be removed by washing. For example, ionic interactions can be reduced by increasing salt (0.1–0.5 M) or changing pH values, and hydrophobic interactions can be removed by decreasing salt, altering pH, or adding surfactants (such as Triton X-100). Low amounts of competitive reagents can also be used to remove contaminants with weak affinity to the ligands. The flow rate and the volume (e.g., 5–10 column volumes) of the wash buffer should also be carefully determined for maximum removal of contaminants with minimum loss of target.
5.3. Elution Elution of bound target from the resin is essentially the reverse process of binding, where conditions are optimized to reduce the Ka, that is, weakening the interaction between target and ligand. The elution condition should
Overview of Affinity Chromatography Methods
not denature the target protein, unless such conditions are compatible with downstream applications. There are two different types of elution methods, namely, specific and nonspecific elution. In specific elution, the target protein–ligand complex is challenged by agents that will compete for either the ligand or the target thereby releasing the target protein into solution. The concentration and amount (volume) of competitive reagent used for elution will depend on their affinity relative to that of the immobilized complex, where weaker competitive reagents require higher concentration and more volume as compared to higher affinity additives. A good starting point for weak competitors is to use concentration 10-fold higher than that of the ligand. The specific elution is usually milder and proteins are more likely to retain their activity, but the slow elution, broad elution peaks, and the need to remove competing agent from the recovered protein are some of the drawbacks of this approach. For nonspecific elution, solvent conditions are manipulated to reduce the association rate constant (Eq. (26.1)), which ideally should approach zero, and to increase the dissociation rate constant, thus, weakening the overall affinity (Ka) resulting in dissociation of the complex. Elution conditions can be optimized according to the mechanism of interaction between the ligand and protein, such as increasing salt concentration to reduce ionic interactions or by altering pH to change the protonation/ionization state, thus modulating the strength of hydrogen bonds, hydrophobic interactions as well electrostatic interactions. An example of elution by changing pH is the elution of antibodies from immobilized protein A or protein G, yet because the affinity of proteinA/G to antibodies is very strong, with Ka in the 108 M 1 range, a combination of different elution conditions may be required for maximum antibody release. In many cases, affinity requires proper three-dimensional folding of a protein so that chaotropic reagents or reagents that will affect protein folding can be used to elute target of interest; however, care must be taken to maintain proper folding of the target after elution by quickly returning to native conditions. When the affinity is weak, binding is achieved at high concentration of the target molecule which is then eluted by dissociation of the complex through dilution. This approach is known as isocratic elution.
REFERENCES Alexander, C., Andersson, H. S., Andersson, L. I., Ansell, R. J., Kirsch, N., Nicholls, I. A., O’Mahony, J., and Whitcombe, M. J. (2006). Molecular imprinting science and technology: A survey of the literature for the years up to and including 2003. J. Mol. Recognit. 19, 106–180. Blumberg, P. M., and Strominger, J. L. (1972). Isolation by covalent affinity chromatography of the penicillin-binding components from membranes of Bacillus subtilis. Proc. Natl. Acad. Sci. USA 69, 3751–3755.
Marjeta Urh et al.
Brena, B., Ovsejevi, K., Luna, B., and Batista-Viera, F. (1993). Thiolation and reversible immobilization of sweet potato beta-amylase on thiolsulfonate-agarose. J. Mol. Catal. 84, 381–390. Campbell, D. H., Luescher, E., and Lerman, L. S. (1951). Immunologic adsorbents: I. Isolation of antibody by means of a cellulose-protein antigen. Proc. Natl. Acad. Sci. USA 37, 575–578. Cecı´lia, A., Roque, A., and Lowe, C. R. (2005). Advances and applications of de novo designed affinity ligands in proteomics. Biotechnol. Adv. 24, 17–26. Chandran, S. P., Hotha, S., and Prasad, B. L. V. (2009). Tunable surface modification of silica nanoparticles through ‘‘click’’ chemistry. Current Science 95, 1327–1333. Cuatrecasas, P., Wilchek, M., and Anfinsen, C. B. (1968). Selective enzyme purification by affinity chromatography. Proc. Natl. Acad. Sci. USA 61, 636–643. Faulmann, E. L., Duvall, J. L., and Boyle, M. D. (1991). Protein B: A versatile bacterial Fc-binding protein selective for human IgA. Biotechniques 10, 748–755. Gauchet, C., Labadie, G. R., and Poulter, C. D. (2006). Regio- and chemoselective covalent immobilization of proteins through unnatural amino acids. J. Am. Chem. Soc. 128, 9274–9275. GE Healthcare (ed.) (2007). Affinity chromotography, principle and methods. GE Healthcare. Guilbault, G. G. (ed.) (1988). In ‘‘Analytical uses of immobilized biological compounds for detection, medical and industrial uses,’’ Dordrecht, Boston, MA. Guss, B., Eliasson, M., Olsson, A., Uhlen, M., Frej, A. K., Jornvall, H., Flock, J. I., and Lindberg, M. (1986). Structure of the IgG-binding regions of streptococcal protein G. EMBO J. 5, 1567–1575. Gustavsson, P.-E., and Larsson, P.-O. (2006). Support materials for affinity chromatography. In ‘‘Handbook of Affinity Immobilization,’’ (D. S. Hage, ed.), pp. 15–34. CRC press/ Taylor and Francis Group, Boca Raton, FL. Hage, D. S. (ed.) (2006). In ‘‘Handbook of affinity chromatography,’’ CRC press/Taylor and Francis Group, Boca Raton, FL. Hage, D. S., Bian, M., Burks, R., Karle, E., Ohnmachi, C., and Wa, C. (2006). Bioaffinity Chromatography. In ‘‘Handbook of Affinity Immobilization,’’ (D. S. Hage, ed.), pp. 101–126. CRC press/Taylor and Francis Group, Boca Raton, FL. Hermanson, G. T. (ed.) (1992). In ‘‘Immobilized affinity ligand techniques,’’ Academic Press, San Diego, CA. Hirabayashi, J. (2008). Concept, strategy and realization of lectin-based glycan profiling. J. Biochem. 144, 139–147. Koneracka, M., Kopcansky, P., Timbo, M., Ramchand, C. N., Saiyed, Z. M., and Trevan, M. (2006). Immobilization of enzymes on magnetic particles. In ‘‘Immobilization of Enzymes and Cells,’’ ( J. M. Guisan, ed.), pp. 217–228. Humana Press. Korpela, T., and Hinkkanen, A. (1976). A simple method to introduce aldehydic function to agarose. Anal. Biochem. 71, 322–323. Kowal, R., and Parsons, R. G. (1980). Stabilization of proteins immobilized on Sepharose from leakage by glutaraldehyde crosslinking. Anal. Biochem. 102, 72–76. Labrou, N., and Clonis, Y. D. (1994). The affinity technology in downstream processing. J. Biotechnol. 36, 95–119. Labrou, N. E. (2000). Dye-ligand affinity chromatography for protein separation and purification. Methods Mol. Biol. 147, 129–139. Labrou, N. E. (2003). Design and selection of ligands for affinity chromatography. J. Chromatogr. B Analyt. Technol. Biomed. Life Sci. 790, 67–78. Labrou, N. E., and Clonis, Y. D. (2002). Immobilized synthetic dyes in affinity chromatography. In ‘‘Theory and Practice of Biochromatograph,’’ (M. A. Vijayalakshimi, ed.), pp. 235–251. Taylor and Francis Group, London.
Overview of Affinity Chromatography Methods
Labrou, N. E., Karagouni, A., and Clonis, Y. D. (1995). Biomimetic-dye affinity adsorbents for enzyme purification: Application to the one-step purification of Candida boidinii formate dehydrogenase. Biotechnol. Bioeng. 48, 278–288. Langone, J. J. (1982). Applications of immobilized protein A in immunochemical techniques. J. Immunol. Methods 55, 277–296. Lerman, L. S. (1953a). A biochemically specific method for enzyme isolation. Proc. Natl. Acad. Sci. USA 39, 232–236. Lerman, L. S. (1953b). Antibody chromatography on an immunologically specific adsorbent. Nature 172, 635–636. Los, G. V., Encell, L. P., McDougall, M. G., Hartzell, D. D., Karassina, N., Zimprich, C., Wood, M. G., Learish, R., Ohana, R. F., Urh, M., Simpson, D., Mendez, J., et al. (2008). HaloTag: A novel protein labeling technology for cell imaging and protein analysis. ACS Chem. Biol. 3, 373–382. Lowe, C. R., Lowe, A. R., and Gupta, G. (2001). New developments in affinity chromatography with potential application in the production of biopharmaceuticals. J. Biochem. Biophys. Methods 49, 561–574. Madoery, R., and Minchiotti, M. (2006). Cibacron Blue-Eupergit, an affinity matrix for soybean (Glycine max) phospholipase A2 purification. Enzyme and Microb. Technol. 38, 869–872. Mallik, R., and Hage, D. S. (2006). Affinity monolith chromatography. J. Sep. Sci. 29, 1686–1704. Mallik, R., Wa, C., and Hage, D. S. (2007). Development of sulfhydryl-reactive silica for protein immobilization in high-performance affinity chromatography. Anal. Chem. 79, 1411–1424. O’Shannessy, D. J., and Wilchek, M. (1990). Immobilization of glycoconjugates by their oligosaccharides: Use of hydrazido-derivatized matrices. Anal. Biochem. 191, 1–8. Oates, M. R., Clarke, W., Marsh, E. M., and Hage, D. S. (1998). Kinetic studies on the immobilization of antibodies to high-performance liquid chromatographic supports. Bioconjug. Chem. 9, 459–465. Ohana, R. F., Encell, L. P., Zhao, K., Simpson, D., Slater, M. R., Urh, M., and Wood, K.V (2009). Halotag7: A genetically engineered tag that enhances bacterial expression of soluble protein and improves protein purification. Protein Expr. Purif. 68, 110–120. Ostrove, S. (1990). Affinity chromatography: General methods. Methods Enzymol. 182, 357–371. Plievaa, F. M., De Seta, E., Galaevb, I. Y., and Mattiasson, B. (2009). Macroporous elastic polyacrylamide monolith columns: Processing under compression and scale-up. Sep. Purif. Technol. 65, 110–116. Porath, J., Axen, R., and Ernback, S. (1967). Chemical coupling of proteins to agarose. Nature 215, 1491–1492. Renkin, E. M. (1954). Filtration, diffusion, and molecular sieving through porous cellulose membranes. J. Gen. Physiol. 38, 225–243. Saiyed, Z., Telang, S., and Ramchand, C. (2003). Application of magnetic techniques in the field of drug discovery and biomedicine. Biomagn. Res. Technol. 1, 2. Starkenstein, E. (1910). Ferment action and the influence upon it of neutral salts. Biochem. Z. 24, 210–218. Stellwagen, E. (1990). Chromatography on immobilized reactive dyes. Methods Enzymol. 182, 343–357. Urh, M., Hartzell, D., Mendez, J., Klaubert, D. H., and Wood, K. (2008). Methods for detection of protein-protein and protein-DNA interactions using HaloTag. Methods Mol. Biol. 421, 191–209. Vijayendran, R. A., and Leckband, D. E. (2001). A quantitative assessment of heterogeneity for surface-immobilized proteins. Anal. Chem. 73, 471–480.
Marjeta Urh et al.
Wilchek, M., Knudsen, K. L., and Miron, T. (1994). Improved method for preparing N-hydroxysuccinimide ester-containing polymers for affinity chromatography. Bioconjug. Chem. 5, 491–492. Wilchek, M., Miron, T., and Kohn, J. (1984). Affinity chromatography. Methods Enzymol. 104, 3–55. Zachariou, M. (ed.) (2007). In ‘‘Affinity Chromatography, Methods and Protocols,’’ Humana Press, Totowa, NJ.