[29] AFFINITY CHROMATOGRAPHY: GENERAL METHODS 357

to the design and synthesis of new reactive dyes outside of the commer-
cial dye industry. It was anticipated that the blue dye could better occupy
the coenzyme site if the anthraquinone and the triazine rings were further
separated by insertion of an ethylene bridge as shown in Fig. le, This
immobilized dye, 24 in contrast to all other immobilized dyes and biomole-
cules, is able to resolve the purified alcohol dehydrogenase into two com-
ponents of different activity, with the low activity form having a covalent
modification on a lysine side chain. It is anticipated that additional de-
signed dyes will increase the capability of immobilized dyes in protein
purification. The chemistry employed in the synthesis of reactive dyes
which is necessary for preparation of designed dyes is described in detail
in two texts. 25,26
Finally, it should be noted that the solid fluorocarbons developed by
du Pont de Nemours & Co. (Wilmington, DE) afford a promising new
matrix for immobilized dye chromatography. A reactive dye is first sub-
jected to a perfluoralkylation and then the perfluoroalkylated dye is essen-
tially irreversibly adsorbed to a fluorocarbon surface. 27 Accordingly, this
matrix should prevent slow bleeding of dye into proteins purified by im-
mobilized dye chromatography.
Most aspects of immobilized reactive dye-protein interaction have
been reviewed by several authors 1.2,7,8,18and the concerned investigator is
urged to pursue them for access to more detailed information.

24 C. R. Lowe, S. J. Burton, J. C. Pearson, and Y. D. Clonis, J. Chromatogr. 376, 121

(1986).
22 W. F. Beech, "Fibre-Reactive Dyes." Logos Press, London, 1970.
~6 K. Venkataraman, "The Chemistry of Synthetic Dyes," Vol. 6, Academic Press, New
York, 1972.
27 j. V. Eveleigh, Abstr. Biotechnol. Microsymp. Macromol. Interact. Affinity Chromatogr.
Technol., Mogilany, Pol. (1988).

[29] A f f i n i t y C h r o m a t o g r a p h y : G e n e r a l M e t h o d s
By STEVEN OSTROVE

Affinity chromatography is one of the most powerful procedures that

can be applied to protein purification. Over the years there have been
many good books on the subject and many reviews of the theory of
affinity chromatography.l,2 This chapter will not be a further review of the
I j. Turkova, ed., "Affinity Chromatography." Elsevier, New York, 1978.
2 p. Mohr and K. Pommerening, "Affinity Chromatography: Practical and Theoretical As-
pects." Dekker, New York, 1985.

Copyright© 1990by AcademicPress, Inc.

METHODS IN ENZYMOLOGY,VOL. 182 All rightsof reproductionin any formreserved.
358 PURIFICATION PROCEDURES: CHROMATOGRAPHIC METHODS [29]

theory behind this technique nor will it review all of the proteins that have
been separated by this technique since there are so many. Discussion will
center on practical ideas and general considerations behind choosing a
matrix, coupling a ligand, and some of the more common problems often
encountered with the implementation of this procedure. Suggestions in
this chapter should allow for the easier and more productive use of affinity
chromatography. As the specific purification system for the sample of
interest is developed, expect modification of these procedures. This chap-
ter serves only as a guide to performing separations of biomolecules by
affinity chromatography. Areas where the manufacturer of the affinity
product can be contacted for specific recommendations on the use of the
product are indicated throughout. Refer specific questions to the manu-
facturer.
Affinity chromatography as a biospecific technique began only about
20 years ago even though it had been used as an experimental separation
procedure for many years. 3 This procedure takes advantage of one or
more biological properties of the molecule(s) being purified. These inter-
actions are not due to the general properties of the molecule such as
isoelectric point (pl), hydrophobicity, or size. This highly specific method
of separation utilizes the specific reversible interactions between biomole-
cules.
Some of the biological properties that can be exploited to effect a
separation include specific shapes (that "recognize" other molecules
such as receptors or enzymes), specific changes in conformation after
changes in pH, or certain subareas or regions of the molecule that can
interact or bind to other molecules (e.g., epitopes of antibodies).
When developing a separation scheme keep in mind that the sample of
interest is not the only component in the sample mixture that can be
bound to an affinity matrix (gel). One affinity matrix may be specific for
the sample of interest while others may be more specific for other compo-
nents in the mixture (contaminating proteins). Just as the sample of inter-
est can be bound to an affinity matrix, the contaminating proteins may
also be specifically bound. An affinity gel could be chosen to bind the
contaminating proteins, allowing the sample of interest to pass through
the gel in the wash volume. This method of separation could result in a
great saving of time.

Matrix
Choice of the proper matrix is a very important step in any chromato-
graphic process. A good matrix for affinity chromatography should have
the following properties:
3 p. Cuatrecasas and M. Wilchek, Biochem. Biophys. Res. Commun. 33, 235 (1968).
[29] AFFINITY CHROMATOGRAPHY: GENERAL METHODS 359

1. Hydrophilic: Reduce the nonspecific interactions.

2. Large pores: Allow all areas of the matrix to be available to most of
the molecules in the mixture. Some matrices allow binding only to the
outer surface. This latter type of matrix is useful in separating very large
molecules, cells, or viruses.
3. Rigid: The matrix must withstand the pressures of packing and
solvent flow during elution or washing.
4. Inert: The matrix should not contribute to the separation.
5. Chemical stability: The matrix must be stable to all solvents used in
the separation.
Base the choice of an affinity gel on both the ligand and the sample.
There are two major types of affinity gels: group-specific gels and cova-
lent coupling gels. The former are usually supplied ready to use. Table I
provides examples of ligands that are group specific in action and can be
used to isolate whole families of biomolecules which share common prop-
erties. Covalent coupling gels (Table II) require more chemistry and some
specific considerations. First, consider the length and type of the spacer
arm; second, the coupling chemistry.

TABLE I
LIGAND SPECIFICITY

Ligand Specificity

NAD, NADP Dehydrogenases

Lectins Polysaccharides
Poly(U) Poly(A)
Poly(A) Poly(U)
Histones DNA
Protein A Fc antibody
Protein G Antibodies
Lysine rRNA, dsDNA, plasminogen
Arginine Fibronectin, prothrombin
Heparin Lipoproteins, DNA, RNA
Blue F3G-A NAD ÷
Red HE-3B NADP ÷
Orange A Lactate dehydrogenase
Benzamidine Serine proteases
Green A CoA proteins, HSA, dehydrogenases
Gelatin Fibronectin
Polymyxin Endotoxins
2' ,5'-ADP NADP ÷
Calmodulin Kinases
Boronate cis-Diols, tRNA, plasminogen
Blue B Kinases, dehydrogenases, nucleic acid-binding proteins
360 PURIFICATION PROCEDURES" CHROMATOGRAPHIC METHODS [29]

TABLE II
COUPLING CHEMISTRY

Ligand Spacer Active

Linkage group length pH Specificity

CNBr NH2 8-10 Proteins,peptides

Thiolpropyl SH Equivalentto about 9-11 Sulfhydryls
13 carbons
Thio SH 9-13 Sulfhydryls
Epoxy NH2 Equivalentto about 9 Proteins,peptides
OH 11 carbons 10 Carbohydrates
SH 11 Sulfhydryls
Tresyl NH2 8-10 Proteins,peptides
Aminohexyl COOH 6 Amino acids, proteins
Carboxyihexyl NH2 6 Carboxylic acids

Solvents
The solvent system chosen for the entire affinity chromatography sep-
aration is also a critical factor to a good separation. The solvent should
not degrade the sample. Unfortunately, avoiding denaturing solvents is
not always possible. For example, separation of an antibody (IgG)-anti-
gen complex requires some very harsh conditions. Dissociation at a low
pH or use of a strong chaotropic agent are the most commonly used
methods. Minimizing the time of contact with these agents is vitally im-
portant. One method used to reduce the contact time with harsh reagents
(e.g., low pH) is to add Tris base (dry) to the collecting tubes. This will
rapidly increase the pH and help to protect the sample.
Try to choose an elution buffer specific for the sample (e.g., a buffer
containing an analog to the sample). The elution buffer should release the
sample safely and rapidly. Again, the buffer should not denature the
sample, nor cause any change in its specific activity or function. Optimi-
zation of sample binding and elution conditions is usually by trial and
error. When choosing a buffer system try to avoid using one that has a
pKa at or near the pI of the sample. This will help prevent precipitation of
the sample on the column. However, when starting a separation read the
literature, as it will often provide a good starting point. Even a related
separation can serve as a starting point for selecting the separation condi-
tions.

Spacer Arms
Choosing a gel with or without a spacer arm depends on the ligand, the
sample, and the binding chemistry. A spacer arm is used to keep the
[29] AFFINITY CHROMATOGRAPHY: GENERAL METHODS 361

TABLE III
SPACER LENGTH CONSIDERATIONS

Best
Ligand Protein spacer arm

Small Small Short

Small Large Long
Large Small None
Large Large None

ligand away from the matrix so that the active site of the ligand is avail-
able to the sample. This is especially important with small ligands. As a
general rule if the ligand is large and the sample is small (low molecular
weight) this spacing may be unnecessary. With samples of high molecular
weight a spacer arm can be used to limit steric hindrance and increase the
availability of the active site (Table III).
A wide variety of spacer arm lengths are available. If unsure of the
required spacer arm length, start with one that is equivalent to about six
carbon atoms. This seems to be a good length for many affinity applica-
tions. 4 Shorter arms give less flexibility so the ligand will not " w a v e "
around in the medium.
Predicate the spacer arm length on the amount of steric hindrance
deemed tolerable. As spacers are evaluated, remember that the spacer
molecule itself can cause steric hindrance by blocking adjacent active
sites on the gel; thus, longer is not always better.

Gel Preparation
After choosing the affinity gel type, prepare the resin (gel) for use. The
manufacturer will usually supply the instructions needed to prepare the
gel correctly. However, short of those instructions, following these gen-
eral guidelines will help ensure a successful preparation.
First, calculate the amount of gel that needs to be packed into a
column (or flask for batch work) by the capacity of the gel for the sample.
That is, x units of gel bind y units of sample. Next, calculate the volume:
(total sample/sample units) × gel volume per sample unit. This value
should be multiplied by a factor of 2-3 and this factored amount of gel
should be used. For example, if I ml of protein A-Sepharose binds 20 mg
IgG and you are using 40 mg of sample (with contaminants), divide the
total (40 mg) by the gel capacity (20 rag) and multiply by the gel volume

4 C. R. Lowe, Biochem. J. 133, 499 (1973).

362 PURIFICATION PROCEDURES: CHROMATOGRAPHIC METHODS [29]

(1 ml). Multiply this figure by 2-3, giving a final gel volume of 4-6 ml, for
best results.
If the gel is supplied in a preswollen state, reconstituting the gel is
unnecessary to obtain the full swollen volume. A wash is all that is usually
required. The swollen gel is typically supplied in glycerin or similar mate-
rial which is used to help in the gel preparation and to stabilize the ligand
or activated coupling complex. Wash on a sintered glass filter of medium
grade (#3) or on a membrane-type filter that has a low protein-binding
capacity. This allows easy removal of the washed gel with a minimum of
loss due to sticking to the filter. A wash ratio of about 200 : 1 (buffer to gel)
works well. For the safest wash buffer use either distilled water or the
starting buffer (unless otherwise directed). By definition the starting
buffer is that buffer used to initially prepare the matrix for the addition of
the sample. It creates an environment on the gel so that the sample will
bind specifically to the attached ligand.
If the gel needs to be swollen to regain full working volume, then use a
swelling buffer prior to washing. This buffer is often a low concentration
phosphate buffer (0.1 M) at or near neutral pH. Swelling times vary
between 15 min and 1 hr. After swelling, wash the gel either in the buffer
solution used for swelling, distilled water, or starting buffer.
Since washing and swelling buffers are generally pulled through the gel
under a low vacuum, it is critical that the gel does not become dry at this
stage. Following the reswelling and the wash, the ligand can be bound to
the gel or loaded into a column if no ligand is to be added (i.e., group-
specific gels).

Coupling or Linkage Chemistry

Before using a covalent coupling gel, the ligand-binding (linkage)
chemistry needs to be decided. There are a variety of linking groups
available, such as cyanogen bromide (CNBr), tresyl, epoxy, and triazine.
The linkage chemistry may be available either in an activated or nonac-
tivated form. Activated means " r e a d y " to use without additional chemi-
cal activation steps (washing is still necessary). The nonactivated gels
require some additional chemical activation step, such as carbodiimide
treatment, to prepare them for binding the ligand. Leaching, or loss of
ligand, after binding is inevitable. The trick is to minimize this loss.
CNBr-type linkages commonly leach more than do tresyl or epoxy
linkages.
Other types of ligand linkages are also possible by using C~---C, C ~ O ,
and other available bonds. Nucleic acids and sugars can also be bound
through their amine or hydroxyl groups. Table II lists some examples of
[29] AFFINITY CHROMATOGRAPHY: GENERAL METHODS 363

the active groups that can be used to link ligands to the matrix, and the
types of sample that can bind to these ligands.

Coupling the Ligand

Next, couple the ligand to the matrix, assuming that you have a spe-
cific ligand that needs to be attached. Consider the steric effects of the
ligand, the spacer arm, and the sample. High concentrations of small
ligands can block some active sites on the matrix, causing a lower binding
efficiency. Large ligands can also block adjacent sites, again resulting in
lower binding efficiency. Spacer arms and ligands can also cause some
blockage of adjacent sites as the ligand " w a v e s " back and forth. As a
general rule use about 10 mg ligand/ml of gel. This amount applies also to
proteins having an average size of (50 kDa). A lower amount should be
applied for larger molecules such as IgG (5 mg/ml), or IgM (1 mg/ml), or
molecules with low dissociation constant (KD) values.
Mix the ligand and the matrix together using a rotating motion. Avoid
magnetic stirrers at all times since they can damage the matrix. The
volume ratio of binding buffer (with ligand) to gel matrix should be about
2:1 for best results. Carefully control the pH, ionic strength, and ion
content during this stage of coupling.
Coupling times of 2-4 hr at room temperature or up to 16 hr in the cold
(4°) are commonly used. The choice of time and temperature is deter-
mined by the stability of the ligand and the amount of time available. The
time available is important since there should be no processing interrup-
tions from the time the gel is activated until the excess ligand is washed
out. This minimizes loss of coupling activity. Once again users should
consult the manufacturer's instructions for the optimum coupling condi-
tions.
Coupling of the ligand to the matrix can be by a single point or multi-
point attachment. An example of single-point attachment is the binding of
a single primary amine via CNBr coupling. This type of linkage offers the
best flexibility to the ligand and thus the most accessibility of the active
site to the sample. Single-point attachment is possible only if secondary
and tertiary amines are blocked. Multipoint attachments are stronger than
single-point attachments and are less likely to leach during the run. How-
ever, this type of coupling often causes blockage of the active site of the
ligand.

Blocking Unreacted Groups

After incubation of the ligand with the matrix, remove the excess
ligand and block the unreacted sites on the matrix. When coupling a
364 PURIFICATION PROCEDURES" CHROMATOGRAPHIC METHODS [29]

ligand to the matrix some sites on the matrix remain unreacted. These
unreacted sites are potential sites for nonspecific interactions with the
sample or contaminants in the solution. Blocking these sites is most easily
accomplished with reagents that have an opposite charge or can be cova-
lently linked to the matrix. For example, when a carboxyl group (COO-)
is used to couple a ligand, such as an amino group (NH2), use a Tris or
ethanolamine solution (0.1-I.0 M) as the blocking agent. When NH2
groups are the coupling sites for ligand containing COO-, acetic acid can
be used as the blocking agent.
The concentration of the blocking agent should be in excess of the
total reactive site concentration on the matrix. This assures complete
blockage of all unreacted sites. Normally, a 5- to 10-fold excess over the
ligand concentration is sufficient.
Control of the pH of the blocking agent is another critical factor impor-
tant to good affinity separations. A pH that is either too high or too low
may prevent complete blocking or even destroy the matrix or the bound
ligand. The blocking reaction is usually done at room temperature for 2-4
hr, but can also be done in the cold (4°) for longer periods of time. Wash
out the excess blocking agent and equilibrate the column with 5 to 10
column volumes of starting buffer before sample application.
The coupling process can be summarized as follows:
1. Swell the matrix in swelling buffer (15-60 min).
2. Wash the matrix (200: 1, buffer:gel).
3. Add ligand and incubate with mixing (2-4 hr, 2 : 1 buffer : gel).
4. Wash out unused ligand and buffer (200: 1, buffer:gel).
5. Block unreacted sites on the matrix (2-4 hr, room temperature).
6. Wash the matrix (200: 1, buffer:gel).
7. Use or store the gel under appropriate conditions (4-8°).

Monitoring Coupling Efficiency

The extent of ligand coupling determines both the efficiency of the
separation and the amount of purified sample that can be prepared. The
amount of ligand bound can be determined in several ways:
1. Measure the difference in UV absorption before and after coupling:
a. A2s0 is best for proteins; however, other wavelengths specific for
other ligands which can be coupled should be chosen. For exam-
ple, A4o5 for heme groups can be used. This technique is best
accomplished when the concentration of ligand is not very high,
since a small amount of ligand binding will not be detected. In
[29] AFFINITY CHROMATOGRAPHY: GENERAL METHODS 365

addition, avoid buffer components that absorb at this wavelength

[e.g., dithiothreitol (DTT)], or high concentrations of Tris buffer.
b. Perform a colorimetric protein assay (see [6]).
c. Perform a fluorescence detection assay (more sensitive than the
colorimetric assays).
. Dissolve a portion of the gel containing the ligand and do a protein
assay, amino acid analysis, or assay for total nitrogen.
3. Perform an activity test (a small binding experiment).
4. Perform a radiolabeled ligand or RIA test (assumes that the radiola-
beled ligand binds at the same rate, and to the same extent, as
nonradiolabeled ligand).

Binding the Sample

The binding of proteins to a ligand, through the carboxyl or the amino
groups available, is based on the specific affinity of the protein for a
particular ligand. As indicated previously, the ligand should not be cou-
pled in such a way as to block or interfere with the availability of the
active site on either the ligand or the protein. Binding between the ligand
and the protein is generally noncovalent. Although the binding is specific,
the forces involved are general chemical interactions, such as hydrogen
bonds. The buffer conditions that are used to load the protein on to the
column are often phosphate or Tris buffers (0.1-0.2 M) containing salts
such as sodium chloride (0.5 M). The choice of buffer and concentration
is predicated on minimizing nonspecific interactions and maximization of
the specific attraction between the ligand and the protein. Load the sam-
ple in the normal downward direction such that it will bind to the upper
half to upper third of the matrix.

Flow Rates
Different flow rates are used in the various stages of every affinity
chromatography run: (1) the loading of the sample, (2) the wash step to
remove nonspecifically bound material, (3) the elution phase, where the
protein of interest is removed from the gel, and (4) the regeneration of the
matrix for the next run.
The flow rate commonly used for loading of the sample is often about
10 cm/hr. The notation cm/hr refers to linear flow rate of the buffer. To
calculate the volume flow rate, which is the rate that is used for the run,
multiply the linear flow rate by the cross-sectional area of the column.
The flow rate used is dependent upon the kinetics of the binding of the
366 PURIFICATION PROCEDURES: CHROMATOGRAPHIC METHODS [29]

desired protein to the ligand. Factors such as temperature, concentration,

and the KD influence the interaction of the protein and ligand. If the
protein has a high affinity for the ligand (KD < 10-6), the flow can be faster
since the protein still can bind effectively. On the other hand, if the
affinity is low, use a much slower flow rate.
During the wash step the flow rate can be considerably faster since the
wash serves to remove nonspecifically adhering material (assumed for
this discussion not to be of interest). The flow rate at this stage can be
increased to about 20 to 50 cm/hr in order to effect a rapid cleaning of the
matrix. However, if the protein of interest is loosely bound to the ligand, a
lower flow rate is better.
Perhaps most important is the flow rate during the elution phase. This
flow could also be faster then the loading rate. Elution flow rates depend
on the strength of the elution buffer as related to the affinity of the sample.
The goal is to use a buffer that will easily strip the desired protein from the
ligand without damaging it, the ligand, or the matrix. The elution flow rate
can be as high as the wash flow, but is always lower than the flow rate
used for packing the matrix.
The flow rate during reequilibration can also be very rapid. At this
point, only the coupled ligand should be left on the matrix. Therefore,
flow rates up to the packing rate can be used to save time.
Flow rates should not exceed about 80% of the flow rate used to pack
the resin. This avoids compression during the chromatography run. An-
other factor that determines the maximum flow rate is the stability of the
matrix. In order to avoid gel compression and deformation of the beads do
not exceed the maximum flow rate recommended for the matrix. Also, try
to avoid turbulence and high shear rates due to rapid buffer flow in the
matrix when loading or eluting. The manufacturer can generally provide
information on the best flow rates for all steps.

Nonspecific Interactions
Nonspecific interactions, if a problem, can usually be avoided by using
a salt concentration between 0.1 to 0.5 M since in this range nonspecific
ionic attractions are greatly reduced. These salt concentrations are usu-
ally not so high as to make hydrophobic interactions between the protein
and the matrix or ligand a problem. As always, the manufacturer of the
matrix usually can supply the specific information needed to prevent non-
specific interactions.
Other methods that may be used to decrease nonspecific interactions
include the addition of agents such as glycerol up to a concentration of
about 10% (no higher due to increased viscosity of the buffer, resulting in
[29] AFFINITY CHROMATOGRAPHY: GENERAL METHODS 367

higher back pressures and lower flow rates). Low levels of detergent
(below the critical micelle concentration) are also useful in reducing non-
specific interactions. Detergents, however, can interfere with both ligand
and protein binding, and not every ligand can be safely used with all
detergents. For example, a ligand may dissociate or denature in the pres-
ence of detergents and some may interact with the active site of the
ligand, lowering its affinity for the protein.

Specific vs. Nonspecific Elution

Specific elution of the protein of interest is always the best method to
use in affinity chromatography (see Table IV). This type of elution is the
result of a competitive action of the eluent for the ligand. An eluent is
chosen that has a greater affinity for the ligand than the protein so that it
will displace the protein from the ligand. The eluent can then be removed
by a more stringent cleaning of the matrix. An example of a specific eluent
is the use of c~-methylglucoside to elute samples from concanavalin A
(ConA)-Sepharose.
If there are no known specific eluents for the protein of interest then
nonspecific elution may be used (e.g., elution using a salt gradient, chang-
ing the pH or temperature). Design conditions so that the protein of
interest is eluted separately from the majority of contaminating proteins.
One procedure is to raise the eluting buffer concentration to a level just
below that at which the desired protein starts to be eluted, followed by an

TABLE IV
ELUT1ON CONDITIONS

Ligand Eluent Specific Nonspecific

Protein A Acetic acid X

Glycine
ConA c~-D-Methylmannoside X
Borate buffer X
C~-D-Methylglucoside X
Lysine Temperature X
Salt X
Blue dye Salt X
Urea X
Gelatin Arginine X
pH X
5'-AMP NAD ÷, NADP ÷ X
Salt X
368 PURIFICATION PROCEDURES: CHROMATOGRAPHIC METHODS [29]

increasing shallow gradient. This results in sharper peaks and greater

purity. If there are any proteins remaining bound to the matrix after the
protein of interest has been completely eluted, the elution buffer strength
can be increased rapidly to remove this remaining material.
Elution is most often done in the "forward" direction, i.e., the same
direction as sample application. Ideally, the sample should bind to the
upper third to half of the column. Molecules with the highest affinity for
the ligand will bind near the top of the column, while the remainder will
bind further down the column as the affinity decreases.
If the protein of interest is bound near the top of the column then the
rest of the proteins can be more easily eluted in the forward direction,
leaving the protein of interest on the column. Even if this protein moves
down during the preliminary elution it will not come off the column. At
this point if the eluent flow can be reversed and a strong eluent used, the
sample can also be eluted off the top of the column in a sharper peak and
in a shorter time, thus limiting exposure of the protein to potentially harsh
conditions. Such a flow reversal can be accomplished by turning the
column upside down, or using valves to allow the eluent to flow from the
bottom of the column to the top. When reversing the direction of flow in a
column always make sure that flow adaptor are used to prevent the loss of
the matrix through the top of the column.
The use of reversed elution to yield a more concentrated sample is
valid only in certain situations. These occur when the desired protein
binds more strongly than the other proteins and when it is bound to the
upper half of the matrix. In all other cases it is still best to elute the sample
in the forward direction.
Measurement of the elution profile is usually done by monitoring pa-
rameters such as the Aaso or fluorescence. Specific assays for the protein
of interest, such as enzyme activity, can yield information on the concen-
tration and the condition of the separated material, and should be used
whenever possible.
Detection of any ligand that has leached off the matrix is usually
difficult and requires specific assays. Radioimmunoassays (RIA) for li-
gand or matrix material are often useful in these situations. Electrophore-
sis (i.e., S D S - P A G E and immunoelectrophoresis) may also be used to
determine protein purity, activity, and the extent of leaching.

Regeneration
Thoroughly clean resins prior to their reuse. Regeneration means that
any material that remains on the resin must be removed and the gel
reequilibrated with starting buffer. If all the material is not removed, and
the ligand not properly prepared, the efficiency of the gel will be impaired.
[29] AFFINITY CHROMATOGRAPHY: GENERAL METHODS 369

This will result in less material binding in successive runs and a concomi-
tant loss in resolution. With proper care an affinity resin can be used
multiple times. The actual number of uses depends on the sample, the
ligand, and the elution conditions.
Clean the gel with either a higher concentration of the specific eluent
or by using a high concentration of a nonspecific agent such as sodium
chloride (e.g., 0.5-1 M). Increasing the salt concentration is usually effec-
tive in removing nonspecifically bound material as well as some of the
specifically bound sample that may be left on the resin. Take care not to
damage the bound ligand or to alter its activity. In some cases high salt
levels cause proteins to change their conformation. 5 If the ligand is a
protein, its active site may be altered, causing it to lose some or all of its
binding capacity or affinity for the sample.
Some procedures are gentle enough for almost all gels. A general
regeneration scheme (recommended by Pharmacia LKB Biotechnology,
Piscataway, NJ) 6 follows. However, if the manufacturer provides a spe-
cific method for regeneration, then follow their advice:
1. Wash with 10 column volumes of Tris-C1 (0.1 M, pH 8.5) containing
0.5 M NaCI.
2. Follow with 10 column volumes of sodium acetate (0.1 M, pH 4.5)
containing 0.5 M NaCI.
3. Reequilibrate with 10 column volumes of starting buffer. Any re-
generation procedure requires buffer volumes up to 10 times the column
volume. This assures that all areas of the resin have been reached and
cleaned.
Be sure those cations or anions needed for ligand stability are added to
the regeneration buffer. These ions are usually present in the start and
elution buffers, but are often overlooked in the regeneration buffer. For
example, ConA requires Ca 2÷ and Mg 2÷ or Mn 2÷ at concentrations of 1-
10 mM to maintain its tertiary structure.
Resins are usually stable to most regeneration buffers, but if in doubt
check with the resin manufacturer.

Sterilization
If the sample is to be kept sterile, the affinity column and gel must also
be sterilized. Take special care throughout the entire process to assure the
maintenance of this sterility. The gel can often be sterilized by autoclav-
5 p. H. von Hippel and T. Schleich, in "Structure and Stability of Biological Macromole-
cules" (S. N. Timasheff and G. D. Fasman, eds.). Dekker, New York, 1969.
6 Pharmacia LKB, "Affinity Chromatography: Principles and Methods," p. 88. Piscataway,
New Jersey, 1983.
370 PURIFICATION PROCEDURES: CHROMATOGRAPHIC METHODS [29]

ing, and is more easily accomplished before ligand attachment. The ligand
can be filter sterilized before coupling to the resin. Coupling can be done
under aseptic conditions such as with sterile buffers, air, and vessels.
Depyrogenation of the matrix with agents such as sodium hydroxide is
also best done prior to ligand attachment.
If sterilization is necessary after the ligand is attached, take care to
avoid altering the ligand or the linkage to the matrix. Autoclaving is not
usually feasible at this stage since proteins and most other biological
material are denatured under these conditions. Possible solutions to the
sterilization of sensitive gels include radiation treatment or chemical ster-
ilization. One gentle method for the sterilization of a sensitive gel-ligand
system follows7:
1. Equilibrate the column with 2% chlorhexidine diacetate and 0.2%
benzoyl alcohol.
2. Let stand for 4 days.
3. Wash with sterile buffer; a neutral phosphate buffer or the start
buffer can be used.
4. Reequilibrate with the chlorhexidine diacetate (2%) and benzoyl
alcohol (0.2%).
5. Rewash with the sterile buffer.
6. Store in 0.5% chlorhexidine diacetate and 0.05% benzoyl alcohol.

Gel Storage
Storage of the gel after preparation is usually quite easy. The actual
conditions used for proper storage are dependent on the ligand that is
bound to the matrix. In general, 4° is the preferred temperature. This
lowers the possibility of bacterial growth and does not harm either the
matrix or the ligand. Avoid freezing since this may rupture the matrix.
It is best not to store a gel in the middle of the coupling process. This is
especially true with CNBr gels, since they will lose activity rapidly at the
pH used for activation. Clean the gel before storage by removing all
residual material that is known to adhere to the column. This will allow
for easier reuse of the matrix.
In general, store the gel at temperatures below 8 ° but not frozen. Store
all affinity resins in the presence of antibacterial agents such as chlorhex-
idine digluconate (or acetate), sodium azide, 20% ethanol, and thimerosal
(do not use this with SH-active ligands, e.g., thiolpropyl-Sepharose).
Base the choice of the antibacterial agent on the stability of the ligand to
7 S. S. Block, ed., "Disinfection, Sterilization, and Preservation." Lea & Fabiger, Philadel-
phia, Pennsylvania, 1977.
[30] AFFINITY CHROMATOGRAPHY" SPECIALIZED TECHNIQUES 371

that agent and the charge characteristics of the gel-ligand combination. Is

it anionic, cationic, or neutral? Is it temperature sensitive? Is it subject to
degradation by enzymatic action?
To maintain the same level of activity after storage as existed previ-
ously, choose a bacteriostatic agent that will not bind to the gel matrix or
ligand, and one that is easily washed out when the gels are reused (e.g.,
ethanol). Carefully remove all of the storage solution prior to reuse to
prevent denaturation of the sample.
Do not freeze the gel at any time. This will disrupt the matrix and can
lead to fine particles that can interfere with the buffer flow. Again, follow
the manufacturer's advice for proper storage.

[30] A f f i n i t y C h r o m a t o g r a p h y : S p e c i a l i z e d T e c h n i q u e s
By STEVEN OSTROVE and SHELLY WEISS

This chapter discusses some specialized affinity chromatography tech-

niques: cell affinity chromatography, metal chelate affinity chromatogra-
phy, covalent affinity chromatography, and other binding techniques and
the scaling up of affinity chromatography. It will be a guide in the use of
these techniques and give a start in understanding the reasons behind
their use. In addition, some of the possible problems and danger areas
associated with these techniques are described.
Not all of the specific methodologies available for separation by affin-
ity chromatography will be reviewed in this chapter, nor will it provide an
exhaustive list of examples for each technique. As you read this chapter,
and try to use the techniques, however, you will find new and different
ways to accomplish your separation task.
Certain assumptions need to be made before we begin: First, that you
are aware of general affinity chromatography procedures; second, that
you know how some parameters such as temperature, pH, ionic strength,
and flow rates affect affinity separations (see [29] in this volume).

Cell Affinity Chromatography

Isolating cells by affinity chromatography requires some special con-
siderations due to the size and sensitivities of the living cell. Cells can be
separated by affinity chromatography in two ways: either by binding the
cell directly to the matrix as one binds a protein, or by binding a protein or

Copyright © 1990 by Academic Press, Inc.

METHODS IN ENZYMOLOGY, VOL. 182 All rights of reproduction in any form reserved.