Affinity chromatography is far more specific than other purification techniques. It relies on the preparation of a matrix to which the protein of interest, and preferably only this protein, will bind reversibly. The matrix is usually beaded agarose (Sepharose or BioGel A), polyacrylamide (e.g. BioGel P) or cross-linked dextran (e.g. Sephacryl) to which a ligand has been covalently attached. The chemical nature of the ligand is determined by the known biological specificity of the protein to be purified. In practice, affinity chromatography has been used to purify proteins such as immunoglobulins, membrane receptors, enzymes, and hormones as well as nucleic acids and even whole cells. In the case of an enzyme, the ligand chosen would probably be a substrate or a reversible inhibitor or activator. One could attempt to use a ligand which is absolutely specific in that it will bind only that enzyme but, failing this, one can use a 'group-specific' ligand. For example, if 5' AMP is used as a ligand, its structural similarity to NAD⁺ causes it to bind to many NAD⁺-dependent dehydrogenases which are therefore co-purified from other proteins.

The procedure for affinity chromatography of proteins is similar to that for the other types of liquid chromatography. The matrix is packed into a column in a buffer that will be optimal for enzyme-ligand binding. Thus the buffer must contain any co-factors such a metal ions that are needed for binding. Usually the buffer has a fairly high ionic strength to minimize non-specific binding of other proteins to the ligand. The sample is applied and washed through the column. Ideally, only the enzyme of interest should bind. It can then be eluted specifically by the addition of a relatively high concentration of substrate or competitive inhibitor or, failing this, by changing the pH and/or ionic strength to disrupt enzyme-ligand interaction.

An alternative protocol can be used if a specific antibody to the protein of interest is available. This procedure is applicable to all proteins irrespective of their functional activities. The antibody is covalently coupled to a suitable matrix and then poured into a column and exposed to the sample protein mixture as described above. Only the required protein will bind to the antibody and can then be eluted by procedures which weaken the antibody antigen interaction.

Affinity chromatography is theoretically capable of purifying a single protein from a complex mixture in a single step. However, even if this theoretical ideal is not achieved, the degree of purification is commonly very good indeed. As such, it is perhaps the most powerful protein purification procedure currently available. However, affinity chromatography can be applied only when the functional activity of the required protein is known and a suitable ligand is available or when a specific antibody to the protein has already been obtained. Unfortunately, in many cases neither condition is satisfied and so protein purification must rely upon more general procedures.

In this simulation, three types of immobilised monoclonal antibody are available for each protein in the mixture. These differ in their affinities for the protein. Low affinity antibodies will not bind the protein sufficiently tightly to prevent its passage through the column. High affinity antibodies will bind the protein so tightly that it cannot be eluted in an active form. So you will need to use the medium-affinity antibodies. You will need to carry out some experiments to find out which they are! An immobilised polyclonal IgG preparation is also available and, for some proteins, an immobilised competitive inhibitor.