The use of affinity tags is now standard in protein purification and such tags greatly simplify the purification of cloned proteins when compared to classical chromatographic techniques. One such affinity tag is the 6x-his tag and several companies (e.g. InVitrogen, Qiagen) now offer expression vectors for N or C terminal fusion of proteins to this tag for expression in different systems (E. coli, yeast, insect cells, tissue culture cells). The 6x-his tag has several advantages over other affinity tags such as small size, conformation independent interaction with the affinity matrix and tight binding allowing reproducible purification. Qiagen markets a full system, the QIAexpress System, that contains the vectors, bacterial strains, affinity matrix and antibodies required for the purification of 6x-his tagged proteins. The QIAexpress System is made up of four components, The QIAexpress Expression System, The QIAexpress Protein Purification System, The QIAexpress Detection System and The QIAexpress Assay System, each one of these components is available seperately.
The basis of 6x-his tag purification is the tight coordination of histidine residues to immobilized metal ions, classically nickel, in a procedure termed immobilized-metal affinity chromatography (IMAC). Qiagen's patented matrix uses an enhanced chelating ligand, nitrilotriacetic (NTA), to complex nickel ions to a matrix as opposed to the older iminodiacetic acid (IDA) functionalities. NTA has one extra metal-chelating site (4 versus 3) for the six possible ligand sites on a nickel ion. As such, metal ions bind more tightly and ion leaching during loading or washing of proteins should not occur. Although I have never used older matrices, I have never experienced leaching with Ni-NTA agarose. Further, I have never seen a column lose its distinctive blue color even after several uses.
The NTA chelator is available in several formats - I have exclusively used the Ni-NTA agarose in which the NTA functional group is attached to Sepharose CL-6B. This is an extremely easy to handle, robust matrix that is excellent for most lab scale applications. I use open columns of various sizes and generally pour them as a 50 % slurry of the matrix in my first purification (lysis) buffer. Despite this rather crude technique I always seem to generate columns that pack well as well as load, wash and elute evenly. I have also tried batch loading, mixing matrix and sample prior to column pouring, but as it is so easy and quick to pour columns, I favor that format.
Once the matrix is chosen and the column poured, 6x-his tagged proteins can be easily purified to a high degree of homogeneity. I purify native proteins - however, because of the previously mentioned conformation independent interaction of the 6x-his tag, denatured proteins can also be purified in buffers containing guanidine HCl.
Bacterial cells are lysed and clarified prior to column loading. The lysis/binding buffer contains 10 mM imidazole (the ring functionality of histidine), which can bind nickel ions. To wash the column and elute non-specifically bound proteins, higher concentrations of imidazole (e.g. 20-50 mM) are included in buffers to compete for binding sites. To elute the tightly bound protein of interest, 250 mM imidazole is included in the buffers. Bacteria apparently contain very few proteins with successive histidine residues and as such, non-specific interactions should be minimal. I find some contaminants if I don't wash columns sufficiently and so tend to use 5-10 column volumes instead of 3. I have also increased the imidazole concentrations in wash steps and obtained good results. For one protein, I needed to include various agents to influence hydrophobic interactions (ethanol, glycerol, Triton) and eventually got pure protein. As such, the purification steps are flexible, but the standard form generally does yield proteins of >95% purity on SDS-PAGE. In terms of yields, the binding efficiency of the matrix is suggested to be 5-10 mg recombinant protein/ml resin and this is typically what I can achieve. I consequently adjust bed volumes so that columns are almost full and this also helps with purity.
Unfortunately, Ni-NTA agarose is not without limitations. Buffer components that would alter the redox state of nickel are obviously not recommended. This list includes reducing agents such as DTT (however b-mercaptoethanol at <20 mM can be used), chelatng agents such as EDTA and EGTA above 1 mM and certain buffers such as Tris, HEPES and MOPS, which contain secondary or tertiary amines. Although I have not experienced too many problems because of these limitations, I can image that the purification of certain proteins would be adversely affected and this will obviously govern the choice of a protein purification system.
To conclude, I am very happy with Qiagen's Ni-NTA agarose and The QIAexpress System as a whole. The ultimate aim is a high degree of protein purity (>95%) and that is easily achieved using 6x-his tagged proteins and this matrix, with only a small degree of optimization (changing column size and wash steps). Further, the matrix is mechanically easy to use, it is easy to pour and the flow is even during the purification steps. Finally, the column regenerates well and can be used several times before disposal. As a footnote, perhaps the two most important reasons that I am happy with Ni-NTA agarose are that I get high level expression of my proteins in bacteria and my proteins perform well in pH 8 phosphate buffers with no DTT or EDTA. If these criteria are not met by your protein, then maybe it would be worth looking into one of the numerous other systems that are currently marketed.
Peter Haggie, PhD
Post-Doctoral Fellow
University of California, San Francisco