X-ray crystallography has been used to determine the structure of a new, improved protein that could be employed in the purification of therapeutic antibodies and to reveal details of its complexes with antibodies. The work represents an improved molecular design based on greater stability and higher affinity of the protein for its antibody target and could cut costs in therapeutic antibody manufacture.

Antibody therapy uses monoclonal antibodies to bind selectively to target cells involved in disease in the hope of stimulating the patient's immune system to attack and destroy those cells. Antibodies can be raised to almost any surface target protein and so they could find utility in treating a vast array of diseases including rheumatoid arthritis, multiple sclerosis and various forms of cancer. They hold much promise, especially given their apparently high efficacy and minimal side effects compared with conventional low-molecular weight drugs.

Unfortunately, therapeutic antibodies are expensive, so there is a significant demand for reducing costs at each stage of production, especially the purification step, which is the most costly. Affinity chromatography accounts for half the purification costs and Protein A from Staphylococcus aureus is used as the affinity ligand in the separation of antibodies from other components. It is expensive so alternative ligands have been keenly sought.

One such ligand is Protein G, which also binds to immunoglobulin and is expressed by streptococcal bacteria. It is its affinity for immunoglobulin that has led to its utility in purifying antibodies in the laboratory through its binding to the Fc region of those molecules. To function, it has to have its albumin-binding region removed otherwise it would simply bind to the serum albumin in antibody sources. As such biomedical researchers are familiar with experiments using recombinant forms of Protein G.

Now, Shinya Honda of the National Institute of Advanced Industrial Science and Technology, Tsukuba, Japan, and their colleagues have modified protein G based on molecular design in order to improve its stability and affinity for antibodies. The modified protein is much more heat resistant and less susceptible to enzyme decomposition than its predecessors. The modification also improved chemical resistance, and pH response.

The new, improved Protein G has could be viable as a purification agent for therapeutic antibodies under reaction moderate conditions and so might be amenable to scale-up beyond the laboratory into an industrial setting.

An additional interesting property of Protein G is that, unlike Protein A, it has affinity for all four forms of the immunoglobulin G (IgG) sub-classes of antibodies: IgG1, IgG2, IgG3, and IgG4. Protein A does not bind to IgG3.

To make use of this affinity for all IgG, the researchers had to find a way to make Protein G as viable as Protein A in the working environment of an industrial-scale production environment. They engineered the protein in two stages. First, the amino acid sequences were designed to stabilize the molecular structure of Protein G (first generation). Then the stabilized Protein G molecules were further modified to improve molecular recognition properties such as affinity (second generation). In the first-generation design, the amino acid sequences were generated by inputting the atomic coordinates of wild-type (unmodified) Protein G structure into the team's molecular design program.

They then synthesised the proteins suggested by the program, several types of engineered Protein G were synthesized and tested. "All engineered Protein G molecules showed increases of 7 to 13 Celsius in heat resistance, 1.4 to 1.6 times increases in chemical resistance against denaturants, and 4 to 14 times increases in resistance against protease destructive enzymes," the team explains.

They determined the three-dimensional structure of one of the engineered proteins using X-ray crystallography. Aside from the engineered region of the modified Protein G, the atomic coordinates were almost identical to those of wild-type Protein G.

In the second-generation step, the computer model structure of a complex of the first-generation engineered Protein G and the antibody IgG1 was built from their 3D structures. From this model they once again modified the amino acid sequence to improve the affinity by conducting computer simulations including that of electrostatic repulsion at the contact interface between the engineered Protein G and IgG1 within the complex. Finally, they synthesised the second-generation Protein G and tested it. The antibody affinity at neutrality was found to be 11 times higher, and the pH response was 18 times greater than that of wild-type Protein G.

Of course, the final test was to see how the second-generation engineered Protein G behaved in antibody purification. The first observation was that this engineered protein shifted the pH of peak separation, elution, towards the neutral, from pH 3.1 to pH 4.1, reducing degradation problems. Moreover, it remained fully functional in separating all sub-classes of human IgG, as had been predicted. The team is now working on the further development of engineered Protein G and looking to transfer the technology into the commercial sector.

The team presented details of their work at the 2nd Annual International Congress of Antibodies, 2010, held in Beijing, China, earlier this year.

Related links:

• J Biol Chem, 2009, 284, 12373-12383 
• Honda page

Article by David Bradley