An artificial technique used to design proteins using natural selection rather than rational design.

The Challenge

How a protein assumes a very specific shape and structure is a long standing challenge in the field of protein folding. As if it were not hard enough to determine what shape an extant protein will assume, some researchers, known as protein engineers, have taken this problem to the next level. They wish to design proteins that can have a specific function or shape. The most challenging approach to this is de novo protein design, which starts from scratch with the basic building blocks, and uses basic principles to construct a protein. Although there has been remarkable progress in this area, we are still far from designing enzymes de novo tailored to specific industrial and medical needs. Much research is still going into building generic scaffolds (tertiary folds) of various types with the hope that these can then be modified to make functional proteins. This technique requires much foresight and understanding of the chemistry and physics of protein interactions. However, nature does not explicitly know the laws of physics. How is it that we have developed the amazing complement of proteins that currently carry out our everyday metabolic and other corporeal needs?

Directed Evolution

An alternate approach, pioneered to a large extent by Frances Arnold, a professor of chemical engineering at Caltech, is that of directed evolution. Also termed forced evolution, this method harnesses the natural processes of adaptation and mutation that organisms normally are subjected to in the environment. By implementing recursive cycles of mutation and selection, one can optimize a protein for a particular process or new environmental parameter.

For example: Enzymes are a useful component of laundry detergent, as they can specifically target components of stains that other chemicals cannot efficiently, or selectively remove. With the high levels of detergent and other chemicals in laundry water, most normal enzymes would fall apart and not be able to function. Imagine if you had a screen that could select for effective cleaning agents - say an enzyme that can attack blood and dissolve it by breaking up the clotting material. This enzyme works fine in regular water when presented with a blood stain. Now, take a plasmid that has the gene for this enzyme and mutagenize it, so you now have a pool of 100 variants of the enzyme. Test these 100 variants on a new sample which has 10% detergent. Keep those variants that work in detergent and discard the rest. Now mutagenize this select pool to generate another 100 variants. Screen these in 20% detergent. Again collect the variants that function and re-mutagenize them... The end result is an enzyme that has built up a set of compensating mutations that allows it to function in the high detergent environment of laundry soap water.


The degree of mutagenesis is critical to the success of directed evolution methods. If the mutagenesis is too heavy, then one will likely cripple the enzymes and the directed evolution will proceed slowly, if at all. Mutagenizing too little will also require a large number of screens to make any progress. Generally, the mutagenesis rate should be one base-pair of DNA per round of mutation. Accumulation of single point mutations will eventually lead to compensatory interactions which benefit the protein in the new environment.

Methods of mutagenesis include error-prone PCR. In brief, PCR is an automated way of amplifying the amount of a specific gene one has. If one runs a PCR process at a temperature or other physical condition where mistakes are introduced while copying the gene, then one has a system that naturally introduces mutations. Variations on this technique, such as DNA shuffling and sexual PCR allow for larger scale mutations, which can be particularly useful when trying to design proteins with new functions. The formation of chimeric proteins, which are made up from parts of other proteins, can sometimes fortuitously yield one protein that inherited the function of multiple sources.


The automation of screening is important. The more variants that can be screened at each step, the more efficient the selection process becomes. This is now possible with phage display, a technique that puts the protein to be designed on the surface coat of a virus, the phage. Since the virus expresses the protein on the surface and contains the DNA that created that protein, you have a direct link between variant and its gene. Thus, you do not have to keep track of these things manually. The protein on the viral surface (hopefully) behaves as it would on its own. Thus, for example, one can test how well this protein binds to blood adhered to a filter. After running a huge population of phage over the filter, one can then wash off those that dont stick and keep the ones that do for the next round of mutagenesis and screening. Once the number of variants is down to a manageable size, one can then use higher accuracy, lower throughput techniques to screen the final enzymes.


Directed evolution has been successfully applied to building proteins that function in extremely high temperatures, high salt concentrations, organic solvents, toxic chemicals, and other non-protein-friendly environments. Its also been used to design proteins that catalyze steps in chemical reactions. This is of great value in the synthesis of pharmaceuticals and complex chemicals. Due to enzymatic specificty, the need for protecting groups in chemical reactions is reduced, increasing product yields during organic synthesis.

Directed evolution takes the problem of protein design and uses the computational ability of natural mutation to select proteins of desired traits. By copying natural mechanims, it is possible to improve current industrial and medical products and possibly identify new enzymes with novel catalytic functions and physical properties.

See also:

Log in or register to write something here or to contact authors.