Screening clones from DNA libraries.

The selection procedure is the heart of the directed evolution technique. Random mutagenesis by PCR will generate a gene library in which different copies contain different nucleotide base-pair mutations in their DNA sequences. The selection process will identify those sequences which encode proteins that present the desired property. Although this may appear trivial, in practice it is frequently not as simple as it might at first appear, and a reliable screening proceedure is critical for the success of a directed evolution project.

Imagine that a screening proceedure has been developed in which the aim is to isolate an enzyme that shows an increased catalytic activity. For example, this might be as simple as identifying a change in the halo of colour  surrounding a bacterial colony in a colorimetric test on agar plates. In this simple case the screening process will select those clones which show an increased conversion of substrates into product. However, the screening will not discriminate between those mutants which show improved substrate binding, improved active site adaptation for the substrate, improvements in catalytic efficiency or in an enhanced rate of product release. So although the simple screening may look attractive, in practice a variety of effects may be selected for. This may not be a problem if the aim is broadly to improve catalytic turnover of an enzyme, since any or all of these properties will achieve the desired goal. However, if the aim of the experiment is to improve thermostability, for example, then the screening at high temperature will identify both thermostable and thermotolerant enzymes. The underlying molecular mechanisms behind these two phenomena may be quite different.  Bearing this in mind, it is worthwhile to think carefully before devising a selection procedure, since "you always get what you screen for".

A random mutant library contains all possible substitutions of all amino acids at every position in the DNA sequence. After translation, this results in a collection of individual proteins each with one or more amino acid substitutions along the protein sequence. In principal, all 20 amino acid substitutions will be possible at all positions in the protein, and the library will include a huge number of protein variants. The actual number will be astronomical, (for a protein of 150 amino acids, there will be 20¹⁵⁰ variants). This raises a serious problem during the screening process, since in the laboratory only a miniscule fraction of this innate variability can be sampled.

The question is therefore how many clones must be screened to give a reasonable sampling of the variability, and which permits the selection of the desired property from the random mutants library. Most screening methods rely on the isolation and characterization of individual clones, therefore the limits for screening will be the rate at which clones individual clones can be characterized. The techniques which are used for this characterization of individual loans fall into two broad categories. The first involves growing individual colonies in 96 well plates, and subsequently screening of the cultures. This has the advantage in that it is readily adaptable to a modern laboratory robotics, however investment in such screening techniques is high. A cheaper alternative is to use agar plates coupled to a suitable chemical reaction which can identify those colonies containing the clones of interest. This second technique has the advantage that it is easy to implant in the laboratory, and thousands of clones can be screened extremely rapidly. The disadvantage, however, is that the sensitivity of the technique may be limiting in some applications.