Library-based methods for identification of soluble expression constructs

When expression or crystallisation of a protein target in its wild-type full-length form proves problematic, a common strategy is to divide it into subconstructs comprising one or more domains. Rational construct design is not always successful, especially with targets for which there are few similar sequences to generate multiple sequence alignments. Even when this is possible, expression constructs may still fail to yield soluble protein, commonly expressing insolubly or at unusable yields. To address this, several new methods have been described that borrow concepts from the field of directed evolution whereby a random library is generated encompassing construct diversity; this is then screened to identify soluble constructs empirically. Here, we review progress in this area.