Objective

The availability of sufficient protein of good quality can be rate limiting for many drug discovery programmes, as a lack of target protein supply can hamper the ability to establish hit screening assays and to gain key structural information. Frequently, the desired activity or structure is supported by a discrete domain, and thus it is often not necessary to express the full-length protein. However, good construct design is critical for a successful outcome as even small variations in N- and C-terminal boundaries can dramatically influence both solubility and expression. While modern bioinformatics techniques can provide excellent predictions of the conserved regions within a gene, these regions do not always translate into the best protein constructs for expression in a recombinant system.

Our approach uses a technique called Combinatorial Domain Hunting (CDH), which couples random fragmentation of a target gene to efficient screening of resultant protein fragments. With expression libraries in excess of 100,000 clones, this method allows rapid and even sampling of the available protein space for soluble protein domains suitable for a range of applications including structural analysis and biochemical and biophysical screening. CDH has been applied to over 50 targets from a range of cytoplasmic protein classes including kinases, epigenetic targets, transcription factors, proteases and deubiquitinylating enzymes (DUBs) with an overall success rate in excess of 90%.

We present here two cases studies: A DUB and Lysine Methyltransferase KMT2D (MLL4)