The ESPRIT technology developed in our laboratory uses exonuclease III/mung bean nuclease protocols to generate unidirectionally or bidirectionally truncated construct libraries. Tens of thousands of clones can then be screened in a colony array format using efficiency of in vivo biotinylation of a fused biotin acceptor peptide to enrich soluble clones from the library. Positive clones are then further validated in 96 well plates by automated affinity chromatography purification. ESPRIT has been used to express a number of challenging proteins for further structural study. There has been no detailed description of library methods being used to express protein complexes directly i.e. incorporating coexpression approaches. Here, we have established a high-throughput automated SCH727965 CDK inhibitor strategy in which a library of constructs is screened for solubility in the presence of an interacting bait protein. As such, it is similar in concept to two-hybrid methods but in the context of recombinant over-expression of multi-milligram quantities of material required for many downstream applications including structural biology and vaccine research. Soluble protein complexes identified by this method can either result from association of prefolded partners or inter-folded polypeptide chains. To demonstrate the isolation of both types of complexes, we used subunits from the heterotrimeric influenza RNA polymerase that comprises three subunits: PA, PB1 and PB2. This complex catalyses the transcription and replication of the viral RNA genome in the nucleus of infected cells. The PB2 subunit has been shown to interact with importin a to achieve nuclear localisation. For many years the polymerase subunits resisted all attempts at soluble recombinant expression due their relatively large sizes and their lack of homology with other proteins which prevented domain identification using multiple sequence alignments. The PB2 subunit was previously studied using the ESPRIT method leading to the expression and subsequent structure solution of a series of novel domains key to viral function reviewed in. Here we screened PB2 gene libraries against bait proteins with the aim of isolating purifiable complexes. Firstly a 59 truncation library of the gene encoding the polymerase PB2 subunit was screened against importin a1 that has been shown to bind the purified C terminus of PB2 when mixed in vitro. Secondly a 39 truncation library of the same subunit was screened against a poorly behaving, marginally soluble C-terminal construct isolated from the PB1 polymerase subunit in an earlier ESPRIT experiment. A similar PB1 construct was recently shown by X-ray crystallography to form an inter-folded complex with a short N-terminal fragment of PB2, explaining its poor behaviour in isolation. In both experiments, a series of soluble complexes were isolated, some of which were similar to structurally validated forms, while others may be of potential interest in future functional studies. The application of ESPRIT in this co-expression format provides a powerful way of R428 purchase identifying wellexpressing soluble complexes for in vitro and in vivo biochemical and structural characterisation, as well as immunisation, high throughput screening and other applications that require multimilligram quantities of material.