**OVERSUBSCRIBED, LONG DELAYS EXPECTED** Please contact firstname.lastname@example.org before submitting an application to Instruct. The Instruct financing will cover the costs of the consumables, but the visitor will be required to contribute towards the costs of flight and accommodation. Please ask D. Hart for more details.
ESPRIT: Screening of tens of thousands of constructs of a single gene to identify well-behaving soluble constructs.
Academic structural biologists often work on proteins that lack accurate domain annotations. When the full-length protein cannot be expressed and a domain-focused approach is necessary, problems arise since it is unclear how to design high yielding, soluble expression constructs. Some proteins have little or no sequence similarity to others and this prevents domain identification using multiple sequence alignments. More often, some functional annotation exists e.g. from mutagenesis or deletion studies, but these regions do not define well the structural boundaries. Even when a soluble construct is obtained, disordered extensions may confound crystallisation attempts. We are all familiar with these situations; in many cases they are what keep our proteins “hot" and out of the PDB.
The ESPRIT technology was developed in the Hart lab at EMBL to express proteins whose domain boundaries are difficult to predict. It does not aim to replace the initial PCR cloning experiments based upon careful inspection of the protein sequence, but provides a rescue strategy when this fails - as it often does. ESPRIT, which stands for “expression of soluble proteins by random incremental truncation", is a directed evolution-type process combining random deletion mutagenesis with high throughput solubility screening Yumerefendi et al. 2010 (PubMed ID: 20206698) (Figure 1 & 2). We use exonuclease to truncate the ends of the target gene sequence in a sequential manner, thereby generating all possible construct termini for downstream testing. Up to 28,000 constructs per gene are isolated using colony picking robots and gridded out to form high density colony arrays for protein expression testing. For detection of soluble constructs in the library, the efficiency of in vivo biotinylation of a fused C-ter peptide is measured using fluorescent probes. Recent developments include adaptation to protein complexes using a coexpression system (An et al., 2011a) and incorporation of a genetic selection to eliminate out-of-frame constructs from the library, a step that greatly enhances the screening power of ESPRIT and its application to more challenging systems (An et al., 2011b).
Over the last ten years, the ESPRIT platform has been visited by European scientists who have brought their problematic targets for screening. Despite all the failures preceding the visit (typically over one or two years), about half of these users have returned home with soluble, purified proteins for further study. A recent highlight was the identification and soluble expression of the terminase domain from HCMV that resulted in its crystal structure (Nadal et al., 2010). Applications beyond structural biology have included vaccination of animals with domains from pathogen proteins and use of the soluble constructs for raising monoclonal antibodies.
Figure 1. ESPRIT process flow diagram. A random truncation library of a target protein is generated with exonuclease III and mung bean nuclease, producing tens of thousands of variant clones after transformation of E. coli. The first screening step is a high-throughput enrichment for putatively soluble, non-degraded constructs in colony array format on nitrocellulose membranes. The best ranked clones are selected for the second step of the ESPRIT screen where clones are prioritised based upon their purification profiles after expression in small-scale liquid cultures. Positive clones are sequenced to define boundaries and then scaled-up to larger volumes for further analysis.
Figure 2. Fabrication and analysis of colony arrays for expression screening. A robot is used to array the library as inocula onto nitrocellulose membranes (left) resulting in colony arrays (centre) in which truncated gene inserts are expressed. Expression of the protein products is assessed using fluorescent conjugates (right panel): in vivo biotinylation levels of the C-terminal biotin acceptor peptide are measured with streptavidin (green) and N-terminal hexahistidine tag signals with antibodies (red). Both channels have been merged into a single image with orange/yellow signals corresponding putatively soluble, undegraded constructs.