Heterologous GPCR expression: a bottleneck to obtaining crystal structures.

G protein-coupled receptors (GPCRs) are an important, medically relevant class of integral membrane proteins. Laboratories throughout all disciplines of science devote time and energy into developing practical methods for the discovery, isolation, and characterization of these proteins. Since the crystal structure of rhodopsin was solved 6 years ago, the race to determine high-resolution structures of more GPCRs has gained momentum. Since certain GPCRs are currently produced at sufficient levels for X-ray crystallography trials, it is speculated that heterologous expression of GPCRs may no longer be a bottleneck in obtaining crystal structures. This Review focuses on the current approaches in heterologous expression of GPCRs and explores the problems associated with obtaining crystal structures from GPCRs expressed in different systems. Although milligram amounts of certain GPCRs are attainable, the majority of GPCRs are still either produced at very low levels or not at all. Developing reliable expression techniques for GPCRs is still a major priority for the structural characterization of GPCRs.

Expression and purification of milligram levels of inactive G-protein coupled receptors in E. coli.

G-protein coupled receptors (GPCRs) are seven transmembrane helical proteins involved in cell signaling and response. They are targets for many existing therapeutic agents, and numerous drug discovery efforts. Production of large quantities of these receptors for drug screening and structural biology remains challenging. To address this difficulty, we sought to express genes for several human GPCRs in Escherichia coli. For most of the receptors, expression was poor, and was not markedly improved even in strains designed to compensate for differences in codon bias between human and E. coli genes. However, the gene for human NK(1) receptor (hNK(1)R) was expressed in large quantities as inclusion bodies in E. coli. The inclusion bodies were not soluble in chemical denaturants such as guanidine chloride or urea, but were soluble in ionic detergents such as SDS, and the zwitterionic detergent fos-choline. Using immobilized metal affinity chromatography, we purified milligram amounts of hNK(1)R. Although inactive in ligand-binding assays, purified hNK(1)R in fos-choline micelles appeared to have a high content of alpha-helix, and was well-behaved in solution. Thus this protein is suitable for additional biophysical characterization and refolding studies.

Bound solute dialysis.

We used the thermodynamic principles governing bound solute dialysis, commonly referred to as "albumin dialysis" or "sorbent dialysis" and practiced clinically with the Molecular Adsorbent Recirculating System (MARS) and Biologic-DT approaches, respectively, to develop a comprehensive understanding of the process. Dimensionless parameters emerging from the thermodynamic analysis that govern bound solute dialysis are as follows: (1) lambda, the binding power of the solute binding moiety; (2) kappa, the dialyzer mass transfer/blood flow rate ratio; (3) alpha, the dialysate/blood flow rate ratio; (4) beta, the dialysate/blood binding moiety concentration ratio, and (5) psi, the solute/binding moiety concentration ratio in the blood. Results from a mathematical model of countercurrent bound solute dialysis for phi = 0.9 indicate that for a given binding moiety (fixed lambda), the most important parameter for achieving high removal rates is the dialyzer mass transfer ratio for free (unbound) solute. The results also show solute removal approaching an asymptote with increasing beta that is dependent on kappa and independent of alpha. More importantly, results indicate that once a dialysis membrane is chosen, solute removal is virtually independent of blood flow rate, dialysate flow rate, and amount of binding moiety in the dialysate, provided the amount is greater than approximately 90% of that required to reach the asymptote. Experimental observations over a range of blood flow rates (100-400 ml/ minute), dialysate flow rates (50-400 ml/minute), and dialysate/blood albumin concentration ratios (beta = 0-0.3) corroborate the model predictions and indicate that < 4 g/L albumin in the dialysate solution is required for effective bound solute dialysis. The experimental results also show evidence of enhanced mass transfer once the dialysis membrane pore structure surface saturates with albumin.