Surface patches on recombinant erythropoietin predict protein solubility: engineering proteins to minimise...

Protein solubility characteristics are important determinants of success for recombinant proteins in relation to expression, purification, storage and administration. Escherichia coli offers a cost-efficient expression system. An important limitation, whether for biophysical studies or industrial-scale production, is the formation of insoluble protein aggregates in the cytoplasm. Several strategies have been implemented to improve soluble expression, ranging from modification of culture conditions to inclusion of solubility-enhancing tags.

Surface patch analysis has been applied to predict amino acid changes that can alter the solubility of expressed recombinant human erythropoietin (rHuEPO) in E. coli, a factor that has importance for both yield and subsequent downstream processing of recombinant proteins. A set of rHuEPO proteins (rHuEPO E13K, F48D, R150D, and F48D/R150D) was designed (from the framework of wild-type protein, rHuEPO WT, via amino acid mutations) that varied in terms of positively-charged patches. A variant predicted to promote aggregation (rHuEPO E13K) decreased solubility significantly compared to rHuEPO WT. In contrast, variants predicted to diminish aggregation (rHuEPO F48D, R150D, and F48D/R150D) increased solubility up to 60% in relation to rHuEPO WT.

These findings are discussed in the wider context of biophysical calculations applied to the family of EPO orthologues, yielding a diverse range of calculated values. It is suggested that combining such calculations with naturally-occurring sequence variation, and 3D model generation, could lead to a valuable tool for protein solubility design.