Because of specificity for both heparin/heparan sulfate and the receptor complex on epithelial cells relative to other fibroblast growth factor (FGF) homologues, there is considerable interest in clinical and commercial applications of FGF7 (also called keratinocyte growth factor or KGF) that require large quantities at reasonable cost. Production of recombinant FGF7 from bacteria suffers from lower yields and recovery relative to FGF1 and FGF2. Fusion of FGF7 at the N-terminus with glutathione-S-transferase (GST) followed by removal of GST by proteolysis while bound to natural ligand heparin improved the intrinsically low yields from Escherichia coli hosts to 3.2 mg per liter per OD600, which was still only 10% of that for FGF1. Yield of the GST-FGF7 fusion product was improved to about 17mg per liter per OD600 in strain BL21(DE3)pLysS by inclusion of 10-100 mM magnesium chloride (MgCl2) in the culture medium. This improved by about five times the yields of fully active (54)ser-FGF7 after proteolytic excision of the GST portion from GST-FGF7 immobilized on heparin-Sepharose. This simple enhancement improves the cost-effectiveness of production of recombinant FGF7 in bacteria for clinical and commercial applications. (C) 2003 Elsevier Inc. All rights reserved.