Abstract Collagen prolyl 4-hydroxylase (C-P4H) plays a central role in the biosynthesis of collagens by hydroxylating proline residues. The enzyme has been a subject of intense interest as a target enzyme for drug development. The recombinant expression of human C-P4H in prokaryotes has not yet been described. This work reports on the development of an expression system for human C-P4H in E. coli.

The vertebrate C-P4H enzymes are α2β2 tetramers, consisting of two β subunits which are identical to protein disulphide isomerase (PDI), aside from the two α subunits which have the catalytic activity. The function of PDI is to keep the α subunit in a soluble and active state. Therefore, the expression system should assure the expression of the β subunit in the cell before the α subunit by using two different promoters. An active C-P4H tetramer was obtained in the periplasm of E. coli. However, further optimization for production by stepwise regulated coexpression of its subunits in the cytoplasm of a thioredoxin reductase and glutathione reductase mutant E. coli strain resulted in large amounts of human C-P4H tetramer. The exchange of four rare E. coli codons of the pdi gene and the optimized distance between ribosome binding site and translation initiation, resulted in 50-fold P4H-activity and 25 mg/l purified enzyme.

Comparison of the expression level of mRNA from the α and β subunits by Sandwich hybridization identified single induction with anhydrotetracycline in fed-batch fermentations as a limiting parameter. This caused an insufficient expression level of mRNA and thereby a low yield of C-P4H. A maximum yield was obtained by repeated addition of anhydrotetracycline that led to higher mRNA levels and increased productivity.

A newly developed stochastic simulation model of translational ribosome traffic in bacteria assesses the effect of codon usage to ribosome traffic and to the overall translation rate and mRNA stability. Using human PDI, it was shown that substitution of four 5' codons of the human PDI sequence that are rare in E. coli sequences, by synonymous codons preferred in E. coli led to a 2-fold increase of total PDI amount and even to a 10-fold increase of soluble PDI amount.