Synthetic Two-Cistron Expression System

Abstract

Expression systems for the production of heterologous proteins are now available in many different host organisms such as Escherichia coli, yeast, baculovirus and mammalian cells. The choice of an appropriate host organism is dictated by the particular protein to be expressed, and by requirements for posttranslational modifications, solubility and yield. The primary advantages that E. coli offers as a host organism include the ease and speed with which recombinant strains can be constructed and expectations that the protein yield will be high when compared with other organisms. This is owing to extensive knowledge about mechanisms that control gene expression in E. coli, availability of well-characterized plasmids and phages that can serve as cloning and expression vehicles, and relatively inexpensive and efficient large-scale fermentation properties. Many proteins have been expressed at very high levels in E. coli, representing in some cases up to 50% of total cell protein (1). Typically, to achieve such high levels of expression, multicopy plasmids, which carry strong promoters to optimize transcription and strong ribosome binding sites to optimize translation, have been used. However, there are instances where little or no protein expression is achieved even though all the required elements appear to have been optimized and incorporated into the expression plasmids. The most accepted explanation for this observation is inefficient translation of mRNA containing eukaryotic coding sequences (2–6). It has been speculated that the cause for inefficient translation of mRNA is the potential for secondary structure formation involving the 5′ ends of mRNA, which encompass the ribosome binding site, the translational start codon, and the 5′ end of the coding sequence. Presumably, such secondary structures can lower translation initiation whenever the Shine-Dalgarno (SD) sequence and/or the translation initiation codon are sequestered in a double-stranded structure that prevents the binding of the 30S ribosomal subunit to the mRNA. This explanation is based largely on the empirical observation that changes within the 5′ end of mRNA have led to improved protein production (6). However, this approach is not very efficient because the bases at the 5′ ends of coding sequences are different for each gene, and it is difficult if not impossible to predict which bases to change in order to minimize the propensity of mRNA to form secondary structures and thus achieve optimal expression. Further, any base changes within the 5′ ends of coding sequences are limited by the need to retain the proper amino acid sequence at the amino terminus of the desired protein product.