Negative effect of sequential serine codons on expression of foreign genes in Escherichia coli.

Herpes simplex virus encodes a 1298-residue protein designated ICP4 that regulates transcription of viral genes. Structural and functional analyses of ICP4 have been facilitated by production of portions of ICP4 in Escherichia coli. We previously observed that expression of most truncated forms of ICP4 in E. coli was relatively efficient, with the exception of portions of the ICP4 gene approximately between codons 160 and 220. We have now localized the portion of ICP4 that inhibits expression to a serine-rich region from position 176 to 199. Our experimental results suggest that codons within the serine-rich domain do not induce termination of transcription, do not alter the intrinsic stability of mRNA, and do not create a proteolytically sensitive site in this portion of ICP4. Silent mutations that alter codon usage of many of the 19 serine codons in this region had no effect on expression. However, we observed that the level of protein expression was inversely proportional to the number of serine codons in this region. The results are consistent with a model in which the serine-rich domain induces premature termination of translation. This effect is not due to any specific secondary structure in the mRNA or lack of sufficient seryl-tRNA synthetase. It remains to be determined whether premature termination can result from insufficient seryl-charged tRNAs. Our results suggest that foreign genes with more than 20 consecutive serine codons may be poorly expressed in E. coli.