Stretching the genetic code : Incorporation of selenocysteine at specific UGA codons in recombinant proteins produced in Escherichia coli

Abstract: Selenocysteine (Sec) exists in all domains of life and represents the 21st naturally occurring amino acid. A Sec residue is co-translationally incorporated at a predefined opal (UGA) codon. UGA codons normally encode for translational stop via the protein release factor 2 (RF2). The incorporation mechanism of Sec into the selenoprotein involves complex machineries, dependent on several specific factors that differ between organisms. For both eukaryotes and prokaryotes, an mRNA secondary structure, called a Sec insertion sequence (SECIS) element, is required. Sec insertion systems for eukaryotes are different from that of bacteria. Due to the differences between species, recombinant expression of eukaryotic selenoproteins in E. coli is not a trivial task. However, our group has previously been able to overcome this species barrier and successfully expressed the mammalian selenoprotein thioredoxin reductase 1 (TrxR1) and other selenoproteins in E. coli. In this thesis, we have further developed this recombinant selenoprotein production system. We have also further characterized the recombinantly expressed rat TrxR1. We have studied growth conditions affecting yield of the recombinant selenoprotein when expressing rat TrxR1, using various levels of the selenoproteinencoding mRNA and growth in different types of medium. Guided by Principal Component Analysis (PCA), we discovered that the most efficient bacterial selenoprotein production conditions were obtained using high-transcription levels in the presence of the selA, selB and selC genes, with induction of production at late exponential phase. We also constructed an E. coli strain with the endogenous chromosomal promoter of the gene for relase factor 2 (RF2), prfB, replaced with the titrable PBAD promoter. In a turbidostatic fermentor system the simultaneous impact of prfB knockdown on growth and on recombinant selenoprotein expression was studied, using rat TrxR1 as the model selenoprotein. This showed that lower levels of RF2 correlate directly to an increase of Sec incorporation specificity, while also affecting total selenoprotein yield concomitant with a slower growth rate. Recombinant rat TrxR is expressed as a mixture of full-length and twoamino acid truncated subunits. Phenylarsine oxide (PAO) Sepharose can be used to enrich the Sec-containing protein. We investigated the mechanism of this purification by extensively purifying recombinant rat TrxR1, which gave an enzyme with about 53 U/mg in specific activity, which was higher than ever reported. Surprisingly, onlyabout 65% of the subunits of this TrxR1 preparation contained Sec, which revealed a theoretical maximal specific activity of about 80 U/mg for TrxR with full Sec content. The high specific activity revealed that the inherent turnover capacity of rat TrxR1 must be revised, and that the efficiency of bacterial Sec incorporation may be lower than previously believed. We also discovered and characterized tetrameric forms of recombinant TrxR1, having about half the specific activity compared to the dimeric protein in relation to Sec content. In conclusion, this thesis describes limiting factors for recombinant selenoprotein production in E. coli and shows how this production system can be optimized for higher yield and specificity. The results may prove to be of importance for the further development of E. coli as a useful source for synthetic selenoproteins. Results are also presented and discussed regarding the catalytic capacity of rat TrxR1 and novel multimeric states of the protein, which could represent unknown regulatory features of TrxR having potential physiological importance.