Comparison of IRES and F2A-based locus-specific multicistronic expression in stable mouse lines.

Efficient and stoichiometric expression of genes concatenated by bi- or multi-cistronic vectors has become an invaluable tool not only in basic biology to track and visualize proteins in vivo, but also for vaccine development and in the clinics for gene therapy. To adequately compare, in vivo, the effectiveness of two of the currently popular co-expression strategies - the internal ribosome entry site (IRES) derived from the picornavirus and the 2A peptide from the foot-and-mouth disease virus (FDMV) (F2A), we analyzed two locus-specific knock-in mouse lines co-expressing SRY-box containing gene 9 (Sox9) and enhanced green fluorescent protein (EGFP) linked by the IRES (Sox9(IRES-EGFP)) or the F2A (Sox9(F2A-EGFP)) sequence. Both the constructs expressed Sox9 and EGFP proteins in the appropriate Sox9 expression domains, with the IRES construct expressing reduced levels of EGFP compared to that of the F2A. The latter, on the other hand, produced about 42.2% Sox9-EGFP fusion protein, reflecting an inefficient ribosome 'skipping' mechanism. To investigate if the discrepancy in the 'skipping' process was locus-dependent, we further analyzed the FLAG(3)-Bapx1(F2A-EGFP) mouse line and found similar levels of fusion protein being produced. To assess if EGFP was hindering the 'skipping' mechanism, we examined another mouse line co-expressing Bagpipe homeobox gene 1 homolog (Bapx1), Cre recombinase and EGFP (Bapx1(F2A-Cre-F2A-EGFP)). While the 'skipping' was highly efficient between Bapx1 and Cre, the 'skipping' between Cre and EGFP was highly inefficient. We have thus demonstrated in our comparison study that the efficient and close to equivalent expression of genes linked by F2A is achievable in stable mouse lines, but the EGFP reporter may cause undesirable inhibition of the 'skipping' at the F2A sequence. Hence, the use of other reporter genes should be explored when utilizing F2A peptides.