Fig. 2.
Synthetic uORFs specify and tune protein expression levels. (A) Schematic of engineered mRNA elements and fluorescent protein reporter construct. 5′m, 5′ methylated cap of mRNA; AAA, poly-A tail; NNNG, bases preceding the GFP ORF; (NNN)S, non-AUG start codon; NNNu, bases preceding uORF; uORF, upstream ORF with sequence AUGGGUUGA where AUG and UGA are start and stop codons, respectively. All ORFs shown contain a G at the +4 position. (B–F) In PD31 cells, GFP/RFP levels as a measure of translation. (B) Effect of different TIS sequences without use of uORFs (construct 1). The sequence GCCACCAUGG (positions −6 to +4) was used by one construct, represented by GCCACC. One construct, RFP only, contained no GFP gene. (C) Effect of different numbers of uORFs (constructs 1–4). (D) Effect of distance (n) between upstream and GFP ORFs, where n is the number of bases after the uORF stop codon and before the GFP start codon (construct 5). (E) Variation of the three bases preceding the uORF and GFP ORF (construct 6) where GFP uses an AUG start codon. (F) Use of non-AUG start codons (in parentheses) to express GFP, where uORFs and bases preceding uORFs were varied (constructs 7 and 8). (G) GFP expression (not normalized to RFP) in different cell lines. Except for the transiently transfected 293T cells in G, all expression constructs were stably integrated into the genome. Unless noted, a five-base spacer exists between the uORF stop codon and GFP start codon, i.e., when n = 5 in D. Error bars represent SD from three experimental replicates.