The Highly Conserved Codon following the Slippery Sequence Supports -1 Frameshift Efficiency at the HIV-1 Frameshift Site

Mathew, Suneeth F.; Crowe-McAuliffe, Caillan; Graves, Ryan; Cardno, Tony S.; McKinney, Cushla; Poole, Elizabeth S.; Tate, Warren P.

Initially, a strategy was used to evaluate the significance of eRF1 depletion on well-characterised translational systems. First, we tested whether down-regulating eRF1, using RNAi, would affect the natural competition that occurs between eRF1 and near-cognate tRNAs at stop codons in vivo. Dual luciferase constructs where the downstream luciferase sequence was in the same frame as the upstream luciferase reporter but separated by a stop signal were used for this purpose. We established that at the weakest stop signals, where the near-cognate tRNAs more effectively compete with the release factor, up to 3% readthrough can occur even with endogenous levels of eRF1. Cellular levels of eRF1 mRNA and protein were decreased two-fold 120 h after transfection of the vector expressing shRNAs (α-eRF1), compared with the shRNA negative control (−) (Fig. 4A & B) . Cellular eRF1 protein was depleted using two vector-expressed shRNA sequences, si90 and si1186 (adapted from si1187) known to target eRF1 mRNA [43] . Both shRNAs reduced eRF1 mRNA and protein expression similarly (see S2 File for si1186 data). For both shRNAs, the physiological significance of reducing eRF1 was demonstrated with increased readthrough of *4.5-fold at the weak termination signal, UGA CUG (P < 0.001) (Fig. 4C ). This reduction in eRF1 resulted in a *30% increase in frameshift efficiency (P < 0.01) at the natural antizyme frameshift element where frameshifting naturally occurs with a UGA intercodon (Fig. 4D ). This suggested that the cell was responding to the lowered concentration of eRF1 as expected.