A general strategy to inhibiting viral -1 frameshifting based on upstream attenuation duplex formation

Hu, Hao-Teng; Cho, Che-Pei; Lin, Ya-Hui; Chang, Kung-Yao

To see if there is a minimal requirement for duplex length of an effective upstream attenuation duplex, three antisense variants (restore DNA23, restore DNA18 and restore DNA13) with the potential of forming RNA-DNA duplex of 23, 18 and 13 bp were designed with spacing to the 0frame E site being kept as 0 to prevent E-site invasion occurring (38) (Supplementary Figure S1C) . The −1 PRF attenuation activity of the shortest upstream duplex declined dramatically while being compared with that of the longest upstream duplex (both mediated by 10 M of antisenses) (Figure 2A and B) . Although the proximity requirement of a functional upstream −1 PRF attenuation duplex in Figure 1A and B suggests that the observed −1 PRF attenuation is not the result of ribosomal drop-off during translational elongation of a ribosome, the loss of observed −1 PRF attenuation activity by a shorter upstream duplex in Figure 2A and B implies that the observed '−1 PRF attenuation' by restore DNA23 could have been caused by drop-off effect mediated by the longer upstream duplex (39) . In particular, the gel based assay may not resolve the potential upstream duplex-mediated drop-off product from the 0-frame translation product. Eventually, this could result in the amount of 0-frame product being overestimated, leading to underestimation of −1 PRF efficiency. To address this issue, we created a new construct (6BPGC5 WT-SARSPK2) with the 0-frame stop codon being moved further downstream of the slippery site (Supplementary Figure S1D) to help distinguishing 0-frame products from the potential drop-off products that should appear within upstream mRNA sequences targeted by the antisenses. The trend in −1 PRF attenuation efficiency for upstream duplexes of different lengths using the 6BPGC5 WT-SARSPK2 construct is similar to that of the one used in Figure 2B in the presence of 10 M of antisenses ( Figure 2C and D) , indicating upstream duplexes of sufficient length do act as a −1 PRF attenuator. We noted that the attenuation efficiencies of antisensemediated upstream duplexes in Figures 1 and 2 were not dramatic (but statistically significant) in the presence of 1 and 10 M of antisenses because they were calculated from constructs carrying 6BPGC5 WT hairpin that possesses residual −1 PRF attenuation activity. Nevertheless, these results confirm that the −1 PRF attenuation effect can be generated in-trans by an antisense DNA-mediated upstream duplex proximal to the slippery site. Figure S2) in the presence of different amounts of corresponding antisense DNA targeting upstream sequences. The stimulator used in (A) is MMTV pseudoknot while the stimulators used in (B) and (C) are SRV pseudoknot and SRV hairpin, respectively. The p2luci vector was used as the −1 frame product control in all three cases. (D-F) Relative frameshifting activity of (A-C). Relative frameshifting activity was calculated with the frameshifting efficiency of reporters without antisense DNA addition being treated as 1 for comparison. Value for each bar is the mean of three independent experiments with standard error of the mean. P-values were determined by a student's t-test with P-value < 0.0001 designated by an '*' and referring to the comparison with the construct without the addition of an antisense or restore DNA23.