HIV, ADAR Editing, The Rev-RRE Pathway and Antisense RNA-mediated Inhibition

In this study we demonstrate that a 287-bp long dsRNA panhandle structure can form in a 937-nucleotide HIV1 Antisense (AS) RNA that is expressed in a lentivirus vector which was previously shown to efficiently inhibit HIV expression. The presence of the panhandle showed a ten-fold better inhibition than an AS RNA that lacked the panhandle. Significant inhibition was also achieved with a vector that maintained only the 287-bp HIV panhandle, but substituted a heterologous EGFP sequence in the loop between the panhandle. Furthermore, expression of an AS RNA with an antisense/sense EGFP panhandle, rather than the one derived from HIV, led to efficient inhibition of both viral production and EGFP expression from the NL4-3-derived proviral clone pHIG, which contains an EGFP sequence inserted into the HIV genome. In all cases, efficient inhibition was only observed when the AS RNA trafficked on the Rev/RRE pathway. Thus our experiments demonstrate that the Rev/RRE pathway can be utilized to achieve efficient inhibition using long AS RNAs, especially when the generated AS RNA can form a double-stranded panhandle between sense and antisense sequences.

Sanger and Next Generation Sequence (NGS) analysis of the AS RNA expressed from the Lentivirus vector expressing the 937-nt AS env RNA with the panhandle showed that multiple A-to-G changes could be detected in the panhandle region in cDNA obtained from AS RNA in cells transfected with the vector alone. This was likely to be the result of editing of the dsRNA by Adenosine Deaminases Acting on RNA (ADAR) enzymes. Removal of the RRE from the vector, expression without Rev or redirection of the AS RNA to the Nxf1 export pathway, resulted in a significant reduction in the observed A-to-G changes in the panhandle. This indicates that RNA that traffics on the Rev/RRE pathway intersects with the ADAR machinery in a way that is different from mRNA trafficking on other export pathways. This is further supported by the fact that over expression of huADAR1 and huADAR2s interferes with Rev/RRE-driven AS inhibition. In the presence of HIV target mRNA, A-to-G changes were also observed in the regions of AS and target mRNA outside of the panhandle, which would be expected to form intermolecular double-stranded hybrids in transfected cells. However, compared to the changes in the AS RNA, only very low levels of editing were observed in the target mRNA. Thus these findings speak against a previously proposed mechanism for AS inhibition that involves retenton of antisense/sense hybrids in the nuclear matrix after A-to-I editing.

Processing of edited AS dsRNA into small RNAs that function as miRNAs/siRNAs represents a novel alternative mechanism that could explain the efficient inhibition. This is consistent with the fact that the partial depletion of the human Tudor Staphylococcal Nuclease (huTudor-SN or SND1), implied in the processing of edited dsRNAs, led to a less effective AS RNA-mediated inhibition. Furthermore, NGS analysis showed that AS RNA-derived miRNAs/siRNAs could be specifically detected in cells during Rev/RRE-mediated AS inhibition. Taken together, our results demonstrate a potential novel mechanism for AS RNA-mediated inhibition dependent on RNA trafficking on the Rev/RRE pathway.