Recognition of cellular RNAs by the S9.6 antibody creates pervasive artefacts when imaging RNA:DNA hybrids

Abstract

The contribution of RNA:DNA hybrid metabolism to cellular processes and disease states has become a prominent topic of study. The S9.6 antibody recognizes RNA:DNA hybrids with a subnanomolar affinity, making it a broadly used tool to detect and study RNA:DNA hybrids. However, S9.6 also binds double-stranded RNA in vitro with significant affinity. Though frequently used in immunofluorescence microscopy, the possible reactivity of S9.6 with non-RNA:DNA hybrid substrates in situ, particularly RNA, has not been comprehensively addressed. Furthermore, S9.6 immunofluorescence microscopy has been methodologically variable and generated discordant imaging datasets. In this study, we find that the majority of the S9.6 immunofluorescence signal observed in fixed human cells arises from RNA, not RNA:DNA hybrids. S9.6 staining was quantitatively unchanged by pre-treatment with the human RNA:DNA hybrid-specific nuclease, RNase H1, despite experimental verification in situ that S9.6 could recognize RNA:DNA hybrids and that RNase H1 was active. S9.6 staining was, however, significantly sensitive to pre-treatments with RNase T1, and in some cases RNase III, two ribonucleases that specifically degrade single-stranded and double-stranded RNA, respectively. In contrast, genome-wide maps obtained by high-throughput DNA sequencing after S9.6-mediated DNA:RNA Immunoprecipitation (DRIP) are RNase H1-sensitive and RNase T1- and RNase III-insensitive. Altogether, these data demonstrate that the S9.6 antibody, though capable of recognizing RNA:DNA hybrids in situ and in vitro, suffers from a lack of specificity that precludes reliable imaging of RNA:DNA hybrids and renders associated imaging data inconclusive in the absence of controls for its promiscuous recognition of cellular RNAs.