DDX17: Structured RNA recognition drives diverse outputs

DEAD-box helicases control nearly every aspect of RNA metabolism. They are characterized by 12 conserved motifs, including the signature D-E-A-D motif, and have been implicated in transcription, splicing, decay, ribosomal RNA biogenesis and miRNA processing. Beyond these canonical functions, recent data suggest that these proteins have also been co-opted by the immune system to fight invading pathogens.

To discover novel antiviral DEAD-box helicases, we performed a RNAi screen using Drosophila as a model system.¹ We silenced 23 distinct and conserved DEAD-box genes in Drosophila cells and infected them with Rift Valley fever virus (RVFV), a mosquito-transmitted virus in the bunyavirus family. This screen identified 3 genes that restrict RVFV replication, including the helicase Rm62, which has 2 closely related homologs in mammals, DDX5 (p68) and DDX17 (p72). Rm62-deficient flies were more susceptible to RVFV infection, demonstrating that Rm62 controls RVFV replication in vivo. This antiviral restriction was specific, as Rm62 was dispensable for immunity against arthropod-borne viruses from other classes such as the alphavirus Sindbis virus (SINV). Lastly, depletion of DDX17 but not DDX5 in human cells increased RVFV replication, revealing a conserved antiviral function for DDX17 in evolutionarily distant hosts.

DDX5 and DDX17 have been identified as components of the Microprocessor complex that cleaves pri-miRNA transcripts to generate pre-miRNAs.² These helicases regulate the processing of a subset of miRNAs, although the exact mechanism remains obscure.⁴ In our study, Rm62 was not required for miRNA- and siRNA-mediated silencing, suggesting that its antiviral mechanism is RNAi-independent.

Therefore, as DDX17 is an RNA-binding protein, we characterized the RNAs bound to DDX17 using cross-linking immunoprecipitation with high-throughput sequencing (CLIP-Seq) in human cells in the presence and absence of infection. We found that DDX17 showed a bias for CA- and CT-repeat elements on mature cellular mRNAs, which were enriched for genes involved in cell adhesion and signaling, such as the MAPK pathway. In addition, we identified 160 miRNAs with significant CLIP-seq peaks, consistent with the known role of DDX17 in miRNA maturation. These miRNAs exhibited extensive overlap with previous studies that characterized DDX17-regulated miRNAs.³ DDX17-bound miRNAs were not enriched for the CA- or CT-repeat motif, suggesting that unlike cellular mRNAs, these primary sequence elements are not the determinant for DDX17-miRNA binding. Notably, the CLIP-seq peaks were depleted at the predicted miRNA loop and preferentially concentrated at the immediate 5′ and 3′ regions, indicating that DDX17 selectively binds the structured stem. Interestingly, a recent study³ identified a ([GTA]CATCC[CTA]) motif in the 3'-flanking segment of DDX17-regulated miRNAs, which was enriched in our dataset. DDX17 also cooperates with heterogeneous nuclear ribonucleoprotein (hnRNP) H/F, which bind G-quadruplex RNA structures, to control alternative splicing and cell differentiation.⁴ Taken together, these data suggest that DDX17 utilizes both primary sequence and secondary structure for optimal binding energy to facilitate RNA processing from splicing to small RNA biogenesis.

RVFV RNAs also exhibited several DDX17 peaks, particularly on the S genomic segment. One of these peaks was restricted to a known and essential stem loop on the RVFV RNA. Biotinylated RVFV RNA bearing this stem loop, but not other RNA sequences, precipitated DDX17 from cell lysates, confirming that DDX17 binds this RNA. Cloning this stem loop into SINV sensitized the virus to DDX17-mediated restriction, revealing that DDX17 binding to structured viral RNA confers its antiviral activity. The known roles of DDX17 were in the nucleus, and DDX17 resides in this compartment under basal conditions. We found that upon viral infection, DDX17 relocalizes to the cytoplasm where it can interact with these viral RNAs. Therefore, our data suggest that DDX17 has distinct functions in the nucleus and cytoplasm, both of which depend on structured RNA recognition (Fig. 1). However, early descriptions of Rm62 mutants reported increased abundance of blood and copia retroposon transcripts and enhanced retroelement insertion.⁵ Thus, DDX17 may have evolved to detect and destroy foreign RNAs in the nucleus as well. The signals that drive DDX17 cytoplasmic localization are unknown, but studies have shown posttranslational modifications on DDX17 that may play a role.⁶

Figure 1. DDX17 recognizes both primary sequence elements and secondary structures of RNAs. CLIP-seq analysis of DDX17-bound RNAs revealed that DDX17 recognizes mature cellular mRNAs enriched for CA- and CT-repeat elements. DDX17 is a member of the Microprocessor complex and facilitates pri-miRNA processing, but not through these repeat elements. Rather, DDX17 is localized to the stem region and cooperates with a previously described [GTA]CATC[CTA] element, and together may determine DDX17 binding to miRNAs. In the cytoplasm, DDX17 binds to viral RNAs by recognizing structured regions such as the intergenic stem loop on the RVFV S genomic RNA to inhibit viral RNA replication.

How DDX17 binds to diverse RNA sequences and structures remains unclear, and may be elucidated by additional structural analysis. DDX5 and DDX17 also cooperate with other binding partners that dictate their functions in specific contexts, and so it will be important to define DDX17-protein interactions during infection, particularly within the cytoplasm. Of note, DDX17 has been linked to the cellular RNA degradation machinery that mediates either 3′-5′ and 5′-3′ degradation, suggesting that DDX17 may facilitate viral RNA decay through either recognition or unwinding RNA structures. We suspect that subcellular localization may be a critical factor in defining these varied DDX17 roles and the ability of DDX17 to recognize certain types of RNAs. Clarification of the mechanism underlying DDX17 binding to cellular and viral RNAs will enhance our understanding of DDX17 in immunity, as well as in tumor proliferation, cell cycle regulation and differentiation where DDX17 also plays a fundamental role.