PROJECT SUMMARY Specific recognition of double-stranded RNA (dsRNA) structures underlies key biological processes from development to antiviral immune responses, and involves a host of dsRNA- binding proteins. Biochemical and structural studies have delineated how these proteins interact with model ligands, yet dsRNAs that they encounter in cells are far more complex. Moreover, little is known about how cells regulate dsRNA recognition. The goal of my lab is to understand how dsRNA-binding proteins interact with natural dsRNAs in cells, connecting biochemical models of dsRNA recognition to cellular dsRNA recognition and regulation. In the next five years, we will focus on dsRNA recognition in microRNA (miRNA) biogenesis and innate immune sensing. During miRNA biogenesis, Microprocessor recognizes and cleaves hairpin-like structures from long transcripts called primary miRNAs (pri-miRNAs). My previous research built a unifying model of human pri-miRNA and discovered that an optimal miRNA hairpin enhances the cleavage of a suboptimal miRNA hairpin on the same transcript—known as cluster assistance. Because nearly 40% of human miRNAs reside in clusters, with many implicated in diseases and influenced by cluster assistance, we propose to dissect this fundamental regulatory mechanism in miRNA biogenesis using biochemical, single-molecule, and structural approaches. Recognition of dsRNA by innate immune sensors such as RIG-I and MDA5 initiates antiviral responses. Recent studies show that endogenous dsRNAs can also activate these sensors, sometimes resulting in autoinflammatory and autoimmune diseases. The recognition of self dsRNAs therefore needs to be restricted by proteins that modify, degrade or shield endogenous dsRNAs. Building on our expertise in dsRNA recognition and processing, we will 1) develop new biochemical methods to identify dsRNA ligands of innate immune sensors, and 2) develop new genetic screen methods to identify factors that restrict self dsRNA sensing in innate immunity. Together these studies will provide crucial insights into how cellular dsRNAs are sensed and regulated, with therapeutic implications for cancer, autoimmune and infectious diseases.

RESEARCH STRATEGY: Understanding Double-Stranded RNA Recognition in Human Cells — Instrument Supplement Summary of parent grant R35GM150953 Recognition of double-stranded RNA (dsRNA) structures underlie many important biological processes, such as RNA interference, innate immune sensing of foreign RNAs, and post-transcriptional gene regulation. My lab focuses on two areas of dsRNA recognition: microRNA (miRNA) biogenesis and innate immune sensing. During miRNA biogenesis, Microprocessor, consisting of Drosha and DGCR8, recognizes and cleaves hairpin-like structures from long transcripts called primary miRNAs (pri-miRNAs). My previous work has built a unifying model of human pri-miRNA1 and discovered that some natural, suboptimal miRNA hairpins rely on neighboring optimal miRNA hairpins for efficient Microprocessor cleavage—known as “cluster assistance”2. Nearly 40% of human miRNAs reside in clusters, with many implicated in diseases and influenced by cluster assistance. We propose to dissect this fundamental regulatory mechanism in miRNA biogenesis using biochemical, single-molecule, and structural approaches (Project 1). Recognition of dsRNA by innate immune sensors such as RIG-I and MDA5 initiates antiviral responses3. Recent studies show that endogenous dsRNAs can also activate these sensors, sometimes resulting in autoinflammatory and autoimmune diseases4. The recognition of self dsRNAs therefore needs to be restricted by proteins that modify, degrade, or shield endogenous dsRNAs. Building on our expertise in dsRNA recognition and processing, we proposed to 1) develop new biochemical methods to identify dsRNA ligands of innate immune sensors (Project 2), and 2) develop new genetic screen methods to identify factors that restrict self dsRNA sensing in innate immunity (Project 3). Together these studies will provide crucial insights into how cellular dsRNAs are sensed and regulated, with therapeutic implications for cancer, autoimmune and infectious diseases. Recent progress and scientific justification For Project 1, we have made important progress dissecting the mechanism of cluster-assisted pri-miRNA procesing. We had envisioned two models of cluster assistance: in a dimerization model, one Microprocessor binds the helper miRNA hairpin and promotes the binding of a second Microprocessor to the recipient hairpin; in a transfer model, Microprocessor binds and cleaves the helper hairpin and is then transferred to the proximal recipient (Figure 1). These processes could be mediated by ERH which I identified biochemically2, and SAFB proteins that the Herzog lab Figure 1. Working models of cluster assistance. identified through a CRISPR screen5. To test these models biochemically, we utilized an in vitro pri-miRNA cleavage assay, which I previously established using lysates from cells that overexpress Drosha and DGCR8. This in vitro system recapituates cluster assistance, i.e., the defective miR-451 hairpin was cleaved faster in the context of the miR- 144~451 cluster. To test whether cleavage of the miR-144 helper hairpin is required for cluster assistance, we introduced modifications (phosphorothioate linkages flanked by 2′-O methylation) at both cleavage sites of the miR- Figure 2. Blocking miR-144 hairpin cleavage inhibited cluster 144 hairpin, which successfully blocked assistance. The line for each substrate represents the best fit of all of cleavage of the miR-144 hairpin (data not the data to an exponential reaction course, which generated the shown) by Drosha. Importantly, preventing the observed rate constants (k, shown ± 95% confidence intervals). Pri- processing of miR-144 hairpin abolished cluster miR-144*~451 has modifications to block miR-144 hairpin cleavage. assistance (Figure 2). This finding suggests that the dimerization model is unlikely. To strengthen this result, we are testing whether blocking one out of the two Drosha cleavage sites is sufficient to inhibit cluster assistance, and we will also test whether such cleavage-site modifications affect Micrprocessor binding. 1 We have also purified a C terminus-truncated form of SAFB2 and showed that it can enhance the cleavage of miR-451 hairpin in the cluster context in vitro. Therefore, cluster assistance can now be reconstituted using purified Microprocessor, ERH, and SAFB2 (Figure 3). Next, we will map the interaction between SAFB2 and Microprocessor using crosslinking mass spec. We will then introduce mutations that disrupt SAFB2–Microprocessor interactions to identify residues important for cluster assistance. Furthermore, as proposed in the parent award, we will use cryo-EM to provide the structural basis of cluster assistance. For project 2, we proposed to identify endogenous dsRNAs that activate innate immune Figure 3. Reconstitution of cluster assistance using sensors. We have developed a novel approach using purified Microprocessor. Processing assays of pri-miR-451 mouse oocyte Dicer (mDcrO), which misses the and pri-miR-144~451 using purified Microprocessor, with the helicase domain and cleaves long dsRNAs much presence of SAFB2 and ERH as indicated above. The red star better than full-length mammalian Dicer6,7. We indicates 3′ radiolabel. transiently overexpressed mDcrO in HEK293T cells, performed small RNA-sequencing analyses to identify mDcrO-dependent small RNAs, and then predicted the dsRNAs from which these small RNAs are processed by mDcrO. We used MACS2 peak calling8,9 to identify sRNA clusters. Our initial experiments identified 1297 small RNA clusters derived from intermolecular and intramolecular dsRNAs. The majority (66%) come from intramolecular folding of repetitive elements. We identified 60 stem-loop structures in the 5′ UTR and 41 in the 3′ UTR of mRNA transcripts. We also identified 51 intermolecular dsRNAs formed by overlapping transcripts (27 at the 3′ end, and 24 at the 5′ end). We are currently investigating the editing status of these dsRNAs to understand how editing suppresses innate immune sensing of dsRNAs. Our preliminary data suggest that many of the intermolecular dsRNAs formed by 3′ overlaps do not have extensive editing, and we plan to pull-down these dsRNA pairs and identify potential proteins that prevent them from activating MDA5. For project 3, we proposed to identify regulators that restrict sensing of endogenous dsRNA in human cells through CRISPR screens. We have made CRISPRi-compatible, GFP knock-in N/TERT-1 cells that reports IFN-β expression. We have verified this cell line in a mock screen (data not shown), and we are currently conducting a genome-wide screen to identify factors whose loss results in increased IFN-β expression. After obtaining the hits, we will verify and focus on RNA-related genes. We will then use biochemical, genetic, and other approaches to understand the mechanism by which these genes suppress endogenous dsRNA sensing. Rationale for the purchase of an ÄKTA pure micro system: Project 1 is a mechanistic project that relies heavily on protein and RNA purification. Although Projects 2 and 3 are discovery-oriented, we expect to move beyond discovery and conduct mechanistic studies using biochemical approaches. Currently we are using an ÄKTA pure 25 M system in the laboratory of my RNA Therapeutics Institute colleague Dr. Li Li. Dr Li’s system meets our current purification needs, but the ÄKTA pure micro system will be much better suited for our proposed cryo-EM studies because of specific requirement of sample concentration and quality. In addition, we routinely use the mammalian expression system (Expi293) for protein purification. The expression system is expensive and does not efficiently produce large quantities of RNA-binding proteins, so our experiments will benefit from the ability to purify minute quantities of proteins and RNA–protein complexes for analyzing protein–protein interactions, oligomerization states, and RNA–protein interactions. The small injection and elution volumes, improved resolution, and sharper peaks offered by the ÄKTA pure micro is ideal in this respect. I have recently hired a postdoc who has extensive experience with cryo-EM analysis of RNA–protein complexes, and who has used the ÄKTA pure micro system before. Therefore, we are prepared to take full advantage of this system. Plans for service contracts The service contract for the requested ÄKTA pure micro system will cost ~$6,600 per year. I will use my startup provided by the RNA Therapeutics Institute (my host department at UMass Chan Medical School) to cover the service contract for the duration of the project. Moreover, our department has dedicated funding and personnel for instrument maintenance, which is available for this proposed equipment. 2 Reference 1 Fang, W. & Bartel, D. P. The Menu of Features that Define Primary MicroRNAs and Enable De Novo Design of MicroRNA Genes. Mol Cell 60, 131-145, doi:10.1016/j.molcel.2015.08.015 (2015). 2 Fang, W. & Bartel, D. P. MicroRNA Clustering Assists Processing of Suboptimal MicroRNA Hairpins through the Action of the ERH Protein. Mol Cell 78, 289-302 e286, doi:10.1016/j.molcel.2020.01.026 (2020). 3 Kato, H., Takahasi, K. & Fujita, T. RIG-I-like receptors: cytoplasmic sensors for non-self RNA. Immunol Rev 243, 91-98, doi:10.1111/j.1600-065X.2011.01052.x (2011). 4 Chen, Y. G. & Hur, S. Cellular origins of dsRNA, their recognition and consequences. Nat Rev Mol Cell Biol 23, 286-301, doi:10.1038/s41580-021-00430-1 (2022). 5 Hutter, K. et al. SAFB2 Enables the Processing of Suboptimal Stem-Loop Structures in Clustered Primary miRNA Transcripts. Mol Cell 78, 876-889 e876, doi:10.1016/j.molcel.2020.05.011 (2020). 6 Flemr, M. et al. A retrotransposon-driven dicer isoform directs endogenous small interfering RNA production in mouse oocytes. Cell 155, 807-816, doi:10.1016/j.cell.2013.10.001 (2013). 7 Ma, E., MacRae, I. J., Kirsch, J. F. & Doudna, J. A. Autoinhibition of human dicer by its internal helicase domain. J Mol Biol 380, 237-243, doi:10.1016/j.jmb.2008.05.005 (2008). 8 Zhang, Y. et al. Model-based analysis of ChIP-Seq (MACS). Genome Biol 9, R137, doi:10.1186/gb- 2008-9-9-r137 (2008). 9 Feng, J., Liu, T., Qin, B., Zhang, Y. & Liu, X. S. Identifying ChIP-seq enrichment using MACS. Nat Protoc 7, 1728-1740, doi:10.1038/nprot.2012.101 (2012). 3