Flipped Over U: Structural Basis for dsRNA Cleavage by the SARS-CoV-2 Endoribonuclease

Coronaviruses generate double-stranded (ds) RNA intermediates during viral replication that can activate host immune sensors. To evade activation of the host pattern recognition receptor MDA5, coronaviruses employ Nsp15, which is uridine-specific endoribonuclease. Nsp15 is proposed to associate with the coronavirus replication-transcription complex within double-membrane vesicles to cleave these dsRNA intermediates. How Nsp15 recognizes and processes dsRNA is poorly understood because previous structural studies of Nsp15 have been limited to small single-stranded (ss) RNA substrates. Here we present cryo-EM structures of SARS-CoV-2 Nsp15 bound to a 52nt dsRNA. We observed that the Nsp15 hexamer forms a platform for engaging dsRNA across multiple protomers. The structures, along with site-directed mutagenesis and RNA cleavage assays revealed critical insight into dsRNA recognition and processing. To process dsRNA Nsp15 utilizes a base-flipping mechanism to properly orient the uridine within the active site for cleavage. Our findings show that Nsp15 is a distinctive endoribonuclease that can cleave both ss- and dsRNA effectively.


Protein expression and purification
Wild type (WT) and mutant Nsp15 constructs were created as described previously 1,2 . Nsp15 was overexpressed in E. coli C41 (DE3) competent cells in Terrific Broth with 100 mg/L ampicillin. At an optical density (600 nm) between 0.8-1.0, cultures were cooled prior to induction with 0.2 mM Isopropyl β-D-1-thiogalactopyranoside (IPTG). Cells were harvested after overnight expression at 16˚C and stored at -80˚C until use. Nsp15 purification was done as described previously 1,2 . Briefly, cells were resuspended in Lysis Buffer (50 mM Tris pH 8.0, 500 mM NaCl, 5% glycerol, 5 mM b-ME, 5 mM imidazole) supplemented with cOmplete EDTAfree protease inhibitor tablets (Roche) and disrupted by sonication. The lysate was clarified at 26,915 x g for 50 minutes at 4˚C and then incubated with TALON metal affinity resin (Clontech). His-Nsp15 was eluted from the resin with 250 mM imidazole, and buffer exchanged into Thrombin Cleavage Buffer (50 mM Tris pH 8.0, 150 mM NaCl, 5% glycerol, 2 mM b-ME, 2 mM CaCl2) for cleavage at room temperature for 4 hours. The cleavage reaction was repassed over TALON resin and quenched with 1 mM phenylmethylsulfonyl fluoride (PMSF) prior to gel filtration using a Superdex-200 column equilibrated in SEC buffer (20 mM Hepes pH 7.5, 150 mM NaCl, 5 mM MnCl2, 5 mM b-ME).

Cryo-EM sample preparation
RNA oligos were annealed at 250 uM final concentration by incubating equimolar amounts for 5 min at 75˚C and then cooled on the heat block for 1.5 hours. Purified Nsp15 H235A was diluted in a low-salt buffer (20 mM Hepes pH 7.5, 100 mM NaCl, 5 mM MnCl2, 5 mM b-ME) to 0.75 µM and incubated with excess RNA substrate (50 µM) for 1 hour at 4˚C. The purified protein was plunge frozen on homemade customized support grids made of C-flat R1.2/1.3 (Protochips) sputtered with 30nm thick layer of gold on the grid bar side using Leica ACE-600 sputterer. Before specimen deposition, the grids were pretreated on a plasma cleaner (Tergeo, model) in immersion mode with a power of 38W for a period of 75s. The Nsp15/RNA mixture (3 µL) was deposited onto the grids inside the chamber of a Leica EM-GP2 vitrification robot held at 15˚C and RH of 90% humidity. The grid was back-blotted for 3 seconds (Whatman Grade 40 filter paper), and then quickly plunged into liquid ethane kept at 90K. The vitrified sample on the grid was then transferred to liquid nitrogen for storage.

Data collection and processing
Nsp15 images were collected using a Talos Arctica electron microscope at 200 keV with a Gatan K2 Summit detector, and a Titan Krios at 300 keV with a K3 Bioquantum detector. Beaminduced motion and drift were corrected using MotionCor2 3 through Scipion 3 4 . CryoSPARC v2 5 was used in all subsequent image processing. The aligned dose-weighted images were used to calculate CTF parameters using CTFFIND4 6 . For the first dataset, particles were selected using blob and template-based particle picking, downsampled by a factor of 4, extracted with a box size of 64 and subjected to an initial round of 2D classification. Full resolution particle projections from good classes were re-extracted using a box size of 256. Ab initio reconstruction was used to an initial model, followed by several rounds of 3D classification with localized CTF refinement and per particle motion correction. No symmetry was applied.
For the combined Arctica/Krios datasets, the Topaz Extract machine learning algorithm was used to pick particles. The extracted particles were curated and segregated using a combination of 2D classifications, Ab-Initio Model Reconstruction and Heterogeneous Refinement. The curated particles stacks from both Talos Arctica and Titan Krios were then combined. The combined particle stack that yielded the most ideal 3D map was then further refined using unsymmetrized homogeneous, Global CTF, and Non-Uniform Refinements. Maps were re-scaled to optimize RMS fit to core domain residues of reference structure PDBID 6WLC 7 .

Model building
A SARS-CoV-2 Nsp15 cryo-EM structure (PDBID 7K0R) and model dsRNA (PDBID 6BU9) were used as starting models and fit into the cryo-EM maps using rigid body docking in Phenix 8 . A combination of rigid body and real-space refinement in Phenix as well as iterative rounds of building in COOT 9 were used to improve the fit of the models. Given the weaker density of the RNA outside the active site, it was restrained to model RNA parameters during refinement. Molprobity 10 was used to evaluate the model (Table 1). Figures were prepared using Chimera 11 and Chimera X 12 .

Urea-PAGE endoribonuclease assay
DsRNA was prepared by incubating equimolar concentrations of complementary oligos at 75˚C for 5 minutes, followed by cooling on the heat block for 1.5 hrs. To remove any remaining unpaired RNA, samples were purified on a 20% native polyacrylamide gel. The dsRNA band was excised, and the RNA was eluted overnight in 10 mM Hepes pH 7.5, 50 mM NaCl, 0.5 u/µL RNAsin. Double fluorescently-labeled RNA substrates (5′-FI and 3′-Cy5, 500 nM) were incubated with Nsp15 (50 nM) in RNA cleavage buffer (20 mM Hepes pH 7.5, 150 mM NaCl, 5 mM MnCl2, 5 mM DTT, 1 u/µL RNasin ribonuclease inhibitor) at room temperature for 30 minutes, with samples collected at 0, 1, 5, 10, and 30 minutes. The reaction was quenched with 2x urea loading buffer (8M urea, 20 mM Tris pH 8.0, 1 mM EDTA). At least three independent reactions were performed with protein from at least 2 different purifications. Due to the expected size of cleavage products and the size of bromophenol blue, loading buffer without dye was used. To monitor the gel front, a control lane of protein only with bromophenol blue was run. A ladder composed of double and single labeled oligos of different sizes were used to help assign cleavage product identities. Cy5-labeled products run anomalously compared to FI-labeled products 13,14 . The cleavage reactions were separated using 15% TBE-urea PAGE gels and visualized with a Typhoon RGB imager (Amersham) using Cy2 (λex=488 nm, λem=515-535 nm ) and Cy5 (λex=635 nm, λem=655-685 nm) channels. RNA cleavage was measured by disappearance of the uncleaved RNA, and quantified using Image Studio Lite (LI-COR). Prism (Graphpad) was used to calculate significant differences using Dunnett's T3 multiple corrections test. Tables   Supplementary Table 1