key: cord-0906808-3znolsd9
authors: Basu, Souradeep; Mak, Tiffany; Ulferts, Rachel; Wu, Mary; Deegan, Tom; Fujisawa, Ryo; Tan, Kang Wei; Lim, Chew Theng; Basier, Clovis; Canal, Berta; Curran, Joseph F.; Drury, Lucy; McClure, Allison W.; Roberts, Emma L.; Weissmann, Florian; Zeisner, Theresa U.; Beale, Rupert; Cowling, Victoria H.; Howell, Michael; Labib, Karim; Diffley, John F.X.
title: Identification of SARS-CoV-2 Antiviral Compounds by Screening for Small Molecule Inhibitors of the nsp14 RNA Cap Methyltransferase
date: 2021-04-08
journal: bioRxiv
DOI: 10.1101/2021.04.07.438810
sha: 80fabcc835ad4854e50c2fba6f16fbd5a2a0e7f6
doc_id: 906808
cord_uid: 3znolsd9

The COVID-19 pandemic has presented itself as one of the most critical public health challenges of the century, with SARS-CoV-2 being the third member of the Coronaviridae family to cause fatal disease in humans. There is currently only one antiviral compound, remdesivir, that can be used for the treatment of COVID-19. In order to identify additional potential therapeutics, we investigated the enzymatic proteins encoded in the SARS-CoV-2 genome. In this study, we focussed on the viral RNA cap methyltransferases, which play a key role in enabling viral protein translation and facilitating viral escape from the immune system. We expressed and purified both the guanine-N7 methyltransferase nsp14, and the nsp16 2’-O-methyltransferase with its activating cofactor, nsp10. We performed an in vitro high-throughput screen for inhibitors of nsp14 using a custom compound library of over 5,000 pharmaceutical compounds that have previously been characterised in either clinical or basic research. We identified 4 compounds as potential inhibitors of nsp14, all of which also show antiviral capacity in a cell based model of SARS-CoV-2 infection. Three of the 4 compounds also exhibited synergistic effects on viral replication with remdesivir.

The SARS-CoV-2 virus is a novel respiratory pathogen that is able to infect both animals and humans, resulting in the disease COVID-19, which was officially declared a global pandemic by the World Health Organisation (WHO) on 11 th March, 2020 [1] . This is the third novel zoonotic virus belonging to the genus Betacoronavirus since the start of the 21 st century, after the original SARS-CoV-1 in 2003, and MERS-CoV in 2012 [2] . What sets SARS-CoV-2 apart from the previous two is its distinctive virulence on an overall population, as well as its epidemiological dynamics (reviewed in [3] ). Therefore, vastly different detection and containment strategy have been required compared to those that were used successfully with SARS-CoV-1 and MERS-CoV outbreaks. SARS-CoV-2 is a membrane enveloped virus with peplomer-forming spike (S) glycoproteins on its surface, giving it the characteristic "corona" shape when visualised under electronmicroscopy (EM) [4] . The structural components of the virus are highly variable and mutable, with over 10% of the open reading frame mutations identified between December 2019 to April 2020 found to occur in the Spike glycoprotein gene alone [5] . The highly mutative nature of the viral coat therefore poses a problem for producing effective long-term neutralising antibodies through administration of vaccines, leading to viral strains such as B1.351 that escape from the immune response [6] .

However, the structural proteins of the virus actually only constitute a small part of the coding capacity of the viral genome, with only 4 out of 29 proteins encoding viral structural components ( Figure 1A ) [7] . A pair of very large open reading frames (Orf1a and Orf1ab) comprise the first two-thirds of the genome. The encoded polyproteins (pp1a and pp1ab) are autoproteolytically cleaved into 16 distinct non-structural proteins (nsp), generating the enzymes and accessory proteins responsible for viral replication once inside a eukaryotic cell.

Coronavirus RNAs are produced by the viral replicase/transcriptase complex, and are post transcriptionally capped at the 5' end with a cap structure similar to endogenous mRNA caps, which are necessary for efficient translation and RNA stability [8] [9] [10] [11] . Crucially, viral RNA capping has been shown to be essential for the synthesis of viral proteins through eukaryotic translation initiation factor 4E (eIF4E) recognition [12, 13] . In addition, RNA cap methylation is also important for ensuring efficient ribosome binding and engagement of the host translation machinery, as well as avoiding degradation by exoribonucleases [9, 14] .

Aside from reduced translational capacity, uncapped RNA also triggers the host innate immune response leading to the expression of antiviral cytokines, which limit virus replication and shape adaptive immunity [15] . Other host sensor proteins recognise incomplete or absent cap RNA structures on viral RNA, and are responsible for the inhibition of viral translation [16, 17] . Formation of the viral RNA cap structures thus protect the virus against cell intrinsic antiviral effectors and potentiate the infective potential of the virus. For these reasons, viral enzymes involved in the capping the viral RNA are promising targets for antiviral drug development [18] [19] [20] .

In eukaryotes, these methylated cap structures are added co-transcriptionally upon transcription by RNA Pol II in the nucleus (reviewed in [21] ). Because coronavirus replication and transcription occur in the cytoplasm, independently of Pol II, the formation of these cap structures must be catalysed by viral enzymes. Over one-third (6/15) of the coronavirus nsps are necessary to efficiently cap viral RNAs, which exemplifies the complexity of this process as well as its importance for the viral lifecycle. Firstly, following RNA synthesis, nsp13 removes the terminal γ-phosphate from the initiating adenosine nucleotide triphosphate ( Figure 1B ). Nsp12 then acts as an RNA-guanylyltransferase, generating GpppA-capped RNA ( Figure 1B ). Subsequently, nsp14 transfers a methyl group to the N7 position of the terminal cap guanine forming the m7G cap structure, otherwise known as cap-0. Finally, the nsp16/nsp10 complex is responsible for 2'-O methylation of the cap ribose to form cap-1, which is the terminal cap structure on viral RNAs ( Figure 1B ).

The nsp14 enzyme in SARS-CoV-2, similar to other coronaviruses, carries dual functionality as both a methyltransferase and a 3'-5' exoribonuclease. While the exoribonuclease activity is dependent upon the nsp10 cofactor, this cofactor is not required for the N7-MTase function in nsp14 [18, 22] . In contrast, the other methyltransferase enzyme, nsp16, which converts cap-0 to cap-1, requires the presence of nsp10 to act as an allosteric activator to increase RNA substrate binding affinity [22, 23] . Both nsp14 and nsp16/10 are dependent upon the presence of S-adenosyl-L-methionine (SAM), which acts as the methyl donor for the respective methylated cap modifications, and S-adenosyl-L-homocysteine (SAH) is the resulting by-product. Using this property, we adopted a biochemical assay that detects the conversion of SAM to SAH to measure the relative methyltransferase activity. We then proceeded to first investigate the in vitro methyltransferase activity of purified nsp14 enzyme, and screened against a custom library of over 5000 characterised chemical compounds, with the aim of identifying potential antiviral drugs.

To generate untagged nsp14 we used the His-SUMO tag [24] , which can be completely removed by the SUMO protease, Ulp1. His 14 -SUMO-nsp14 was expressed in Escherichia coli cells after induction with IPTG overnight. Clarified cell extract was then passed over a Ni-NTA column and His 14 -SUMO-nsp14 was eluted with imidazole ( Figure 1C ). The His 14 -SUMO tag was cleaved by addition of the Ulp1 ( Figure 1C ) and was further purified by gel filtration chromatography, where untagged nsp14 eluted as a single peak ( Figure 1D ). This methodology allowed the expression and subsequent purification of nsp14 in high yield.

To ensure that purified nsp14 was functional, we examined the ability of the enzyme to catalyse methylation of the G(5')pppG(5') cap in a capped RNA substrate. Radiolabelled 32 P-RNA substrate was incubated with nsp14, the viral nsp16 methyltransferase, human CMTR1 (Cap-specific mRNA 2'-O-Methyltransferase 1) and RNMT (RNA Guanine-N7 Methyltransferase) in complex with its activating co-factor RAM (RNMT Activating Miniprotein). Whilst nsp14 and RNMT-RAM are guanine-N7 methyltransferases that use GpppG-RNA as a substrate [25] , CMTR1 and viral nsp16 are 2'-O-methyltransferases that utilise previously guanine-N7 methylated me 7 GpppN-RNA as a substrate. Following incubation with methyltransferase, RNA substrate was digested with Nuclease-P1, which digests RNA but leaves RNA cap structures intact. This digested mixture was then analysed by thin layer chromatography. As expected nsp16 and CMTR1 failed to utilise the GpppG-RNA substrate, whereas both viral nsp14 and human RNMT-RAM efficiently methylated the GpppG-RNA to form me 7 GpppG-RNA (Figure 2A ). To confirm that the methylation was strictly dependent on the catalytic activity of nsp14 we mutated aspartate 331, which resides in the catalytic core of nsp14 and is predicted to be essential for methyltransferase activity, to alanine (D331A). The purified nsp14 D331A protein was inactive as a methyltransferase in this assay ( Figure 2B ), confirming that the nsp14 of SARS-CoV-2 functions as a methyltransferase in accordance with the function of nsp14 enzymes from other coronaviruses [26] , and demonstrating that our purified wild-type nsp14 is catalytically active.

Thin layer chromatography is not amenable to high-throughput screening for nsp14 inhibitors, and therefore we turned to a fluorescence-based readout of methyltransferase activity. We employed a commercially available homologous time-resolved fluorescence (HTRF) assay, which monitors the activity of S-adenosyl methionine (SAM)-dependent methyltransferases. SAM acts as a methyl donor for methyltransferases, yielding S-adenosyl homocysteine (SAH) as a final product ( Figure 3A , left). The HTRF assay revolves around a Terbium (Tb) cryptate conjugated to the Fc region of an anti-SAH antibody. This antibody is able to bind SAH which has been conjugated to the d2 fluorophore (SAH-d2) through its variable (Fab) region. In a system without newly produced SAH, the Tb cryptate is able to form a FRET pair with the d2 fluorophore present on the bound SAH-d2 ( Figure 3A , right upper). The production of SAH from successful methyltransferase reactions competes with and displaces SAH-d2 from the variable region of the antibody, causing a reduction in signal ( Figure 3A , right lower).

A robust reduction in HTRF signal was seen after incubation of nsp14, SAM and GpppA-RNA, which was lost if any one of the three components necessary for methyltransfer was omitted ( Figure 3B ). SAM dependent methyltransferases are susceptible to inhibition by sinefungin, which acts as a competitive inhibitor (with respect to SAM) of methyltransferases. Figure 3B shows that sinefungin inhibited nsp14 activity in a dosedependent manner. Monitoring over time, the reaction reaches a plateau after around 20 min ( Figure 3C ). This plateau is unlikely to be due to time-dependent enzyme inactivation since we observed a shallow but continuous decrease of HTRF signal over the course of a 45minute reaction in the presence of sinefungin ( Figure 3C ). Therefore, the rapid plateau reached likely represents a lower boundary for the HTRF assay under our experimental conditions. We next characterised the range of substrates that nsp14 is capable of methylating. Unlike other guanine-N7 methyltransferases, coronavirus nsp14 has previously been reported to be able to methylate free GpppA cap analogue without attached RNA, as well as free nucleotide GTP [25] . When titrating GpppA-capped RNA, GpppA cap analogue, or nucleotide GTP, we were able to see a reduction in HTRF signal, indicating that SARS-CoV-2 nsp14 retains this unusually broad substrate repertoire, and is able to catalyse methyltransfer to all three substrates ( Figure 4A ). nsp14 is, however, unable to catalyse methyltransfer to already guanine-N7 methylated me 7 GpppA cap analogue ( Figure 4A ). This suggests that, despite wide substrate specificity, SARS-CoV-2 nsp14 acts specifically as a guanine-N7 directed methyltransferase. Following this, we were able to quantify the Michaelis constants (K m ) for all of the known substrates of the nsp14 methyltransferase.

nsp14 showed similar K m values for both GpppA and GpppA-RNA substrates ( Figure 4B ,C), but demonstrated slightly reduced affinity towards nucleotide GTP ( Figure 4D ).

Next, we adapted the HTRF assay conditions to conduct a high-throughput screen against a custom compound library to discover novel nsp14 inhibitors. Primary screening for inhibitors of nsp14 was conducted using a custom compound library containing over 5000 compounds at a concentration of 3.125 μM. Given that nsp14 showed similar K m values between the GpppA-RNA substrate and GpppA cap analogue, we conducted the screen using the cap analogue substrate which did not add the complications of working with an RNA substrate.

Library compounds were resuspended in DMSO, and negative controls wells also included DMSO to a final concentration of 0.03125 % (v/v). Sinefungin at 3.125 μM was included in several wells on each plate to serve as a positive control, thereby allowing determination of screen quality.

After screening, we calculated the Z' factor of our screen to be 0.625, indicating a highquality screen ( Figure S1 ). We therefore calculated Z-scores for all compounds, and ranked them by increasing Z-score ( Figure 5A ). We firstly selected 'hit' compounds from the list of compounds with a Z-score of above 3.0, but made exceptions for some compounds if their Z-score was >2.5 and either clinically relevant, or a previously characterised methyltransferase inhibitor. This left 83 compounds, which was narrowed by removing compounds which likely represented screening errors (see methods), resulting in a list of 63 hits. These hits, and their associated Z-scores, are listed in Table S1 . We subsequently selected compounds to take forward for validation based on if they were available to purchase commercially, if they were rapidly available, if they were known to be tolerated in humans, and favoured particularly those that were already licenced therapeutics. This was done to focus our hits to compounds that may be of clinical relevance to the treatment of COVID-19, and resulted in a list of 15 compounds to carry forward for validation (Table S1 ).

Using the HTRF-based assay, we found that 4 of these 15 compounds (PF-03882845, Inauhzin, Lomeguatrib, and Trifluperidol) consistently inhibited nsp14 when assayed individually over a wider range of concentrations ( Figure 5B and S2). Among these compounds, PF-03882845 was the most potent inhibitor (HTRF 50 = 1.1 μM), followed by Trifluperidol (HTRF 50 = 12.9 μM), Inauhzin (HTRF 50 = 23.0 μM), and finally Lomeguatrib (HTRF 50 = 59.8 μM). The remaining 11 compounds did not show significant inhibition, and were excluded from further analysis ( Figure S2 , Table S1 ).

To test the specificity of the four validated compounds towards nsp14 methyltransferase activity, we checked for cross-inhibition of the other viral methyltransferase nsp16/nsp10. To test the role of these 4 compounds on the nsp16 methyltransferase activity, we purified nsp16 fused to its cofactor nsp10. Fusion of nsp14 to its cofactor nsp10 has been shown an effective strategy to obtain active recombinant nsp14 exonuclease (Canal et al, this issue).

Following this strategy, we decided to fuse the methyltransferase nsp16 with its cofactor nsp10 that, similarly to for nsp14, would ensure stoichiometric expression of both subunits as well as their association. In the nsp10-16 fusion protein, nsp10 was placed in the Nterminus followed by the linker GSGSGS and nsp16, and the protein was purified using an N-terminus His 14 -SUMO tag. We purified the nsp10-16 fusion protein to homogeneity from E. coli cells in a manner similar to nsp14 ( Figure 6A , see methods). We then confirmed that the nsp10-16 fusion 2'-O-methyltransferase was able to give a dose-dependent reduction in HTRF signal when catalysing methyltransfer to its endogenous substrate me 7 GpppA-RNA ( Figure 6B ). Nsp10-16 was able to methylate me 7 GpppA-RNA efficiently at substantially lower substrate concentrations than nsp14, and demonstrated a Km value ~100x lower for me 7 GpppA-RNA than nsp14 for GpppA-RNA ( Figure 6C ). Finally, we then used the nsp10-16 methyltransferase assay to test specificity of our 4 validated compounds towards the inhibition of nsp14 methyltransferase activity. When using this assay, none of the nsp14 inhibitors were able to inhibit nsp10-16, however, the SAM-competitor sinefungin was able to inhibit methyltransfer by nsp10-16 ( Figure 6D ). Therefore, all four compounds appear to be specific inhibitors of nsp14 methyltransferase.

In order to check if our compounds identified in vitro had any inhibitory effects on viral replication in mammalian cells, we utilised a SARS-CoV-2 viral infection assay using VERO E6 cells that are a model cell line for viral infection assays. We assayed viral replicative capacity by infecting cells with a constant amount of SARS-CoV-2 in the presence of inhibitor. We then quantified viral replicative capacity in fixed cells 22 hours post-infection using a fluorescent anti-Nucleoprotein antibody. This method allows the co-determination of cell viability through DNA staining, and ensures that our compounds were not reducing viral load through cytotoxicity rather than inhibiting viral replication ( Figure 7A ).

All four compounds showed antiviral activity at or below 80 μM, with limited cytotoxic effects ( Figure 7B ). The most effective compounds were PF-03882845 (EC 50 = 10.97 μM), Inauhzin (EC 50 = 12.96 μM) and Trifluperidol (EC 50 = 14.9 μM). Lomeguatrib was less effective at inhibiting viral replication, with an EC 50 = 59.84 μM. Concentrations at which Lomeguatrib was effective also came with slight cytotoxicity, whereas all three of PF-03882845, Inauhzin, and Trifluperidol show little to no cytotoxicity at their EC 50 concentrations ( Figure 7C ). This establishes that the compounds we identified in vitro have antiviral activity in mammalian cells with limited cytotoxic effects.

At the moment, remdesivir is the only antiviral compound that can be taken both prophylactically and therapeutically for the treatment of COVID-19. Other treatments for COVID-19 are solely therapeutics that either modulate the immune response (such as dexamethasone) or are antibody-based therapeutics (such as bamlanivimab). Combination therapy, the use of two or more drugs with different modes of action, is a tested therapeutic strategy for the treatment of some diseases. In addition to achieving better physiological outcomes, combination therapies may also be an effective strategy for limiting antiviral drug resistance [27] . Therefore, we wished to see if the identified compounds from our screen demonstrated synergistic effects with remdesivir, which might warrant exploration for combination therapies and prophylaxis.

Although remdesivir is effective at inhibiting SARS-CoV-2 replication in cellular models, it is known that remdesivir has limited capacity to reduce viral titre in VERO E6 cells at 1 μM and below ( Figure 8A ) (25) . Therefore, we conducted similar viral infection assays as described previously, but in additional presence of remdesivir at a concentration of 0.5 μM ( Figure 8B ). Trifluperidol EC 50 +Rem.= 5.05 μM Lomeguatrib EC 50 +Rem. = 44.14 μM), indicating a potential synergy between our identified nsp14 inhibitors and remdesivir ( Figure 8C ).

The SARS-CoV-2 RNA methyltransferases have been somewhat overlooked as a therapeutic target, currently with no characterised inhibitors in vitro or in vivo. Here we describe the purification and characterisation of both RNA cap methyltransferases encoded by the SARS-CoV-2 genome. In addition, we describe the successful discovery of novel inhibitors of nsp14 methyltransferase activity in vitro, which are not only effective in cellbased assays at low micromolar concentration, but demonstrate synergy with the only approved SARS-CoV-2 therapeutic, remdesivir. We show for the first time that SARS-Cov-2 nsp14 in vitro inhibitors are effective at cessation of viral replication, strongly suggesting that the methyltransferase activity of nsp14 is essential for coronavirus replication.

We initially demonstrated that a commercially available HTRF assay was able to detect the methyltransferase activity of both nsp14 and an nsp10-16 fusion protein. This level of quantitation allowed us to accurately determine the K m values of nsp14 against: GTP, free GpppA, and GpppA-capped RNA; as well as the K m for me 7 GpppA-RNA for nsp10-16. This revealed that whilst nsp14 preserved the ability to methylate guanine without need for the complete cap structure or attached RNA, that the K m values for these substrates are significantly (~100x) higher than the K m of me 7 GpppA-RNA for nsp10-16. Thus, it appears that nsp14 is able to methylate a broad variety of substrates at the expense of a higher K m , whereas nsp10-16 has a lower K m but is known to only methylate capped RNAs [28].

We identified and validated four antiviral compounds that were potential inhibitors of nsp14 carried forward for phase III trials, they may be safe for human use. Trifluperidol is a licenced therapeutic currently in use for the treatment of psychoses including schizophrenia but has relatively severe side effects that limit its potential use as a prophylactic treatment for COVID-19, but may warrant its exploration as a post-infection antiviral. Finally, Inauhzin has previously been characterised as an inhibitor of SIRT1 (IC 50 ~ 1 μM) [33], but has not yet been taken forward into human trials. Although all are able to inhibit nsp14, these compounds have no obvious common chemical similarities ( Figure S3 ), raising the potential of multiple inhibitory binding modes that may be exploited to generate future, more potent, nsp14 inhibitors.

The potency of inhibitors in mammalian cells was similar to the in vitro inhibition of Nsp14 for Lomeguatrib and Trifluperidol. The efficacy of PF-03882845 was significantly reduced in cells compared to the biochemical assays, which might suggest issues with cell permeability or that the drug is actively metabolised into a non-inhibitory form within cells. Inauhzin presented an interesting result as it appeared to be more potent in cells than in the biochemical assays. Perhaps Inauhzin, as a sirtuin inhibitor, for example, has other cellular effects aside from the inhibition of Nsp14 that may contribute to a reduction in viral infectivity. Although we cannot rule out that inhibition of other cellular pathways might be contributing to the observed reduction of viral load in cells, aside from the inhibition of Nsp14, we note that the IC 50 values for the known functions of Lomeguatrib, Trifluperidol, and PF-03882845 are far below the IC 50 values for viral load reduction. Therefore, for these three compounds, it is likely that they exert their effects on reduction of viral load through the inhibition of Nsp14. Readings were taken with a lag time of 60 μs after excitation at λ=337nm. Readings were taken emission wavelengths of λ=665nm and λ=620nm. The experimental HTRF ratio (HTRF exp ) was then calculated as ratio of emission intensities: λ=665/λ=620. To reach the normalised HTRF ratio, HTRF ratio measurements were also taken of wells without enzyme (E 0 ) and without SAH-d2 (d2 0 ), representing the maximum and minimum achievable HTRF values, respectively. The normalised HTRF ratio was then calculated as a linear transformation of the experimental HTRF ratio, the E 0 ratio, and the d2 0 ratio:

The buffer used for the nsp16 assay was 40 mM HEPES-KOH, pH7.6, 1 mM DTT, 5 mM MgCl 2 , 10% glycerol, 0.02% Tween-20, 1.3 mM substrate and 100 nM enzyme, unless stated otherwise. The inhibitor was pre-incubated with enzyme for 10 min at RT. EPIgenous Methyltransferase kit (Cisbio Bioassays) was used for the detection of methyltransferase activity as with nsp14.

High-throughput screening was performed using a custom compound collection assembled from commercial sources (Sigma, Selleck, Enzo, Tocris, Calbiochem, and Symansis). 2.5 nl of a 10mM stock of the compounds dissolved in DMSO were arrayed and pre-dispensed into the assay plates using an Echo 550 (Labcyte), before being sealed and stored at -80C until screening day. When selecting 'hit' compounds, we removed aberrant high Z-scores due to screening errors. These screening errors arose due to systematic dispensation errors, and occurred within the same repeated wells, or cluster of wells within the screen. These wells were either discarded completely from all plates, or hits within these wells were discarded after the initial Z-score cut-off. This resulted in approximately 1% data loss. The initial Z' factor to assess screen quality was conducted including these outlier wells, and therefore provides a conservative estimate of screen quality. Recombinant mAb production) and cellular DNA using DRAQ7 (ABCAM). Whole-well imaging at 5x was carried out using an Opera Phenix (Perkin Elmer) and fluorescent areas and intensity calculated using the Phenix-associated software Harmony (Perkin Elmer). 

All data associated with this paper will be deposited in FigShare (https://figshare.com/). b) nsp14 and nsp14 D331A were incubated with radiolabelled GpppG as above, with a negative control given in which no enzyme was added. Errors given are 95% confidence ranges. 

Normalised HTRF values for other screen hit compounds in validation. All reactions were run with 10 nM nsp14, 1 μM SAM, and 0.11 mM GpppA cap analogue for 20 minutes at room temperature before quenching with NaCl (see methods)

Chemical structures of nsp14 inhibitors validated in HTRF assays and in cellular infection models.

List of 63 compounds that were considered "hits" from primary screening, and their associated Z-scores. 

WHO Declares COVID-19 a Pandemic

Origin and evolution of pathogenic coronaviruses

Clinical, molecular, and epidemiological characterization of the SARS-CoV-2 virus and the Coronavirus Disease 2019 (COVID-19), a comprehensive literature review

Coronavirus biology and replication: implications for SARS-CoV-2

CoV-GLUE: A Web Application for Tracking SARS-CoV-2 Genomic Variation

Evidence of escape of SARS-CoV-2 variant B.1.351 from natural and vaccine-induced sera

The coding capacity of SARS-CoV-2

mRNA capping: biological functions and applications

Shatkin, 5'-termini of reovirus mRNA: ability of viral cores to form caps post-transcriptionally

Structure, mechanism, and evolution of the mRNA capping apparatus

Post-transcriptional modifications of mRNA. Purification and characterization of cap I and cap II RNA (nucleoside-2'-)-methyltransferases from HeLa cells

Mutagenesis of S-Adenosyl-l-Methionine-Binding Residues in Coronavirus nsp14 N7-Methyltransferase Demonstrates Differing Requirements for Genome Translation and Resistance to Innate Immunity

Cytoplasmic foci are sites of mRNA decay in human cells

Processing the message: structural insights into capping and decapping mRNA

Innate immune recognition of viral infection

Ribose 2'-O-methylation provides a molecular signature for the distinction of self and non-self mRNA dependent on the RNA sensor Mda5

2'-O methylation of the viral mRNA cap evades host restriction by IFIT family members

Functional screen reveals SARS coronavirus nonstructural protein nsp14 as a novel cap N7 methyltransferase

IFIT1: A dual sensor and effector molecule that detects non-2'-O methylated viral RNA and inhibits its translation

Coronavirus non-structural protein 16: evasion, attenuation, and possible treatments

mRNA cap regulation in mammalian cell function and fate

RNA 3'-end mismatch excision by the severe acute respiratory syndrome coronavirus nonstructural protein nsp10/nsp14 exoribonuclease complex

Crystal structure and functional analysis of the SARScoronavirus RNA cap 2'-O-methyltransferase nsp10/nsp16 complex

Key steps in ERAD of luminal ER proteins reconstituted with purified components

Characterization of the guanine-N7 methyltransferase activity of coronavirus nsp14 on nucleotide GTP

Structural basis and functional analysis of the SARS coronavirus nsp14-nsp10 complex

Combinatorial drug therapy for cancer in the post-genomic era

Author Contributions

Conceptualization, Methodology, Validation, Formal analysis, Investigation, Resources, Writing -Original Draft, Writing -Review & Editing, Visualisation

Conceptualization, Methodology, Validation, Formal analysis, Investigation, Resources, Writing -Original Draft, Writing -Review & Editing. Rachel Ulferts: Methodology, investigation. Mary Wu: Methodology, resources, investigation. Tom Deegan: Methodology, resources, investigation, Writing -Review & Editing. Ryo Fujisawa: Methodology, resources, investigation, Writing -Review & Editing

Methodology, Investigation, Formal analysis, Validation, Resources, Writing -Review & Editing. Chew Theng Lim: Methodology, Investigation, Formal analysis, Validation, Resources, Writing -Review & Editing. Clovis Basier: Investigation. Berta Canal: Resources, Writing -Review &

Resources. Lucy Drury: Investigation. Allison W. McClure: Resources, Writing -Review & Editing. Emma L. Roberts: Resources. Florian Weissmann: Resources

Cowling: Supervision, project administration, investigation, resources, visualisation, formal analysis, writing -review & editing. Michael Howell: Supervision, project administration, writing -review & editing. Karim Labib: Supervision, project administration, writingreview & editing

We are grateful to Anne Early, Tobias Fuller, and Jingkun Zeng for their assistance. This work was supported by the Francis Crick Institute, which receives its core funding from