Identification of a guanine-specific pocket in the protein N of SARS-CoV-2

1 Identification of a guanine-specific pocket in the protein N of SARS-CoV-2 1 J. Rafael Ciges-Tomas*, María Luisa Franco, Marçal Vilar* 2 Instituto de Biomedicina de Valencia-CSIC Spanish National Research Council 3 C/Jaime Roig, 11. 46010-Valencia, Spain. 4 5 6 7 8 9 10 * Address correspondence to: 11 J. Rafael Ciges-Tomas 12 [email protected] 13 Marçal Vilar 14 m[email protected]sic.es 15 16 17 18 19 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted June 21, 2022. ; https://doi.org/10.1101/2022.06.21.496991doi: bioRxiv preprint .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted June 21, 2022. ; https://doi.org/10.1101/2022.06.21.496991doi: bioRxiv preprint .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted June 21, 2022. ; https://doi.org/10.1101/2022.06.21.496991doi: bioRxiv preprint .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted June 21, 2022. ; https://doi.org/10.1101/2022.06.21.496991doi: bioRxiv preprint 2 ABSTRACT 20 The SARS-CoV-2 nucleocapsid protein (N) is responsible for RNA binding. Here 21 we report the crystal structure of the C-terminal domain (NCTD) in open and closed 22 conformations and in complex with guanine triphosphate, GTP. The crystal 23 structure and biochemical studies reveals a specific interaction between the 24 guanine, a nucleotide enriched in the packaging signals regions of coronaviruses, 25 and a highly conserved tryptophan residue (W330). In addition, EMSA assays with 26 SARS-CoV-2 derived RNA hairpin loops from a putative viral packaging sequence 27 showed the preference interaction of the N-CTD to RNA oligonucleotides 28 containing G and the loss of the specificity in the mutant W330A. Here we propose 29 that this interaction may facilitate the viral assembly process. In summary we have 30 identified a specific guanine-binding pocket in the N protein that may be used to 31 design viral assembly inhibitors. 32 33 INTRODUCTION 34 The severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has emerged 35 as a pandemic virus causing a global human health crisis. SARS-CoV-2 shares 36 ≈ 80% genome identity to SARS-CoV 1. At the 5´-terminus two thirds of its genome 37 there are located the ORF1a and ORF1a/b encoding sixteen non-structural 38 proteins (NSPs 1-16), which are responsible of establishing the cellular conditions 39 favourable for viral infection and viral mRNA synthesis 2,3. The 3´-terminus one 40 third of the genome contains the ORFs for the four structural proteins: Spike (S), 41 Envelope (E), Membrane (M) and Nucleocapsid (N) and other accessory proteins 42 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted June 21, 2022. ; https://doi.org/10.1101/2022.06.21.496991doi: bioRxiv preprint 3 3. S, E and M are transmembrane proteins that are incorporated into the viral lipid 43 envelope 4. The protein N is a multifunctional protein that packs the genomic RNA 44 and enhances the efficiency of virus transcription and assembly 5–13. N contains an 45 N-terminal domain, involved in RNA binding, and a C-terminal (NCTD) involved in 46 RNA binding, protein dimerization, and interaction with the viral membrane protein 47 (M) 14,15. In the case of the Murine Hepatitis coronavirus, MHV (a betacoronavirus) 48 the NCTD has been implicated in the interaction with a conserved packaging signal 49 (PS) necessary for viral RNA packaging 16. 50 To date, eleven crystallographic structures of the SARS-CoV-2 NCTD have been 51 deposited at the protein databank presenting minor differences between them 52 (PDB: 7N0I, 7F2B, 7F2E, 7C22, 7CE0, 6ZCO, 6YUN, 6WZO, 6WZQ, 6WJI, 7DE1, 53 7F2B, 7F2E). Some of them has been already published and others are pending 54 publication, 17–2122, however none of these contain RNA or nucleotides that provide 55 any insight on the RNA recognition process. The SARS-CoV-2 NCTD folds into a 56 helical core with a protruding β -hairpin employed for protein dimerization. In this 57 study, we report three SARS-CoV-2 NCTD structures that bring novel features and 58 provide a co-crystal structure with GTP that might provide an insight on the 59 specificity of the PS recognition. 60 RESULTS 61 Structure of SARS-CoV-2 NCTD in open and closed conformations. 62 The crystal structure of NCTD was solved by molecular replacement to a resolution 63 of 1.94 Å (Table I). The crystal asymmetric unit contains two homodimers of NCTD 64 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted June 21, 2022. ; https://doi.org/10.1101/2022.06.21.496991doi: bioRxiv preprint 4 with almost identical conformation (rmsd 0.14 Å). As described previously, it folds 65 into five α -helices ( α 1α 5), two 310 helices ( η 1η 2) and two β -strands ( β 1β 2), 66 presenting the following Nto Csequence for the structural elements: η 1α 1α 267 α 3α 4β 1β 2α 5η 2 17–21. (Figure 1a and Supplementary Figure 1a). Two 68 monomers are intertwined through the β -hairpin forming a four antiparallel β -sheet 69 on one face of the dimer (Figure 1). In striking contrast to the previously described 70 structures, the superposition of the subunits in our structure shows a ≈ 5.5 Å 71 movement of the β -hairpin (Figure 1b). One subunit presents the β -hairpin in an 72 extended and previously unseen conformation that we named “open”. Whereas the 73 other subunit shows a β -hairpin in a flexed conformation that we named “closed”, 74 similar to other structures of this protein (Figure 1b and Supplementary Figure 1b). 75 The alternative conformation of the β -hairpin is associated with the structural 76 displacement of the loop between α 1 and α 2 (residues 280-283) (Figure 1b). 77 In the closed conformation the β -hairpin interacts with the C-terminal proline 78 residue (P364) of a symmetry-related protein (Figure 1c,d and Supplementary 79 Figure 1b). This interdimeric interaction between W330 and P364 through a π - π 80 stacking (face-to-face rings interaction), and making hydrogen bonds with T325 81 and S327 (Figure 1d), is a common feature with seven NCTD structures determined 82 previously. But strikingly, our structure shows that in the open conformation, the 83 side chain of the residue W330 is rotated 180º towards the β -hairpin making 84 hydrophobic contacts with the side chain of the residues T325 and S327 (Figure 85 1c). Therefore, our structure shows that upon a side chain arrangement the β - 86 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted June 21, 2022. ; https://doi.org/10.1101/2022.06.21.496991doi: bioRxiv preprint 5 hairpin changes from an open conformation to a closed conformation that might 87 favour the interdimeric interaction. 88 Previous structures of SARS-CoV-2 NCTD showed that the β -hairpin that does not 89 interact with the symmetry-related molecule have a molecule of acetate (PDB 90 7C22) 20 or sulphate (PDB 6WZQ) 19 interacting with the side chain of W330 91 (Supplementary Figure 2a,b). 92 Structure of SARS-CoV-2 NCTD bound to GTP 93 Based on the structural data we hypothesized that W330 could be a suitable 94 residue for RNA recognition, as W330 could interact with the phosphate moiety, 95 similarly to the sulphate, or via π - π stacking with the nitrogen base. To test our 96 hypothesis, we attempted to individually co-crystallize NCTD with the oxynucleotides 97 UTP, ATP, CTP or GTP. Although crystals were obtained in all the mixtures, only 98 the GTP co-crystallized with NCTD. The structure of the binary complex NCTD-GTP 99 was solved in two different space groups, P21 and P1, to a resolution of 1.8 Å and 100 2 Å, respectively (Table I). Both crystal forms contain two protein dimers in the 101 asymmetric unit. However, in the space group P21 there is one molecule of GTP 102 bound to one of the dimers, whereas in the P1 the asymmetric unit contains two 103 molecules of GTP, each bound to one dimer (Figure 2). The electron density map 104 is well defined for the three GTP molecules, which present temperature factors that 105 increase from the guanine moiety to the phosphates (Figure 2 and Supplementary 106 Figure 3a), suggesting that the guanine is well anchored to the protein and the 107 phosphates are more flexible. In fact, the phosphates β and γ in the crystal P1 108 show two alternative dispositions, being one identical to the crystal P21 (Figure 2). 109 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted June 21, 2022. ; https://doi.org/10.1101/2022.06.21.496991doi: bioRxiv preprint 6 In the two crystal forms the GTP binds to a cleft between the two subunits of the 110 dimer, and adjacent to the β -hairpin in the closed conformation (Figure 3a). The 111 cleft corresponds to a cavity adjacent and perpendicular to W330, reducing the 112 accessibility of the tryptophan to the solvent (Figure 3a and Supplementary Figure 113 3b). 114 One side of the guanine ring establishes a π - π T-shaped interaction (edge-to-face) 115 with the indole ring of W330 and hydrophobic interactions with K338 (Figure 3a). 116 The other side of the guanine ring stacks over the guanidinium group of R259, 117 which makes hydrogen bonds with the OHgroups of the ribose ring (Figure 3a). 118 R259 and R262 are responsible for multiple hydrogen-bonding contacts with the β - 119 and γ - phosphates (Figure 3a). At the bottom of the cleft the guanine moiety 120 interacts with the M317 and with the main chain of the residues K338 and A336 121 (Figure 3a). These interactions suggest high specificity for guanine, as other 122 nitrogen base have not as much favourable interactions as the guanine 123 (Supplementary Figure 4). For instance, adenine lacks the C6 oxygen that makes 124 interactions with the M317, K338 and A336 in the case of guanine (Supplementary 125 Figure 4). All protein-ligand contacts are summarized in the Supplementary Table I. 126 Importantly, the C-terminal tail of the symmetry-related dimer also contributes to 127 the GTP binding, covering the cleft as a lid, with Van der Waals and hydrophobic 128 interactions between T362´ and the ribose and with the π - π stacking between 129 P364´ and W330 (Figure 3a). The structure suggests that the GTP binding favours 130 swapping of the β -hairpin and interdimeric interaction. 131 132 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted June 21, 2022. ; https://doi.org/10.1101/2022.06.21.496991doi: bioRxiv preprint 7 Characterization of GTP binding 133 To confirm our structural results, we studied the binding of GTP in solution using 134 the intrinsic fluorescence of tryptophan. The protein contains two tryptophan 135 residues: W330, located in the β -hairpin and exposed to the solvent, and W301 136 partially buried in the protein core (Supplementary Figure 5a). NCTD shows a 137 fluorescence peak profile with a maximum fluorescence emission centred at 340 138 nm that decreases concomitantly with the increasing concentration of acrylamide, 139 confirming the tryptophan fluorescence is quenchable (Figure 3b and 140 Supplementary Figure 5b). 141 We also made the mutant NCTD-W330A, replacing W330 to alanine, and confirmed 142 that the fluorescence of the remaining W301 is also quenchable (Figure 3b). We 143 determined the quenching of the fluorescence in absence or presence of 0.5 mM 144 GTP. Whereas GTP decreased significantly the quenching slope of NCTD (from 145 6.31 ± 0.08 to 5.17 ± 0.09; p-val ****, n=7 and n=4), the quenching of NCTD-W330A 146 (1.99 ± 0.06; n=4) is not affected by the presence of GTP (1.74 ± 0.06; n=4), 147 indicating that only the accessibility of W330 is affected by GTP (Figure 3b and 148 Supplementary Tables II and III). This effect is not observed in the presence of 149 ATP or UTP, but it is shown in the presence of CTP in a much lesser degree (p-val 150 **) (Supplementary Figure 6 and Supplementary Table III). We hypothesized that 151 the binding of GTP to the homodimerization interface could stabilize the protein 152 dimer. To confirm this hypothesis, we perform differential scanning fluorimetry 153 (DSF) experiments with NCTD and NCTD-W330A in presence or absence of 5 mM GTP 154 (Figure 3c). NCTD presents a sigmoidal trajectory in response to temperature, with a 155 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted June 21, 2022. ; https://doi.org/10.1101/2022.06.21.496991doi: bioRxiv preprint 8 melting temperature (Tm) of 48.96 ºC (±0.03) that increased to 49.97 ºC (±0.03) in 156 presence of GTP (Figure 3c). This difference is statistically significant (Aikake test, 157 see Methods). In contrast, the mutant NCTD-W330A shows a Tm= 47.98 ºC (±0.07) 158 unaffected by the presence of GTP, Tm= 47.67 (±0.07) (Figure 3c). 159 We also measure the binding affinity (Kd) for GTP of NCTD and NCTD-W330A using 160 Microscale Thermophoresis (MST) (Figure 3d and Supplementary Table IV). NCTD 161 presents a Kd value of 196 μ M for GTP, which increases to 858 μ M in the mutant 162 NCTD-W330A. This result confirms the participation of W330 for the GTP binding in 163 solution (Figure 3d). Although the affinity for GTP is in μ M range, it should be taken 164 into account that this Kd value corresponds to the affinity of one nucleotide of a 165 protein that binds RNA, in which a synergistic effect of multiple nucleotide 166 interactions would favour the avidity (see below for RNA hairpin binding affinity). 167 168 W330 confers specificity of the N-CTD to the guanine-containing RNA 169 oligonucleotides 170 Phylogenetic analysis suggest the existence of conserved short structured repeats, 171 termed repetitive structural motifs or RSM in the gRNAs from coronaviruses, that 172 may facilitate the genome packaging by specific interaction to protein factors 23. 173 These RNA elements are called stem-loop (SL) 1 to 6 and SL5 is the most 174 interesting as displays a conserved sequence 5’-UUYCGU-3’ in the tripartite apical 175 substructures, SL5a-c. Interestingly these sequences contains a highly conserved 176 G (in bold) in nearly all alphaand beta-coronavirus 23. 177 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted June 21, 2022. ; https://doi.org/10.1101/2022.06.21.496991doi: bioRxiv preprint 9 We analysed the binding of the N-CTD to an RNA oligonucleotide derived from the 178 apical region of the stem loop 5a (SL5a) of SARS-CoV-2 gRNA (Oligo-G, Figure 179 4). Incubation of increasing concentrations of NCTD to the Oligo-G showed the 180 binding and the protein-RNA complex formation, Kd=32 ± 5 nM, (Figure 4). 181 Interestingly mutation of the only G to U, Oligo-U, induces a decrease in the 182 binding affinity, Kd=140 ± 29 nM (Figures 4a and b). To study the role of the 183 W330, we generated and purified the mutant NCTD-W330A. EMSA assays showed 184 that the mutant W330A binds less efficiently to the oligonucleotide Oligo-G (Kd = 185 130 ± 1 nM, n= 3 ) than the NCTD-wt to Oligo-G (Kd=32 ± 5 nM, n=3), and with a 186 similar affinity to Oligo-U (W330A:Oligo-U Kd=90 ± 19 nM, n=3 and NCTD187 wt:Oligo-U Kd = 140 ± 34 nM, n=3 ), (Figure 4a and b). This supports the specific 188 recognition of the guanine in the RNA hairpin by the N-CTD and that the mutation 189 W330A losses this specific recognition. 190 DISCUSSION 191 In coronavirus assembly of the genomic RNA, gRNA, is a fundamental problem as 192 infection produces several copies of subgenomics RNAs, sgRNA. The specific 193 interaction of N to PS sequences will assure the packaging of gRNA versus other 194 sgRNAs, mRNAs from the host and other RNAs from the infected cells 24. In the 195 alphaand beta-coronavirus, like SARS-CoV, SARS-CoV-2 and MERS-CoV 196 phylogenetic data suggest the conservation of several of the RNA stem loops (SL) 197 such as SL5a-c that might constitute regions that participate as authentic PS 23. 198 Although we did not succeed to co-crystallize N-CTD with the SL5a hairpin, we found 199 that the NCTD is able to specifically recognize the SL5a apical sequence from 200 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted June 21, 2022. ; https://doi.org/10.1101/2022.06.21.496991doi: bioRxiv preprint 16 added to the mixtures and gels were run at 4º C for 20 min at 100V. Gels were 336 revealed in an Odyssey Imaging System (LI-COR) and images analyzed using 337 ImageStudio. Quantification was done by normalizing the signal from the free oligo 338 to 1.0 at 0 μ M protein and plotting the free oligo versus the concentration of the N339 CTD protein (in Log scale). Data was adjusted to a nonlinear curve of the form One 340 site equation (Prism 6 software). Data of at least three independent experiments 341 were plotted. Error bars represent the standard error of the mean. 342 Statistics and Reproducibility 343 Statistical analyses were used in all graphs represented in the Figure 3 and 4. 344 Figure 3b was analyzed using a one-way ANOVA with Tukey’s multiple 345 comparison test. Figure 3c was analyzed by comparing the nonlinear fitting curve 346 using the Aikake test (AIC) with a >99.9% probability that the Tm of N and Tm of 347 N+GTP is different, with an AIC of 221.1 (using GraphPad prism 6.0). In Figure 4 348 statistics was analyzed by comparing the nonlinear fitting curve (one site equation 349 in GraphPad Prism 6.0) using the Aikake test (AIC) with a >99.9% probability that 350 the Kd of the binding of N-CTD to Oligo-G is statistically different from the Kd of the 351 binding of N-CTD to Oligo-U is with a AIC of 24.1. 352 REFERENCES 353 1. Zhou, P. et al. A pneumonia outbreak associated with a new coronavirus of 354 probable bat origin. Nature 579, 270–273 (2020). 355 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted June 21, 2022. ; https://doi.org/10.1101/2022.06.21.496991doi: bioRxiv preprint 17 2. Prentice, E., McAuliffe, J., Lu, X., Subbarao, K. & Denison, M. R. 356 Identification and characterization of severe acute respiratory syndrome 357 coronavirus replicase proteins. J. Virol. 78, 9977–9986 (2004). 358 3. Wu, F. et al. A new coronavirus associated with human respiratory disease 359 in China. Nature 579, 265–269 (2020). 360 4. Neuman, B. W. & Buchmeier, M. J. Supramolecular architecture of the 361 coronavirus particle. Adv. Virus Res. 96, 1–27 (2016). 362 5. Cong, Y. et al. Nucleocapsid Protein Recruitment to Replication363 Transcription Complexes Plays a Crucial Role in Coronaviral Life Cycle. J. 364 Virol. 94, (2020). 365 6. Surjit, M., Liu, B., Chow, V. T. K. & Lal, S. K. The nucleocapsid protein of 366 severe acute respiratory syndrome-coronavirus inhibits the activity of cyclin367 cyclin-dependent kinase complex and blocks S phase progression in 368 mammalian cells. J. Biol. Chem. 281, 10669–10681 (2006). 369 7. Surjit, M., Liu, B., Jameel, S., Chow, V. T. K. & Lal, S. K. The SARS 370 coronavirus nucleocapsid protein induces actin reorganization and apoptosis 371 in COS-1 cells in the absence of growth factors. Biochem. J. 383, 13–18 372 (2004). 373 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted June 21, 2022. ; https://doi.org/10.1101/2022.06.21.496991doi: bioRxiv preprint 18 8. Zeng, Y. et al. The nucleocapsid protein of SARS-associated coronavirus 374 inhibits B23 phosphorylation. Biochem. Biophys. Res. Commun. 369, 287– 375 291 (2008). 376 9. Zhang, L. et al. SARS-CoV nucleocapsid protein induced apoptosis of 377 COS-1 mediated by the mitochondrial pathway. Artif. Cells Blood Substit. 378 Immobil. Biotechnol. 35, 237–253 (2007). 379 10. Lu, X., Pan, J., Tao, J. & Guo, D. SARS-CoV nucleocapsid protein 380 antagonizes IFNβ response by targeting initial step of IFNβ induction 381 pathway, and its C-terminal region is critical for the antagonism. Virus Genes 382 42, 37–45 (2011). 383 11. Yasui, F. et al. Prior immunization with severe acute respiratory syndrome 384 (SARS)-associated coronavirus (SARS-CoV) nucleocapsid protein causes 385 severe pneumonia in mice infected with SARS-CoV. J. Immunol. 181, 6337– 386 6348 (2008). 387 12. Zúñiga, S. et al. Coronavirus nucleocapsid protein facilitates template 388 switching and is required for efficient transcription. J. Virol. 84, 2169–2175 389 (2010). 390 13. Mu, J. et al. SARS-CoV-2-encoded nucleocapsid protein acts as a viral 391 suppressor of RNA interference in cells. Sci. China Life Sci. 63, 1413–1416 392 (2020). 393 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted June 21, 2022. ; https://doi.org/10.1101/2022.06.21.496991doi: bioRxiv preprint 19 14. Luo, H. et al. Severe acute respiratory syndrome coronavirus membrane 394 protein interacts with nucleocapsid protein mostly through their carboxyl 395 termini by electrostatic attraction. Int. J. Biochem. Cell Biol. 38, 589–599 396 (2006). 397 15. Lu, S. et al. The SARS-CoV-2 nucleocapsid phosphoprotein forms mutually 398 exclusive condensates with RNA and the membrane-associated M protein. 399 Nat. Commun. 12, 502 (2021). 400 16. Kuo, L., Koetzner, C. A. & Masters, P. S. A key role for the carboxy-terminal 401 tail of the murine coronavirus nucleocapsid protein in coordination of 402 genome packaging. Virology 494, 100–107 (2016). 403 17. Peng, Y. et al. Structures of the SARS-CoV-2 nucleocapsid and their 404 perspectives for drug design. EMBO J. 39, e105938 (2020). 405 18. Zinzula, L. et al. High-resolution structure and biophysical characterization 406 of the nucleocapsid phosphoprotein dimerization domain from the Covid-19 407 severe acute respiratory syndrome coronavirus 2. Biochem. Biophys. Res. 408 Commun. 538, 54–62 (2021). 409 19. Ye, Q., West, A. M. V., Silletti, S. & Corbett, K. D. Architecture and self410 assembly of the SARS-CoV-2 nucleocapsid protein. Protein Sci. 29, 1890– 411 1901 (2020). 412 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted June 21, 2022. ; https://doi.org/10.1101/2022.06.21.496991doi: bioRxiv preprint 20 20. Zhou, R., Zeng, R., von Brunn, A. & Lei, J. Structural characterization of the 413 C-terminal domain of SARS-CoV-2 nucleocapsid protein. Mol. Biomed. 1, 2 414 (2020). 415 21. Yang, M. et al. Structural Insight Into the SARS-CoV-2 Nucleocapsid 416 Protein C-Terminal Domain Reveals a Novel Recognition Mechanism for 417 Viral Transcriptional Regulatory Sequences. Front. Chem. 8, 624765 (2020). 418 22. Ye, Q., Lu, S. & Corbett, K. D. Structural Basis for SARS-CoV-2 419 Nucleocapsid Protein Recognition by Single-Domain Antibodies. Front. 420 Immunol. 12, 719037 (2021). 421 23. Chen, S.-C., Olsthoorn, R. C. L. & Yu, C.-H. Structural phylogenetic 422 analysis reveals lineage-specific RNA repetitive structural motifs in all 423 coronaviruses and associated variations in SARS-CoV-2. Virus Evol. 7, 424 veab021 (2021). 425 24. Masters, P. S. Coronavirus genomic RNA packaging. Virology 537, 198– 426 207 (2019). 427 25. Yu, I.-M., Oldham, M. L., Zhang, J. & Chen, J. Crystal structure of the 428 severe acute respiratory syndrome (SARS) coronavirus nucleocapsid 429 protein dimerization domain reveals evolutionary linkage between corona430 and arteriviridae. J. Biol. Chem. 281, 17134–17139 (2006). 431 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted June 21, 2022. ; https://doi.org/10.1101/2022.06.21.496991doi: bioRxiv preprint 21 26. Nguyen, T. H. V. et al. Structure and oligomerization state of the C-terminal 432 region of the Middle East respiratory syndrome coronavirus nucleoprotein. 433 Acta Crystallogr. D Struct. Biol. 75, 8–15 (2019). 434 27. Szelazek, B. et al. Structural characterization of human coronavirus NL63 N 435 protein. J. Virol. 91, (2017). 436 28. Wienken, C. J., Baaske, P., Rothbauer, U., Braun, D. & Duhr, S. Protein437 binding assays in biological liquids using microscale thermophoresis. Nat. 438 Commun. 1, 100 (2010). 439 29. Powell, H. R., Johnson, O. & Leslie, A. G. W. Autoindexing diffraction 440 images with iMosflm. Acta Crystallogr. D Biol. Crystallogr. 69, 1195–1203 441 (2013). 442 30. Battye, T. G. G., Kontogiannis, L., Johnson, O., Powell, H. R. & Leslie, A. 443 G. W. iMOSFLM: a new graphical interface for diffraction-image processing 444 with MOSFLM. Acta Crystallogr. D Biol. Crystallogr. 67, 271–281 (2011). 445 31. Evans, P. R. & Murshudov, G. N. How good are my data and what is the 446 resolution? Acta Crystallogr. D Biol. Crystallogr. 69, 1204–1214 (2013). 447 32. Winn, M. D. et al. Overview of the CCP4 suite and current developments. 448 Acta Crystallogr. D Biol. Crystallogr. 67, 235–242 (2011). 449 33. McCoy, A. J. et al. Phaser crystallographic software. J. Appl. Crystallogr. 450 40, 658–674 (2007). 451 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted June 21, 2022. ; https://doi.org/10.1101/2022.06.21.496991doi: bioRxiv preprint 22 34. Murshudov, G. N. et al. REFMAC5 for the refinement of macromolecular 452 crystal structures. Acta Crystallogr. D Biol. Crystallogr. 67, 355–367 (2011). 453 35. Emsley, P., Lohkamp, B., Scott, W. G. & Cowtan, K. Features and 454 development of Coot. Acta Crystallogr. D Biol. Crystallogr. 66, 486–501 455 (2010). 456 457 Data availability 458 The coordinates and structure factors of the crystallographic structures have been 459 deposited in the Protein Data Bank with accession codes PDB 7O05, 7O35 and 460 7O36. Graph raw data and gel images are deposited at Mendeley Data at the 461 address https://data.mendeley.com/datasets/8cfsrmd8by/draft?a=a0d0ad46-dfb5462 474e-b9c5-37c59410e6a3. 463 Acknowledgments 464 We would like to thank to Dr. Encarnación Martínez-Salas from CBM-CSIC for her 465 suggestions on the EMSA assays, and Dr. Santiago Ramón-Maiques and Dr. Francisco 466 Del Caño-Ochoa from IBV-CSIC, for their technical support and critical reading of the 467 manuscript and the staff of the synchrotron for technical assistance. The X-ray diffraction 468 experiments were performed in XALOC beamline at ALBA Synchrotron (Barcelona; 469 Spain). This work was supported by the COVID research grant COV20/01265 awarded to 470 M.V. from the Instituto de Salud Carlos III (ISCIII; Spain) and by the European 471 Commission–NextGenerationEU (Regulation EU 2020/2094), through CSIC's Global 472 Health Platform (PTI Salud Global). X-ray diffraction data collection was supported by the 473 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted June 21, 2022. ; https://doi.org/10.1101/2022.06.21.496991doi: bioRxiv preprint 23 Spanish Synchrotron Radiation Facility ALBA through the COVID Proposal 2020074407 474 awarded to J.R.C.-T. and M.V. 475 Author contributions 476 Conceptualization: J.R.C.-T and M.V.; Methodology: J.R.C.-T, M.L.F, and M.V.; 477 Investigation: J.R.C.-T and M.V.; writing-original draft, J.R.C.-T and M.V.; funding 478 acquisition: M.V. 479 Corresponding authors 480 Correspondence to J.Rafael Ciges-Tomas or Marçal Vilar. 481 Competing interest 482 The authors declare NO competing interests. 483 Figure Legends 484 485 Figure 1. Crystal structure of SARS-CoV-2 NCTD in open and closed 486 conformations. 487 a) Cartoon representation of the NCTD dimer. Each monomer is colored in blue and 488 orange, respectively. The β -hairpin and the loop connecting the helices α 1α 2 are 489 highlighted in dark tones. Secondary structural elements and residues are 490 numbered and labeled in order from N to C terminus: the symbol η corresponds to 491 310 helix; α to α -helix and β to β -strand. b) Superimposition of the NCTD monomers. 492 The common helical core is colored in grey. c,d) Detailed view of the open (c) and 493 the closed (d) conformations. The side chain of key residues is shown in sticks with 494 carbon atoms colored according the monomer to which they belong. Hydrophobic 495 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted June 21, 2022. ; https://doi.org/10.1101/2022.06.21.496991doi: bioRxiv preprint 24 and polar interactions are represented as dashed black lines. The C-terminal of the 496 symmetric molecule (sym) is colored in magenta. The symmetric Pro residue is 497 indicated with apostrophe. Nitrogen and oxygen atoms are colored in blue and red, 498 respectively. 499 Figure 2. Structures of SARS-CoV-2 NCTD in complex with GTP. 500 Cartoon representation of the two dimers contained in the asymmetric unit of the 501 crystal in the space group P21 (a) and P1 (b). Each monomer is colored in white 502 and black for one dimer, and orange and blue for the second dimer. GTP is 503 represented in sticks with carbon atoms colored in green. The electron density map 504 2Fo-Fc ( σ =1) of the GTP is represented in blue. The moieties guanine, ribose and 505 phosphates are labelled. Nitrogen, oxygen and phosphorus atoms are colored in 506 blue, red and orange, respectively. The temperature factors (B-factors) of nitrogen 507 and oxygen atoms of the nucleoside and phosphorus atoms of phosphates are 508 indicated below. 509 Figure 3. Crystal structure of SARS-CoV-2 NCTD in complex with GTP. 510 a) Upper panel, the NCTD dimer in complex with GTP is represented in cartoon. 511 Each monomer is colored in blue and orange, respectively. The β -hairpins of the 512 dimer are labeled and highlighted in dark tones. The GTP molecule is shown in 513 sticks with its electron density map 2Fo-Fc ( σ =1) in green color. The close view of 514 the GTP binding site is represented below. The side chain of key residues and the 515 GTP molecule are shown in sticks with carbon atoms colored according the 516 monomer to which they belong. H-bond interactions of the ligand are represented 517 as dashed black lines. The C-terminal of the symmetric molecule (sym) is colored 518 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted June 21, 2022. ; https://doi.org/10.1101/2022.06.21.496991doi: bioRxiv preprint 25 in magenta with residues indicated with apostrophe. Secondary structural elements 519 are numbered and labeled in order from N to C terminus. Nitrogen, oxygen and 520 phosphorus atoms are colored in blue, red and orange, respectively. b) Quenching 521 of fluorescence (Fo/F) represented against increasing concentration of acrylamide 522 for the NCTD and NCTD-W330A proteins in presence or absence of GTP. Trajectories 523 are fitted to a linear regression. Statistical differences are indicated with asterisks 524 (**** p <0.0001, n.s. p >0.05). c) Thermal unfolding curves of NCTD and NCTD-W330A 525 in presence or absence of GTP. Trajectories are fitted to a sigmoid. The 526 corresponding melting temperature (Tm) is indicated. d) GTP binding quantified 527 thermophoretically. GTP is titrated to a constant amount of fluorescently labeled 528 NCTD and NCTD-W330A. The binding affinity (Kd) is indicated. Error bar represent the 529 standard error of the mean of at least four experiments. 530 Figure 4. Binding of NCTD to the apical region of SL5a RNA. 531 (a) Ag-EMSA gels of titration experiments. Binding of NCTD and NCTD-W330A to a 532 ssRNA oligonucleotide from the apical part of the SL5a region of SARS-CoV-2 533 gRNA with a conserved guanine (arrowhead, Oligo-G) or with a G to U mutation 534 (Oligo-U, lower panels). (b) ssRNA oligonucleotide shift with increasing 535 concentration of protein. Protein-ssRNA combination are represented in different 536 colours. Trajectories are fitted to a sigmoidal one-site binding model. Error bar 537 represent the standard error of the mean of at least four experiments. 538 Figure 5. Conservation of the Guanine binding pocket in human 539 coronaviruses. 540 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted June 21, 2022. ; https://doi.org/10.1101/2022.06.21.496991doi: bioRxiv preprint 4 a) b) W330 R259 R262 A336 T282 K338 M317 F274 I337 Q260 GTP SARS-CoV-2 a) MERS c) F338 R261 R264 A344 L284 K346 M319 F276 I345 H262 TMO SARS-CoV-1 b) W331 R260 R263 A337 T283 K339 M318 F275 I338 Q261 NL63 d) Q303 R235 R238 K309 N256 L311 D290 F250 M310 W236 Human Coronaviruses until 2003 2012 2003 η1 α2 α3 α1 α5 β1 β2 α4 α5 η2 Wuhan, 2019 e) 5