key: cord-0950420-8cam6vv7
authors: Tugaeva, Kristina V.; E. D. P. Hawkins, Dorothy; L. R. Smith, Jake; Bayfield, Oliver W.; Ker, De-Sheng; Sysoev, Andrey A.; Klychnikov, Oleg I.; Antson, Alfred A.; Sluchanko, Nikolai N.
title: The mechanism of SARS-CoV-2 nucleocapsid protein recognition by the human 14-3-3 proteins
date: 2021-02-05
journal: J Mol Biol
DOI: 10.1016/j.jmb.2021.166875
sha: 058e80cc81c36023e2d846a4a3845713b7f1573e
doc_id: 950420
cord_uid: 8cam6vv7

The coronavirus nucleocapsid protein (N) controls viral genome packaging and contains numerous phosphorylation sites located within unstructured regions. Binding of phosphorylated SARS-CoV N to the host 14-3-3 protein in the cytoplasm was reported to regulate nucleocytoplasmic N shuttling. All seven isoforms of the human 14-3-3 are abundantly present in tissues vulnerable to SARS-CoV-2, where N can constitute up to ∼1% of expressed proteins during infection. Although the association between 14-3-3 and SARS-CoV-2 N proteins can represent one of the key host-pathogen interactions, its molecular mechanism and the specific critical phosphosites are unknown. Here, we show that phosphorylated SARS-CoV-2 N protein (pN) dimers, reconstituted via bacterial co-expression with protein kinase A, directly associate, in a phosphorylation-dependent manner, with the dimeric 14-3-3 protein, but not with its monomeric mutant. We demonstrate that pN is recognized by all seven human 14-3-3 isoforms with various efficiencies and deduce the apparent KD to selected isoforms, showing that these are in a low micromolar range. Serial truncations pinpointed a critical phosphorylation site to Ser197, which is conserved among related zoonotic coronaviruses and located within the functionally important, SR-rich region of N. The relatively tight 14-3-3/pN association could regulate nucleocytoplasmic shuttling and other functions of N via occlusion of the SR-rich region, and could also hijack cellular pathways by 14-3-3 sequestration. As such, the assembly may represent a valuable target for therapeutic intervention.

The new coronavirus-induced disease, COVID19, has caused a worldwide health crisis with more than 90 million confirmed cases and 1.9 million deaths as of January 2021 [1] . The pathogen responsible, Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2), is highly similar to the causative agent of the SARS outbreak in 2002-2003 (SARS-CoV) and, to a lesser extent, to the Middle East Respiratory Syndrome Coronavirus (MERS-CoV) [2, 3] . Each is vastly more pathogenic and deadly than human coronaviruses HCoV-OC43, HCoV-NL63, HCoV-229E, and HCoV-HKU1 which cause seasonal respiratory diseases [4] . Like SARS-CoV and HCoV-NL63, SARS-CoV-2 uses angiotensin-converting enzyme 2 (ACE2) as entry receptor [5] . The ACE2 expression roughly correlates with the evidenced SARS-CoV-2 presence in different tissue types, which explains the multi-organ character of the disease [6] (Fig. 1) .

In contrast to multiple promising COVID19 vaccine clinical trials [7] [8] [9] , treatment of the disease is currently limited by the absence of approved efficient drugs [10] . The failure of several leading drug candidates in 2020 warrants the search for novel therapeutic targets including not only viral enzymes, but also heterocomplexes involving viral and host cell proteins. Unravelling mechanisms of interaction between the host and pathogen proteins may provide the platform for such progress.

The positive-sense single-stranded RNA genome of SARS-CoV-2 coronavirus encodes approximately 30 proteins which enable cell penetration, replication, viral gene transcription and genome assembly amongst other functions [11] . The 46-kDa SARS-CoV-2 Nucleocapsid (N) protein is 89.1% identical to SARS-CoV N. Genomic analysis of human coronaviruses indicated that N might be the major factor conferring the enhanced pathogenicity to SARS-CoV-2 [4] . N represents the most abundant viral protein in the infected cell [12] [13] [14] , with each assembled virion containing approximately one thousand molecules of N [15] . Given that each infected cell can contain up to 10 5 virions (infectious, defective and incomplete overall) [14] , the number of N molecules in an infected cell can reach 10 8 , accounting for ~1% of a total number of cellular proteins (~10 10 [16] ).

The N protein interacts with viral genomic RNA, the membrane (M) protein and selfassociates to provide for the efficient virion assembly [17] [18] [19] . It consists of two structured domains and three regions predicted to be disordered ( Fig. 2A) , including a functionally important central Ser/Arg-rich region [20] [21] [22] and a set of potential protein-binding sites ( Fig. 2B ). Such organization allows for a vast conformational change, which in combination with positively charged surfaces [23] , facilitates nucleic acid binding [24] .

Indeed, the crystal structure of the N-terminal domain (NTD) reveals an RNA binding groove [25] [26] [27] , while crystal structures of the C-terminal domain (CTD) show a highly interlaced dimer with additional nucleic acid binding capacity [28, 29] . The N protein shows unusual properties in the presence of RNA, displaying concentration-dependent liquid-liquid phase separation [22, 23, 30, 31] that is pertinent to the viral genome packaging mechanism [32, 33] . In human cells, the assembly of condensates is downregulated by phosphorylation of the SR-rich region [30, 34] . SARS-CoV-2 N protein is a major target of phosphorylation by host cell protein kinases, with 22 phosphosites identified in vivo throughout the protein (Supplementary table 1 and [13, 35] ). Functions of N and viral replication can be regulated by a complex, hierarchical phosphorylation of the SR-rich region of SARS-CoV-2 N by a cascade of protein kinases [36] . Nevertheless, the potential functional role of N phosphorylation at each specific site is not understood.

Using immunofluorescence, immunoprecipitation, siRNA silencing and kinase inhibition, it has been shown that SARS-CoV N protein shuttles between the nucleus and the cytoplasm in COS-1 cells [37] . This process is regulated by N protein phosphorylation by several protein kinases including glycogen synthase kinase-3, protein kinase A, casein kinase II and cyclin-dependent kinase [37] [38] [39] . Consequently, phosphorylated N associates with 14-3-3 proteins in the cytoplasm [37] . Notably, treatment with a kinase inhibitor cocktail eliminated the N/14-3-3 interaction, whereas inhibition of 14-3-3θ expression by siRNA led to accumulation of N protein in the nucleus [37] . These data suggest that 14-3-3 proteins directly shuttle SARS-CoV N protein in a phosphorylationdependent manner: a role which may be universal for N proteins of all coronaviruses, including SARS-CoV-2. However, the molecular mechanism of the 14-3-3/N interaction remains ill-defined.

14-3-3 proteins are amongst the top 1% of highest-expressed human proteins in many tissues, with particular abundance in tissues vulnerable to SARS-CoV-2 infection including the lungs, gastrointestinal system and brain [40, 41] (Fig. 1 ). 14-3-3 proteins recognize hundreds of phosphorylated partner proteins involved in a magnitude of cellular processes ranging from apoptosis to cytoskeleton rearrangements [42, 43] . Human 14-3-3 proteins are present in most of the tissues as seven conserved "isoforms" (β, γ, ε, ζ, η, σ, τ/θ) (Fig. 1) , with all-helical topology, forming dimers possessing two identical antiparallel phosphopeptide-binding grooves, located at ~35 Å distance from each other [44] . By recognizing phosphorylated Ser/Thr residues within the structurally flexible (R/K)X 2-3 (pS/pT)X(P/G) consensus motif [44, 45] , 14-3-3 binding is known to regulate the stability of partner proteins, their intracellular localization and interaction with other factors [46] . In addition to their high abundance in many tissues susceptible to SARS-CoV-2 infection (Fig. 1) and a detectable increase of expression of some 14-3-3 isoforms upon SARS-CoV-2 infection [12] , 14-3-3 proteins were reported as one of nine key host proteins during SARS-CoV-2 infection [47] . These data indicate a potential association of 14-3-3 with viral proteins.

In this work, we dissected the molecular mechanism of the interaction between SARS-CoV-2 N and human 14-3-3 proteins. SARS-CoV-2 N protein containing several phosphosites reported to occur during infection, was produced using the efficient Escherichia coli system that proved successful for the study of polyphosphorylated proteins [48] . We have observed the direct phosphorylation-dependent association between polyphosphorylated SARS-CoV-2 N and all seven human 14-3-3 isoforms and determined the affinity and stoichiometry of the interaction. Series of truncated mutants of N localized the key 14-3-3-binding site to a single phosphopeptide residing in the functionally important SR-rich region of N. These findings suggest a topology model for the heterotetrameric 14-3-3/pN assembly occluding the SR-region, which presents a feasible target for further characterization and therapeutic intervention.

Host-expressed SARS-CoV-2 N protein represents a phosphoprotein, harboring multiple phosphorylation sites scattered throughout its sequence. The most densely phosphorylated locus is the SR-rich region (Fig. 2B, Supplementary table 1, Supplementary data file 1 and [13, 35] ). Remarkably, this region is conserved in N proteins of several coronaviruses [39, 49] , including SARS-CoV (Fig. 2B) . Although a number of protein kinases have been implicated in SARS-CoV N phosphorylation [36, 37, 39, 50] , the precise enzymes responsible for identified phosphosites and the functional outcomes are largely unknown. Of note, many of the reported phosphosites within the SARS-CoV-2 N protein are predicted to be phosphorylated by protein kinase A (PKA), (Supplementary table 1) . Hence, PKA was used for production of phosphorylated SARS-CoV-2 N in E.coli [48] , using the same approach that was successfully applied for production of several phosphorylated eukaryotic proteins [48, [51] [52] [53] [54] including the polyphosphorylated human tau competent for specific 14-3-3 binding [48] .

Indeed, co-expression with PKA yielded a heavily phosphorylated SARS-CoV-2 N ( Fig. 2C ) containing more than 20 phosphosites according to LC-MS analysis (Supplementary table 1 and Supplementary data file 1) . Especially dense phosphorylation occurred within the SR-rich region, involving recently reported in vivo sites Ser23, Thr24, Ser180, Ser194, Ser197, Thr198, Ser201, Ser202, Thr205 and Thr391 [13, 35] (Fig. 2B , D and Supplementary data file 1), implicating the success of the PKA co-expression at emulating native phosphorylation. Interestingly, due to the frequent occurrence of Arg residues, it was possible to characterize the polyphosphorylation of the SR-rich region only with the use of an alternative protease such as chymotrypsin, in addition to datasets obtained separately with trypsin (4 independent experiments overall, see Supplementary data file 1). Due to high conservation between SARS-CoV and SARS-CoV-2 N proteins, many phosphosites identified in the PKA-co-expressed N are likely shared by SARS-CoV N (Fig. 2B) . Importantly, many identified phosphosites lie within the regions predicted to be disordered ( Fig. 2A) , and contribute to predicted 14-3-3-binding motifs, albeit deviating from the optimal 14-3-3-binding sequence RXX(pS/pT)X(P/G) [44] (Fig. 2B and   Supplementary table 1) .

Of note, the bacterially expressed SARS-CoV-2 N protein avidly binds random E.coli nucleic acid, which results in a high 260/280 nm absorbance ratio in the eluate from the nickel-affinity chromatography column. This is unchanged by polyphosphorylation The skewed Mw distribution across the SEC peak indicated the propensity of the SARS-CoV-2 N to oligomerization (Fig. 2F) , which was instigated by the addition of nucleic acid (Supplementary Fig. 1A and B). Using tRNA isolated from E.coli DH5ɑ cells, SEC and agarose gel electrophoresis, we showed that our nucleic acid-free, polyphosphorylated N preparation is capable of avid binding of noncognate nucleic acid (Supplementary Fig. 1A -C). In sharp contrast, co-expression of SARS-CoV-2 N with PKA, and subsequent polyphosphorylation (Fig. 2) , allows for tight complex formation between pN.1-419 and human 14-3-3γ. This is evident from the peak shift and corresponding increased Mw from ~95 to 150.7 kDa (Fig. 3E) , perfectly matching the addition of the 14-3-3 dimer mass (calculated dimer Mw 56.6 kDa) to the pN.1-419 dimer (calculated dimer Mw 96 kDa).

The presence of both proteins in the complex was confirmed by SDS-PAGE (Fig. 3F ).

Collectively these data pointed toward the equimolar binding upon saturation. Given the dimeric state of both proteins in their individual states ( Fig. 2F and Fig. 3 ) and the Mw of the 14-3-3γ/pN.1-419 complex, the most likely stoichiometry is 2:2. The ratio of the proteins does not change across the peak of the complex (Fig. 3F) , implying that they form a relatively stable complex with the well-defined stoichiometry.

We then questioned whether the interaction with pN is preserved for other human 14-3-3 isoforms. Analytical SEC clearly showed that the phosphorylated SARS-CoV-2 N can be recognized by all seven human 14-3-3 isoforms, regardless of the presence of a His-tag or disordered C-terminal tails on the corresponding 14-3-3 constructs (Fig. 4A) .

However, the efficiency of complex formation differed for each isoform. Judging by the repartition of 14-3-3 between free and the pN-bound peaks, the apparent efficiency of pN binding was higher for 14-3-3γ, 14-3-3η, 14-3-3ζ and 14-3-3ε, and much lower for 14-3-3β, 14-3-3τ and 14-3-3σ, in a roughly descending order (Fig. 4A ). The interaction also appeared dependent on the oligomeric state of 14-3-3, since the monomeric mutant form of 14-3-3ζ, 14-3-3ζm-S58E [56] (apparent Mw 29 kDa) showed virtually no interaction relative to the wild-type dimeric 14-3-3ζ counterpart (apparent Mw 58 kDa), (Fig. 4B ).

This apparent variation in binding efficiency (Fig. 4A ) justified more detailed assessment of the binding parameters between pN.1-419 and selected 14-3-3 isoforms.

In light of the relative positions of the two proteins, separately and complexed, on SEC profiles ( Fig. 3 and 4) , we used analytical SEC to track titration of a fixed concentration of pN.1-419 (around 10 μM) against increasing quantities (0-100 μM) of either of two selected full-length human 14-3-3 isoforms, γ and ε ( Fig. 5A and B) . A saturation binding curve showed the maximal concentration of bound 14-3-3γ to asymptotically approach 10 μM (Fig. 5C ), supporting 2:2 stoichiometry. The apparent K D of the 14-3-3γ/pN.1-419 complex was estimated as 1.5 ± 0.3 μM. A similar binding mechanism could be observed for 14-3-3ε, however in this case we could achieve pN.1-419 saturation only at much higher 14-3-3ε concentrations ( Fig. 5B and C), and the resulting apparent K D was ~7 times higher than for 14-3-3γ (Fig. 5C ). Nevertheless, once again the stoichiometry was close to 2:2. These findings strongly disfavor the earlier hypothesis that 14-3-3 binding affects dimerization status of N [57] .

We further asked what are the specific regions of SARS-CoV-2 N that are responsible for interaction with human 14-3-3.

Among multiple phosphosites identified in our pN.1-419 preparations (Supplementary table 1) , the two most interesting regions are located in the intrinsically disordered or loop segments ( Fig. 2A) , normally favored by 14-3-3 proteins [45] . The first represents the C-terminally located RTA[pT 265 ]KAY site, which is predicted by the 14-3-3-Pred webserver [58] as the 14-3-3-binding site within the loop region immediately preceding the CTD. The second, SR-rich region features multiple experimentally confirmed phosphosites including several suboptimal predicted 14-3-3-binding sites (Fig.  2B ). To narrow down the 14-3-3-binding locus we used several constructs representing its N-and C-terminal parts (N.1-211 and N.212-419, respectively). The individual CTD included the C-terminal phosphosite around Thr265 (N.247-364), and the longer Nterminal construct extended toward the C-terminus to include the predicted NES sequence (N.1-238) (Fig. 6A ). This contains Asp225 and a cluster of Leu residues which together resemble the unphosphorylated 14-3-3-binding segments from ExoS/T [59] and therefore could be important for 14-3-3 binding.

As for the wild-type protein, the truncated SARS-CoV-2 N constructs were cloned and expressed in the absence or presence of PKA to produce unphosphorylated or In contrast, N-terminal constructs including the RNA-binding domain, such as N.1-211, are monomeric ( Supplementary Fig. 3 ). Thus, our data align with the low-resolution structural model of SARS-CoV and SARS-CoV-2 N proteins, in which the CTD (residues 247-364), largely responsible for dimerization, and NTD, involved in RNA-binding, are only loosely associated [24, 28, 29] .

Of the constructs analyzed, the N-terminal constructs N.1-211 and N.1-238 both interacted with 14-3-3γ by forming distinct complexes on SEC profiles, and this interaction was strictly phosphorylation-dependent (Fig. 6) . Given the similarity of the elution profiles for pN.1-211 and pN.1-238, it may be concluded that the presence of NES in the latter is dispensable for the 14-3-3 binding.

Under similar conditions, only a very weak interaction between the dimeric pN.212-419 construct and 14-3-3γ could be observed, whereas the phosphorylated dimeric CTD (pN.247-364) displayed virtually no binding (Fig. 6 ). Neither unphosphorylated construct interacted with 14-3-3. Thus, the Thr265 phosphosite can be broadly excluded as the critical binding site. Separate phosphosites outside the CTD, for instance, in the last ~30 C-terminal residues (Supplementary table 1) likely account for residual binding of the pN.212-419 construct. Only its SEC profile showed a significant positional peak shift with phosphorylation ( Fig. 6B and C) , indicating a potential change in the oligomeric state. It is tempting to speculate that such phosphorylation outside the CTD could affect higher order oligomerization associated with the so-called N3 C-terminal segment (Fig. 2B) [20, 21, 49, 60] .

According to SEC-MALS, the 14-3-3γ dimer (Mw 57 kDa) interacts with the pN.1-238 monomer (Mw of 31 kDa) by forming a ~82 kDa complex (Fig. 7A ) with an apparent 2:1 stoichiometry. It is remarkable that despite a moderate molar excess of pN.1-238 a 2:2 complex (one 14-3-3 dimer with two pN monomers) is not observed. The well-defined 2:2 stoichiometry of the 14-3-3γ complex with the full-length pN (Fig. 3) suggests that the dimeric pN is anchored using two equivalent, key 14-3-3-binding sites, each located in a separate subunit of N. It is tempting to speculate that, in the absence of the second subunit, the interaction involves the key phosphosite and an additional phosphosite which is separated by a sufficiently long linker (≥ 15 residues [61] ), to secure occupation of both phosphopeptide-binding grooves of 14-3-3 ( Our finding that the minimal N-terminal construct pN.1-211 exhibits firm binding to 14-3-3γ indicates that the key 14-3-3-binding phosphosite(s) is/are located exclusively within this region. Given the presence of numerous candidate 14-3-3-binding sites within its most C-terminal part, i.e., the SR-rich region, we further focused on the 1-211 sequence in search of the 14-3-3-binding phosphosite(s).

Further N mutants were designed to disrupt the most probable 14-3-3-binding phosphosites. These are located in the intrinsically flexible phosphorylatable SR-rich segment centered at positions 197 and 205 (Fig. 8A ). Both represent suboptimal 14-3-3binding motifs SRN[pS 197 ]TP and SRG[pT 205 ]SP in lacking an Arg/Lys residue in position -3 (bold font) relative to the phosphorylation site (squared brackets). However, each also features a Pro residue in position +2 (bold underlined font), which is highly favorable for 14-3-3 binding [44] and absent from the other potential 14-3-3-binding phosphosites in the SR-rich region (Fig. 2B) . These conflicting factors complicate predictions for the true 14-3-3 binding site. Moreover, even beyond the SR-rich region the NTD is predicted to host further possible 14-3-3-binding phosphosites, including the RRA[pT 91 ]RR site, which is the highest-scoring in 14-3-3-Pred [58] prediction (Supplementary table 1) .

We conceived stepwise truncations to remove the most probable 14-3-3-binding phosphosites, aiming to identify the iteration at which binding (observed for the pN.1-211 construct) ceased. 14-3-3 can bind incomplete consensus motifs at the extreme Cterminus of some proteins [62] , so truncations were designed to remove the critical phosphorylated residue. However, upstream residues of each candidate 14-3-3-binding site were preserved, in light of the sheer number of overlapping potential binding motifs in the SR-rich region (Fig. 8A) . Supplementary Fig. 3) , consistent with the proposed architecture of the N protein [21, 24] .

None of the truncated mutants interacted with 14-3-3γ in the unphosphorylated state ( Fig. 8B ). More importantly, no binding to 14-3-3γ was detected with phosphorylated N.1-179 (Fig. 8C) and only very limited binding could be observed with phosphorylated N.1-196 (Fig. 8C ). This strongly indicated that all phosphosites of the 1-196 segment (including at least three phosphosites within the SR-rich region, i.e., Ser180, Ser188 and Ser194, see Fig. 8A ) are dispensable for 14-3-3 binding and at most could contribute only as auxiliary sites (as suggested by the scheme in Fig. 7B ). This narrowed the 14-3-3binding region within SARS-CoV-2 N down to 15 residues from 196 to 211, leaving only two possible sites centered at Ser197 and Thr205 (Fig. 8A ).

By contrast, pN.1-204 showed only a very slightly altered interaction with 14-3-3γ compared to pN.1-211 (Fig. 8C) . Although this does not exclude that Thr205 phosphosite may contribute to 14-3-3 binding in the context of the full-length pN (particularly if pSer197 is absent or mutated), pSer197 appears to be critical for 14-3-3 recruitment. Intriguingly, in contrast to Thr205, Ser197 is preserved in most related coronavirus N proteins (see Fig. 2B, Fig. 9 ).

In this work, we investigated the molecular association between the SARS-CoV-2 N protein and human phosphopeptide-binding proteins of the 14-3-3 family. The former is the most abundant viral protein [12, 13] , the latter is a major protein-protein interaction hub involved in multiple cellular signaling cascades, expressed at high levels in many human tissues including those susceptible to SARS-CoV-2 infection (Fig. 1) [40] . SARS-CoV-2 N is heavily phosphorylated in infected cells [13, 35, 36] , which poses a significant challenge for proteomic approaches: the densely phosphorylated SR-rich region alone, functionally implicated in numerous viral processes [34, 49, 63] , hosts seven closely spaced Arg residues (Fig. 2B) . These arginines restrict the length of tryptic phosphopeptides and decrease the probability of their unambiguous identification and phosphosite assignment [64] . The multiplicity of implicated protein kinases (including GSK-3, protein kinase C, casein kinase II and mitogen-activated protein kinase [36, 37, 50] ) further hinders study of specific phosphorylations. Assuming that the mechanistic implication of a specific phosphorylation is independent of the acting kinase, we produced polyphosphorylated N protein (pN) via bacterial co-expression with a catalytic subunit of PKA. A combination of orthogonal cleavage enzymes and LC-MS phosphoproteomics mapped > 20 phosphosites (Supplementary table 1) including Ser23, Thr24, Ser180, Ser194, Ser197, Thr198, Ser201, Ser202, Thr205 and Thr391 reported recently at SARS-CoV-2 infection [13, 35] . At least six of the identified phosphosites are located in the unstructured regions and represent potential 14-3-3-binding sites ( Fig. 2A and B) .

Biochemical analysis confirmed that polyphosphorylated N is competent for binding to all seven human 14-3-3 isoforms (Fig. 3 and 4) , but revealed remarkable variation in binding efficiency between them (Fig. 4) . This was supported by the quantified affinities to two selected isoforms, 14-3-3γ (K D of 1.5 µM) and 14-3-3ε (K D of 10.7 µM) (Fig. 5 ). Our observations are in line with the recent finding that 14-3-3γ and 14-3-3η systematically bind phosphopeptides with higher affinities than 14-3-3ε and 14-3-3σ [55] .

The low micromolar-range K D values compare well to those reported for other physiologically relevant partners of 14-3-3 [65] [66] [67] , indicating a stable and specific interaction. Meanwhile, the well-defined 2:2 stoichiometry of the ~150-kDa 14-3-3/pN complex, supported by titration experiments and SEC-MALS analysis ( Fig. 3 and 5) , excludes the possibility that 14-3-3 binding disrupts pN dimerization [57] . It is reasonable to assume that the principally bivalent 14-3-3 dimer [44, 46] recognizes just one phosphosite in each pN subunit because a bidentate 14-3-3 binding to different phosphosites within a single pN subunit would inevitably alter the observed 2:2 stoichiometry. Identification of the single phosphosite responsible for 14-3-3 recruitment proved challenging, as none of the potential sites were a perfect match to the currently known optimal 14-3-3-binding motifs [65] .

To restrict the search, we analyzed the interaction of various N constructs with 14-3-3γ. This eliminated the high-scoring potential 14-3-3-binding phosphosite RTApT 265 KAY, present in the C-terminal N fragments, as the true site of interaction despite its conservation in many related coronavirus N proteins (Supplementary table 2 and 3). The residual binding of pN.212-419 to 14-3-3γ suggested the existence of auxiliary 14-3-3-binding sites located outside the folded CTD (residues 247-364). Both pN.1-211 and pN.1-238 bound 14-3-3 with equivalent efficiency suggesting the binding lies between amino acid 1 and 211. Interestingly, the N-terminal constructs existed as monomers ( Supplementary Fig. 3) , which could potentially lower the binding affinity to 14-3-3 dimers in light of the 2:2 stoichiometry. Nonetheless, sufficient phosphorylationdependent binding was clearly observed between the dimeric 14-3-3γ and the monomeric N-terminal constructs ( Fig. 6 and 7) .

Truncation of the SR-rich region streamlined the search for the 14-3-3-binding site to the 15-residue stretch of amino acids 196-211. This sequence hosts two principally similar potential 14-3-3-binding sites, RNpS 197 TP and RGpT 205 SP ( Fig. 2B and D) .

Importantly, the proximity of these sites rules out their simultaneous bidentate binding to the 14-3-3 dimer: 14-3-3 binds in an antiparallel manner requiring a minimum of [13] [14] [15] residues between phosphosites on a single peptide [44, 68] . Thus, binding to Ser197 and Thr205 sites must be mutually exclusive.

The markedly different binding between pN.1-204 and pN.1-196 to 14-3-3 (Fig. 8) prompted us to propose Ser197 as the critical phosphosite. This finding aided the design of a topology model for the complex (Fig. 9A) , in which the 14-3-3 dimer is anchored by two identical Ser197 phosphosites from the SR-rich region in the two equivalent pN chains. Noteworthily, the RN(pS/pT) 197 TP site is conserved in not only SARS-CoV and SARS-CoV-2 but also in N proteins from several bat and pangolin coronaviruses (Fig.   9B ). Meanwhile, plausible phosphorylation of residue 205 is possible in a smaller subset of coronaviruses (Fig. 9B) .

The model shown in Fig. 9A notably does not exclude the possibility of 14-3-3 binding to alternative phosphosites under significantly different phosphorylation conditions. Moreover, we speculate that hierarchical phosphorylation within the SR-rich region [36] would alter 14-3-3 binding with phosphorylation at adjacent Ser/Thr residues, likely to inhibit the interaction [69] . The SR-rich region is phosphorylated by both Prodirected and non-Pro-directed protein kinases [36, 37, 39, 50] . Since 14-3-3 proteins typically reject peptides with a proline adjacent to the phosphorylated residue, the interplay between these aforementioned kinases could be regulatory. In theory, this could create a phosphorylation code and conditional binding of 14-3-3, as has been discussed recently for alternative polyphosphorylated 14-3-3 partners such as LRRK2, CFTR and tau protein [48, 67, 70, 71] .

The recruitment of the 14-3-3 dimer is expected to occlude the SR-rich region of N by masking 10-20 residues surrounding the Ser197 phosphosite within the complex.

Apart from the likely effects on the properties of N and its ability to phase separate and bind RNA, the 14-3-3 binding at the SR-rich region, triggered by phosphorylation, can potentially interfere with N binding to the M protein, an event clearly relevant to the virion assembly [17] . In support of this hypothesis, the 14-3-3-occluded area reported here overlaps with the N region (residues 168-208) proposed to mediate its association with the M protein in SARS-CoV [19] .

The presence of SR-rich regions in many viral N proteins suggests a more broad interaction of 14-3-3 with N proteins. Indeed, using 14-3-3-Pred prediction [58] Given the reasonable threat that other zoonotic coronaviruses may ultimately enter the human population [72, 73] , the relevant N proteins are highly likely to undergo phosphorylation and 14-3-3 binding, as seen for SARS-CoV-2 N. This is particularly likely, given the high concentration of both proteins in the infected cells (see above).

Our findings underline the essential role of the SR-rich region in the biology of N proteins and host-virus interactions [49] . Unrelated proteins with similar domains also tend to show RNA-binding capability (e.g., the splicing factors) [74, 75] and are subject to multisite phosphorylation [76] . Such proteins are often associated with phase separation as a means to regulate membraneless compartmentalization within the cytoplasm.

Likewise, SARS-CoV-2 N protein has been shown to undergo phase separation in vitro upon RNA addition: a phenomenon dependent on the concentration of salt, presence of divalent ions, phosphorylation state of N and on RNA sequence [30, 34, 77, 78] .

Furthermore, the N protein has been shown to recruit the RNA-dependent RNApolymerase complex, and granule-associated heterogeneous nuclear ribonucleoproteins forming phase separated granules which can aid SARS-CoV-2 replication [30, 34] . Thus

also affect the accessibility of the NES sequence located nearby (Fig. 2B ). This in turn may impact nucleocytoplasmic shuttling of N, as seen for SARS-CoV N [37] .

Conclusively, 14-3-3 binding to the SR-rich region of N holds potential to regulate multiple host cell processes affected by N. 14-3-3 binding to pN may present a cell immune-like response to the viral infection aimed at arresting or neutralizing N activities [57] . On the other hand, in light of the abundance of N protein in the infected cell [12, 13] , pN may instead arrest 14-3-3 proteins in the cytoplasm and indirectly disrupt cellular processes involving 14-3-3. For example, 14-3-3ε and 14-3-3η each play a role in the innate immune response via RIG-1 and MDA5 signaling respectively [79, 80] . The N protein:14-3-3 interaction would modulate these and other signaling pathways involving 14-3-3 proteins. Intriguingly, two 14-3-3 isoforms, ζ and ε, have been detected in purified particles of infectious bronchitis coronavirus [81] . The 14-3-3 protein take up could be mediated by its interaction with N, potentially resulting in the 14-3-3 transmission between coronavirus hosts.

As such, understanding the molecular mechanism of pN association with 14-3-3 proteins may inform the development of novel therapeutic approaches and paves the road for structural studies.

The tagless SARS-CoV-2 N gene coding for Uniprot ID P0DTC9 protein sequence was commercially synthesized (GeneWiz) and cloned into pET-SUMO bacterial expression vectors using NdeI and HindIII restriction endonuclease sites. To obtain constructs of N carrying the N-terminal His 6 -tag cleavable by HRV-3C protease (N.1-419,   N.1-211, N.1-238, N.212-419) , the N gene was PCR amplified using primers listed in Supplementary table 4 and cloned into the pET-YBSL-Lic-3C plasmid vector using an established ligation-independent cloning procedure [82] . After 3C cleavage each construct contained extra GPA residues at the N terminus. The N.247-364 construct carrying an N-terminal uncleavable His 6 -tag and corresponding to the previously crystallized folded CTD dimer [28] , was kindly provided by Prof. W. Baumeister's laboratory. Untagged full-length human 14-3-3γ (Uniprot ID P61981) and 14-3-3ζ (Uniprot ID P63104) cloned in a pET21 vector, and full-length human 14-3-3ε (Uniprot ID P62258) carrying an N-terminal His 6 -tag cleavable by Tobacco Etch Virus (TEV) protease, and cloned into a pRSET-A vector have been previously described [83] [84] [85] . Truncated versions of human 14-3-3η (Uniprot ID Q04917), 14-3-3β (Uniprot ID P31946), 14-3-3ε (Uniprot ID P62258), 14-3-3τ (Uniprot ID P27348), 14-3-3γ (Uniprot ID P61981) and 14-3-3σ (Uniprot ID P31947) devoid of the short disordered segment at the C terminus, cloned into pProExHtb vector and carrying TEV-cleavable His 6 -tags on their N-terminal ends were obtained as described previously [66, 86] . The monomeric mutant form of untagged full-length human 14-3-3ζ carrying monomerizing amino acid substitutions and pseudophosphorylation in the subunit interface (14-3-3ζm-S58E) was obtained as before [56] .

14-3-3γ, 14-3-3ζ and 14-3-3ζm-S58E were expressed and purified using ammonium sulfate fractionation and column chromatography as previously described [83] . His 6 -tagged N.247-364 was also expressed and purified as before [28] . All other proteins carrying cleavable His 6 -tags were expressed in E.coli BL21(DE3) cells and purified using subtractive immobilized metal-affinity chromatography (IMAC) and gelfiltration. For the N protein and its various constructs the eluate from the first IMAC column, performed at 1 M NaCl, harbored a significant quantity of random bound nucleic acid (the 260/280 nm absorbance ratio of 1.7-2.0). This necessitated an additional long on-column washing step with 3 M NaCl (50 column volumes) to ensure the 260/280 nm absorbance ratio of ~0.6 and removal of all nucleic acid. Tagless N protein was purified after treatment of cell lysate with RNAse, using heparin and subsequent ion-exchange chromatography on a sulfopropyl Sepharose (elution gradients 100-1000 mM NaCl) followed by gel-filtration. Here, It proved challenging to entirely eliminate nucleic acid and the typical 260/280 nm absorbance ratio was 0.7-0.9.

Protein concentration was determined by spectrophotometry at 280 nm on a N80 Nanophotometer (Implen, Munich, Germany) using sequence-specific extinction coefficients calculated using the ProtParam tool in ExPASy (see Supplementary table 5) .

The DH5ɑ cells were incubated in 30 ml of liquid medium (LB without antibiotics) for 16 h at 37 °C with maximum aeration and then harvested by centrifugation at 7000 g for 10 min. The pellet was gently resuspended in RNAse-free Tris-acetate buffer, followed by alkali-SDS lysis and neutralization by cold ammonium acetate. The suspension was incubated on ice for 5 min and centrifuged at 21000 g at 4 °C for 10 min. The resulting supernatant (8 ml) was incubated with 12.5 ml isopropanol for 15 min at 25 °C and centrifuged at 12100 g for 5 min. The pellet was resuspended in 800 µl of 2 M ammonium acetate, incubated for 5 min at 25 °C and centrifuged (12100 g, 10 min). RNA was precipitated from supernatant by 800 µl of isopropanol, incubated for 5 min at 25 °C and centrifuged (12100 g, 5 min). After supernatant removal the pellet was washed with 70% ice-cold ethanol, dried and dissolved in 100-200 µl milliQ-water.

For phosphorylation in cells, SARS-CoV-2 N was bacterially co-expressed with a catalytic subunit of mouse protein kinase A (PKA), as described previously [48] . PKA was cloned into a low-copy pACYC vector [48] which ensured that the target protein was expressed in excess of kinase.

PKA and SARS-CoV-2 N were co-transformed into E.coli BL21(DE3) cells against Kanamycin and Chloramphenicol resistance. Cells were grown in LB to an OD 600 reading of 0.6 before inducing with 500 µM of IPTG. After induction, cultivation was continued for 

Sample treatment for proteomics analysis. For phosphopeptide mapping, the SARS-CoV-2 N protein co-expressed with PKA for 4 h at 30 ⁰C was purified as above.

An aliquot (35 μg) was subjected to enzymatic hydrolysis "in solution" either with trypsin (Sequencing Grade Modified Trypsin, Promega) or with chymotrypsin (Analytical Grade, Sigma-Aldrich). Briefly, the sample preparation was as follows. The sample was reduced with 2 mM Tris(2-carboxyethyl)phosphine (TCEP) and then alkylated with 4 mM S-Methyl methanethiosulfonate (MMTS); a protein:enzyme ratio was kept at 50:1 (w/w); digestion was performed overnight at 37 ⁰C and pH 7.8 (for chymotrypsin, 10 mM Ca 2+ was added to the reaction solution). The reaction was stopped by adjusting it to pH 2 with formic acid.

The resulting peptides were purified on a custom micro-tip SPE column with Oasis HLB (Waters) as a stationary phase, using elution with an acetonitrile:water:formic acid mix 

The source LC-MS data on SARS-CoV-2 N phosphoproteomics are available along with the paper as a Supplementary data file 1. Fig. 1. 14-3-3 proteins are highly abundant in human tissues with SARS-CoV-2 presence. Correlation of ACE2 expression levels (*) and SARS-CoV-2 reported presence (**) in various tissues of COVID19 patients based on the data from [6] , shown with abundances (indicated in ppm, part per million, i.e., one molecule of a given protein per 1 million of all proteins from a given tissue) of the seven human 14-3-3 isoforms, extracted from the PAXdb database [40] . Different tissues are shown in the order corresponding to the SARS-CoV-2 presence, starting, at the top, from the highest virus presence [6] . The shown relative scale of ACE2 expression is also taken from [6] . The total abundance of the seven human 14-3-3 isoforms in a given tissue and the average abundance of an isoform in 12 selected tissues are also indicated. The latter values were used for ordering the data for 14-3-3 isoforms, left to right, from the highest average abundance (14-3-3ζ; 2423 ppm, or ~0.24%) to the lowest average abundance (14-3-3η; 575 ppm, or ~0.06%). Note that the average abundance of all seven 14-3-3 proteins in three tissues with the highest SARS-CoV-2 presence (oral cavity, gastrointestinal tract, lungs) reaches 1.21% of all proteins. 5 . Binding affinity. pN at a fixed concentration (~10 μM) was titrated against either of the two human 14-3-3 isoforms and monitored by analytical SEC with serial sample loading. A. Titration of pN with 14-3-3γ. B. Titration of pN with 14-3-3ε. Changes of the elution profiles associated with the increasing 14-3-3 concentration are shown by arrows. C. Binding curves used for apparent K D determination. Note that, regardless of the observed differences in the affinities, the maximum 14-3-3 concentration complexed with pN equals the concentration of pN (~10 μM) for both 14-3-3 isoforms. Typical results from two independent experiments are shown. Note that the difference shown by a two-headed arrow roughly corresponds to a pN.1-238 monomer mass. B. Schematic representation of the possible bidentate binding mode where the 14-3-3 dimer interacts with the tentative key 14-3-3-binding phosphosite and another sterically allowed phosphosite separated by a sufficiently long linker.

A. Sequence of the SR-rich region showing potential 14-3-3binding sites (green and blue font) and the truncations designed to exclude one or two main 14-3-3-binding sites (blue font). # denotes the designed C-terminus. B, C. Analysis by SEC of the interaction between human 14-3-3γ and the truncated mutants of SARS-CoV-2 N in their unphosphorylated (B) or phosphorylated forms (C). Each graph contains SEC profiles for the individual 14-3-3 (black line), individual N or pN construct (red line), and the 14-3-3γ N/pN mixture (blue line) where 50 μM of each protein was used throughout. Inserts on panel C show the upward shift on SDS-PAGE as the result of phosphorylation (U, unphosphorylated; P, phosphorylated protein, arrowheads indicate the shift). Note that only the phosphorylated mutants interacted with 14-3-3γ and that only pN.1-211 and pN.1-204 formed a defined complex with 14-3-3γ. Column: Superdex 200 Increase 5/150, flow rate: 0.45 ml/min. The experiment was repeated twice and the most typical results are presented. Fig. 9 . Association of SARS-CoV-2 N with human 14-3-3. A. A topology model for the complex of SARS-CoV-2 N protein dimer with the dimer of human 14-3-3 illustrating the occlusion of the SR-rich region. Although Ser197 is critical for the 14-3-3 binding, other phosphosites (e.g., the semiconserved Thr205) may play a secondary role. B. Local alignment of SR-rich regions of the most similar coronavirus N proteins in order of descending sequence identity (s.id.) determined using entire N protein sequences. Alignment was performed using Clustal omega and visualized using Mview (https://www.ebi.ac.uk/Tools/msa/mview/). Residues identical to those in the SARS-CoV-2 sequence are shadowed by grey, phosphorylatable residues in positions 197 and 205 are in green color, residues blocking phosphorylation in position 205 are in red. 

TAt  K  A  Y  N  V  TQAFGR

WHO Coronavirus Disease (COVID-19) Dashboard

Three Emerging Coronaviruses in Two Decades

Unraveling the Epidemiology, Geographical Distribution, and Genomic Evolution of Potentially Lethal Coronaviruses (SARS, MERS, and SARS CoV-2). Frontiers in cellular and infection microbiology

Genomic determinants of pathogenicity in SARS-CoV-2 and other human coronaviruses

ACE2: Evidence of role as entry receptor for SARS-CoV-2 and implications in comorbidities. eLife

On the whereabouts of SARS-CoV-2 in the human body: A systematic review

ChAdOx1 nCoV-19 vaccination prevents SARS-CoV-2 pneumonia in rhesus macaques

Evaluation of the mRNA-1273 Vaccine against SARS-CoV-2 in Nonhuman Primates

Phase I/II study of COVID-19 RNA vaccine BNT162b1 in adults

COVID-19 update: The race to therapeutic development. Drug resistance updates : reviews and commentaries in antimicrobial and anticancer chemotherapy

The Genome sequence of the SARS-associated coronavirus

Proteomics of SARS-CoV-2-infected host cells reveals therapy targets

The Global Phosphorylation Landscape of SARS-CoV-2 Infection

The total number and mass of SARS-CoV-2 virions in an infected person

SARS-CoV-2 (COVID-19) by the numbers. eLife

What is the total number of protein molecules per cell volume? A call to rethink some published values

The coronavirus nucleocapsid is a multifunctional protein

Molecular Architecture of the SARS-CoV-2 Virus

Characterization of proteinprotein interactions between the nucleocapsid protein and membrane protein of the SARS coronavirus

Transient oligomerization of the SARS-CoV N protein--implication for virus ribonucleoprotein packaging

Architecture and self-assembly of the SARS-CoV-2 nucleocapsid protein

The SARS-CoV-2 nucleocapsid protein is dynamic, disordered, and phase separates with RNA. bioRxiv

Liquid-liquid phase separation by SARS-CoV-2 nucleocapsid protein and RNA

Biochemical characterization of SARS-CoV-2 nucleocapsid protein

Crystal structure of SARS-CoV-2 nucleocapsid protein RNA binding domain reveals potential unique drug targeting sites

Structural insights into the mechanism of RNA recognition by the N-terminal RNA-binding domain of the SARS-CoV-2 nucleocapsid phosphoprotein. Computational and structural biotechnology journal

Structure of the SARS coronavirus nucleocapsid protein RNA-binding dimerization domain suggests a mechanism for helical packaging of viral RNA

High-resolution structure and biophysical characterization of the nucleocapsid phosphoprotein dimerization domain from the Covid-19 severe acute respiratory syndrome coronavirus 2

Structural characterization of the C-terminal domain of SARS-CoV-2 nucleocapsid protein

SARS-CoV-2 nucleocapsid protein phase-separates with RNA and with human hnRNPs

A proposed role for the SARS-CoV-2 nucleocapsid protein in the formation and regulation of biomolecular condensates

The Coronavirus Nucleocapsid Protein

Coronavirus genomic RNA packaging

Nucleocapsid protein of SARS-CoV-2 phase separates into RNA-rich polymerase-containing condensates

Characterisation of the transcriptome and proteome of SARS-CoV-2 reveals a cell passage induced in-frame deletion of the furin-like cleavage site from the spike glycoprotein

The FDAapproved drug Alectinib compromises SARS-CoV-2 nucleocapsid phosphorylation and inhibits viral infection in vitro

The severe acute respiratory syndrome coronavirus nucleocapsid protein is phosphorylated and localizes in the cytoplasm by 14-3-3-mediated translocation

Glycogen synthase kinase-3: A putative target to combat severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) pandemic. Cytokine & growth factor reviews

Phosphorylation of the arginine/serine dipeptide-rich motif of the severe acute respiratory syndrome coronavirus nucleocapsid protein modulates its multimerization, translation inhibitory activity and cellular localization

PaxDb, a database of protein abundance averages across all three domains of life

Human 14-3-3 protein: radioimmunoassay, tissue distribution, and cerebrospinal fluid levels in patients with neurological disorders

14-3-3 proteins: key regulators of cell division, signalling and apoptosis

14-3-3 proteins: a historic overview

The structural basis for 14-3-3:phosphopeptide binding specificity

Intrinsic disorder is a key characteristic in partners that bind 14-3-3 proteins

Structural basis of 14-3-3 protein functions

Looking for pathways related to COVID-19 phenotypes: Confirmation of pathogenic mechanisms by SARS-CoV-2 -Host interactome

Bacterial co-expression of human Tau protein with protein kinase A and 14-3-3 for studies of 14-3-3/phospho-Tau interaction

SR/RS Motifs as Critical Determinants of Coronavirus Life Cycle. Frontiers in molecular biosciences

Glycogen synthase kinase-3 regulates the phosphorylation of severe acute respiratory syndrome coronavirus nucleocapsid protein and viral replication

Chimeric 14-3-3 proteins for unraveling interactions with intrinsically disordered partners

Concatenation of 14-3-3 with partner phosphoproteins as a tool to study their interaction

Design, expression, purification and crystallization of human 14-3-3zeta protein chimera with phosphopeptide from proapoptotic protein BAD

Molecular basis for the recognition of steroidogenic acute regulatory protein by the 14-3-3 protein family

Recognition of highrisk HPV E6 oncoproteins by 14-3-3 proteins studied by interactomics and crystallography

Hidden disorder propensity of the N-terminal segment of universal adapter protein 14-3-3 is manifested in its monomeric form: Novel insights into protein dimerization and multifunctionality

Mutations in the phosphorylation sites of SARS-CoV-2 encoded nucleocapsid protein and structure model of sequestration by protein 14-3-3

14-3-3-Pred: improved methods to predict 14-3-3-binding phosphopeptides

14-3-3 proteins activate Pseudomonas exotoxins-S and -T by chaperoning a hydrophobic surface

Oligomerization of the carboxyl terminal domain of the human coronavirus 229E nucleocapsid protein

Bioinformatic and experimental survey of 14-3-3-binding sites

C-terminal binding: an expanded repertoire and function of 14-3-3 proteins

The SR-rich motif in SARS-CoV nucleocapsid protein is important for virus replication

Identification of phosphorylation sites in the nucleocapsid protein (N protein) of SARS-coronavirus. International journal of mass spectrometry

Association of Multiple Phosphorylated Proteins with the 14-3-3 Regulatory Hubs: Problems and Perspectives

Structural Basis for the Interaction of a Human Small Heat Shock Protein with the 14-3-3 Universal Signaling Regulator

Characterization and small-molecule stabilization of the multisite tandem binding between 14-3-3 and the R domain of CFTR

Recognition of an intra-chain tandem 14-3-3 binding site within PKCepsilon

Binding of the Human 14-3-3 Isoforms to Distinct Sites in the Leucine-Rich Repeat Kinase 2

Reading the phosphorylation code: binding of the 14-3-3 protein to multivalent client phosphoproteins

Structural interface between LRRK2 and 14-3-3 protein

CoV-2: Zoonotic origin of pandemic coronavirus

Isolation and characterization of a bat SARS-like coronavirus that uses the ACE2 receptor

A genome-wide survey of RS domain proteins

Phase separation in biology; functional organization of a higher order

Nterminus of the protein kinase CLK1 induces SR protein hyperphosphorylation

Genomic RNA elements drive phase separation of the SARS-CoV-2 nucleocapsid

Phosphoregulation of Phase Separation by the SARS-CoV-2 N Protein Suggests a Biophysical Basis for its Dual Functions

The 14-3-3eta chaperone protein promotes antiviral innate immunity via facilitating MDA5 oligomerization and intracellular redistribution

The mitochondrial targeting chaperone 14-3-3epsilon regulates a RIG-I translocon that mediates membrane association and innate antiviral immunity

Proteomic analysis of purified coronavirus infectious bronchitis virus particles

Higher-throughput approaches to crystallization and crystal structure determination

Small heat shock protein Hsp20 (HspB6) as a partner of 14-3-3gamma

Identification of a novel ATPase activity in 14-3-3 proteins--evidence from enzyme kinetics, structure guided modeling and mutagenesis studies

Modulation of 14-3-3/phosphotarget interaction by physiological concentrations of phosphate and glycerophosphates

Structural insights of the MLF1/14-3-3 interaction

Some phosphosites (red spheres and labeled) are predicted as suboptimal 14-3-3 binding sites (green arrows). C. A Phos-tag gel showing that bacterial co-expression of the full-length N with PKA yields polyphosphorylated protein. D. A fragmentation spectrum of the representative phosphopeptide carrying phosphorylation at Ser197 and Thr205. Z-(red) and c-series (blue) of ETD fragmentation are shown. Error did not exceed 5 ppm. E. Absorbance spectra show that both recombinant unphosphorylated and PKAphosphorylated N proteins elute from the Ni-affinity column bound to random E.coli nucleic acid. On-column washing with 3 M NaCl (50 column volumes) eliminates bound nucleic acid. F. Analysis of the oligomeric state of pN using size-exclusion chromatography on a Superdex 200 Increase 10/300 column at 200 mM NaCl, with multiangle light scattering detection (SEC-MALS) and SDS-PAGE of the eluted fractions. Flow rate was 0.8 ml/min

Investigation Methodology Visualization Dorothy E. D. P. Hawkins. Investigation Writing -review & editing Jake L. R. Smith. Investigation Oliver W Bayfield. Investigation Writing -review & editing De-Sheng Ker. Investigation Writing -review & editing Andrey A Sysoev. Investigation Oleg Klychnikov. Investigation Methodology Validation Alfred A Antson. Conceptualization Funding acquisition Supervision Resources Writing -review & editing

Conceptualization Funding acquisition Resources Methodology Investigation Supervision Validation Visualization Formal analysis Writing -original draft Writing -review & editing

☒ The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.☐The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: