Ebola Virus NP Binding to Host Protein Phosphatase-1 Regulates Capsid Formation

The Ebola virus (EBOV) transcriptional regulation involves host protein phosphatases PP1 and PP2A, which dephosphorylate the transcriptional cofactor of EBOV polymerase VP30. The 1E7–03 compound, which targets PP1, induces VP30 phosphorylation and inhibits EBOV infection. This study aimed to investigate the role of PP1 in EBOV replication. When EBOV-infected cells were continuously treated with 1E7–03, the NP E619K mutation was selected. This mutation moderately reduced EBOV minigenome transcription, which was restored by the treatment with 1E7–03. Formation of EBOV capsids, when NP was co-expressed with VP24 and VP35, was impaired with NPE 619K. Treatment with 1E7–03 restored capsid formation by NP E619K mutation, but inhibited capsids formed by WT NP. The dimerization of NP E619K, tested in a split NanoBiT assay, was significantly decreased (~ 15-fold) compared to WT NP. NP E619K bound more efficiently to PP1 (~ 3-fold) but not B56 subunit of PP2A or VP30. Cross-linking and co-immunoprecipitation experiments showed fewer monomers and dimers for NP E619K which were increased with 1E7–03 treatment. NP E619K showed increased co-localization with PP1α compared to WT NP. Mutations of potential PP1 binding sites and NP deletions disrupted its interaction with PP1. Collectively, our findings suggest that PP1 binding to the NP regulates NP dimerization and capsid formation, and that NP E619K mutation, which has the enhanced PP1 binding, disrupts these processes. Our results point to a new role for PP1 in EBOV replication in which NP binding to PP1 may facilitate viral transcription by delaying capsid formation and EBOV replication.


INTRODUCTION
The Ebola virus (EBOV) is a non-segmented, negative strand RNA virus responsible for a deadly human disease [1]. The 2013-2016 outbreak in West Africa was the most severe on record, with 28,000 cases and 11,000 deaths [2]. The most recent outbreak in the Democratic Republic of the Congo in 2018-2020 resulted in over 3,400 con rmed cases and 2,280 deaths [3], followed by additional outbreaks and cases [4]. Consequently, it is essential to create new drugs and treatments against EBOV.
A subsequent study showed that serine-arginine protein kinase 1 (SPRK1) interacts with the R 26 xxS 29 motif of VP30 and phosphorylates VP30 Ser- 29 [8]. PP1 and PP2A belong to the phosphoprotein phosphatase (PPP) superfamily. The PP1 holoenzyme is composed of a catalytic subunit (PP1α, PP1β or PP1γ) and a regulatory PP1-interacting protein that directs PP1 to a speci c location in the cell and determines its activity and substrate speci city [9]. Over 200 validated PP1 interactors bind to the PP1 catalytic subunits using multiple binding motifs such as RVxF, SpiDoc, SILK, MyPhoNE, ΦΦ and NIPP1-helix [10,11]. Small molecules that can compete with these docking motifs can be used to block the binding of the interactors to PP1, thus functionally disrupting distinct PP1 holoenzymes and inhibiting PP1-mediated processes, such as the replication of EBOV by changing the phosphorylation of viral or host proteins.
In the current study, we analyzed the effects of 1E7-03 on EBOV replication and virus mutation during a continuous long-term virus culture. We observed a single mutation, NP E619K, after passage 4 in viruses puri ed from 1E7-03 treated cultures. To explore the involvement of PP1 in EBOV replication further, we examined the effects of this mutation on viral replication/transcription, viral capsid formation, interaction of NP with PP1 and PP2A, and on NP dimerization. Analysis using transmission electron microscope (TEM) showed that while Wild-type (WT) NP e ciently formed helical capsids with VP35 and VP24, no capsids were detected for the NP E619K mutant. However, NP E619 was able to form capsids in 1E7-03 treated cells, while capsid formation by the WT NP was inhibited by 1E7-03. We used the split NanoBiT system to assess the dimerization of NP and its binding to PP1, PP2A, and VP30. The NP E619K mutant dimerized with less e ciency than the WT NP, yet it bound PP1 with greater a nity than WT NP. By contrast, the NP E619K mutant bound to PP2A and VP30 with similar intensity. To investigate whether NP E619K can oligomerize/polymerize and the effect of 1E7-03, we conducted DPS-crosslinking experiments that revealed fewer monomers and dimers in the soluble fraction for the NP E619K mutant. Treatment with 1E7-03 increased both monomers and dimers for NP E619K, but not for the WT NP. We employed uorescent microscopy to determine co-localization of PP1 with NP in live cells using NP-RFP and PP1-  21,22]. To further investigate the mechanism of 1E7-03-mediated inhibition of EBOV replication, we conducted a continuous viral culture study (Fig. 1A for work ow). First, Vero-E6 cells were pre-treated with 3 µM 1E7-03 for 24 hrs, and then infected with a recombinant EBOV that expresses eGFP (EBOV-eGFP) [23] at a multiplicity of infection of 0.01 PFU/cell. The cells were treated with 3 µM 1E7-03 every 24 hrs for 4 days (Supplemental Fig. 1). After 4 days, the supernatants were collected, virus was titrated, and used in the following round of infection. As the viral titer from the rst passage was low, the maximum available amount (15 PFU) was used to infect the cells in the next passage. Passage 2 was performed in the same way, except that the treatment was extended to 11 days (Supplemental Fig. 2). At 11 days post infection, the supernatants were collected, titrated, and used to infect the monolayers for passage 3. The third passage began with a MOI 0.01 PFU/cell and was continued for 11 days with daily 1E7-03 treatments (Supplemental Fig. 3). Over the course of 10 days, the treatment was repeated. On day 11, the virus titers from passages 2 and 3 were nearly equal, while the titer from passage 4 was lower. (4.8 log 10 reduction, Fig. 1B). Following passage 4, three viral samples were isolated from both 1E7-03 treated and untreated cultures. Deep sequencing of the viral RNA was performed to identify mutations present in the viral genome of the samples from the treated 1E7-03 cultures. Subsequently, several mutations in these proteins were found (Supplemental Table 1). In contrast to the treated cultures, the samples obtained from the untreated ones only exhibited mutations in the VP24 protein (Supplemental Table 2). These mutations coincided with VP24 alterations seen in the treated cultures. When the three samples were compared, the NP E619K mutation (Fig. 1C) had the most signi cant p-value in sample 2 and was the only single mutation observed in sample 3. Therefore, we focused our subsequent analysis on the NP E619K mutation.
The NP E619K mutation is located within the unstructured linker that connects the N-terminal structured lobes and the C-terminal tail of NP [24] (Fig. 1D). To investigate whether NP E619K mutation had an effect on NP expression, it was introduced in the NP expression vector and the expression was analyzed in Vero-E6 cells. Immunoblotting analysis showed similar expressions of both the WT NP and NP E619K (Fig. 1E). We then evaluated the impact of the NP E619K mutation on EBOV replication/transcription using an EBOV minigenome system (Fig. 1F). Vero-E6 cells were cultured in 24-well plates and transfected with the EBOV minigenome and plasmids expressing components of the EBOV polymerase complex (L, VP35, and VP30) under the control of T7 polymerase and a T7 polymerase-expressing vector [6]. Additionally, WT NP or NP E619K mutants were expressed from co-transfected plasmids with CMV promoter controlling the expression. Renilla luciferase activity was measured 48 hrs after transfection.
When compared with the WT NP, the mutation resulted in a less than two-fold decrease in the minigenome reporter gene signal (Fig. 1F), indicating that NP E619K was functional in EBOV transcription. The reintroduction of the NP E619K mutation to EBOV-eGFP caused a slight delay in replication kinetics at day 5 post infection compared to WT EBOV-eGFP. However, 1E7-03 still effectively inhibited both viruses ( Fig. 1G and Supplemental Fig. 4), indicating that the adaptation was insu cient to fully overcome the inhibition.
The NP E619K mutation impairs EBOV capsid formation but facilitates capsid formation in cells treated with 1E7-03 Using transmission electron microscopy (TEM), we tested the effect of NP mutation on EBOV capsid formation. We expressed either WT NP or an NP E619K mutant fused with RFP in HEK 293 cells along with VP35 and VP24 proteins. The WT NP was able to facilitate capsid structures formation ( Fig. 2A). In contrast, no capsid formation was observed when the NP E619K mutant was used (Fig. 2B). Flow cytometry analysis revealed similar expression levels of WT NP and NP E619K (Supplemental Fig. 5), indicating that the NP E619K mutant might be de cient in capsid formation. To assess the effect of 1E7-03 treatment on capsid formation, TEM images of HEK 293 cells expressing NP, VP35 and VP24 were analyzed after treatment with DMSO vehicle or 1E7-03. While capsids were absent in 1E7-03-treated cells expressing WT NP (Fig. 2C), capsids were observed in cells expressing the NP E619K mutant and treated with 1E7-03 (Fig. 2D). We found that the linear size of EBOV capsids in cells expressing the NP E619K mutant and treated with 1E7-03 was 1.4-fold smaller than that of the capsids observed in the cells expressing WT NP (Fig. 2E). These results suggest that the NP E619K mutation facilitates the virus adaptation to 1E7-03 treatment and promotes capsid formation under the drug treatment.

NP E619K mutation prevents NP dimerization and increases binding of PP1
Recently we have used a Split NanoBiT-based system for studying the interaction of PP1 with its regulatory partners [22]. By fusing PP1 to the large bit (LgBit) and PP1 binding peptides (such as the central domain of NIPP1 (cdNIPP1)) to the small bit (SmBit), we were able to assess the effect of the NP E619K mutation on NP dimerization and PP1 binding (see Materials and Methods for vector details) (Fig. 3A). WT NP formed dimers e ciently, as indicated by the strong NanoBit signal for the NP-NP interaction (Fig. 3B). In contrast, the self-interaction of the NP E619K mutant was 1,650-fold lower than that of WT NP, suggesting a decrease in dimerization (Fig. 3B). The binding of VP30 to both WT NP and the NP E619K mutant was similar (Fig. 3B). However, the NP E619K bound PP1 three-fold more strongly than the WT NP (Fig. 3C). Both WT and NP E619K mutant bound equally well PP2A B56 subunit (Fig. 3C). Taken together, these results suggest that the NP E619K mutant was impaired in dimerization, but had increased binding to PP1, suggesting that PP1 binding to NP might negatively affect NP dimerization.
Next, we tested the effect of 1E7-03 on NP-NP and NP-PP1 interaction. We rst assayed the effect of 1E7-03 on PP1 binding to its known interactor, cdNIPP1, and compared it to cdNIPP1 RVxF mutant (cdNIPP1rata, V201A/F203A mutation) that was de cient in PP1 binding as indicated by the reduced NanoBiT signal (70-fold reduction, Fig. 3D). For this analysis, 293T cells were co-transfected with the speci ed NanoBiT plasmids for 24 hrs and then treated for 6 hrs with serial dilutions of 1E7-03 (1.3-14 µM). We found that 1E7-03 disrupted the PP1-cdNIPP1 binding (IC 50 = 2.1 µM) (Fig. 3D), and also affected NP-NP interaction to a lesser extent (IC 50 = 15 µM) (Fig. 3E). This suggests that 1E7-03 may not have a direct effect on NP homodimerization. As was previously mentioned, the interaction between NP E619K-NP E619K was weaker than that of WT NP-WT NP, and 1E7-03 had no effect on this weak interaction (Fig. 3E). However, 1E7-03 was able to disrupt both PP1-NP interaction (IC 50 = 0.8 µM) and the PP1-NP E619K interaction (IC 50 = 2.2 µM) (Fig. 3F). As PP1-NP E691K binding was stronger, the quantity of PP1 bound to NP E619K mutant even in the cells treated with 3 µM 1E7-03 was higher than that of untreated NP-PP1 (Fig. 3F, dashed line). This indicates that the NP E619K mutation allowed for adequate PP1 binding even under the presence of 3 µM 1E7-03. Additionally, 1E7-03 had an equal suppressing effect on the interactions between WT NP and NP E619K with the PP2A B56 subunit (Fig. 3G). Split NanoBiT system experiments revealed that both PP1 and NP were co-expressed equally well (Supplemental Fig. 6A). 1E7-03 treatment had no effects on PP1 expression, but slightly reduced NP expression (Supplemental Fig. 6B and C). This suggests that the effects of 1E7-03 were not due to the signi cant reduction of PP1 or NP protein expression.
To validate the NanoBiT results, we next analyzed the effect of 1E7-03 treatment on the oligomerization of the WT NP and the NP E619K mutant using dithiobis[succinimydylpropionate] (DSP) crosslinking followed by non-reducing SDS gel electrophoresis [25]. Crosslinking with DSP allowed us to detect both monomers and dimers of the WT NP and NP E619K. After DSP cross-linking, WT NP formed diners and higher oligomeric forms compared to non-DSP-treated WT NP (Fig. 4A). In contrast, DSP cross-linking of NP E619K resulted in the formation of signi cantly fewer monomers and dimers compared to WT NP, despite having equal amounts of monomers the input not treated with DPS ( Fig. 4A and B). This suggests that the NP E619K mutant is compromised in its ability to form oligomers. Treatment with 1E7-03 signi cantly increased the amount of NP E619K monomers in DSP-treated non-reduced samples and also had a trend toward increasing the amount of dimers ( Fig. 4A and B). In contrast, 1E7-03 treatment had no effect on the amount of NP WT monomers and dimers ( Fig. 4A and B). These observations suggest that the inability of NP E619K to form dimers might be due to the sequestration of the mutant NP within an insoluble complex within the cells. This could reduce the amount of soluble NP E619K monomers available for oligomerization. Treatment with 1E7-03 seems to release the mutant NP into the soluble fraction, thus facilitating its oligomerization.
To further con rm the interaction between NP and PP1, we analyzed the co-precipitation of NP with PP1. 293T cells were co-transfected with vectors expressing V5-tagged PP1α and Flag-tagged NP. The cells were treated with either vehicle (DMSO) or 10 µM 1E7-03. PP1α was immunoprecipitated with anti-V5 antibodies, resolved on 10% SDS polyacrylamide gel and immunoblotted with anti-Flag antibodies to detect NP protein and anti-V5 antibodies to detect PP1 (Supplemental Fig. 7). NP E619K showed a slightly higher a nity for PP1 (Supplemental Fig. 7B). Treatment with 1E7-03 reduced PP1 binding to the WT NP, whereas PP1 binding was increased with NP E619K (Supplemental Fig. 7).
Overall, these observations indicate that the NP E619K mutation is defective in dimerization and has a stronger binding a nity to PP1 compared to the WT NP.
NP co-localizes with PP1α To determine which PP1 isoform might interact with NP in cultured cells, we examined the co-localization of NP with PP1 by utilizing live-cell uorescence imaging of co-expressed NP-RFP and PP1-eGFP fusion proteins (Fig. 5A). We observed robust co-localization of NP with PP1α, but not with PP1 β/δ or PP1γ (Fig. 5A). We then calculated Mander's coe cient to validate and differentiate the co-localization of WT and mutant NP. There was a signi cant increase in the co-localization of the NP E619K mutant with PP1α, as well as an inclination towards increased co-localization with PP1 β/δ, but no change in colocalization with PP1γ when compared to WT NP (Fig. 5B). Thus, NP is likely to bind PP1α.

Analysis of PP1 binding sites on NP
To identify PP1 binding sites on NP, we analyzed the presence of potential PP1 binding motifs within the NP sequence and found ve potential binding sites (Fig. 6A). To evaluate their functionality, we mutated them (Fig. 6B) and tested their effect of mutations on PP1-NP interaction in the split NanoBit system (Fig. 6C). Motifs 1, 2 and 3 are located within the N-terminal NP core (Fig. 6C), while motifs 4 and 5 are located within the disorder linker and the C-terminal tail domain, respectively (Fig. 6C). Mutations of motifs 2 and 3 had a strong negative effect on PP1-NP interaction (Fig. 6D), while mutation of motif 5 showed an intermediate effect (Fig. 6D). In contrast, mutations in motifs 1 and 4 had no adverse effect on PP1-NP interaction (Fig. 6D). We further tested these mutations in the EBOV minigenome and found that the mutations in either motif 2, 3, or 5 signi cantly impacted EBOV transcription (Fig. 6E).
We next analyzed the effect of NP deletions on PP1 binding. We generated six NP deletion mutants that spanned the entire NP sequence (Fig. 7A). NP deletion mutants 1 and 2, which contained only part of the N-terminal domain (1-133 aa and 1-266 aa, respectively), did not bind PP1 (Fig. 7B). NP lacking the rst 133 amino acids also did not bind PP1 (Fig. 7B). NP deletion mutants containing 1-401 amino acids or 1-533 amino acids could bind PP1 to the level of WT NP (Fig. 7B). The NP mutant containing 1-666 amino acids also retained the ability to bind PP1. Structures of the NP deletion mutants 1, 2 and 6, which were unable to bind PP1, are shown in Fig. 7C. It is likely that these deletions interfere with the interaction of NP and RNA, suggesting that the NP-PP1 interaction may involve RNA binding.
Taken together, our data indicate that PP1 interacts with NP, and that the E619K mutation in NP enhances this interaction, while decreasing NP-NP binding and compromising EBOV capsid formation but facilitating NP capsid formation in the presence of 1E7-03.

DISCUSSION
In this study, we extended our previous analysis of the role of PP1 in EBOV replication by investigating the long-term effects of 1E7-03 treatment in viral culture. We identi ed the NP E619K mutation that developed during the long-term treatment of EBOV infected Vero-E6 cells. We observed that, although this mutation had only a mild effect on EBOV transcription resulting in less than 50% decrease in minigenome replication, it might still facilitate EBOV capsid formation when cells expressing NP E619K were treated with 1E7-03. Our analysis of viral capsid assembly indicated that there was a striking difference between the WT NP and the NP E619K mutant. In cells that expressed WT NP, VP35, and VP24, we observed the formation of viral capsids, whereas no capsids were detected in the cells that expressing NP E619K mutant, VP35 and VP24. Furthermore, no capsids were found in WT NP, VP35 and VP24 -expressing cells after treatment with 1E7-03, while capsids were observed in NP E619K, VP35 and VP24 expressing cells treated with 1E7-03, although the capsid were shorter in size. This observation suggests that EBOV might have adapted to 1E7-03 by being able to form capsids under drug pressure. To obtain further insight into the effect of the NP E619K mutation, we analyzed NP dimerization, as well as NP E619K binding to PP1, PP2A B56 and VP30, using split NanoBiT system and crosslinking experiments. We found that the NP E619K mutation impaired NP's ability to dimerize, reducing dimerization e ciency by more than 1,000fold. NP E619K mutant bound to PP1 more e ciently but had the same binding of PP2 B56 and VP30 as the WT NP. Analysis of the effect of 1E7-03 in the split NanoBiT system showed that at 3 µM 1E7-03 concentration, which was used in viral passaging experiments, NP E619K mutant retained PP1 binding comparable to untreated WT NP-PP1. Crosslinking with DSP showed that the NP E619K mutant had reduced amounts of monomers and dimers compared to WT NP. Remarkably, treatment with 1E7-03 increased both monomers and dimers, suggesting that PP1 may aid in sequestration of NP, which hinders its ability to dimerize. Co-precipitation of PP1α with NP demonstrated that the NP E619K mutant binds more PP1 and that this enhanced binding was preserved in the presence of 1E7-03. On the contrary, WT NP binds less PP1 and its weaker binding was also sustained even in the cells treated with 1E7-03.
Together, these results support the idea that NP E619K is more e cient at inducing capsid formation than WT NP in the presence of 1E7-03. Furthermore, NP was found to co-localize with PP1α, with the colocalization being particularly pronounced for NP E619K. This agrees with the ndings from the split NanoBiT system, which showed a stronger interaction between NP E619K and PP1. Our analysis of PP1's interaction with NP revealed at least three potential PP1 binding sites. However, deletion analyses highlighted the importance of an intact N-terminal domain for the binding of PP1 to NP, suggesting RNA binding might be involved. Consequently, PP1 might bind to NP indirectly through a PP1 regulatory subunit capable of binding RNA. As PP1 interacting regulatory proteins include more than 200 validated members [10,11], it is possible that one of these subunits with an RNA binding capability is involved in this interaction. Further analysis with the use of techniques such as proximity labelling is required to identify this subunit and to better understand PP1 binding.
Our study points to the hitherto unrecognized role of PP1 in regulating NP capsid assembly (summarized in Fig. 8). We propose that NP binds PP1 and this binding is requisite for NP oligomerization and subsequent capsid formation. Treatment with 1E7-03 precludes PP1 binding with WT NP and thus blocks nucleocapsid formation. The increased PP1 binding by NP E619K seems to impede oligomerization. Treatment with the PP1-targeting 1E7-03 compound allows partial PP1 dissociation from NP E619K mutation and thus induces capsid formation. In accord with the previous study [7], we observed strong NP interaction with PP2A, which was not affected by NP E619K mutation. We hypothesize that NP might interact with both PP2A and PP1, and that PP1 and PP2A might work in concert. An example of concert recruitment and ne-tuning of PP1 and PP2A can be seen in RepoMan, a protein encoded by CDCA2, which acts as a scaffold for both PP1 and PP2A B56 [26,27]. Phosphorylation of RepoMan by Aurora B on Ser-893 and Thr-394 prevents PP1 binding [27,28]. Conversely, phosphorylation of RepoMan by CDK1 on Ser-591 promotes the recruitment of PP2A B56, which reverses Ser-893 phosphorylation and enables re-recruitment of PP1 and dephosphorylation of M phase proteins [29][30][31]. In this regard, the NP E619K mutation, which introduces a positive charge instead of a negative charge, may mimic NP dephosphorylation which would facilitate PP1 recruitment. It is thought that the function of PP2A which bound to NP may include dephosphorylation of an unknown NP residue(s) to aid PP1 recruitment, prevent capsid formation and induce transcription. The N-terminal part of NP ( rst 450 aa) is necessary for NP-NP interaction and, as well as the following 150 amino acid residues, is critical for nucleocapsid formation and viral replication [32]. Our previous global phosphoproteomic analysis of EBOV virions identi ed twenty NP phosphorylation sites [33]. Fourteen of these sites (Thr-536, Ser-541, Thr-545, Thr-563, Thr-597, Ser-598, Thr-601, Thr-603, Tyr-686, Thr-687, Tyr-688, Ser-691, Tyr-696 and Thr-701) were found within the large unstructured sequence that connects the N-terminal and C-terminal domains of NP [33]. Previous studies identi ed phosphorylation of Thr-563, Ser-581, Ser-587 and Ser-647 in NP expressed in cultured cells [34]. As existing crystal structures of EBOV NP only capture parts of the Nterminal region [35][36][37], and a C-terminal domain [38], we previously constructed a full-length model of NP using de novo prediction, which showed highly stable globular-like 'structured' sections during an equilibrium MD simulation in a periodic water box [33]. The NP residues 412-645 were highly exible during MD simulation [33], suggesting that these residues are accessible for PP1 or PP1 regulatory subunit binding. We also identi ed ten phosphorylation sites on VP35 that was packaged in EBOV virions [33]. Interestingly, dephosphorylation of Thr-210 blocked EBOV transcription and prevented VP35 binding to NP [33], suggesting that NP-associated phosphatases can also control EBOV transcription by dephosphorylating VP35.
The NP E619K mutation resides near the VP30-binding PPxPxY motif that was also found in the host RBBP6 protein [39]. Recently, a mass spectrometry approach identi ed additional PPxPxY-containing host proteins, including hnRNP L, hnRNPUL1, and PEG10 that all strongly interact with VP30 [40]. While we did not observe any changes in VP30 binding to NP E619K, which was extremely weak in our split NanoBiT system, it is possible that NP E619K mutation affects the VP30 exchange between host proteins and NP. Further analysis is needed to test the binding of NP E619K with host proteins and other factors such as VP35.
Taken together, our ndings suggest that host PP1 has a unique role that involves interacting with NP and controlling NP dimerization and EBOV capsid formation.

Chemicals and reagents
Page 11/30 1E7-03 (purity above 98%) was synthesized by Enamine (Kyiv, Ukraine) as previously described [14].  To prepare WT NP-mCherry and NP E619K-mCherry expression vectors, pcDNA3.1(-) plasmid was digested with Not1 and Kpn1 restriction enzymes and puri ed on the agarose gel. The mCherry fragment was ampli ed by PCR from pcDNA3.1-mCherry plasmid with forward CATGGCAATCCTGCAACATCATCAGAAGggcgaggaggataacatggccatc and reverse GTTTAAACTTAAGCTTGGTACCTACTTGTACAGCTCGTCCATGCCGCCGGTGG primers. PCR fragments of WT NP and NP E619K were ampli ed with forward AACGGGCCCTCTAGACTCGAGCGGCCGCATGGATTCTCGTCCTCAGAAAATCTGG and reverse GATGGCCATGTTATCCTCCTCGCCCTTCTGATGATGTTGCAGGATTGCCATG primers from pCEZ NP and pCEZ NP E619K expression vectors described above. All fragments were combined in one vector with Gibson assembly kit (New England BioLab, Ipswich, MA; cat# E2611S) according to the manufacturer's protocol.
NanoBiT vectors. We evaluated all combinations of expression constructs to determine the best combination and orientation for fusions of tested proteins to LgBiT and SmBiT, as recommended by

NP molecular presentation
The previously described de novo NP structure [31] was used to prepare NP images in Chimera 1.14 (https://www.cgl.ucsf.edu/chimera).

Statistical analysis
All graphs were prepared using GraphPad Prism 6 software. The data were presented as mean ± SD or standard error of the mean (SEM) as indicated in the gure legends. Statistical comparison was done with Student's t test. Where indicated, non-linear regression analysis was performed to determine IC 50 using GraphPad Prism 6 built-in algorithms.     Pearson's correlation analysis. Prior to the correlation analysis, the image colors were split, and the threshold parameters were adjusted.  Model of NP-PP1 interaction and the effect of 1E7-03 on capsid formation. NP binding to PP1 leads to NP dimerization and EBOV nucleocapsid formation. The addition of 1E7-03 prevents PP1 binding to WT NP and blocks nucleocapsid formation. The enhanced association of NP E619K with PP1 prevents its dimerization and capsid formation. Treatment with 1E7-03, dissociates the excess of PP1 and allows NP E619K to dimerize and form nucleocapsids.