Crystal Structure of the Marburg Virus Nucleoprotein Core Domain Chaperoned by a VP35 Peptide Reveals a Conserved Drug Target for Filovirus

ABSTRACT

Filovirus nucleoprotein (NP), viral protein 35 (VP35), and polymerase L are essential for viral replication and nucleocapsid formation. Here, we identify a 28-residue peptide (NP binding peptide [NPBP]) from Marburg virus (MARV) VP35 through sequence alignment with previously identified Ebola virus (EBOV) NPBP, which bound to the core region (residues 18 to 344) of the N-terminal portion of MARV NP with high affinity. The crystal structure of the MARV NP core/NPBP complex at a resolution of 2.6 Å revealed that NPBP binds to the C-terminal region of the NP core via electrostatic and nonpolar interactions. Further structural analysis revealed that the MARV and EBOV NP cores hold a conserved binding pocket for NPBP, and this pocket could serve as a promising target for the design of universal drugs against filovirus infection. In addition, cross-binding assays confirmed that the NP core of MARV or EBOV can bind the NPBP from the other virus, although with moderately reduced binding affinities that result from termini that are distinct between the MARV and EBOV NPBPs.

IMPORTANCE Historically, Marburg virus (MARV) has caused severe disease with up to 90% lethality. Among the viral proteins produced by MARV, NP and VP35 are both multifunctional proteins that are essential for viral replication. In its relative, Ebola virus (EBOV), an N-terminal peptide from VP35 binds to the NP N-terminal region with high affinity. Whether this is a common mechanism among filoviruses is an unsolved question. Here, we present the crystal structure of a complex that consists of the core domain of MARV NP and the NPBP peptide from VP35. As we compared MARV NPBP with EBOV NPBP, several different features at the termini were identified. Although these differences reduce the affinity of the NP core for NPBPs across genera, a conserved pocket in the C-terminal region of the NP core makes cross-species binding possible. Our results expand our knowledge of filovirus NP-VP35 interactions and provide more details for therapeutic intervention.

INTRODUCTION

Marburg virus (MARV), a member of the family Filoviridae, can cause severe disease in humans, including hemorrhagic fever (1). Since its first identification in 1968, MARV has caused sporadic infections in small numbers of individuals in Africa for decades (2). Two large outbreaks occurred in the Democratic Republic of the Congo (DRC) in 1998 to 2000, with ∼83% lethality (1), and in Angola in 2004 and 2005, with ∼90% lethality (3). Thus, MARV represents a tremendous threat to public health. Currently, there are no clinically approved vaccines or therapeutics available to treat MARV infections, and therefore, research on MARV is urgently needed.

Like other members of the family Filoviridae, MARV has a nonsegmented, negative-strand (NNS) RNA genome that is encapsidated by a viral nucleoprotein (NP). Filoviruses belong to the order Mononegavirales, which includes two other genera, Ebolavirus (members, Ebola virus [EBOV], Sudan virus [SUDV], Reston virus [RESTV], Taï Forest virus [TAFV], and Bundibugyo virus [BDBV]) and Cuevavirus (Lloviu virus [LLOV]). Filoviruses share a genome organization that encodes seven viral proteins: an NP, the viral proteins VP35 and VP40, the surface glycoprotein (GP), the viral proteins VP30 and VP24, and a polymerase (L) (4). In MARV, NP plays a central role in virion maturation. As an indispensable component of the viral replication machinery (5), NP interacts with other viral proteins, including VP24 (6), VP40 (7), VP35, and VP30 (8), to organize the replication process. It also assembles into a helical tubular structure as the scaffold of nucleocapsid formation (9–12). Sequence homology (13) shows that NP contains a conserved N-terminal region that is sufficient for self-assembly and single-stranded RNA (ssRNA) packaging (9), as well as a largely unstructured C-terminal region (14) that contains a domain required for virion budding (7).

Structural studies of filovirus NP mainly focus on EBOV NP, while the structure of MARV NP has yet to be solved. The structure of a globular domain in the NP C-terminal region from EBOV was recently determined (15), followed by structures from other members of the genus Ebolavirus (14). However, the homologous region in MARV is nonconserved and appears to exist as a molten globule, according to nuclear magnetic resonance (NMR) data (14).

Crystallographic studies have also been reported for the N-terminal globular domain of EBOV NP (16–18). In these studies, a peptide from VP35 termed NP binding peptide (NPBP) was found to bind the N-terminal core domain of NP and obstruct the oligomerization of NP (16, 18). In both EBOV and MARV, VP35 is a polymerase cofactor in the RNA-dependent RNA polymerase (RdRp) complex, serving as a bridging protein between NP and polymerase L (8, 19). It is also an interferon (IFN) antagonist in viral immune evasion (20).

Previous studies have demonstrated that the C-terminal interferon-inhibitory domain (IID) of EBOV VP35 (21) and two coiled-coil motifs in MARV NP (amino acids [aa] 320 to 348 and 371 to 400) are essential for the interaction between NP and VP35 (22). There is high affinity between the VP35 N-terminal peptide and the NP core region, and this binding indicates that VP35 may chaperon NP to maintain it in a monomeric and RNA-free form until it reaches the RNA-binding niche (16, 18).

Here, we identify the NPBP from MARV VP35 and show that it binds to the core region of the N terminus of MARV NP. We then solve the complex structure of NPBP and the NP core region and compare it to the EBOV NP core/NPBP structure (16, 18). Most of the interactions observed in these two structures are conserved, except the discrepant contacts between the NPBP termini and the NP core. Biochemical analysis of the cross-binding reaction also demonstrated dynamic variation when the same NPBP binds to the NP core region from either EBOV or MARV. Furthermore, a conserved pocket in the NP core is responsible for the mutual cross-binding of NPs and NPBPs from EBOV and MARV and may be a good target for broad-spectrum drug development.

RESULTS

Identification of the MARV NPBP.

A previous study revealed that residues 20 to 48 of EBOV VP35 are conserved in the genus Ebolavirus, and this peptide is essential for binding to NP, whereas residues 1 to 19 are largely dispensable for VP35 activity (16). We found a mostly conserved but truncated N-terminal region in VP35 of different stains in the genus Marburgvirus through comparison of EBOV VP35 by sequence alignment (Fig. 1A). Residues 1 to 19 in EBOV VP35 are absent in MARV VP35, and residue M20 in EBOV VP35 corresponds to the first residue (M1) of MARV VP35. We chose a peptide from MARV VP35 consisting of residues 1 to 28, termed MARV NPBP, to test for binding to NP. Although there are variations in the N and C termini of the NPBP peptides between EBOV and MARV, most of the key residues (including I29, S30, E31, L33, M34, T35, G36, I40, V4,3 and F44 [EBOV VP35 numbering]) for NP binding are conserved between the viruses (Fig. 1A).

FIG 1.

MARV NPBP from VP35 binds the NP core domain. (A) Sequence alignment of the conserved VP35 N-terminal peptides of EBOV (AAD14582.1) and MARVs (MARV Ci67, ABS17556.1; MARV Pop, Q03039.1; MARV Ozo, Q6UY68.1; MARV Mus, P35259.1; RAVV [Ravn virus, a member of the species Marburg marburgvirus] Ravn, Q1PDC9.1; MARV DRC99, ABE27090.1; MARV Angola, Q1PD52.1; MARV Leiden, AEW11934.1). The alignment is numbered with respect to the EBOV VP35 sequence, and the secondary structure of EBOV NPBP is shown above the alignment. Conserved residues involved in contacts with EBOV NP core are indicated with asterisks below. (B) Size exclusion analysis of MARV NP core monomer (red), NPBP (yellow), and the MARV NP core/NPBP complex (blue). The chromatographs of the three prepared protein species on a calibrated Superdex 75 column (10/300 GL), as well as the separation profiles of each pooled sample (peaks 1, 2, 3) on SDS-PAGE, are shown. Lane M is the standard protein ladder used as marker. Only NP core could be observed on the gel; NPBP is too small to detect. (C) ITC assay showing the high-affinity binding of the MARV NPBP (VP35 1-to-28 peptide) to the MARV NP core (NP 18 to 344).

We used a truncated NP construct termed MARV NP core, including residues 18 to 344 of the MARV NP, for biochemical assays. The rationale for this was that the N-terminal residues of EBOV NP induce NP oligomerization and precipitation (16–18), which makes NP purification difficult. This ∼37-kDa MARV NP core construct retains most of the conserved N-terminal region, while the N-terminal oligomerization domain is excluded. Expression and purification of NP core were achieved by following a previously described method (17) with some optimization. Size exclusion chromatography indicated that the purified protein exists as a monomer in solution without nucleic acids (Fig. 1B).

Next, we mixed the NP core monomer with excess NPBP and found a moderate shift of the main peak, representing complex formation (Fig. 1B). Furthermore, isothermal titration calorimetry (ITC) assays were conducted to confirm the binding, demonstrating that NPBP binds NP core with high affinity (equilibrium dissociation constant [K_D] = 7.35 ± 3.40 nM) (Fig. 1C).

Complex structure of MARV NP core and NPBP.

The MARV NP core/NPBP complex was then successfully crystallized, and the crystal structure was solved to a resolution of 2.6 Å by molecular replacement using the EBOV NP core/NPBP complex structure (PDB ID 4YPI) (16) as a search model (Table 1). Three complex molecules were observed in the asymmetric unit. According the unambiguous electron density, residues 21 to 336 of MARV NP were built into the model, and residues 5 to 27 of NPBP were included. Like the EBOV NP core domain, the MARV NP core contains an N-terminal head lobe (K21 to T221) and a C-terminal foot lobe (L227 to L336) connected by a flexible hinge, and both domains are principally composed of α-helices (Fig. 2A and B). The head lobe consists of 11 α-helices and two parallel β-strands, while the foot lobe contains four α-helices and two short antiparallel β-strands.

TABLE 1.

Data collection and refinement statistics

Statistic	Value for MARV NP core/NPBPa
Data collection
Wavelength (Å)	0.97922
Space group	P 32
Cell dimensions
a, b, c (Å)	101.32, 101.32, 96.33
α, β, γ (°)	90.00, 90.00, 120.00
Resolution (Å)	50.00–2.60 (2.69–2.60)
Rmergeb	0.141 (2.307)
I/σI	14.937 (1.209)
Completeness (%)	100.0 (100.0)
Redundancy	6.3 (6.3)
Wilson B factor	40.76
Refinement
Resolution (Å)	34.91–2.60
No. of reflections	30,340
Rwork/Rfreec	0.214/0.244
No. of atoms
Protein	7,871
Ligand/ion	0
B factors
Protein	57.5
Ligand/ion
RMSd deviations
Bond lengths (Å)	0.005
Bond angles (°)	0.990
Ramachandran plot
Favored (%)	94.49
Allowed (%)	5.31
Outliers (%)	0.20

^aThe values in parentheses are for the highest-resolution shell.

^bR_merge = Σ_iΣ_hkl |I_i − <I>|/Σ_iΣ_hklI_i, where I_i is the observed intensity and <I> is the average intensity from multiple measurements.

^cR_work = Σ‖F_o| − |F_c‖/Σ|F_o|, where F_o and F_c are the structure factor amplitudes from the data and the model, respectively. R_free is the R factor for a subset (5%) of reflections that were selected prior to refinement calculations and were not included in the refinement.

^dRMS, root mean square.

FIG 2.

Overall structure of the MARV NP core/NPBP complex. (A) NPBP/NP core complex in cartoon format. The head lobe of NP core (residues 21 to 221) is dark green, and the foot lobe (residues 222 to 336) is light green, while NPBP is magenta. (B) Topological diagram of the NP core and NPBP. The hinge region (α7 to α11) of the NP core is shown.

The peptide consisting of 23 residues from VP35 binds to one side of the foot lobe, which is formed by the α12 and α13 helices of NP core (Fig. 2A). Residues 7 to 16 and 21 to 24 of NPBP form two orthogonal α-helices, and an abrupt turn is located between the helices (Fig. 2A and B).

We then compared our complex structure with two complex structures of EBOV NP core bound to NPBP (PDB ID 4YPI and 4ZTA), reported by two independent groups (16, 18). Although these two structures have different truncated C termini of EBOV NP, the NPBP binding modes are almost the same. We found several differences between MARV and EBOV NPBPs (Fig. 3). First, though MARV NPBP has a truncated N terminus compared to EBOV NPBP due to the “missing” residues, it has a similar extended α1 helix with a differently oriented N terminus. Second, the C-terminal region following the α2 helix of MARV NPBP is shorter than that in EBOV NPBP and has an orientation that is distinct from EBOV's. While EBOV NPBP wraps around the β hairpin of NP, MARV NPBP just adjoins one side of the protruding β-strands and forms an open angle buried in the NP core.

FIG 3.

Overall comparison of the MARV NP core/NPBP complex with the two EBOV NP core/NPBP complexes. Two EBOV NP core/NPBP complexes (PDB ID 4YPI and 4ZTA) were used for comparison. The NPBPs are highlighted, showing the extended first α-helix and the differently oriented C terminus in MARV NPBP.

The NPBP binds to the MARV NP core via multiple contacts.

Examination of the complex structure showed that the distances of 16 out of the 23 NPBP residues are within 4.5 Å of NP core residues (Table 2), which suggests the existence of a large binding interface between MARV NPBP and NP core. Surface electrostatics analysis indicated that charge complementarity exists at three sites in the binding interface, including both termini and the first α-helix of NPBP (Fig. 4A). For the N terminus of NPBP (site I), positively charged residues of NPBP surround the negatively charged E234 from the NP core α14 helix. In site II, a negatively charged region from the first helix of NPBP binds to positively charged residues connecting the two lobes of NP core. In site III, at the C terminus of NPBP, the side chains of R251 and K263 from NP core are in close contact with the negatively charged NPBP terminus (Fig. 4B).

TABLE 2.

Contacts between NPBP and the NP truncations of EBOV and MARV^a

EBOV			MARV
NP core	No. of contacts	VP35	VP35	No. of contacts	NP core
K257, L354, Q355, A358	5, 8, 4 (1), 2	P21
L255, Q256, K257, E351, L354, Q355	3, 7, 10 (2), 1, 2, 2	G22
D252, H253, L255, V323, H327, E351	4, 4, 4, 3, 5,3	P23
D252, L255, V262	13 (1), 5 (1), 1	E24	S5	1	E234
H253	4	L25
K257	2	S26	M7	1, 5, 4, 5, 1	E234, L237, Q238, K239, V244
K248, D252	4 (1), 7	G27	Q8	5, 3	K230, E234
L255, V262	3, 5	I29	V10	2, 1	L237, V244
K248, L251, D252, L255	11 (1), 6, 7 (1), 3	S30	S11	5, 6, 4, 1	K230, L233, E234, L237
R240, F241, K248	4 (1), 8, 6	E31	E12	4 (1), 10, 4	R222, F223, K230
L251, L255, L284, A288	3, 2, 2, 2	L33	L14	3, 1, 1	L233, L237, A270
F241, L244, K248, L251, L284, L287, A288, A294	5, 2, 7, 1, 2, 8, 6, 2	M34	M15	3, 1, 2, 7, 2, 2, 9, 4, 2	F223, L226, V229, K230, L233, L266, L269, A270, A276
F241, A288, G291, E292, A294, P295	1, 2, 13, 1, 3, 2	T35	T16	2, 12, 3, 2	A270, G273, A276, P277
A288	8	G36	G17	5	A270
K281, L284, S285	4, 6, 5	V40	I21	4, 5, 5	K263, L266, S267
			Q23	2	V244
L255, V262, L264, R269	4, 2, 4, 2 (1)	I43	V24	1, 5, 4, 5	L237, V244, T245, L246
L251, I254, L264, V277, F280, K281, L284	2, 1, 6, 2, 4, 13 (1), 3	F44	F25	7, 5 (1), 3, 1, 11, 1	L246, R251, V259, F262, K263, L266
R269	5 (1)	C45	G26	3 (1)	K263
K281	2	D46	A27	2 (1)	K263

^aResidues from NPBP were aligned, and the conserved residues are shaded. The corresponding conserved residues from NP are shown in boldface. The numbers of hydrogen bonds are indicated in parentheses.

FIG 4.

Electrostatic surface view of the MARV NP core and NPBP complex. (A) Overall view of MARV NP core and NPBP, showing the contact interfaces. The charge-complementary sites are labeled. The chains are colored by electrostatic potential at neutral pH from −2 K_BT/e (red) to +2 K_BT/e (blue) using PyMOL. (B) Comparison of the surface electric potentials of the MARV NP core/NPBP complex and the EBOV NP core/NPBP complex (PDB ID 4ZTA). The charge-complementary sites in MARV and the corresponding regions in EBOV are labeled.

Further evaluation of the interaction also demonstrated that NPBP residues in these sites form several important hydrogen (H) bonds with NP core residues (Fig. 5A). In site II, E12 in the first α-helix of NPBP is engaged in a H bond with NP R222, and F25 of NPBP forms H bonds with R251 of NP core. Similarly, G26 and A27 of NPBP H bond with K263 in site III (Fig. 5B). Aside from the charge complementarity and H bonds, hydrophobic contacts also play a key role in the interaction between NPBP and NP core (Table 2 and Fig. 5C), and a primarily hydrophobic surface was observed between the NPBP peptide and NP core.

FIG 5.

Hydrogen bonds and nonbonded contacts in the MARV NP/NPBP complex. (A) Residue-specific hydrogen bonds between MARV NP core and NPBP are shown as dotted lines. (B) Comparison of hydrogen bonds in EBOV and the MARV complex. (C) LigPlot⁺ diagram showing hydrophobic interactions and hydrogen bonds between NP core and NPBP. Protein side chains are shown as balls and sticks. Hydrogen bonds are shown as red dashed lines. Nonbonded contacts are shown as spoked arcs.

We then compared the detailed binding modes between MARV and EBOV NP core domain bound to NPBP. The high-resolution (2.4-Å) structure of EBOV NP core bound to NPBP (PDB ID 4ZTA) was used for detailed analysis. Differences were revealed by intermolecular-contact analysis in the binding details between the two structures. Residues 21 to 28 of EBOV VP35, corresponding to 2 to 9 of MARV NPBP, are not conserved, and some of these residues form H bonds and nonpolar contacts with the surrounding EBOV NP residues. G27, the initiation site of the first α-helix of EBOV NPBP, is one of these residues, and it forms a H bond with K248 in EBOV NP. However, this residue is replaced by a glutamine (Q8) in MARV NPBP, which does not H bond with MARV NP core (Fig. 5B). Moreover, the side chain of E234 in MARV NP core flips because of the N terminus of MARV NPBP and is no longer engaged in H bond formation due to steric clashes. Nonetheless, MARV NP E234 forms contacts with the nonconserved residues in NPBP N-terminal site I by charge complementarity that do not exist in EBOV (Fig. 4B). In addition, the last two visible residues of MARV NPBP (G26 and A27) are nonconserved sites. Both of them can form H bonds with K263 in the middle of the α13 helix in MARV NP, whereas the residue C45 in the EBOV NPBP C terminus forms a H bond with R269 in EBOV NP and contributes to the different NPBP C terminus orientation (Fig. 4B and 5B).

The filovirus NP cores hold a conserved binding pocket for VP35.

We then analyzed the NPBP binding pockets of EBOV and MARV NPs. At both NPBP binding sites, residues from the hinge region, α12, α13, and the intervening β-hairpin, are involved in contacts with the conserved middle portion of NPBP (Table 2). Further sequence alignment of NP proteins from MARVs and EBOVs revealed that 20 residues in the binding site of NP core are conserved across the different MARV strains, and most of them are highly conserved in the filovirus family (Fig. 6A). Structural analysis of our MARV NP core/NPBP complex and EBOV NP core/NPBP demonstrated that these conserved residues form a similar binding pocket buried by VP35 peptides (Fig. 6B).

FIG 6.

The conserved NPBP binding pocket in the NP core foot lobe. (A) Sequence alignment of the NPBP binding pocket in NPs of MARVs (MARV Ci67, ABS17555.1; MARV Pop, P35263.1; MARV Ozo, Q6UY69.1; MARV Mus, Q1PD53.1; RAVV Ravn, Q1PDD0.1; MARV DRC99, ABE27089.1; MARV Angola, Q1PD53.1; MARV Leiden, AEW11935.1) and EBOVs (ZEBOV, AAD14590.1; BDBV, YP_003815432.1; RESTV, NP_690580.1; SUDV, YP_138520.1; TAFV, YP_003815423.1). The alignment is numbered with respect to the MARV NP sequence, and the secondary structure of MARV NP core is shown above the alignment. Conserved residues involved in contacts with NPBP in EBOV and the MARV complex are indicated with asterisks below. (B) Surface view of the NPBP binding pocket in EBOV and MARV. The conserved residues involved in contacts are shown (pale yellow in MARV and salmon in EBOV). The NPBP peptides are displayed as a cartoon.

Filovirus NPBPs bind better to their own NP cores.

Given the conserved NPBP binding site in MARV and EBOV NPs, we sought to examine whether the MARV or EBOV NP core proteins could bind each other's NPBPs. ITC assays were performed to measure the affinity between MARV NPBP and EBOV NP core (residues 36 to 351) (17), as well as the affinity between EBOV NPBP (residues 20 to 48 from EBOV VP35) (16) and MARV NP core.

A previous study reported a remarkably strong interaction (K_D = 1.1 ± 0.2 nM) between the EBOV VP35 N-terminal peptide (residues 1 to 80) and a truncated EBOV NP N-terminal protein (residues 34 to 367) (18). Our ITC assays revealed high affinity between EBOV NPBP and NP core but with a slightly lower value (K_D = 6.37 ± 1.31 nM) (Fig. 7A), which is roughly equal to that of the MARV NP core/NPBP couple. This may be due to discrepancies between the constructs and the reaction conditions used. Next, MARV NPBP was used titrated into solution with EBOV NP core, resulting in 4.5-fold reduction in affinity relative to the MARV NP core/NPBP couple (Fig. 7B). The affinity of EBOV NPBP binding to MARV NP core was measured, and it was 2-fold reduced relative to the EBOV NP core/NPBP couple (Fig. 7C). In conclusion, both MARV NPBP and EBOV NPBP display the best binding to their cognate NP core proteins.

FIG 7.

ITC results of the cross-binding assays of NP core and NPBP from MARV and EBOV. (A) Results of the binding of EBOV NP core (residues 36 to 351) and EBOV NPBP (VP35 residues 20 to 48). (B) Results of the binding of EBOV NP core and MARV NPBP. (C) Results of the binding of MARV NP core and EBOV NPBP.

DISCUSSION

Previous studies demonstrated that NP, VP35, and L are necessary to support filovirus replication (5). Because L is not associated with the nucleocapsid without VP35, VP35 is regarded as the bridge between NP and L, providing a promising target for antiviral drugs (8). The mechanism by which VP35 binds to NP had been ambiguous, though researchers have shown that the C-terminal region of EBOV VP35 interacts with NP (21). Recently, the novel NPBP peptide in the N-terminal region of EBOV VP35 was shown to bind to NP with high affinity (18), and it is thought that the peptide plays an important role in the VP35-NP interaction. Otherwise, removing VP35 was thought to be necessary before NP oligomerization for RNA binding and nucleocapsid packaging (18). The mechanism to release the VP35 peptide from interaction with NP is still unknown, and future studies in this field are needed to shed light on the assembly mechanism of filovirus replication.

While these breakthroughs arose during research on EBOV, little was known about the NP-VP35 interaction in MARV. Sequence alignments suggested that the NP core regions are conserved between EBOV and MARV, but MARV has a truncated VP35 N terminus and a somewhat different NPBP sequence. A recent study of the VP35 oligomerization domain also suggested a distinction between the N-terminal regions of EBOV and MARV VP35s (23). Our work on MARV NP core and NPBP demonstrated that MARV VP35 still contains the high-affinity NP binding peptide at its N terminus, despite “missing” residues 1 to 19 of EBOV VP35 in MARV. Our crystal structure of the MARV NP core/NPBP complex revealed that residues in NP core and NPBP form a large hydrophobic binding surface and that the interaction is composed of charge complementarity, H bonds, and nonbonded contacts. How the extremely high-affinity binding of NPBP and NP core impacts the replication of virus is still a controversial issue. Based on their analysis of the affinity between EBOV NPBP and NP oligomers, Kirchdoerfer et al. suggest that VP35 can prevent premature NP oligomerization and RNA binding via NPBP binding, while an unknown mechanism separates NP and VP35 before viral RNA binding to NP at the proper site (18). Because the MARV NP core/NPBP complex retains a structure similar to that of the EBOV complex, the same process must also exist in MARV, indicating a universal mechanism for filoviruses.

Comparing the structures of the MARV NP core/NPBP complex and the EBOV NP core/NPBP complex, we found that the N and C termini of the NPBPs are different. The termini of NPBP include nonconserved residues, and they form different contacts with NP in complex. Biochemical assays demonstrated that the binding of NP core and NPBP across genera occurs, but the binding is significantly weaker than the binding between cognate NP core/NPBP pairs. The affinity of the NP core/NPBP interaction seems to be affected by the distinctions between the NPBP termini, suggesting that various intermolecular interactions have evolved in different filoviruses. However, the precise biological function of diverse N and C termini of the NPBPs in different filoviruses deserves further investigation. The existence of cross-species binding of NP core and NPBP may be due to a binding pocket in NP core. The pocket consists of residues from the C-terminal foot lobe of NP core that are conserved in MARV and EBOV, and these residues can form potent contacts with the middle portion of NPBP.

NP interactions have been successfully targeted by antiviral drugs for influenza virus (24, 25). Similarly, the binding of 18β-glycyrrhetinic acid, a compound isolated from crude extracts of Chinese licorice, was recently reported to trigger EBOV NP oligomerization and then disrupt viral RNA association (26), which may have some implications for our future work. For filoviruses, both NP-NP and NP-VP35 interactions are essential for efficient replication, and binding and subsequent separation of NP core and VP35 NPBP is hypothesized to play a central role in the replication process. Therapeutic drugs either blocking or stabilizing the interaction between NP core and NPBP may be effective in the treatment of filovirus infections. Our study supports the development of a universal drug against both EBOV and MARV by targeting the NPBP binding pocket in filovirus NP, and differences in binding details between these two viruses should be considered in drug design. This will provide prophylactic and therapeutic drugs for the control of possible future outbreaks (27, 28).

MATERIALS AND METHODS

Gene cloning, expression, and protein purification.

The gene of the Lake Victoria marburgvirus NP (GenBank accession no. ABE27040.1) was synthesized with Genewiz after codon optimization for Escherichia coli. Then, the NP core region (residues 18 to 344) was subcloned into the NdeI and XhoI sites of the pET-21a expression vector. The accuracy of the inserts was verified by sequencing.

The recombinant plasmid was transformed into E. coli strain BL21(DE3). The cells were cultured at 37°C in 2 liters of LB medium containing 100 μg/ml ampicillin and then transferred to 12°C when the optical density at 600 nm (OD₆₀₀) reached 0.6 and later induced for 18 to 20 h with 0.25 mM isopropyl-β-d-1-thiogalactopyranoside (IPTG). Cells were harvested by centrifugation at 7,000 × g for 15 min at 4°C, resuspended in lysis buffer (20 mM Tris-HCl and 500 mM NaCl, pH 8.0), and then homogenized with a low-temperature ultra-high-pressure cell disrupter (JNBIO, China). The lysate was clarified by centrifugation at 27,000 × g for 30 min at 4°C. The supernatant was loaded twice onto a Ni Sepharose (GE Healthcare) column preequilibrated with lysis buffer. The resin was washed with wash buffer (20 mM Tris-HCl, 500 mM NaCl, and 50 mM imidazole, pH 8.0) and eluted with elution buffer (20 mM Tris-HCl, 500 mM NaCl, and 300 mM imidazole, pH 8.0). The eluted protein was finally purified on a Superdex 75 10/300 gl (GE Healthcare) equilibrated with Tris-NaCl buffer. Fractions from the major peak were concentrated for later assays. The purified protein was >95% pure according to SDS-PAGE analysis and had an A₂₈₀/A₂₆₀ ratio of >1.6. The EBOV NP core (NCBI reference sequence NP_066243.1, residues 36 to 351) used in the ITC assays was also cloned, expressed, and purified using the same protocol.

The VP35 NPBP peptides (residues 1 to 28 of MARV VP35 [GenBank accession no. ADM72996.1]; residues 20 to 48 of EBOV VP35 [NCBI reference sequence NP_066244.1]) were synthesized by Scilight, China. Peptide powder (>99% pure) was dissolved in Tris-NaCl buffer and then purified on a Superdex 75 10/300 gl.

Crystallization.

NP core and NPBP for crystallization were, respectively, purified in buffers containing 20 mM Tris-HCl and 150 mM NaCl, pH 8.0. They were mixed at a molar ratio of 1:3, incubated on ice for 2 h, and then subjected to gel filtration to obtain the complex. Fractions from the survival peak were pooled and concentrated to 10 mg/ml for crystallization.

NP core and NPBP complexes were crystallized using sitting-drop vapor diffusion techniques with commercial crystallization kits (Hampton Research and Molecular Dimensions) at 18°C. Normally, 0.7 μl protein was mixed with 0.7 μl reservoir solution and then sealed to equilibrate against 90 μl reservoir solution at 4 or 18°C. Diffractable crystals appeared in 0.2 M sodium citrate tribasic monohydrate, pH 8.3, and 20% (wt/vol) polyethylene glycol (PEG) 3350 at 18°C. The crystals were cryoprotected in a reservoir solution containing 30% (wt/vol) PEG 3350, followed by flash-freezing in liquid nitrogen for X-ray data collection.

Data collection and structure determination.

X-ray diffraction data were collected under cryogenic conditions (100 K) at the Shanghai Synchrotron Radiation Facility (SSRF) beamline BL17U and indexed, integrated, and scaled with HKL2000 (29).

The MARV NP core/NPBP structure was solved by the molecular-replacement method using Phaser (30) from the CCP4 program suite (31), with the structure of EBOV NP core/NPBP complex (PDB ID 4YPI) as the search model (16). Initial restrained rigid-body refinement and manual model building were performed using REFMAC5 (32) and COOT (33), respectively. Further refinement was performed using Phenix (34). The final statistics for data collection and structure refinement are presented in Table 1.

ITC.

The NP core and NPBP proteins of MARV and EBOV for ITC assays were, respectively, prepared in a buffer containing 20 mM Tris-HCl and 500 mM NaCl, pH 8.0 (the NP core protein of EBOV was not stable in the low-salt buffer). Their binding affinities were measured with an ITC200 microcalorimeter (MicroCal/GE health). NP core was loaded into the ITC cell, and NPBP peptide was loaded into the syringe. Reactions were run by performing 20 injections of NPBP into the cell at the indicated temperatures. Data were processed using Origin (OriginLab).

Structure figure generation and analysis.

Structure figures were prepared using PyMOL (http://pymol.org/). Protein-protein interactions were analyzed using LigPlot⁺ (35). Topology diagrams were generated with TopDraw (36). Sequence alignment was performed using ClustalW (37) and prepared using ESPript 3.0 (38).

Accession number(s).

The coordinates and structure factors have been deposited in the Protein Data Bank with accession code 5XSQ.