Structure of the Fusion Core and Inhibition of Fusion by a Heptad Repeat Peptide Derived from the S Protein of Middle East Respiratory Syndrome Coronavirus

Cell lines.

293T cells (for pseudotyped virus generation) and human hepatoma cells (Huh7; for viral infection assay) were cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum. All cell lines were maintained at 37°C in humidified air containing 5% CO₂.

Gene construction.

Two constructs with variably truncated HR1 and HR2 sequences were made: one involves MERS-CoV spike (GenBank accession number JX869059) residues E992 to L1040 for HR1 and I1246 to L1286 for HR2, and the other covers amino acids E992 to I1054 for HR1 and L1252 to L1286 for HR2. The fusion core was then prepared as a single chain by linking the HR1 and HR2 domains via a 22-amino-acid linker (LVPRGSGGSGGSGGLEVLFQGP). This flexible linker has been shown to work successfully in the SARS-CoV fusion core (13). The coding fragments were synthesized by Generay Biotech Co., Ltd., and inserted into the NdeI and XhoI restriction sites of the pET-21a vector. The hexahistidine tag coding sequence and the stop codon in the pET-21a vector were used for both constructs.

The construct used for pseudovirus production was cloned by inserting the full-length coding sequence of MERS-CoV spike into the pCAGGS vector to yield a recombinant plasmid named pCAGGS-MERS-S. The construct was then verified by direct DNA sequencing.

Protein expression and purification.

For protein expression, expression vectors were transformed into Escherichia coli strain BL21(DE3) competent cells. A single colony was inoculated into 50 ml of Luria-Bertani (LB) medium containing 100 μg/ml of ampicillin and incubated overnight at 37°C. Then, the overnight culture was transferred to 2 liters of fresh LB medium for large-scale protein production by growing at 37°C. When the culture density (optical density at 600 nm [OD₆₀₀]) reached 0.6, protein overexpression was induced with 0.2 mM isopropyl-β-d-thiogalactoside (IPTG), and the cells were grown for an additional 10 h at 16°C before harvesting via centrifugation.

The collected bacterial cell pellet was resuspended in phosphate-buffered saline (PBS) and homogenized by sonication. The suspension was then centrifuged at 12,000 rpm for 15 min at 4°C. The supernatant was collected and then loaded onto a nickel-nitrilotriacetic acid (Ni-NTA) column (Qiagen). After removal of impurities by thorough washing using PBS, the target protein was eluted with a buffer of 300 mM imidazole in PBS and then purified by gel filtration using a Superdex 200 10/300 GL column (GE) running on an Äkta Explorer fast-performance liquid chromatography (FPLC) system in a buffer composed of 20 mM Tris-HCl, pH 8.0, and 100 mM NaCl. The protein fractions were collected and analyzed on a 15% tricine SDS-PAGE gel. The molecular weights of the peak fractions were estimated by comparison with protein standards run on the same column.

Crystallization, data collection, and structure determination.

The purified protein was concentrated to 5 mg/ml. Crystals of good diffracting quality were obtained after 14 days using the sitting drop vapor diffusion method by equilibrating a 2-μl drop (protein solution mixed 1:1 with reservoir solution) against a 100-μl reservoir containing 0.1 M bis-tris, pH 6.5, and 25% (wt/vol) polyethylene glycol 3350. For data collection, a single crystal was picked up using a nylon loop and immersed in a cryoprotectant solution containing 67% reservoir solution and 16% glycerol for ∼30 s. The crystal was then remounted and flash-cooled in liquid nitrogen. The diffraction data were collected at 100 K at beamline BL17U of the Shanghai Synchrotron Radiation Facility (SSRF). Raw data were processed using HKL2000 (21).

The structure of the MERS-CoV HR1/HR2 complex was determined by molecular replacement with Phaser (22) using the structure of the SARS-CoV fusion core (Protein Data Bank [PDB] code, 1WNC) as the search model. The initial model was first refined with Refmac5 in the CCP4 suite (23) using rigid-body refinement and maximum likelihood procedures. The model was then completed by iterative cycles of manual rebuilding in coot (24) and refinement with Phenix.refine (25). During the course of model building and refinement, PROCHECK (26) was used to monitor the stereochemistry of the structure. The detailed statistics are summarized in Table 1. All of the structural figures were generated using PyMOL (http://www.pymol.org/).

CD spectroscopy.

Circular dichroism (CD) spectra were measured on a ChiraScan CD spectrometer (AppliedPhotophysics). Freshly prepared MERS-CoV HR1/HR2 complex proteins were adjusted to 0.2 mg/ml in PBS before data collection. Wavelength spectra were recorded at 20°C using a 0.1-cm-path-length cuvette. Each scan, in the range of 195 to 260 nm, was obtained by taking data points every 0.5 nm with a bandwidth of 2 nm.

Pseudovirus preparation and titration.

The MERS pseudovirus was prepared by cotransfecting 293T cells with a plasmid encoding an Env-defective, luciferase-expressing HIV-1 genome (pNL4-3.luc.RE) (27) and the pCAGGS-MERS-S expression plasmid using Lipofectamine 2000 (Invitrogen). The pseudovirus-containing supernatant was harvested 48 h following transfection, clarified by centrifugation, and then filtered through a 0.45-μm sterilized membrane. Single-use aliquots (1.0 ml) were stored at −80°C. The 50% tissue culture infectious dose (TCID₅₀) for each pseudovirus preparation was determined by infection of Huh7 cells as previously described (28).

Peptide synthesis.

The peptides used in this study were synthesized by ChinaPeptides Co., Ltd., with a purity of >95%. Each peptide was dissolved in 100% dimethyl sulfoxide (DMSO) at 100 mM and stored at −80°C before use. One peptide (denoted P1; LTQINTTLLDLTYEMLSLQQVVKALNESYIDLKEL) covers the full length of the HR2 sequence, while the other one (denoted P2; LTYEMLSLQQVVKALNESYIDLKELGN) removes those residues in HR2 adopting extended conformations and includes two extra amino acids at the C terminus because the equivalent peptide in SARS-CoV exhibits a much higher inhibitory efficacy than the HR2 peptides of longer length (29). We also included the SARS-peptide (29) (denoted SARS-pep; IQKEIDRLNEVAKNLNESLIDLQELGK) in the pseudovirus entry inhibition assay as a specificity control.

Pseudovirus entry inhibition assay.

For the inhibition assay, 100 TCID₅₀s of each pseudovirus was incubated with 10-fold serially diluted peptides from 0.01 nM to 100 μM at 37°C for 30 min. The virus-peptide mixture was then transferred to 96-well plates seeded with Huh7 cells. Each concentration was tested in octuplicate. After a 5-h incubation, the medium was replaced, and the sample was incubated for an additional 48 h at 37°C. The cells were then collected, lysed, and measured for luciferase activity using a GloMax 96 Microplate luminometer (Promega). The 50% effective concentration (EC₅₀) and 95% confidence interval values were calculated using Prism.

Protein structure accession number.

The atomic coordinates and the related structure factors have been deposited in the Protein Data Bank with the accession code of 4MOD.

Characterization of the binding between MERS-CoV HR1 and HR2.

The HR regions exhibit characteristic sequence features with normally hydrophobic residues at the “a” and “d” positions. With this criterion, we first positioned the HR sequences of MERS-CoV to amino acids Y978 to E1062 for the HR1 region and amino acids E1234 to Y1298 for the HR2 region, using the successful SARS-CoV sequence as a reference (13, 20). Both regions are located downstream of a predicted S1/S2 cleavage site between R751 and S752, which can potentially be recognized by the host furin protease (Fig. 1A). We then refined the boundaries of each HR unit by either sequence homology comparison with SARS-CoV HRs or bioinformatics analysis using the LearnCoil-VMF program (30). Despite the overall sequence identity between the MERS-CoV and SARS-CoV S proteins being <28%, we noted that the two proteins exhibit >50% identities for the above-defined HR regions (Fig. 1B). It therefore seemed feasible to use the construction strategies for SARS-CoV HRs as a reference to characterize the MERS-CoV fusion core. Multiple studies have been performed to characterize the fusion core structure of SARS-CoV (13, 20, 31), and one of these with truncations from E900 to L948 for HR1 and I1145 to L1185 for HR2 was successful (13). The corresponding HR regions in MERS-CoV S2 are E992 to L1040 and I1246 to L1286, respectively. Alternatively, we also made truncations based on a LearnCoil-VMF prediction result, selecting E992 to I1054 for HR1 and L1252 to L1286 for HR2 (Fig. 1B). In each case, HR1 and HR2 were connected via a 22-residue linker (sequence, LVPRGSGGSGGSGGLEVLFQGP), which has been shown to work successfully with the HR domains of SARS-CoV (13).

The recombinant proteins, designated HR complex 1 (based on the sequence alignment results) and HR complex 2 (based on the LearnCoil prediction results), were then expressed in E. coli cells, purified, and analyzed by an analytical gel filtration assay. Both proteins were eluted at approximately 15.5 ml on a calibrated Superdex 200 10/300 column, corresponding to a molecular mass of ∼36 kDa. As the single HR1-linker-HR2 chain is only ∼12 kDa, both HR complex 1 and HR complex 2 therefore assembled into a trimer in solution, as expected (Fig. 2A and C). We further tested the proteins by CD spectroscopy. Double minima at 208 and 222 nm were observed for both HR complexes 1 and 2 (Fig. 2B and D), demonstrating that the two proteins folded into typical α-helical structures.

Structure of the MERS-CoV fusion core.

As both HR complexes 1 and 2 appeared to be correctly folded, both proteins were subjected to intensified crystal screening trials, resulting in X-ray-diffractable crystals for HR complex 2. Its structure was determined by molecular replacement using the SARS-CoV fusion core structure (PDB code, 1WNC) as the search model and was refined to an R_work of 0.1847 and an R_free of 0.2135 (Table 1). Within the asymmetric unit, two HR1-linker-HR2 chains, related by a 2-fold axis, are present. The two chains are essentially the same structure, superimposition of which yielded a root mean square deviation (RMSD) of only 0.048 Å for all Cα pairs. Clear electron densities could be traced for residues N993 to E1039 in HR1 and Q1254 to D1282 in HR2. HR1 folds into a 12-turn α-helix, whereas HR2 adopts a much more extended conformation for the first half and a canonical α-helical structure for the second half. Overall, HR2 folds back and aligns with HR1 in an antiparallel manner, forming a helical hairpin and causing the chain N and C termini to end at the same side (Fig. 3A).

By simple symmetry operations, each chain in the crystallographic asymmetric unit could yield a trimeric structure of a typical coiled-coil fold. In this trimer of hairpins, three HR1 helices aligned in parallel with each other to form the central coiled-coil core, whereas three HR2 polypeptides bind to the side grooves of the HR1 core. This leads to a 6-helix bundle of approximately 13 Å in semidiameter and approximately 68 Å in height (Fig. 3B).

Overall, the fusion core structure of MERS-CoV is very similar to those reported for other coronaviruses, such as SARS-CoV (13) and MHV (15). Superimposition of these three structures revealed an RMSD of ∼0.905 Å for 204 Cα atoms between MERS-CoV and SARS-CoV (Fig. 3C) and an RMSD of ∼0.575 Å for 204 Cα atoms between MERS-CoV and MHV (Fig. 3D). All of the helical elements and a majority of the extended HR2 loops could be well aligned. Only some terminal residues exhibit conformational variance. In addition, the MHV fusion core is longer than that of MERS-CoV by about two helix turns (Fig. 3D).

Atomic details of the fusion core formation.

We further characterized the atomic details mediating the formation of the hairpin trimer. Consistent with the HR features, the hydrophobic residues at positions “a” and “d” in HR1 are aligned on one side of the helix, forming an interface of strong hydrophobicity. Three HR1 helices thereby pack against each other and stack the hydrophobic helical laterals in the center of the coiled-coil core. The major stacking forces are contributed by the side chains of residues F1012, F1019, V1026, and L1033 at position “a” and residues F1001, M1008, T1015, V1022, and L1036 at position “d.” They are not evenly allocated along the helix (Fig. 4A). A majority of these contacting amino acids are in the middle portion of the helical structure, while the remaining residues clinch the coil at both ends. We also observed clear alternative conformations for residue M1008, which we believe could increase the buried surface area among HR1 chains. It is also noteworthy that no polar contacts, such as H bonds or salt bridges, are observed in this coiled-coil formation.

HR2 binds to the side grooves of the center core and simultaneously contacts two HR1 chains. Along the groove, HR2 first folds into an extended loop and then proceeds to form an α-helix of about five turns. Much more extensive interactions than those for HR1 assembly are observed, involving not only the “a” and “d” amino acids but also residues at other positions. In total, 17 amino acids, as listed in Fig. 4B, were identified to be important for the HR1/HR2 interaction. In addition to hydrophobic contacts contributed by the apolar and bulky side chains, about two-thirds (11 out 17) of the interface residues also H-bond with HR1 via their main-chain/side chain amino or oxygen groups (Fig. 4B). This results in a strong engagement network, tightly tying HR2 to HR1.

Concerning the hydrophobic groove formed by HR1 helices, the N-terminal half is relatively deep, accommodating the C-terminal helix of HR2. Three pockets were identified in this deep groove, which are occupied by three “d” residues (L1262, L1269, and L1276) after HR2 engagement. In contrast, the C-terminal half of the groove is shallow and accommodates the HR2 extended N-terminal loop (Fig. 4B). Therefore, in this fusion core structure, the HR1 and HR2 polypeptides are highly complementary in both shape and chemical properties.

In vitro inhibition of MERS-CoV infection by an HR2-based peptide.

According to previous studies, HR2-based peptides normally display a higher inhibitory efficacy than those derived from HR1 (14, 16). Therefore, we synthesized two peptides based on the HR2 sequence and tested their inhibitory effects in the pseudotyped virus system. Gradient concentrations of each peptide were incubated with the pseudotyped viruses, and the percentage of virus entry inhibition in cultured Huh7 cells was determined. As shown in Fig. 5, an obvious inhibitory effect was observed for peptide P1 but not for peptide P2 or the SARS peptide, even at the highest concentration (100 μM), indicating the inhibitory specificity of P1. The 50% effective dose (EC₅₀) for inhibition of MERS-CoV infection by P1 was calculated to be ∼3.013 μM (95% confidence interval, 1.815 to 5.001 μM).