Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2017 Aug 29;114(35):E7348-E7357.
doi: 10.1073/pnas.1707304114. Epub 2017 Aug 14.

Immunogenicity and Structures of a Rationally Designed Prefusion MERS-CoV Spike Antigen

Affiliations
Free PMC article

Immunogenicity and Structures of a Rationally Designed Prefusion MERS-CoV Spike Antigen

Jesper Pallesen et al. Proc Natl Acad Sci U S A. .
Free PMC article

Abstract

Middle East respiratory syndrome coronavirus (MERS-CoV) is a lineage C betacoronavirus that since its emergence in 2012 has caused outbreaks in human populations with case-fatality rates of ∼36%. As in other coronaviruses, the spike (S) glycoprotein of MERS-CoV mediates receptor recognition and membrane fusion and is the primary target of the humoral immune response during infection. Here we use structure-based design to develop a generalizable strategy for retaining coronavirus S proteins in the antigenically optimal prefusion conformation and demonstrate that our engineered immunogen is able to elicit high neutralizing antibody titers against MERS-CoV. We also determined high-resolution structures of the trimeric MERS-CoV S ectodomain in complex with G4, a stem-directed neutralizing antibody. The structures reveal that G4 recognizes a glycosylated loop that is variable among coronaviruses and they define four conformational states of the trimer wherein each receptor-binding domain is either tightly packed at the membrane-distal apex or rotated into a receptor-accessible conformation. Our studies suggest a potential mechanism for fusion initiation through sequential receptor-binding events and provide a foundation for the structure-based design of coronavirus vaccines.

Keywords: X-ray crystallography; coronavirus; cryo-EM; neutralizing antibody; peplomer.

Conflict of interest statement

Conflict of interest statement: J.P., N.W., K.S.C., R.N.K., H.L.T., C.A.C., B.S.G., A.B.W., and J.S.M. are inventors on US patent application no. 62/412,703, entitled “Prefusion Coronavirus Spike Proteins and Their Use.” L.W., W.S., W.-P.K., and B.S.G. are inventors on US patent application no. PCT/US2016/019395, entitled “Middle East Respiratory Syndrome Coronavirus Immunogens, Antibodies and Their Use.”

Figures

Fig. 1.
Fig. 1.
Structure-based engineering of MERS-CoV and SARS-CoV S proteins. (A) Domain architecture of the HCoV-HKU1 S protein and sequence alignment of the helix-turn-helix between heptad repeat 1 (HR1) and the central helix (CH). The two residues colored red are those mutated to proline to retain S2 in the prefusion conformation. FP, fusion peptide; HR2, heptad repeat 2; TM, transmembrane domain. (B) Structure of HCoV-HKU1 S2. Residues shown in sticks in magnified region are those mutated to proline in the 2P variants. (C) Gel-filtration profiles of WT (dashed lines) and 2P-engineered (solid lines) S protein ectodomains from MERS-CoV (blue) and SARS-CoV (red). Each protein was produced from a 1-L transient transfection. All four proteins were expressed with a C-terminal T4 fibritin trimerization domain. The S1/S2 furin site was mutated in MERS S-WT and MERS S-2P. (D) Two-dimensional class averages of negative stained MERS S-WT, MERS S-2P, SARS S-WT, and SARS S-2P. All particles are included. For WT constructs both the prefusion (blue boxes) and postfusion (red boxes) conformations are visible, whereas for the 2P mutants only the prefusion conformation is observed.
Fig. S1.
Fig. S1.
Proline substitutions in S2 increase expression levels of coronavirus S proteins. (A and B) Expression levels of WT and proline-substituted S proteins from MERS-CoV (A) and SARS-CoV and HCoV-HKU1 (B) as assessed by SDS/PAGE. (C) Gel-filtration profiles of WT HCoV-HKU1 S (HKU1 S-WT) and stabilized HCoV-HKU1 S (HKU1 S-2P). Each protein was produced from 500 mL FreeStyle 293-F cells.
Fig. 2.
Fig. 2.
Characterization of MERS S-2P. (A) MERS-CoV pseudoviruses encoding a luciferase reporter gene were generated with WT (S WT, blue) or 2P (S-2P, red) S proteins. Mock pseudoviruses (gray), expressing no S protein, served as a control. Infectivity in Huh7.5 cells was determined by measuring RLU. The dotted line represents background RLU. (B) Binding of cell-surface expressed MERS-CoV WT and 2P S proteins, as well as membrane-tethered RBD, to polyclonal sera (Poly) and monoclonal antibodies measured by flow cytometry; 101F is an RSV F-specific antibody. (C) SDS/PAGE analysis of copurified complexes of untagged MERS S-2P and monoclonal antibodies. AM14 is an RSV F-specific antibody. (D) Binding of soluble DPP4 to immobilized MERS S-2P measured by surface plasmon resonance. Best fit of the data to a 1:1 binding model is shown in red.
Fig. 3.
Fig. 3.
Immunogenicity of MERS S-2P in mice. (A) Reciprocal serum IC90 neutralizing activity against autologous MERS England1 pseudotyped lentivirus reporter plotted against vaccine dose. (B) Reciprocal serum IC90 neutralizing activity against multiple homologous MERS-CoV pseudoviruses of sera from mice immunized with 0.1 μg of protein. For both panels, the geometric mean IC90 titer (GMT) of each group is represented by (A) symbols or (B) bars. Error bars represent geometric SDs. P values denoted as *P < 0.05 and **P < 0.01. The limit of detection for the assay is represented by dotted lines; for sera below the limit of detection a reciprocal IC90 titer of 10 was assigned.
Fig. 4.
Fig. 4.
Structure of MERS-CoV S-2P in complex with G4 Fab. (A) Structure of MERS S-2P ectodomain in complex with G4 Fab as viewed along (Left) and above (Right) the viral membrane. A single protomer of the trimeric S protein is shown in ribbon representation and colored as in the primary structure diagram. The two remaining protomers are shown as molecular surfaces and colored white and gray. CD, connector domain; CH, central helix; Fd, Foldon trimerization domain; SD-1, subdomain 1; SD-2, subdomain 2. (B and C) Magnified view of the S1/S2 (B) and S2′ (C) protease sites. Dashed lines represent disordered residues. Arrows indicate position of protease cleavage.
Fig. S2.
Fig. S2.
Cryo-EM data processing. Data were sorted and refined according to the flowchart. In particular, to resolve the heterogeneity in RBD configurations in our dataset we used a method of data sorting that involves subtraction of constant density regions from our raw data followed by local classification of the resulting projection images.
Fig. S3.
Fig. S3.
Crystal structure of MERS-CoV S1 NTD. (A) The structure of the MERS-CoV S1 NTD can be separated into top, core and bottom regions (86). The two β-sheets of the NTD core are colored green and magenta, and the rest of the NTD is colored cyan. (B) The NTD structure is colored as a rainbow from blue to red, N to C terminus, respectively. (C) Extracted ion chromatogram (extracted theoretical mass 442.15 at 5 ppm), MS, and MS/MS spectra of ligand isolated from purified MERS-CoV S NTD. Concentrated NTD was heated at 98 °C for 10 min, and the denatured protein was pelleted via centrifugation. The supernatant was desalted and analyzed as previously described (87) on a Q Exactive Plus mass spectrometer with the modification that +1 ions were also selected for fragmentation. Data analysis was performed in QualBrowser. (D) Extracted ion chromatogram (extracted theoretical mass 442.15 at 5 ppm), MS, and MS/MS spectra of purified folic acid (Sigma) resuspended in 5% methanol and 1% formic acid and analyzed as described above. (E) The chemical structure of folic acid. (F) Crystal structure of MERS-CoV S NTD shown as a ribbon with the Fo-Fc map before placement and refinement of folic acid shown at 2.0 sigma (gray mesh). Folic acid (shown in sticks) was modeled into this density and used for further structural refinement. Oxygen atoms are red and nitrogen atoms are blue.
Fig. 5.
Fig. 5.
G4 recognizes a variable loop in the S2 connector domain. (A) Structure of G4 Fab bound to a variable loop contained within the S2 subunit. Residues 1171–1187 of MERS S-2P are shown as a ribbon, with the side chain of Asn1176 and two attached N-acetylglucosamine moieties shown as sticks. The variable domains of G4 are shown as a molecular surface. (B) G4 binding interface. Side chains of interacting residues are shown as sticks, with residues substituted in G4-escape variants colored orange. Black dotted line indicates a salt bridge. (C) Sequence alignment of MERS-CoV isolates (green) and other lineage C betacoronaviruses (tan). Bold font indicates N-linked glycosylation sites. (D) Side views of one S2 protomer bound to G4 Fab. On the right, S2 is shown as a molecular surface and colored according to sequence conservation as determined by the ConSurf server using 66 diverse coronavirus sequences (85).
Fig. S4.
Fig. S4.
G4 Fab occupancy. (A and B) The MERS S-2P/G4 data contained two stoichiometric G4 classes. A majority of the data (69.6%) contained two G4 Fabs bound to the MERS S-2P spike (A), and the remaining data (30.4%) contained three bound G4 Fabs (B). (C and D) Stereo images of the G4 epitope. Epitopes that have G4 Fabs bound (C) are well-ordered whereas epitopes that remain unbound (D) exhibit a disordered loop.
Fig. S5.
Fig. S5.
G4 Fab affinity as measured by isothermal titration calorimetry. (AC) ITC data for the binding of G4 Fab to MERS S-2P (A), deglycosylated MERS S-2P (B), and variable-loop-mutated MERS S-2P (C). Red lines represent the best fit of the data to a single-binding-site model. The sequence and secondary structure schematics for the variable loop are shown above the titrations.
Fig. 6.
Fig. 6.
RBD conformations observed in the MERS-CoV S protein. (A) Top and side views of the four MERS S-2P structures determined by single particle cryo-EM. Each has a unique arrangement of the three RBDs (green). The percentage of particles in the dataset belonging to each of the four structures is shown below the structures. (B) Interaction between an RBD (green) and the S2 helix–loop–helix spanning the central helix (orange) and HR1 (yellow). The two prolines introduced into S2 are shown as sticks, as are the side chains of interacting residues in the RBD and HR1. Electron density is shown as a transparent surface. (C) Superposition of one protomer with the RBD “in” and another protomer with the RBD “out.” (D) Superposition of the RBD-DPP4 crystal structure (PDB ID code 4KR0) onto trimers with three RBDs in (Left) or two RBDs in and one RBD out (Right). Substantial clashes prevent DPP4 from binding until the RBD rotates outward.
Fig. S6.
Fig. S6.
MERS-CoV S RBD conformations. (A) MERS-CoV S assumes both open (DPP4-accessible) and closed (DPP4-inaccessible) conformations. A small fraction of the data (5.4%) exhibited the closed conformation characterized by having all three RBDs “in.” The majority of our data (94.6%) exists in an open conformation characterized by having at least one RBD “out.” Only a very small fraction of the data (0.3%) has all three RBDs “out.” (B) Density corresponding to the RBD is colored green and shown “in.” (C) Density corresponding to the RBD is colored green and shown “out.” (D) Superposition of one protomer with the RBD “in” and another protomer with the RBD “out.”
Fig. 7.
Fig. 7.
Simplified model of DPP4 binding leading to MERS-CoV S triggering. The model posits that all three RBDs are in a state of equilibrium between the receptor-accessible “out” conformation and the tightly packed, receptor-inaccessible “in” conformation. DPP4 binding acts as a molecular ratchet that locks the RBD in the “out” conformation until all three RBDs are bound. This open conformation of the trimer is unstable and the S1 subunits ultimately dissociate from S2. Once the S2 subunits are no longer constrained by S1, membrane fusion can proceed by way of a prehairpin intermediate. MERS-CoV S protomers are colored pink, blue, and green and the unresolved HR2 region is depicted as a dashed line. Dimeric DPP4 is colored orange.

Similar articles

See all similar articles

Cited by 68 articles

See all "Cited by" articles

Publication types

MeSH terms

LinkOut - more resources

Feedback