Crystal structure of the receptor binding domain of the spike glycoprotein of human betacoronavirus HKU1

Abstract

Human coronavirus (CoV) HKU1 is a pathogen causing acute respiratory illnesses and so far little is known about its biology. HKU1 virus uses its S1 subunit C-terminal domain (CTD) and not the N-terminal domain like other lineage A β-CoVs to bind to its yet unknown human receptor. Here we present the crystal structure of HKU1 CTD at 1.9 Å resolution. The structure consists of three subdomains: core, insertion and subdomain-1 (SD-1). While the structure of the core and SD-1 subdomains of HKU1 are highly similar to those of other β-CoVs, the insertion subdomain adopts a novel fold, which is largely invisible in the cryo-EM structure of the HKU1 S trimer. We identify five residues in the insertion subdomain that are critical for binding of neutralizing antibodies and two residues essential for receptor binding. Our study contributes to a better understanding of entry, immunity and evolution of CoV S proteins.

Figure 1. Inhibition of HKU1 virus infection by 1A-S310-677aa.

(a) Schematic diagram of HKU1 S protein, modified from Kirchdoerfer et al. NTD, N-terminal domain; CTD, C-terminal domain; SD-1, subdomain-1; SD-2, subdomain-2; RBD, receptor-binding domain; FP, fusion peptide; HR-N, N-terminal heptad repeat; HR-C, C-terminal heptad repeat; TMD, transmembrane domain; cleavage site, furin cleavage site between S1 and S2. The amino acid numbers are from the S protein of genotype A HKU1 virus. (b) ELISA was performed using purified 1A-S310-677aa and purified antibodies. The experiments were done in triplicate and the s.d.‘s (n=3) were shown as error bar. The experiments were performed twice and one representative is shown. Differentiated HTBE cells were incubated with 20 μM of 1A-310-677aa protein at 37 °C for 1 h. HKU1 viruses were diluted into the same amount of proteins and added onto HTBE cells for 4 h. Forty-eight hours after inoculation, cells were fixed and stained with polyclonal rabbit anti HKU1 S antibodies 1814 at a dilution at 1:100 (c), and released viruses at 24 h post inoculation from apical wash were analysed using real-time PCR (d). The amount of HKU1 viral RNA from BSA control was set as 100%. Scale bar, 50 μm. The experiments were done in triplicate and the s.d.‘s (n=3) were shown as error bar. The experiments were performed twice and one representative is shown.

Figure 2. Crystal structure of 1A-310-677aa.

(a, top) Ribbon model of 1A-310-677aa; bottom: schematic diagram of 1A-310-677aa. The core subdomain is shown in blue, the insertion loop is shown in magenta, and SD-1 subdomain is shown in green. Twelve pairs of disulfite bridges are numbered and shown in yellow and stick model. The amino acid numbers are from the S protein of the genotype A HKU1 virus. (b) Schematic illustration of the topological graph of 1A- 310-677aa. β-strands are shown with brown arrows, α-helices are shown with bright blue cylinders.

Figure 3. Structural comparison of CTDs of different β-CoV S proteins.

The core subdomains are blue, and the insertion loops are magenta. (a) HKU1 CTD; (b) MHV CTD (PDB: 3JCL); (c) (PDB: 2AJF); (d) MERS-CoV RBD (PDB: 4L72); (e) HKU4 RBD (PDB: 4QZV) and (f) HKU9 RBD (PDB: 5GYQ).

Figure 4. Integrated model of HKU1 S protein trimer with the crystal structure of 1A-S310-677aa.

(a) The trimeric model of HKU1 S protein with the crystal structure of 1A-S310-677aa. (left) side view; (right) top view. The model is built by superimposing the crystal structure of 1A-S310-677aa into the Cryo-EM structure of the HKU1 S protein trimer based on the structure alignment of the core subdomain. The cryo-EM structure of the trimeric HKU1 S protein is shown in grey. The three crystal structures of 1A-S310-677aa are labelled in blue, magenta and green, respectively. The red spheres are sugar moieties found in the crystal structure of 1A-S310-677aa. (b) Superimposition of 1A-S310-677aa onto the cryo-EM density of the HKU1 S protein. The crystal structure of 1A-S310-677aa is shown in blue and the cryo-EM density map is shown in orange. N-linked glycan moieties are shown in red stick.

Figure 5. Epitope mapping for neutralizing mHKUS-2 and mHKUS-3 antibodies in 1A-S310-677aa.

(a, left) Structure of 1A-S310-677aa; middle, schematic diagram of different 1A/B CTD chimeras; right, binding of different mHKUS antibodies to 1A/B chimeric proteins measured by ELISA. +++, positive signal with antibody concentration <10 ng ml⁻¹ ++, positive signal with antibody concentration 10–100 ng ml⁻¹; +, positive signal with antibody concentration 100–1,000 ng ml⁻¹; +/−, positive signal with antibody concentration >1,000 ng ml⁻¹; -, no signal with antibody concentration at 10 μg ml⁻¹. The amino acid residue numbers are from S protein of genotype A HKU1 virus. Experiments were done twice. (b) ELISAs were performed using purified proteins and purified antibodies. The experiments were done in triplicate and the s.d.‘s (n=3) were shown as error bar. The experiments were performed twice and one representative is shown.

Figure 6. Inhibition of HKU1 virus entry by mutant 1A-S310-677aa proteins.

(a) The purified mutant proteins were assayed for their affinity to monoclonal antibodies mHKUS-2, mHKUS-4, and mHKUS-6 by ELISA. Differentiated HTBE cells were incubated with 20 μM of 1A-S310-677aa, 1B-S310-676aaor mutant CTD proteins at 37 °C for 1 h. HKU1 viruses were diluted into the same amount of proteins and added onto the HTBE cells for 4 h. The released viruses from apical wash at 24 h post inoculation were analysed using real-time PCR (b), and cells were fixed and stained with polyclonal rabbit anti HKU1 S antibodies 1814 at a dilution of 1:100 at 48 h post inoculation (c). Scale bar, 50 μm. The amount of HKU1 viral RNA in the BSA control was set as 100%. The experiments were done in triplicate, and the s.d.‘s (n=3) were shown as error bar. The experiments were performed twice and one representative is shown.