Impact of temperature on the affinity of SARS-CoV-2 Spike glycoprotein for host ACE2

Abstract

The seasonal nature of outbreaks of respiratory viral infections with increased transmission during low temperatures has been well established. Accordingly, temperature has been suggested to play a role on the viability and transmissibility of SARS-CoV-2, the virus responsible for the COVID-19 pandemic. The receptor-binding domain (RBD) of the Spike glycoprotein is known to bind to its host receptor angiotensin-converting enzyme 2 (ACE2) to initiate viral fusion. Using biochemical, biophysical, and functional assays to dissect the effect of temperature on the receptor-Spike interaction, we observed a significant and stepwise increase in RBD-ACE2 affinity at low temperatures, resulting in slower dissociation kinetics. This translated into enhanced interaction of the full Spike glycoprotein with the ACE2 receptor and higher viral attachment at low temperatures. Interestingly, the RBD N501Y mutation, present in emerging variants of concern (VOCs) that are fueling the pandemic worldwide (including the B.1.1.7 (α) lineage), bypassed this requirement. This data suggests that the acquisition of N501Y reflects an adaptation to warmer climates, a hypothesis that remains to be tested.

Keywords: ACE2; COVID-19; N501Y; RBD; SARS-CoV-2; Spike glycoproteins; coronavirus; neutralization; temperature; variants of concern.

Figure 1

Enhanced binding of ACE2 to SARS-CoV-2Spike at low temperatures.A and B, LOGO depictions of the frequency of SARS-CoV-2 Spike residues known to (A) contact with ACE2 or (B) corresponding to B.1.1.7 defining mutations. Worldwide sequences deposited in the GISAID database in 2019 to 2020 and in 2021 (up to June 18th, 2021) were aligned using the COVID CoV Genetics program. The height of the letter indicates its frequency over total sequences. Residues corresponding to the WIV04 reference sequence are shown in black and residues corresponding to VOCs are shown in violet. A box with a cross inside (☒) indicates the presence of a residue deletion. C–E, cell-surface staining of transfected 293T cells expressing SARS-CoV-2 Spike (WT, D614G, Furin KO, D614G Furin KO, D614G N501Y, or B.1.1.7 variant) or SARS-CoV-1 Spike (WT) using (C) CV3-25 mAb or (D and E) ACE2-Fc. F, cell-surface staining of Vero E6 or primary human AECs from two healthy donors infected with authentic SARS-CoV-2 viruses (D614G or B.1.1.7 variant) using ACE2-Fc. C–F, the graphs shown represent the binding of primary antibodies performed at (C and D) 37 °C, 22 °C, and 4 °C or (E and F) at 37 °C and 4 °C. ACE2-Fc binding was normalized to CV3-25 binding in each experiment at each temperature. The graphs shown represent the median fluorescence intensities (MFI). Error bars indicate means ± SEM. These results were obtained in at least three independent experiments. Statistical significance was tested using (C–E) one-way ANOVA with a Holm–Sidak posttest or (F) a paired t test (∗p < 0.05; ∗∗∗p < 0.001; ∗∗∗∗p < 0.0001; ns, nonsignificant).

Figure 2

SARS-CoV-2 viral attachment and infectivity is higher at low temperatures.A, pseudoviruses encoding the luciferase gene (Luc+) and bearing SARS-CoV-2 Spike (D614G or D614G N501Y) were tested for virus capture by ACE2-Fc at 37 °C or 4 °C. RLU obtained using ACE2-Fc was normalized to the signal obtained with the CV3-25 mAb. B, pseudoviruses Luc+ bearing SARS-CoV-2 Spike (D614G, D614G N501Y or B.1.1.7), or VSV-G as a control, were used to infect 293T-ACE2 cells. Virions were incubated at 37 °C, 22 °C, or 4 °C for 1 h prior infection of 293T-ACE2 cells for 48 h at 37 °C. C, cell-to-cell fusion was measured between 293T effector cells expressing HIV-1 Tat and SARS-CoV-2 Spike (D614G or B.1.1.7), or HIV-1 Env_JRFL as a control, which were incubated at 37 °C or 4 °C for 1 h prior coculture with TZM-bl-ACE2 target cells. B and C, RLUs were normalized to the signal obtained with cells preincubated at 37 °C. D and E, pseudoviruses Luc+ bearing SARS-CoV-2 Spike (WT, D614G or B.1.1.7) were used to infect 293T-ACE2 cells in presence of increasing concentrations of sACE2 at 37 °C for 1 h prior infection of 293T-ACE2 cells. Fitted curves and IC₅₀ values were determined using a normalized nonlinear regression. F, authentic SARS-CoV-2 D614G virus was used to infect reconstituted human airway epithelia. Viral attachment was performed at 37 °C or 4 °C and cells were further cultured at 37 °C for 96 h. Viral titers (RNA copies/ml) were monitored at 24 h and 96 h postinfection using one-step qRT-PCR. Viral titer values were normalized to the signal obtained with virions adsorbed to the cells at 37 °C. Error bars indicate means ± SEM. These results were obtained in at least three independent experiments. Statistical significance was tested using (A, C and F) an unpaired t test or (B) one-way ANOVA with a Holm–Sidak post-test (∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001; ∗∗∗∗p < 0.0001; ns, nonsignificant).

Figure 3

Enhanced affinity of SARS-CoV-2 RBD for ACE2 at low temperatures.A, the thermodynamic parameters of sACE2 binding to SARS-CoV-2 RBD WT or N501Y measured by ITC at 10 °C, 15 °C, 25 °C, and 35 °C. The graphs shown represent the affinity (K_D), enthalpy (ΔH), and entropy (−TΔS) values obtained at the different temperatures. All ITC titration curves and thermodynamics values are shown in Fig. S3. B and C, binding kinetics between SARS-CoV-2 RBD (WT or N501Y) and sACE2 assessed by BLI at different temperatures. B, biosensors loaded with RBD proteins were soaked in twofold dilution series of sACE2 (500 nM–31.25 nM) at different temperatures (10 °C, 15 °C, 25 °C, or 35 °C). Raw data are shown in blue and fitting model is shown in red. C, graphs represent the affinity constants (K_D), on rates (K_on) and off rates (K_off) values obtained at different temperatures and calculated using a 1:1 binding model. All BLI data are summarized in Table S1.

Figure 4

SARS-CoV-2Spike trimer “opens” at low temperatures.A and B, binding of sACE2 to SARS-CoV-2 Spike (A) D614G or (B) B.1.1.7 expressed on 293T cells was measured at 37 °C or 4 °C by flow cytometry. Cells were preincubated with increasing amounts of sACE2 and its binding was detected using an anti-ACE2 staining. The Hill coefficients were determined using GraphPad software. C, cell-surface staining of transfected 293T cells expressing SARS-CoV-2 Spike (WT or Furin KO) using the CR3022 mAb when performed at 37 °C, 22 °C, or 4 °C. A–C, the graphs shown represent the median fluorescence intensities (MFI). Error bars indicate means ± SEM. These results were obtained in at least three independent experiments. Statistical significance was tested using (C) one-way ANOVA with a Holm–Sidak posttest (∗p < 0.05; ∗∗∗∗p < 0.0001; ns, nonsignificant). D, binding kinetics between RBD WT and CR3022 mAb assessed by BLI at 10 °C, 25 °C, or 35 °C. Biosensors loaded with RBD were soaked in twofold dilution series of CR3022 (100 nM–6.25 nM). Raw data are shown in blue and fitting model (1:1 binding model) is shown in red. All BLI data are summarized in Table S1. E, snapshot of SARS-CoV-2 Spike ectodomain (PDB 6VXX) (14) with one RBD indicated in transparent surface and one protomer’s RBD-to-trimer center-of-mass distance indicated with a cylinder. F, traces of the RBD-to-trimer distances from three replicas each of all-atom, fully glycosylated, and solvated MD simulations of the closed SARS-CoV-2 S trimer at 4 °C (blues) and 37 °C (reds) with dataset averages shown in heavy traces.