SARS-CoV-2 Infection Depends on Cellular Heparan Sulfate and ACE2

Abstract

We show that SARS-CoV-2 spike protein interacts with both cellular heparan sulfate and angiotensin-converting enzyme 2 (ACE2) through its receptor-binding domain (RBD). Docking studies suggest a heparin/heparan sulfate-binding site adjacent to the ACE2-binding site. Both ACE2 and heparin can bind independently to spike protein in vitro, and a ternary complex can be generated using heparin as a scaffold. Electron micrographs of spike protein suggests that heparin enhances the open conformation of the RBD that binds ACE2. On cells, spike protein binding depends on both heparan sulfate and ACE2. Unfractionated heparin, non-anticoagulant heparin, heparin lyases, and lung heparan sulfate potently block spike protein binding and/or infection by pseudotyped virus and authentic SARS-CoV-2 virus. We suggest a model in which viral attachment and infection involves heparan sulfate-dependent enhancement of binding to ACE2. Manipulation of heparan sulfate or inhibition of viral adhesion by exogenous heparin presents new therapeutic opportunities.

Keywords: COVID-19; SARS-CoV-2; coronavirus; heparan sulfate; heparan sulfate-binding proteins; heparin; lung epithelial cells; pseudotyped virus; spike proteins.

Graphical abstract

Figure 1

Molecular Modeling of the SARS-Cov-2 Spike RBD Interaction with Heparin (A) A molecular model of SARS CoV-2 S protein trimer (PDB: 6VSB and 6M0J) rendered with Pymol. ACE2 is shown in blue and the RBD open conformation in green. A set of positively charged residues lies distal to the ACE2 binding site. (B) Electrostatic surface rendering of the SARS-CoV-2 RBD (PDB: 6M17) docked with dp4 heparin oligosaccharides. Blue and red surfaces indicate electropositive and electronegative surfaces, respectively. Oligosaccharides are represented using standard CPK format. (C) Mesh surface rendering of the RBD (green) docked with dp4 heparin oligosaccharides (red). (D) Number of contacts between the RBD amino acids and a set of docked heparin dp4 oligosaccharides from (A and B). (E) Calculated energy contributions of each amino acid residue in the RBD that can interact with heparin. (F) Amino acid sequence alignment of the SARS-CoV-1 and SARS-Cov-2 RBD. Red boxes indicate amino acid residues contributing to the electropositive patch in (A and B). Identical residues are shaded dark gray. Conservative substitutions have backgrounds in blue. Non-conserved residues have a white background (G) Structural alignment of SARS-CoV-1 (cyan; PDB: 3BGF) and SARS-CoV-2 (red; PDB: 6M17) RBD. (H) Electrostatic surface rendering of the SARS-CoV-1 and SAR-CoV-2 RBDs. See also Figure S1.

Figure S1

Location of the Putative Heparin/HS Binding Site in the Spike Protein RBD from SARS-CoV-2, Related to Figure 1 PDB files 6VSB and 6M0J were used to model the spike protein. The residues colored pink on the three RBDs (444+509+346+354+356+357+355+466+ 347+348+349+353+450+448+451+352) make up a potential binding site for heparin and heparan sulfate.

Figure S2

Analysis of Recombinant Spike Proteins and Receptor-Binding Domain, Related to Star Methods, Protein Production (A) SDS-PAGE gel of recombinant SARS-CoV-2 spike ectodomain protein produced in ExpiCho cells and commercial recombinant SARS-CoV-2 RBD. (B) Transmission electron micrographs of recombinant SARS-CoV-2 spike ectodomain protein. (C) Size exclusion chromatography of recombinant SARS-CoV-2 spike ectodomain protein on a Superose 6 column. (D) SDS-PAGE gel of recombinant SARS-CoV-2 RBD produced in ExpiHEK cells. (E) Size exclusion chromatography of recombinant SARS-CoV-2 RBD on a Superdex200 column.

Figure 2

SARS-CoV-2 Spike Binds Heparin through the RBD (A) Recombinant trimeric SARS-CoV-2 spike and RBD proteins were bound to heparin-Sepharose and eluted with a gradient of sodium chloride. (B) RBD protein from SARS-CoV-1 and SARS-CoV-2 binding to immobilized heparin-BSA. (C) Binding of spike protein or ACE2 to heparin-BSA. Insert shows SARS-CoV-2 spike protein binding to heparin-BSA in the presence of ACE2. (D) Binding of spike protein in the active RBD open (Mut7) and inactive RBD closed (Mut2) conformation to heparin-BSA. (E) Binding of ACE2 to spike protein in active RBD open (Mut7) and inactive RBD closed (Mut 2) conformation. (F) ACE2 binding to spike protein immobilized on heparin-BSA. The broken line represents background binding. Statistical analysis was by one-way ANOVA. (ns: p > 0.05, ^∗: p ≤ 0.05, ^∗∗: p ≤ 0.01, ^∗∗∗: p ≤ 0.001, ^∗∗∗∗: p ≤ 0.0001). (G and H) Negative stain-electron microscopy analysis of binding of heparin and ACE2 to spike protein. ACE2 binding to spike protein increases in the presence of heparin. 3D class averages of SARS-CoV-2 spike bound to zero, one, two, or three ACE2 (white, orange, blue, or gray) when complexed with and without a heparin dp20. The incubation was done for 15 min (G) or 60 min (H). The percentage of particles belonging to each class is shown in pie charts. See also Figure S3.

Figure S3

Binding of Spike Protein to Heparin and ACE2 and Electron Micrographs of the Spike-ACE2 Complexes, Related to Figure 2 (A) SARS-CoV-2 spike binding to immobilized heparin or BSA. (B) ACE2 binding to immobilized spike protein. (C) Transmission electron micrographs of stabilized spike protein treated with ACE2 and with or without dp20 for 15 min or 1 h. (D) 2D classes averages for each condition.

Figure 3

SARS-CoV-2 Spike Ectodomain Binding to Cells Is Dependent on Cellular HS (A) Titration of recombinant SARS-CoV-2 spike protein binding to human H1299 cells with and without treatment with a mix of heparin lyases I, II, and III (HSase). (B) Recombinant SARS-CoV-2 spike protein binding (20 μg/mL) to H1299, A549, and Hep3B cells with and without HSase treatment. (C) SARS-CoV-2 S RBD protein binding (20 μg/mL) to H1299, A549, and Hep3B cells with and without HSase treatment. (D) SARS-CoV-2 spike protein binding (20 μg/mL) to H1299 and A375 cells with and without HSase treatment. (E) Anti-HS (F58-10E4) staining of H1299, A549, Hep3B, and A375 cells with and without HSase treatment. (F) Binding of recombinant SARS-CoV-2 spike protein (20 μg/mL) to Hep3B mutants altered in HS biosynthesis enzymes. Specific enzymes that were lacking in the mutants are listed along the x axis. All values were obtained by flow cytometry. Graphs shows representative experiments performed in technical triplicate. The experiments were repeated at least three times. Statistical analysis by unpaired t test (ns: p > 0.05, ^∗: p ≤ 0.05, ^∗∗: p ≤ 0.01, ^∗∗∗: p ≤ 0.001, ^∗∗∗∗: p ≤ 0.0001). See also Figure S4.

Figure S4

Binding of RBD Protein to Hep3B Mutants, Related to Figure 3 Binding of SARS-CoV-2 S RBD protein (20 μg/mL) to Hep3B mutants. Binding was measured by flow cytometry. Statistical analysis by unpaired t test. (ns: p > 0.05, ^∗: p ≤ 0.05, ^∗∗: p ≤ 0.01, ^∗∗∗: p ≤ 0.001, ^∗∗∗∗: p ≤ 0.0001).

Figure 4

SARS-CoV-2 Spike Ectodomain Protein Binding to Cells Is Differentially Affected by HS from Different Organs and Potently Inhibited by Heparinoids (A) LC-MS/MS disaccharide analysis of HS isolated from human kidney, liver, tonsil, and lung tissue. (B) Inhibition of binding of recombinant SARS-CoV-2 S RBD protein to H1299 cells, using tissue HS. Analysis by flow cytometry. (C) Inhibition of recombinant trimeric SARS-CoV-2 protein (20 μg/mL) binding to H1299 cells, using CHO HS, heparin, MST heparin, and split-glycol heparin. Analysis by flow cytometry. (D) Similar analysis of A549 cells. Curve fitting was performed using non-linear regression and the inhibitor versus response least-squares fit algorithm. IC₅₀ values are listed in Table 1. Graphs show representative experiments performed in technical duplicates or triplicates. (ns: p > 0.05, ^∗: p ≤ 0.05, ^∗∗: p ≤ 0.01, ^∗∗∗: p ≤ 0.001, ^∗∗∗∗: p ≤ 0.0001).

Figure S5

ACE2 Expression and Mutations, Related to Figure 5 (A) qRT-PCR analysis of ACE2 expression. (B) DNA sequencing of ACE2 mutant alleles in A375 mutants. (C) DNA sequencing of ACE2 mutant alleles in A549 mutants.

Figure 5

ACE2 and Cellular Heparan Sulfate Are Both Necessary for Binding of SARS-CoV-2 Spike Ectodomain (A) Western blot shows overexpression of ACE2 in the A375 and A375 B4GALT7^−/− cells. A representative blot is shown. (B) Binding of SARS-CoV-2 spike protein to cells with and without ACE2 overexpression. Note that binding is reduced in the cells deficient in HS. (C) Gene targeting of ACE2 in A549 using CRISPR-Cas9. The bars show spike binding to two independent ACE2 CRISPR-Cas9 knockout clones with and without HSase treatment. Note that binding depends on ACE2 expression and that residual binding depends in part on HS. All analyses were done by flow cytometry. The graphs show representative experiments performed in triplicate technical replicates. Statistical analysis by unpaired t test. (ns: p > 0.05, ^∗: p ≤ 0.05, ^∗∗: p ≤ 0.01, ^∗∗∗: p ≤ 0.001, ^∗∗∗∗: p ≤ 0.0001). See also Figure S5.

Figure 6

SARS-CoV-2 Pseudovirus Infection Depends on Heparan Sulfate (A) Left, SARS-CoV-2 spike protein (20 μg/mL) binding to Vero cells measured by flow cytometry with and without HSase. Right, heparin and split-glycol heparin inhibit SARS-CoV-2 spike protein (20 μg/mL) binding to Vero cells by flow cytometry. Statistical analysis by unpaired t test. (B) Western blot analysis of ACE2 expression in Vero E6 cells compared to A549, H1299, and A375 cells. A representative blot of three extracts is shown for each strain. (C) Infection of Vero E6 cells with SARS-CoV-2 spike protein expressing pseudotyped virus expressing GFP. Infection was done with and without HSase treatment of the cells. Insert shows GFP expression in the infected cells by imaging. Counting was performed by flow cytometry with gating for GFP-positive cells as indicated by “infected.” (D) Quantitative analysis of GFP-positive cells. (E) Infection of Vero E6 cells with SARS-CoV-2 S protein pseudotyped virus expressing luciferase, as measured by the addition of Bright-Glo and detection of luminescence. The figure shows infection experiments done at low and high titer. (F) HSase treatment diminishes infection by SARS-CoV-2 S protein pseudotyped virus (luciferase) at low and high titer. (G) Heparin (0.5 μg/mL) blocks infection with SARS-CoV-2 S protein pseudotyped virus (luciferase). (H) Effect of HSase treatment of Vero E6 cells on the infection of both SARS-CoV-1 S and SARS-CoV-2 S protein pseudotyped virus expressing luciferase. (I) Infection of Hep3B with and without HSase and in Hep3B cells containing mutations in EXT1, NDST1, and HS6ST1/HS6ST2. Cells were infected with SARS-CoV-2 S protein pseudotyped virus expressing luciferase. All experiments were repeated at least three times. Graphs shows representative experiments performed in technical triplicates. Statistical analysis by unpaired t test. (ns: p > 0.05, ^∗p ≤ 0.05, ^∗∗p ≤ 0.01, ^∗∗∗p ≤ 0.001, ^∗∗∗∗p ≤ 0.0001). See also Figure S6.

Figure S6

Infection of Hep3B by MERS Pseudovirus, Related to Figure 6 Infection of Hep3B cells with SARS-CoV-2 and MERS-CoV S protein pseudotyped viruses carrying luciferase with and without treatment with heparin lyases.

Figure 7

Manipulation of Cellular Heparan Sulfate Decreases Infection of Authentic SARS-CoV-2 Virus (A,) Flow cytometry analysis of SARS-CoV-2-infected (red) or uninfected (black) Vero cells stained with antibodies against SARS-CoV-2 nucleocapsid and spike protein. (B) SARS-CoV-2 infection of Vero cells performed in the absence and presence of HSase, or with incubation with different concentrations of unfractionated heparin (UFH). The extent of infection was analyzed by flow cytometry as in (A). The graph shows a composite of five separate experiments, each performed in triplicate. The MOI was 0.5, but the extent of infection varied. The MOI in the experiment shown in maroon and blue was 0.2. The mean data from the individual experiments are colorized to allow for separate visualization (C) Same data as in (B), but with the experimental data normalized to the mock infection for each respective experiment. (D) SARS-CoV-2 infection of Hep3B mutants altered in HS biosynthesis enzymes. Cells were infected for 1 h and incubated 48 h, allowing for new virus to form. The resulting viral titers in the culture supernatants were determined by plaque assays on Vero E6 cells. Average values with standard error mean are shown, along with the individual data points. The experiment was initially optimized and then performed in triplicate. (E) Flow cytometry analysis of SARS-CoV-2-infected (red) or uninfected (black) human bronchial epithelial cells at an air-liquid interface stained with antibodies against SARS-CoV-2 nucleocapsid. (F) SARS-CoV-2 infection of human bronchial epithelial cells at an air-liquid interface was performed in the absence and presence of HSase, or with incubation UFH. The extent of infection was analyzed by flow cytometry. The graph shows a composite of three separate experiments, each performed in triplicate. The mean data from the individual experiments are colorized to allow for separate visualization. (G) Same data as in (F), but with each experimental dataset normalized to the uninfected control. Statistical analysis by one-way ANOVA (B, C, and G) or unpaired t test (D); ns: p > 0.05, ^∗: p ≤ 0.05, ^∗∗: p ≤ 0.01, ^∗∗∗: p ≤ 0.001, ^∗∗∗∗: p ≤ 0.0001). See also Figure S7.

Figure S7

Heparin and Heparin Lyases Have No Effect on ACE2 Expression or Cell Viability, Related to Figure 7 (A) Western blot analysis of ACE2 expression in Vero E6 cells treated with heparin lyases or 100 μg/mL UFH. (B and C) Assessment of cell viability after treatment with heparin lyase or 100 μg/mL UFH for 16 h in Vero (B) and human bronchial epithelial cells (HBEC) (C). Cell viability was measured using CellTiter-Blue. Statistical analysis by unpaired t test. (ns: p > 0.05, ^∗: p ≤ 0.05, ^∗∗: p ≤ 0.01, ^∗∗∗: p ≤ 0.001, ^∗∗∗∗: p ≤ 0.0001).