Potent SARS-CoV-2 neutralizing antibodies directed against spike N-terminal domain target a single supersite

Abstract

Numerous antibodies that neutralize SARS-CoV-2 have been identified, and these generally target either the receptor-binding domain (RBD) or the N-terminal domain (NTD) of the viral spike. While RBD-directed antibodies have been extensively studied, far less is known about NTD-directed antibodies. Here, we report cryo-EM and crystal structures for seven potent NTD-directed neutralizing antibodies in complex with spike or isolated NTD. These structures defined several antibody classes, with at least one observed in multiple convalescent donors. The structures revealed that all seven antibodies target a common surface, bordered by glycans N17, N74, N122, and N149. This site-formed primarily by a mobile β-hairpin and several flexible loops-was highly electropositive, located at the periphery of the spike, and the largest glycan-free surface of NTD facing away from the viral membrane. Thus, in contrast to neutralizing RBD-directed antibodies that recognize multiple non-overlapping epitopes, potent NTD-directed neutralizing antibodies appear to target a single supersite.

Keywords: COVID-19; N-terminal domain; SARS-CoV-2; antibody class; antigenic supersite; multi-donor antibody; neutralizing antibody.

Graphical abstract

Figure 1

NTD-directed neutralizing antibodies derived from the VH1-24 gene define a multi-donor antibody class (A) Sequence alignment of VH1-24-derived NTD-directed antibodies showing paratope residues, somatic hypermutations, and gene-specific substitution profile (GSSP) showing somatic hypermutation probabilities for VH1-24 gene. Antibody positions are assigned using the Kabat scheme, the CDRs are assigned by IMGT scheme. Paratope residues are highlighted by underscoring and colored by interaction types. Amino acids in GSSP are colored by chemical property. (B) Cryo-EM reconstructions for spike complexes with antibodies 1-87, 1-68, and 2-51. NTD is shown in orange, RBD in green, and glycans in red, with antibody heavy chains in magenta and light chains in gray. (C) Expanded view of 1-87 interactions with NTD showing overall interface (left), recognition by CDR H3 (middle), and recognition by CDR H1 (right). NTD regions N3 (residues 141–156) and N5 (residues 246–260) are colored in shades of orange; CDR H1, H2, and H3 are colored in shades of magenta; CDR L1, L2, and L3 are colored in shades of gray. Nitrogen atoms are colored in blue and oxygen atoms in red; hydrogen bonds (distance < 3.2 Å) are represented as dashed lines. (D) Crystal structure of antibody 2-51 complexed with NTD, colored as in (B). (E) Expanded view of 2-51 interactions with NTD showing overall interface (left), recognition by CDR H3 (middle), and recognition by CDR H1 (right), colored as in (C). See also Figures S1–S7 and Tables S1 and S2.

Figure 2

NTD-directed neutralizing antibodies derived from the closely related VH3-30 and VH3-33 genes show distinct binding modes (A) Sequence alignment of VH3-30-derived (4-18) and VH3-33-derived (5-24) NTD-directed antibodies showing paratope residues, somatic hypermutations, and gene-specific substitution profile (GSSP) showing positional somatic hypermutation probabilities for VH3-30 gene. Substitutions between VH3-30 and VH3-33 germline genes are highlighted in green. (B) Cryo-EM reconstruction for spike complex with antibody 4-18 from two orthogonal views; NTD is shown in orange, RBD in green, and glycans in red, with antibody heavy chain in blue and light chain in gray. (C) Cryo-EM reconstruction for spike complex with antibody 5-24 from two orthogonal views; NTD is shown in orange, RBD in green, and glycans in red, with antibody heavy chain in brown and light chain in gray. (D) Expanded view of 4-18 interactions with NTD showing the overall interface (left), recognition in CDR H2 (middle), and recognition in CDR L3 (right). NTD regions N1 (residues 14–26), N3 (residues 141–156), and N5 (residues 246–260) are shown in shades of orange; CDR H1, H2, and H3 are shown in shades of blue; CDR L1, L2, and L3 are shown in shades of gray. (E) Expanded view of 5-24 interactions with NTD showing the overall interface (left), recognition in CDR H3 (middle), and recognition in CDR H1 (right), colored as in (D) except for CDR H1, H2, and H3, which are colored in shades of brown. See also Figures S1, S3, and S6, and Table S1.

Figure 3

NTD-directed neutralizing antibodies derived from the closely related VH1-69^∗01 and VH1-69^∗02 genes show distinct binding modes (A) Sequence alignment for VH1-69^∗01-derived (2-17) and VH1-69^∗02-derived (4-8) NTD-directed antibodies showing somatic hypermutations and paratope residues, with gene-specific substitution profile (GSSP) showing positional somatic hypermutation probabilities for VH1-69. Residues that differ between VH1-69^∗01 and VH1-69^∗02 alleles are highlighted in green. (B) Cryo-EM reconstruction for spike complex with antibody 4-8; NTD is shown in orange, RBD in green, glycans and in red, with antibody heavy chain in teal and light chain in gray. Heavy and light chain footprint on NTD (right). (C) Cryo-EM reconstruction for spike complex with antibody 2-17 (left); NTD is shown in orange, RBD in green, and glycans in red, with antibody heavy chain in dark green and light chain in gray. Heavy and light chain footprint on NTD (right, NTD shown with template-based modeling). (D) Expanded view of 4-8 interactions with NTD showing the overall interface (left), recognition in CDR H3 (middle), and recognition in CDR L2 (right). NTD regions N1 (residues 14–26), N3 (residues 141–156), and N5 (residues 246–260) are shown in shades of orange; CDR H1, H2, and H3 are shown in shades of teal; CDR L1, L2, and L3 are shown in shades of gray. (E) Sequence alignment of light chain of VH1-69-derived antibodies showing diverse germline gene usage (2-17 and COV2-2676 utilizes kappa light chain; 4-8 utilizes lambda light chain); IGKV3-15^∗01 is used as reference. Paratope residues of 4-8 are colored as in (A). See also Figures S1, S3, and S6, and Table S1.

Figure 4

Angles of approach for NTD-directed neutralizing antibodies (A) Overall approach of NTD-directed neutralizing antibodies to spike with angles defined with red arrows. The 3-fold axis is indicated by a black triangle. Antibodies are represented by long axes of the Fabs and colored by heavy chain colors defined in Figures 1, 2, and 3. (B) Latitudinal and longitudinal angles of approach. (C) Angles of recognition for antibodies grouped by VH gene. Notably, only those from VH1-24 show a consistent orientation. (D) Heavy-light chain orientations show graphically (left) and quantitatively (right).

Figure 5

NTD-directed antibodies induce conformational changes in NTD and spike (A) Conformational changes in NTD induced by binding of neutralizing antibodies. Antibody-bound NTDs are shown in cartoon representation and colored by per-residue Cα movements compared to unliganded NTD. Antibodies are shown in gray cartoon. Major NTD loops interacting with antibodies are labeled. (B) Sequence of NTD highlighting antibody contact and conformational change. Epitope residues for each antibody are marked with a number representing Cα movements (Å) from unliganded NTD; the symbol “X” indicates movement 10 Å and above. Potential glycosylation sites on NTD are highlighted in green (distances are shown for antibodies with sufficiently resolved interfaces; antibody 2-17 was only at 4.4 Å, and the interface of antibody 1-68 showed extensive mobility. (C) Epitope regions on NTD (red) and their conformational change. Glycans on NTD are shown as green spheres.

Figure 6

A structurally plastic antigenic supersite in the distal-loop region of NTD revealed by comparison of antibodies derived from the four multi-donor classes (A) Epitopes of NTD-targeting antibodies colored by potency (cryo-EM structures were of sufficient resolution to define all epitopes, except for 2-17 and 1-68, which utilized polyAla-based template modeling and homology modeling, respectively). Epitope residues are listed in Table S3. (B) The supersite of vulnerability on NTD. Supersite residues are listed in Table S5. (C) Glycan coverage of the spike. The NTD supersite is surrounded by glycans at N17, N74, N122, and N149. (D) NTD structural properties and antibody potency. Epitope surfaces of different antibodies were overlaid onto NTD with shades of red representing potency (left). Electrostatic potential on NTD (middle). Structural variation of NTD bound by NTD-directed antibodies (right). See also Figures S1, S6, and S7, and Tables S3–S6.

Figure 7

NTD supersite on MERS betacoronaviruses (A) Epitopes of MERS NTD antibodies target a site closer to the trimer axis. Borders of epitopes of antibody G2 and 7d10 are colored teal and cyan, respectively. SARS-CoV-2 NTD supersite is show as red boundary line. Glycans are shown as green spheres. (B) Spike sequence entropy between betacoronaviruses. (C) NTD of HKU1 spike is substantially glycosylated.