Molecular determinants of the interaction between HSV-1 glycoprotein D and heparan sulfate

Abstract

Literature has well-established the importance of 3-O-sulfation of neuronal cell surface glycan heparan sulfate (HS) to its interaction with herpes simplex virus type 1 glycoprotein D (gD). Previous investigations of gD to its viral receptors HVEM and nectin-1 also highlighted the conformational dynamics of gD's N- and C-termini, necessary for viral membrane fusion. However, little is known on the structural interactions of gD with HS. Here, we present our findings on this interface from both the glycan and the protein perspective. We used C-terminal and N-terminal gD variants to probe the role of their respective regions in gD/HS binding. The N-terminal truncation mutants (with Δ1-22) demonstrate equivalent or stronger binding to heparin than their intact glycoproteins, indicating that the first 22 amino acids are disposable for heparin binding. Characterization of the conformational differences between C-terminal truncated mutants by sedimentation velocity analytical ultracentrifugation distinguished between the "open" and "closed" conformations of the glycoprotein D, highlighting the region's modulation of receptor binding. From the glycan perspective, we investigated gD interacting with heparin, heparan sulfate, and other de-sulfated and chemically defined oligosaccharides using surface plasmon resonance and glycan microarray. The results show a strong preference of gD for 6-O-sulfate, with 2-O-sulfation becoming more important in the presence of 6-O-S. Additionally, 3-O-sulfation shifted the chain length preference of gD from longer chain to mid-chain length, reaffirming the sulfation site's importance to the gD/HS interface. Our results shed new light on the molecular details of one of seven known protein-glycan interactions with 3-O-sulfated heparan sulfate.

Keywords: HSV-1; glycoprotein D; heparan sulfate; heparin; herpes.

FIGURE 1

The interaction of the N- and C-terminus modulate gD receptor binding, notably the insertion of W294s side chain in the C-terminus into a crevice formed by L25 and Q27 in the N-terminus. Crystal structure of unliganded gD306_307C (2C36), which lacks electron density from 1–22 and 257–266. The N-terminus (1–40) colored in dark blue, with the crevice-forming residues Q27, L25, and P23 colored in cyan with side chains visible. The C-terminus (267–306) is colored in red. The insert shows the W294 side chain (magenta) inserted into the crevice formed by Q27 and L25. Visualized and colored with PyMOL Molecular Graphics System, Version 2.0 Schrödinger, LLC.

FIGURE 2

The first 22 residues of gD are disposable for heparin binding with gD285 but modulates heparin interaction with gD306. Representative sensorgrams of (A) gD285, (B), gD306, (C) gD23-285 and (D) gD23-306 to a heparin-immobilized streptavidin chip. Concentrations from top to bottom are (A) 500, 200, 100, 50, 10 nM, (B) 20, 10, 5, 1, 0.5 µM, (C) 5, 1, 0.5, 0.25, 0.125 µM, and (D) 20, 10, 5, 2.5, and 1.25 µM. The black lines depict the 1:1 Langmuir kinetic model fit to the raw data. k_a, k_d, and Χ² values are denoted in Supplementary Table S1.

FIGURE 3

2D SV-AUC differentiates the “open” gD285 with “closed” gD306 that cannot be seen in 1D plots, while the difference in binding affinity is evident by gD’s selective glycan complex formation. Sedimentation coefficient [c(s)] distribution analyses for (A) gD285 and (B) gD306 at 5, 10, and 20 µM. The c(s) analysis shows two distinct species, one with a sedimentation coefficient (s) of ∼2.7 and another with 4–4.2 s-values. Based on correlation with SDS-PAGE and MALDI-TOF-MS, this was concluded to be the monomeric (2.7/2.9 s) and dimeric (4–4.3 s) gD species. However, this is not an equilibrium-based dimer, as no there was no shift in the percentage of overall signal from monomer to dimer species proportional to concentration, but an inactive dimeric gD species that does not bind to receptors (unpublished data). Two-dimensional analysis of SV-AUC data with plots of sedimentation coefficient (S) versus frictional ratio f/f ₀ for (C) gD285 apo and (D) gD306 apo. Sedimentation coefficient distributions of (E) 10 µM gD285 or (F) gD306 by itself, with heparin and with low-molecular-weight heparin (LMW) at a 1:1 ratio. gD285 showed distinct complex formation with heparin and LMW heparin, but gD306 only formed a weaker complex with heparin.

FIGURE 4

Competition SPR with chemically de-sulfated heparin and sulfated heparosan oligosaccharides reveals 6-O-sulfation as key to gD-HS interaction. (A) Scheme of competition SPR with de-sulfated (deS) heparin ligands, where binding reflects how important the absent sulfate group is to gD/HS interaction. (B) Normalized, average binding percentage of 0.1 µM gD285 injected onto a heparin-immobilized chip pre-mixed with various de-sulfated (+deNS, +de2S, +de6S) heparin oligosaccharides at 1:1 and 1:2.5 M ratios. No oligosaccharide (gD285) and unmodified heparin were used as negative and positive controls, respectively. (C) Normalized, average binding percentage of 0.1 µM gD285 pre-mixed with various sulfated heparosan oligosaccharides (+NS, +NS2S, +NS6S, +NS6S2S) at 1:1 and 1:2.5 M ratios injected over a heparin immobilized SPR chip. Error bars denote the SD of three replicate flow channels.

FIGURE 5

Glycan microarray illustrates the key role of 6-O- and 3-O-sulfation in gD/HS binding, while also revealing preference for longer-chain ligands with NS/6S/2S pattern that can be shifted to shorter chains if 3-O-sulfated. (A) Raw fluorescent results of the glycan microarray and corresponding glycan code (numbered 1–96, glycan structures in Supplementary Figure S5). (B) Complete and (C) select results of gD285 binding to a glycan microarray and visualized with OG488. Glycans plotted in C (from L → R) are (6-mers) 2, 11, X, 29, 51, 80 (7-mers) 3, 15, 53, 35, 54, 81 (8-mers) 4, 19, 60, 41, 61, 83 (9-mers) 5, 23, 68, 46, 69, X (12-meres) 6, 24, 7, 47, 74, 84 (14-mers) X, X, 75, X, 76, X (18-mers) X, X, 77, 48, 78, X (X denotes that the glycan is not present on the array, reflected as missing bars in the graph). Error bars denote the average fluorescent intensity of 12 replicate spots.