Companion to Cracking the Glycan Shield

The full-density glycan shield.

A crystal structure shows only the first few sugars of each glycan. This page builds complete Man9 glycans onto every glycosylation site of the same 5FYJ trimer¹, so you can see how much of the protein a fully built shield covers.

93 Man9 glycans on the trimer31 sites per protomerPDB 5FYJ← Back to the main report

Not peer reviewedThis page is part of independent computational work and has not been through peer review. I welcome scrutiny. If you work on HIV antibodies, glycan shields, or viral escape and you find an error, a questionable assumption, or a better way to test a result, please write to jezlim.marston@gmail.com.

The viewer

Drag the shield, switch the sugars

Glycans

Footprint

Sites

Model

View

Loading structure…

Man9 glycans (modeled)CD4 binding site targets N197, N276, N386V3-glycan targets N156, N301, N332V2 apex target N160Lighter tints: footprint residues on the protein

This is a snapshot ensemble, not a measurement. Every site carries a Man9 glycan at full occupancy, drawn from simulated poses that avoid the protein. Real Env mixes Man5 to Man9 and complex glycans, and some sites are only partly occupied, so treat this as a dense case rather than the actual shield.

How much do the glycans cover?

Share of each footprint hidden by glycans

Protein surface area of the footprint residues that the glycans cover, on all three protomers (Shrake–Rupley, 200 points²). Modeled columns average seven random draws, with a spread of about 1 to 5 points. A 4 Å probe is a rough stand-in for something bulkier than water reaching the surface, not a model of an antibody.

Footprint	Crystal stumps (main report)		Modeled Man9 25 crystal sites		Modeled Man9 all 31 sequons
Footprint	water	4 Å	water	4 Å	water	4 Å
CD4 binding site	3%	16%	5%	27%	6%	30%
V3-glycan	10%	36%	12%	55%	19%	71%
V2 apex	4%	46%	3%	41%	3%	41%
All of gp120	12%	35%	15%	49%	17%	57%

The V3-glycan footprint is the most hidden once the shield is complete: 36% with crystal stumps, 71% with every site built. All of the jump from 55% to 71% comes from the six extra sites, which sit beside parts of the footprint (residues 135–138 and 415–417).
The V2 apex figure does not rise. It dips from 46% to 41% because the crystal stumps at N160 happen to lie over the apex, while the modeled glycans point in many directions. The stump numbers are therefore a good guide but not a strict floor everywhere.
The CD4 binding site stays the least covered, which is consistent with its position in a recessed pocket.

Method

How the shield was built

Glycans: the GlycoSHIELD conformer library (v6.0, Zenodo record 8414255)^3,4, Man9GlcNAc2 from a molecular dynamics run. I sampled every second of 30,001 frames.
Sites: every N-X-S/T sequon on gp120 and gp41 of 5FYJ (30 per protomer) plus N355, which has an N-X-C sequon and a glycan in the crystal. That is 31 sites per protomer, 93 per trimer. 25 have glycan density in the crystal; the other 6 (N137, N138B, N141, N405, N408, N462) do not.
Placement: at the 25 crystal sites each library glycan is superposed on the crystal's own first GlcNAc ring and Asn side chain (RMSD under 0.9 Å). At the other 6 it is superposed on the Asn backbone.
Clash filter: a pose is kept only if no glycan atom lies within 2.8 Å of protein. At N133 too few poses passed, so the cutoff was relaxed to 2.2 Å there (13 poses kept). N262 is also crowded and kept 24 poses at 2.8 Å.
Draws: for each draw, one pose per site is picked at random while avoiding overlap with glycans already placed (2.4 Å). The viewer shows five draws.

Limits: the trimer and every glycan are rigid within a draw, with no glycan dynamics. Glycan type and occupancy are assumed, not measured. The footprints are the same 5 Å contact sets as in the main report. Library: Glycan grafting and the surface-area calculation used MDAnalysis⁵; the viewer is 3Dmol.js⁶. Library: the GlycoSHIELD Glycan Conformer Library^3,4.

References

Sources

1.Stewart-Jones GBE, et al. Trimeric HIV-1-Env structures define glycan shields from clades A, B, and G. Cell 2016;165:813–826. doi:10.1016/j.cell.2016.04.010. PDB 5FYJ (clade G X1193.c1 SOSIP.664 with PGT122, 35O22 and VRC01).
2.Shrake A, Rupley JA. Environment and exposure to solvent of protein atoms. Lysozyme and insulin. J Mol Biol 1973;79:351–371. doi:10.1016/0022-2836(73)90011-9
3.Tsai YX, et al. Rapid simulation of glycoprotein structures by grafting and steric exclusion of glycan conformer libraries (GlycoSHIELD). Cell 2024;187:1296–1311. doi:10.1016/j.cell.2024.01.034
4.Tsai YX, et al. GlycoSHIELD glycan conformer library, version 6.0 (file gs.27.M.Man9.tar.gz used). Zenodo 2023. doi:10.5281/zenodo.8422323 (concept record doi:10.5281/zenodo.8414255).
5.Michaud-Agrawal N, et al. MDAnalysis: a toolkit for the analysis of molecular dynamics simulations. J Comput Chem 2011;32:2319–2327. doi:10.1002/jcc.21787
6.Rego N, Koes D. 3Dmol.js: molecular visualization with WebGL. Bioinformatics 2015;31:1322–1324. doi:10.1093/bioinformatics/btu829