Control state: no heat-treatment.
The workflow to include glycopeptides in HDX-MS analysis was then tested after inducing a substantial change in spike structure by comparing native and heat treated glycoprotein. Without heat treatment, spike ectodomain (glyco)peptides showed a range of D2O labeling (interpreted as structural dynamics) between ~0% and ~37% that visually correlated with sub-domain location (Figure 6A). Localization of these patterns on a model based upon Protein Data Bank (PDB) files 6VSB and 6VXX (55) included key domains (furin cleavage site, fusion peptide, 630 loop, etc.) often missing from RCSB structures (www.rcsb.org), possibly because disordered regions of spike can have low electron density in cryo-EM analysis (56). The least deuterated area (<5%) was the trimeric core interface (764-782 helix, 997-1020 helix, and 1043-1062 strand in β-sheet), as previously described (4, 13).
The most deuterated regions (37%) were the “hinge” between head and stalk regions (1132-1145), furin cleavage site (672-690), and 630 loop (624-636), the latter also consistent with a previous report (13). Labeling of spike’s “hinge” is a novel HDX-MS result, completely based on the coverage of deuterated glycopeptides containing glycan-modified N1134 (Figures 6 and S2). Robust labeling (24-30%) was observed in the fusion peptide (823-851), N-terminus (19-31), and middle helix of the stalk (1174-1197).
The receptor binding domain (RBD) had the highest deuteration (27%) on a strand forming its hinge with the rest of S1 (319-346, including deuterated glycopeptide coverage at N343) and the two loops of the receptor binding motif (RBM, 442-452 and 471-487) for ACE2 (57). The other strand forming the RBD’s hinge with S1 (516-533) was 20% deuterated. A central helix (784-795), a turn-helix-turn (939-977) and a helix U-turn (753-759) were also moderately deuterated (28-30%). The N-terminal domain (NTD) showed an outer segment (171-176) with moderate labeling (27%).
Several recent reports describe bimodality in isotopic envelopes for certain regions of spike (1, 13, 15); the only apparently bimodal peptide 878-904 in this study was interpreted to be chromatographic co-elution of two precursor peptides with the same charge state and overlapping precursor isotopic envelopes based on observing “bimodality” even at 0 sec (data not shown).