(C) PLOS One
This story was originally published by PLOS One and is unaltered.
. . . . . . . . . .



The structural role of SARS-CoV-2 genetic background in the emergence and success of spike mutations: The case of the spike A222V mutation [1]

['Tiziana Ginex', 'Centro De Investigaciones Biológicas Margarita Salas', 'Cib-Csic', 'Madrid', 'Clara Marco-Marín', 'Instituto De Biomedicina De Valencia', 'Ibv-Csic', 'Valencia', 'Centro De Investigación Biomédica En Red En Enfermedades Raras', 'Ciberer']

Date: 2022-09

In order to assess the evolutionary pressures on S:A222V, we employed the phylogenetically corrected single-likelihood ancestor counting (SLAC) method [ 13 ] to calculate the ratio of nonsynonymous to synonymous substitution rates (dN/dS) estimates for 874 codons in the NTD (N-Terminal Domain) region of the spike. dN/dS can potentially estimate the strength and direction of selection, typically indicating positive selection when dN/dS> 1. Using three thousand haplotypes representative of G-clade, we confirmed previous results [ 14 ] that S:A222V shows signatures of positive selection, with an estimate of dN/dS of 1.63 (p < 0.0001). This result was consistent with the analysis of a wider dataset of eleven thousand haplotypes (dN/dS of 1.62; p < 0.0001). When analysing independently four main periods of the pandemic (see Methods ), we could detect a temporal pattern in point estimates of dN/dS, with values higher than one during all periods analysed ( Fig 1C ). dN/dS showed values of 2.38 and 2.64 when B.1.177 and B.1.1.7 dominated, and a value of 30 when B.1.617.2 appeared, and even a higher dN/dS value for B.1.1.529 period due to dS = 0 ( Fig 1C ), indicating the presence of signals of positive selection in S:A222V during different waves, but especially after B.1.617.2 arose.

S:A222V appeared within B.1.617.2 as early as March 2021, when it accounted for 17% of all B.1.617.2 sequences ( Fig 2A ) . After that, B.1.617.2 sequences with S:A222V increased to 23% in April 2021 but dropped afterwards. This decrease is mainly driven by the decrease of the sequences designated as parental lineage B.1.617.2 ( Fig 2B and Fig A in S1 Text ) . Conversely, since July 2021, the proportion of B.1.617.2 (and S:A222V along with it) has been increasing mainly due to the increase of AY.47 and AY.4.2 ( Fig 2A ) . Among the different appearances of S:A222V within B.1.617.2, it has been successfully transmitted on at least two occasions: one in the last common ancestor of AY.4.2 and the other in the common ancestor of Lineages AY.9, AY.26, AY.27 and AY.47 ( Fig 2B ). Additionally, eight other less abundant lineages which derive from the same common ancestor as AY.9, A.26, AY.27 and AY.47 also have S:A222V ( Fig 2B and Fig B in S1 Text ).

( a) Percentage of sequences designated as parental B.1.617.2 and its sub-lineages (designated as AY) with S:A222V by month. ( b) Maximum likelihood phylogeny of 10,049 sequences from Lineages B.1.617.2 and derivatives. Phylogeny rooted with reference MN908947.3. Red external ring indicates sequences with S:A222V, the colour of branches indicates B.1.617.2 sub-lineages of interest. The scale bar indicates the number of nucleotide substitutions per site. Each circle in branches correlates with the bootstrap value; only bootstraps from 70 to 100 are represented.

To decipher the recent dynamics of S:A222V, we focused on B.1.617.2 in a global dataset of 2,992,547 sequences ( Fig 2A and 2B ). B.1.617.2 sequences with S:A222V belong to different sub-lineages, being characteristic of five sub-lineages: AY.4.2, AY.9, AY.26, AY.27, and AY.47, where it is present in more than 95% of the sequences of each sub-lineage ( Fig 2B ). However, S:A222V is also present in 25% of sequences classified as the parental B.1.617.2 lineage, and in 1.31% of other B.1.617.2 sub-lineages ( Fig 2B ). S:A222V has appeared at least 32 times in the genomic context of B.1.617.2 ( Fig 1A , Fig B in S1 Text ), being detected in at least 171 different sub-lineages ( Fig 2B and Fig B in S1 Text ).

Another seven lineages harbour S:A222V in high or medium frequencies, although those lineages are very rare with very few sequences and are very restricted geographically. Out of 397 sequences classified as B.1.36.31, 93.45% harbour S:A222V (Fig B in S1 Text ). Similarly, 91.66% of 12 sequences classified as B.1.630 harbour S:A222V (Fig B in S1 Text ). Other rare lineages such as B.1.1.325, C.4, AW.1, B.1.1.160.30, and B.1.441 include a percentage of sequences with S:A222V that ranges from 49% to 15% (Fig B in S1 Text , Table A in S1 Text ).

By July 2021, the majority of the non-B.1.177 occurrences appeared in lineage B.1.617.2, where 10.35% of B.1.617.2 contain S:A222V ( Fig 1B ). B.1.617.2 is also known as Variant Of Concern (VOC) Delta and it includes all AY lineages [ 11 ] (see below and Fig A in S1 Text ). Less frequently, S:A222V appears in 0.13% of B.1.1.7 (VOC Alpha), and 0.56% of B.1 overall. The temporal pattern for S:A222V (blue area in Fig 1B ) shows two peaks, the first corresponding to the dynamics of B.1.177 (green line in Fig 1B ) and the second to the expansion of B.1.617.2 (VOC Delta) (red line in Fig 1B ). We did not observe a similar peak in S:A222V during the increase of B.1.1.7 (VOC Alpha; orange line in Fig 1B ), possibly suggesting that epistatic interactions may favour the transmission of S:A222V preferentially in some genetic backgrounds. Sequences with S:A222V began to decrease with the replacement of VOC Delta (B.1.617.2; red line in Fig 1B ) by VOC Omicron [ 12 ] (B.1.1.529; blue line in Fig 1B ). By April 2022, S:A222V only appears in 0.05% of 152,891 sequences classified as B.1.1.529, and due to the predominance of B.1.1.529, S:A222V is almost no detectable (blue area in Fig 1B ).

(a) Global phylogeny of 11,166 sequences belonging to G clade. The red inner circle denotes sequences with S:A222V, the external circle indicates PANGO linages of interest indicated in legend. (b) Percentage of global sequences with S:A222V (blue area) in different VOCs. Sequences belonging to parental lineages designated as B.1.177 appear as a green line, B.1.1.7 as a yellow line, B.1.617.2 as a red line, AY.4.2 as a purple line, and B.1.1.529 as a blue line. (c) The ratio of nonsynonymous to synonymous substitution rates (dN/dS) for the codon 222 of the spike for each period with a predominant variant, the size of the circle indicates–log(p) with a constant value of 10. *dN/dS for B.1.1.529 period is even higher but cannot be represented because dS = 0.

The mutation S:A222V first arose within Lineage B.1.177, described for the first time in Spain in Summer 2020 [ 10 ]. S:A222V has been observed in 447,777 sequences from 123 countries by April 2022, almost exclusively co-occurring with S:D614G (99.86%). Additionally, 9% of 4,993,996 SARS-CoV-2 sequences available until 2022-04-01 include [S:A222V + S:D614G] and only 0.012% harbour S:A222V without S:D614G, indicating that the presence of S:A222V is tightly linked to S:D614G. Interestingly, although S:A222V arose with B.1.177, by September 2021 more than half of the sequences with S:A222V were not in the genomic context of B.1.177 ( Table A in S1 Text and Fig 1A ), indicating that although B.1.177 was replaced by other variants [ 10 ], S:A222V emerged independently in different lineages ( Fig 1A ). Non-B.1.177 sequences with S:A222V were isolated in 141 countries and classified in as many as 364 PANGO lineages ( Table A in S1 Text ), which indicates subsequent and independent appearances of S:A222V in different genetic backgrounds ( Fig 1A ).

The sensitivity to neutralization by convalescent sera of pseudotyped VSV carrying the ancestral Wuhan-Hu-1 (Wuhan), S:D614G, or the [S:A222V + S:D614G] (20E) spike protein was evaluated. No significant differences were detected between any of the sera (n = 6; p>0.05 by Kruskal-Wallis test; p>0.05 by Mann-Whitney test for Wuhan-Hu-1 vs. S:D614G, Wuhan-Hu-1 vs. 20E, or S:D614G vs. Wuhan-Hu-1).

One possibility for the selection of S:A222V could be an improved ability to replicate in immune populations. As vaccination had not yet started in Spain when this mutation was first observed in 2020, any such effect would have to result from selection for escape from existing immunity in convalescent individuals. To test this possibility, we evaluated the ability of VSV pseudotyped with either the ancestral Wuhan-Hu-1 spike protein, S:D614G, or 20E ([S:A222V + S:D614G]) to be neutralized by sera from the first wave of infection in Spain (April 2020) when no S:A222V mutations were circulating ( Fig 1B ). No significant differences were observed in neutralization between the different viruses ( Fig 3 ) .

To functionally characterize the purified protein S variants, we first used thermal shift assays to measure their thermostability and we found that [S:A222V + S:D614G] and S:D614G variants show very similar half-melting temperatures (T 0.5 ) and are both slightly more stable than the original Wuhan-Hu-1 variant ( Table 1 and Fig D in S1 Text ). Then, to see the possible impact of the S:A222V mutation on the spike binding capacity to the protein ACE2, we carried out protein-protein interaction assays by biolayer interferometry (BLI) ( Table 1 and Fig E in S1 Text ). The affinities for ACE2 were somewhat higher for the [S:A222V + S:D614G] and S:D614G variants than for the Wuhan-Hu-1 variant (K D values of 50 nM, 66 nM and 79 nM, respectively), in agreement with previous reports comparing the S:D614G and Wuhan-Hu-1 variants [ 9 ]. Although the dissociation constants for proteins carrying S:D614G and [S:A222V + S:D614G] mutations are similar, we observed a higher k on for the [S:A222V + S:D614G] mutant, perhaps reflecting structural differences related to higher accessibility of the RBD in this mutant compared to that in the S:D614G or the Wuhan-Hu-1 variants. To investigate this possibility further, we determined the cryo-EM structures of both the [S:A222V + S:D614G] and the S:D614G S proteins.

The ectodomain of SARS-CoV-2 protein S corresponding to the Wuhan-Hu-1, S:D614G or [S:A222V + S:D614G] variants, hosting proline substitutions at residues 986 and 987, a mutant non-cleavable sequence at the furin site and a C-terminal foldon trimerization motif, were flash-purified (1 day) identically and used fresh. The three forms were indistinguishable by SDS-PAGE and eluted identically from a SEC column (shown for S:D614G and [S:A222V + S:D614G] spikes in Fig C a-b in S1 Text ), as expected if they had essentially the same size and glycosylation pattern. Their quality, assessed by negative-stain electron microscopy (EM), was also similarly high (shown for S:D614G in Fig C c in S1 Text ).

Structural analysis of the allosteric role of S:A222V and epistatic interactions with S:D614G

To provide a comparative structural analysis of the S:A222V mutation on the background of S:D614G, we combined cryo-EM and classical MD simulations. Samples of spikes with both [S:A222V + S:D614G] and S:D614G, as an internal control, were produced for cryo-EM as described in Methods. Data about cryo-EM on the S:D614G mutant as well as a detailed description of the data sources, the modelled systems, including model validation metrics, the MD protocols and structural stability is reported in Methods and S1 Text.

Structural comparison of the [S:A222V + S:D614G] and S:D614G 1-up mutants. Comparison with the structure of the S:D614G mutant reveals that in the S:A222V mutation the V222 residue is accommodated in a highly hydrophobic environment of the NTD shaped by V36, Y38, F220 and I285 (Fig 5A). As observed in Fig J in S1 Text, this mutation is expected to slightly affect the interaction of the open subunit (chain B; NTD B ) with the neighbouring CTD1 A . Overall, the structures for the two mutants are strikingly similar (RMSD of 1.332 Å over 3193 Cα atoms), especially at the level of the S2 region, although we can observe small conformational changes in the position of the NTDs and RBDs. These differences are particularly relevant for the RBD in the up position (RBD B ), which is involved in the interaction with the host receptor, ACE2. Indeed, superimposition of these two structures with that of a spike in complex with bound ACE2 (PDB: 7DF4) indicates that in these two spike structures small rearrangements of this RBD-up would be beneficial to bind ACE2 [16]. The conformation of subunit C is quite similar in the two structures (RMSD 1.108 Å over 1074 Cα atoms), in contrast with the other two subunits (RMSD of 1.381 Å over 1070 Cα atoms for chain A and 1.389 Å over 1049 Cα atoms for chain B), which differ especially in (i) RBD B , (ii) RBD A , and (iii) NTD B (Fig 5B and S1 Movie). The analysis of these 3 domains allows us to describe the main conformational changes between these two spike structures. RBD A is sandwiched between RBD B and NTD B. We observed that RBD A and NTD B move together as a rigid body (RMSD of these two domains between [S:A222V + S:D614G] and S:D614G structures is only of 0.726 Å), approaching RBD B in the case of [S:A222V + S:D614G] when compared to that in the S:D614G structure. On the other hand, RBD B of [S:A222V + S:D614G] undergoes a rotational movement that also brings it closer to RBD A (Fig 5B and 5C, and S1 Movie). All these movements affect the degree of tightening of the S1 region of the spike (S1 Movie), so that the global conformation of the [S:A222V + S:D614G] mutant seems to be tighter than that of S:D614G. This is mainly a consequence of the tightening of subunit B around RBD A . In this direction, the distance between residues 114 and 381 (Fig 5B), located in NTD B and RBD B at the interface between these domains and the RBD A , are 3.5 Å closer in the [S:A222V + S:D614G] structure (35.3 Å in [S:A222V + S:D614G] vs. 38.8 Å in S:D614G). Although to a lesser extent, the distance between these residues in subunit A is also shorter in [S:A222V + S:D614G] when compared to that in S:D614G (42.4 Å in [S:A222V + S:D614G] vs. 43.9 Å in S:D614G). These tightening movements also increase the surface of contact between the NTD and RBD of the different subunits in [S:A222V + S:D614G] when compared to those in S:D614G (Table C in S1 Text). This effect seems to correlate with neither an increase in the stability of the S protein, as suggested by thermal shift stability assays (see above), nor with a decrease in the flexibility of the protein. In fact, an analysis of the temperature factors (B-factors) of the two structures (Fig 5D) shows a much higher mobility for RBD B in [S:A222V + S:D614G]. Not surprisingly, the regions in direct contact with RBD B (that is, RBD A and the contacting loops of NTD C ) also show increased B-factor values. PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 5. Effects of the S:A222V mutation on the structure of SARS-CoV-2 spike. (a) On the left, cartoon representation of the trimeric spike. On the right, detail of the region of interaction between the NTD from subunit B (salmon) and the CTD1 region from subunit A (blue). The side chains of residues surrounding A222 (upper panel) or V222 (lower panel) are shown as sticks. (b) Comparison of the structures for the S:D614G and [S:A222V + S:D614G] mutants showing main conformational changes observed when S:A222V is present. Detailed view of the trimeric spike surface shown in semi-transparent representation with different colours for the different subunits, and RBD B , RBD A and NTD B domains from the S:D614G and [S:A222V + S:D614G] mutant structures shown as cartoon. Domains corresponding to the S:D614G structure are coloured in yellow. Arrows indicate the direction and magnitude of the observed domain movements. A dashed line represents the distance between the Cα atoms of residues 114 and 381 in chain B, which is ~ 3 Å higher in S:D614G than in [S:A222V + S:D614G]. (c) Schematic representation of the RBD and NTD from different subunits to show the movements observed when comparing the two structures (d) Comparison of the B-factor of RBD B of the two spike mutants. Backbone is coloured and sized according to the B-factor values for S:D614G (left) and [S:A222V + S:D614G] (right). https://doi.org/10.1371/journal.ppat.1010631.g005

[END]
---
[1] Url: https://journals.plos.org/plospathogens/article?id=10.1371/journal.ppat.1010631

Published and (C) by PLOS One
Content appears here under this condition or license: Creative Commons - Attribution BY 4.0.

via Magical.Fish Gopher News Feeds:
gopher://magical.fish/1/feeds/news/plosone/

You are viewing proxied material from magical.fish. The copyright of proxied material belongs to its original authors. Any comments or complaints in relation to proxied material should be directed to the original authors of the content concerned. Please see the disclaimer for more details.