Professional Documents
Culture Documents
BA.2.12.1, BA.4 and BA.5 Escape Antibodies Elicited by Omicron Infection
BA.2.12.1, BA.4 and BA.5 Escape Antibodies Elicited by Omicron Infection
BA.2.12.1, BA.4 and BA.5 Escape Antibodies Elicited by Omicron Infection
The recent emergence and global spread of the SARS-CoV-2 variant with the RBD of BA.1, BA.2 contains three additional mutations, T376A,
Omicron (B.1.1.529) have posed a critical challenge to the efficacy D405N and R408S, and lacks the BA.1 mutations G446S and G496S
of COVID-19 vaccines and neutralizing antibody (NAb) therapy6–9. (Extended Data Fig. 1a). S371L on BA.1 is also substituted with S371F in
Owing to multiple mutations to the spike protein, including in the BA.2. The Omicron variants that have emerged more recently contain
receptor-binding domain (RBD) and N-terminal domain, Omicron BA.1 similar RBD sequences to BA.2, with the addition of L452 and F486
infection can result in substantial NAb evasion3,10–13. Omicron subline- substitutions—L452Q in BA.2.12.1, L452M in BA.2.13 and L452R/F486V in
age BA.2 has rapidly surged worldwide, out-competing BA.1. Compared BA.4 and BA.5—and exhibit a transmission advantage over BA.2. There
d hACE2 + WT RBD hACE2 + Delta RBD hACE2 + BA.1 RBD hACE2 + BA.2 RBD hACE2 + BA.2.12.1 RBD hACE2 + BA.3 RBD hACE2 + BA.4/5 RBD
ka = 3.59 × 105 M–1 s–1 ka = 1.55 × 105 M–1 s–1 ka = 5.81 × 105 M–1 s–1 ka = 7.37 × 106 M–1 s–1 4 –1 –1
600 ka = 4.98 × 10 –4M–1 s ka = 6.77 × 105 M–1 s–1 5 –1 –1
250 ka = 6.48 × 10 M s
600 600 250 250 400
kd = 7.88 × 10–3 s–1 kd = 2.09 × 10–3 s–1 kd = 8.45 × 10–3 s–1 kd = 7.92 × 10–2 s–1 kd = 6.16 × 10 s kd = 1.79 × 10–2 s–1 k = 9.32 × 10–3 s–1
Response (RU)
Fig. 1 | Structural and receptor-binding characteristics of Omicron The binding surface areas between S2 subunits of the variants are calculated in
subvariants. a, Surface representation of S-trimers of BA.1, BA.2, BA.3, BA.2.13, the table on the right. c, Thermoflour analysis for these Omicron variants.
BA.2.12.1 and BA.4/BA.5 (BA.4/5) variants. b, Structural interpretation and Analyses were performed as biological duplicates. d, Binding affinities of RBDs
functional verification of the stability of the spike protein of BA.1, BA.2, BA.3, of Omicron variants for hACE2 measured by SPR. Analyses were performed as
BA.2.13, BA.2.12.1 and BA.4/BA.5 variants. Left, superimposed structures of biological duplicates.
spike protein and the S2 domains of BA.1 (purple), BA.2 (red) and BA.4/BA.5 (blue).
is an urgent and immediate need to investigate the receptor binding three copies of S2 (Fig. 1b). By contrast, BA.1, BA.3 and BA.4/BA.5 spike
and immune-evasion capabilities of these new Omicron variants. possess relatively tight inter-subunit organization with more buried
areas between S2 subunits (Fig. 1b). In line with structural observa-
tions, thermal stability assays also verified that S-trimers from BA.2
Structural analyses of Omicron spike sublineages were the least stable among these variants, which might
We expressed and purified the prefusion-stabilized trimeric ecto- confer an enhanced fusion efficiency (Fig. 1c).
domains of BA.1, BA.2, BA.3, BA.2.12.1, BA.2.13 and BA.4/BA.5 spike Next, we measured the binding affinity between human ACE2
(S-trimer). All the S-trimers contain Gly-Ser-Ala-Ser (GSAS) and 6P muta- (hACE2) and S-trimers of the Omicron variants by surface plasmon
tions along with the T4 fibritin trimerization domain for increased resonance (SPR) (Extended Data Fig. 1c). The BA.4/BA.5 S-trimer showed
stability14,15. We determined the cryo-electron microscopy (cryo-EM) a decreased binding affinity with hACE2 compared with those of the
structures of these S-trimers at overall resolutions of 3.1–3.5 Å. Together other Omicron subvariants; however, this measurement could be
with the previously reported BA.1 structure16, this enabled us to com- misleading, owing to the additional N658S mutation. To exclude the
pare the structural differences across Omicron sublineages (Fig. 1a and potential influence of N658S, we also examined the binding affinities of
Extended Data Fig. 1b). In contrast to the BA.1 S-trimer, which is stably the RBDs of the Omicron variants for hACE2 (Fig. 1d). The RBDs of Delta
maintained in an open conformation with one ‘up’ RBD and two ‘down’ (B.1.617.2) and the circulating Omicron subvariants exhibited similar
RBDs16, BA.2 and BA.2.12.1 spike exhibits two conformational states binding affinities for ACE2, except for the BA.3 RBD, which showed a
corresponding to a closed form, with all three RBDs in the down con- lower affinity, similar to that of the ancestral WT strain. Additionally, the
figuration and an open form with one RBD in the up position. Of note, BA.2 subvariants displayed slightly higher binding affinities for hACE2
one RBD in BA.2.13 was clearly disordered, representing a stochastic than the other Omicron variants. To further explore the molecular
movement, which, together with BA.2 and BA.2.12.1, suggests structural basis for the altered binding affinities of these variants to hACE2, we
heterogeneity in the S-trimers of BA.2 sublineages. Most BA.3 and BA.4 performed molecular dynamics simulations based on the cryo-EM
S-trimers adopt closed or semi-closed forms (Fig. 1a). The differences structures and examined the effects of substitutions in the RBD on the
in the RBD up or down conformation could be allosterically modulated interaction with hACE2 (Extended Data Fig. 1d). The results reveal that
by mutations and deletions in the N-terminal domain or near the furin the lack of G496S in BA.2 sublineages meant that the hydrogen bond
cleavage site, but the detailed mechanism remains unclear. The BA.4/ with hACE2 K353 was regained, increasing their binding capability,
BA.5 spike that we used in our experiments also contains the N658S in line with experimental observations revealed by deep mutational
mutation, which was present in early BA.4/BA.5 sequences but later scanning (DMS) assay17. However, a local conformational perturba-
disappeared owing to the lower transmissibility of this variant, and tion surrounding spike residues 444–448 disrupted the hydrophilic
may correlate with the more closed RBD configurations of the BA.4/ interaction between BA.3 spike (S446) with hACE2 Q42, presumably
BA.5 S-trimer. Of note, S-trimers from the BA.2 sublineage harbour owing to the single mutation G446S rather than double mutations of
relatively less compact architectures in the region formed by the G446S and G496S (Extended Data Fig. 1d). Notably, the F486V mutation
BA 1
.1
.3
BA .2
.2 3
SA BA .1
-C /5
-1
BA 1
.1
.3
BA .2
.2 3
SA BA .1
-C /5
-1
BA 1
.1
.3
BA .2
.2 3
SA BA .1
-C /5
-1
BA 1
.1
.3
BA .2
.2 3
SA BA .1
-C /5
-1
.
BA 2.1
BA 2.1
BA 2.1
BA .2.1
14
14
14
14
BA
.1
BA
BA
BA
.1
BA
BA
BA
.1
BA
BA
BA
.1
BA
BA
2
RS .4
oV
RS .4
oV
RS .4
oV
RS .4
oV
.1
.1
.1
.1
D6
D6
D6
D6
.
.
e IC50 LY- LY- LY- REGN REGN COV2- COV2- BRII- BRII-
S309
DXP-
ADG-2 S2K146
SA58 SA55 LY-CoV016 REGN10933 COV2-2196 BRII-196 SA55
(ng ml–1) CoV016 CoV555 CoV1404 10933 10987 2196 2130 196 198 604 (BD55-5840) (BD55-5514) + LY-CoV555 + REGN10987 + COV2-2130 + BRII-198 + SA58
D614G 32 15 0.7 5.6 5.7 1.6 2.5 53 1,239 74 11 11 17 0.9 11 20 5.0 2.1 81 2.1
BA.1 * * 0.6 * * 5,419 3,007 7,118 1,171 361 285 979 11 4.4 1.7 * * 491 1,890 3.2
BA.1.1 * * 1.8 8,912 * 4,764 * 6,324 * 314 198 991 17 4.5 3.0 * * 8,090 * 3.3
BA.2 * * 0.9 * 590 4,312 6.3 8,530 8,990 918 219 * 20 12 7.2 * 821 8.2 8,610 7.8
BA.3 * * 1.1 * * 5,609 11 7,833 1,687 972 259 6,226 16 8.1 7.1 * * 19 2,190 6.4
BA.2.13 * * 1.0 9,221 417 3,591 6.6 6,902 * 700 148 * 16 4.9 5.9 * 699 7.1 * 4.8
BA.2.12.1 * * 0.8 * 499 5,521 11 7,620 * 989 201 * 13 5.0 5.2 * 714 18 * 5.0
BA.4/BA.5 * * 0.9 * 520 * 23 7,124 * 792 6,264 * 221 3.9 5.0 * 709 40 * 4.5
SARS-CoV-1 * * * * * * * * * 31 * 1.7 108 5.6 4.4 * * * * 4.6
Pangolin-GD 1,125 6.8 8.6 157 84 17 * 13 * * 7.4 5.0 14 296 5.7 10 98 27 33 7.7
RaTG13 * * * * * * * 16 * * 1.1 * 3.9 * 38 * * * 29 49
Fig. 2 | BA.2.12.1, BA.4 and BA.5 exhibit stronger antibody evasion than were calculated using two-tailed Wilcoxon signed-rank tests of paired samples.
BA.2. a–d, Neutralizing titres against SARS-CoV-2 D614G, Omicron subvariants The geometric mean titre is shown above each group of points. e, Neutralizing
and SARS-CoV-1 pseudoviruses in plasma from vaccinated and convalescent activity against SARS-CoV-2 variants and sarbecoviruses by therapeutic NAbs.
individuals. a, Individuals who had received 3 doses of CoronaVac (n = 40). Green, half-maximal inhibitory concentration (IC50) ≤ 30 ng ml−1; white,
b, Individuals who had received 2 doses of CoronaVac and 1 dose of ZF2001 30 ng ml−1 < IC50 < 1,000 ng ml−1; red, IC50 ≥ 1,000 ng ml−1; *, IC50 ≥ 10,000 ng ml−1.
(n = 38). c, Individuals who, after receiving 3 doses of CoronaVac, had been All neutralization assays were performed as biological duplicates. *P < 0.05,
infected with BA.1 and recovered (n = 50). d, People who had recovered from **P < 0.01, ***P < 0.001; NS, not significant (P > 0.05).
SARS and received 2 doses of CoronaVac and 1 dose of ZF2001 (n = 28). P-values
in BA.4/BA.5 spike decreases hACE2 binding activity owing to reduced who had received three doses of CoronaVac before infection (Fig. 2c),
hydrophobic interaction (Extended Data Fig. 1d). We also noted poten- despite their significantly higher neutralization titres against D614G
tial reductions in hydrophilic interactions owing to R493Q reversion. and BA.1 compared with the triple-dosed vaccinees who had not been
Notably, two reports claimed recently that BA.4/BA.5 RBD and spike infected with BA.1 (Extended Data Fig. 2a). The 50% neutralization
(S2P) showed higher binding affinity to hACE2 compared with BA.1 titre (NT50) in plasma of people who had recovered from BA.1 infection
and BA.2 spike, owing to L452R and R493Q reversion18,19. Despite this against BA.2.13, BA.2.12.1 and BA.4/BA.5 was reduced by 2.0-, 3.7- and
discrepancy, we conclude that BA.2 subvariants and BA.4/BA.5 are able 8-fold, respectively, compared with that for BA.1. Plasma from vac-
to maintain high binding affinities for hACE2. cinated individuals who had recovered from SARS infection showed a
marked decrease in neutralization of BA.2 subvariants, BA.3 and BA.4/
BA.5 compared with the other vaccinees (Fig. 2d and Extended Data
NAb evasion by BA.2.12.1, BA.4 and BA.5 Fig. 2b). This suggests that mutations in BA.2 sublineages, BA.3 and
To probe NAb evasion by the recently emerged Omicron sublineages, BA.4/BA.5 may enable escape from broad sarbecovirus-neutralizing
we performed pseudovirus-neutralization assays using D614G, BA.1, antibodies, which are substantially enriched in vaccinated people
BA.1.1, BA.2, BA.3, BA.2.12.1, BA.2.13 and BA.4/BA.5 against plasma previously infected with SARS22. Together, these observations indicate
obtained from individuals who had received three doses of SARS-CoV-2 that the BA.2.12.1 and BA.4/BA.5 display more potent and distinct
vaccine, vaccinated individuals who had recovered from BA.1 infection, humoral immune evasion than BA.1.
and vaccinated individuals who had recovered from severe acute res- Next, we examined the neutralizing activities of therapeutic anti-
piratory syndrome (SARS) (Supplementary Table 1). Plasma samples bodies against new Omicron subvariants (Fig. 2e). All seven tested
were collected four weeks after the booster shot or four weeks after Omicron subvariants displayed substantial evasion against neutrali-
discharge from hospital following COVID-19 illness. In plasma from zation by class 1 and class 2 RBD antibodies: the variants evaded neu-
individuals who had received an inactivated virus (CoronaVac) or RBD tralization by REGN-1093323 (casirivimab), LY-CoV01624 (etesevimab),
protein (ZF2001) booster six months after two doses of CoronaVac, LY-CoV55525 (bamlanivimab), COV2-21965 (tixagevimab) and BRII-19626
BA.1, BA.1.1 and BA.2 showed no significant difference in resistance (amubarvimab), whereas only BA.4/BA.5 evaded DXP-60415,27, which
to neutralization by plasma (Fig. 2a,b), concordant with previous showed reduced but still competitive efficacy against BA.1 and BA.2
reports20,21. However, we found that BA.2 subvariants BA.2.13 and subvariants. Two major differences in antigenicity were observed
BA.2.12.1 showed increased immune-evasion capability over BA.2— between BA.1 and BA.2 subvariants. First, NAbs targeting the linear
with BA.2.12.1 exhibiting greater evasion than BA.2.13—and BA.4/BA.5 epitope 440–4493, such as REGN-1098723 (imdevimab), COV2-21305
exhibiting even greater evasion (Fig. 2a,b). The decrease in neutraliza- (cilgavimab, a component of Evusheld) and LY-CoV1404 (bebtelovimab4)
tion was clearer in plasma obtained from individuals infected by BA.1 could neutralize BA.2 subvariants and BA.4/BA.5. Second, BA.2
0.20
BA.1 RBD–APC
BA.1 RBD–APC
BA.1 RBD–APC
CD20–FITC
CD20–FITC
CD20–FITC
0.15
0
WT–BA.1 BA.1
BA.1 RBD–PE WT RBD–BV605 BA.1 RBD–PE WT RBD–BV605 BA.1 RBD–PE WT RBD–BV605 cross- specific
reactive
c d e BA.1 breakthrough infection Vaccinated
WT convalescent or vaccinees convalescent SARS convalescent
DXP-604 Epitope
S2K146 615
group
LY-CoV016
BD55-1239 REGN10933 A
BD55-5514 B WT convalescent or vaccinees
DH1047
C 102 WT
ADG-2 LY-CoV1404 non-binders
LY-CoV555
D1
COV2-2130
t-SNE2
D2
DXP-593 614
S304 C110 E1 f ACE2 competition level g SARS-CoV-2 D614G SARS-CoV-1
S309 BD55-5840 E2.1
E2.2 Post-vaccination BA.1
FC08 BD-744 infection convalescent
E3
F1
S2H97 F2
411
F3
t-SNE1 Vaccinated
Competition level IC50 (μg ml–1)
SARS convalescent 0.001 0.1 10
Non-competing Competing
h G485
180° F486
E484 N487
A475
D2 D1 C B B A F3
G446 G446 F490 F456
K444 G504
K444 L452 L452 K417
D405
P499 R346
V503
E1 E2.1 E2.2 E3 3
E3 F1 F2
L452 I468
R346
K462 R408
A348 T376
R346
K356 L518 E516
T345
S383 K378
R357 K385
G339 R357
C391 T386
86
Fig. 3 | Isolation, characterization, and comprehensive epitope mapping of that bind WT SARS-CoV-2 RBD. Twelve epitope groups were identified on the
SARS-CoV-2 RBD antibodies. a, FACS analysis of pooled memory B cells basis of DMS of 1,538 antibodies. d,e, Epitope distribution and projection of
(IgM−CD27+) from plasma of individuals who have recovered from BA.1 antibodies from plasma of individuals who had recovered from infection with
breakthrough infection after vaccination, vaccinated individuals and the WT virus, individuals who have recovered from BA.1 breakthrough infection
unvaccinated individuals who have recovered from BA.1 breakthrough infection. after vaccination, and vaccinated individuals who had recovered from SARS.
The percentage of cells recognizing WT or BA.1 RBD are shown. b, The heavy f, ACE2 competition level determined by competition ELISA (n = 1,286) were
chain V domain somatic hypermutation (SHM) rate of BA.1-specific (n = 968) projected onto the t-SNE. g, Neutralizing activity against SARS-CoV-2 D614G
and BA.1–WT cross-reactive (n = 4,782) BCRs obtained from 10X scVDJ-seq from (n = 1,509) and SARS-CoV-1 (HKU-39849; n = 1,457). h, Average mutational
individuals who have recovered from BA.1 breakthrough infection after escape score projection of each epitope group on SARS-CoV-2 RBD
vaccination. Two-tailed Wilcoxon rank-sum test. Boxes show 25th percentile, (Protein Data Bank (PDB): 6M0J). All neutralization assays were performed as
median and 75th percentile, and violin plots show kernel density estimation biological duplicates.
curves of the distribution. c, t-SNE and unsupervised clustering of antibodies
Group D antibodies were most affected by the G446S mutation in BA.1, although group D2 NAbs displayed broad activities, their epitopes are
BA.1.1 and BA.3 (Fig. 4d); these NAbs therefore show higher potency not conserved among sarbecoviruses (Fig. 4d), similar to those of group
against BA.2 (Fig. 4a,b). However, group D1 antibodies showed reduced D1, E2.1 and E2.2 NAbs. This suggests that their breadth may be a result
efficacy against L452 substitutions, with L452M (BA.2.13) causing mild of their rarity in individuals who have recovered from infection with WT
escape, L452Q causing moderate escape (BA.2.12.1) and L452R (BA.4/BA.5) or BA.1 SARS-CoV-2 (Fig. 3f), and these NAbs may be the next target for
causing severe escape (Fig. 4c,d). By contrast, group D2 antibodies, SARS-CoV-2 to escape by evolving specific mutations on their epitopes.
especially those stimulated by BA.1 infection, showed exceptional Group E2 antibodies bind to the chest of the RBD35 (Fig. 4a), and their
broad and potent neutralizing activity against all Omicron subvariants, epitopes are focused around R346, A348, A352, K356, R357 and I468
for example, LY-CoV1404 (Fig. 4b and Extended Data Fig. 5). Notably, (Fig. 4d). Despite similar epitopes, group E2.1 NAbs, especially those
100 100
10–1 10–1
Pseudovirus IC50 (μg ml–1)
10–2 10–2
E2.2
10–3 10–3
G
.1
.1
.3
.2
.1
/5
.1
.1
.3
.2
/5
.1
.1
2.
14
14
BA
.1
BA
BA
BA
.1
BA
BA
.4
.4
346
348
352
354
356
357
426
439
443
444
445
446
447
448
449
450
452
468
484
490
494
496
499
500
.2
.1
.2
.1
BA
BA
BA
BA
D6
D6
BA
.2
BA
.2
BA
BA
Epitope group E2.1 Epitope group E2.2
***
0.33× NS ***
2.6× *
0.98× ***
1.3× ***
3× ***
13× *
1× NS NS NS ***
1.3× ***
1.8× ***
2.9×
WT R A ANKR PNS K VGGNYN L I E F SGP T
101 101 BA.1 R A ANKR PNS K V SGNYN L I A F S S P T
BA.1.1 KA ANKR PNS K V SGNYN L I A F S S P T
SARS-CoV-2
100 100 BA.3 R A ANKR PNS K V SGNYN L I A F SGP T
variants
10–1 10–1 BA.2 R A ANKR PNS K VGGNYN L I A F SGP T
Clade 1b
BA.2.13 R A ANKR PNS K VGGNYNM I A F SGP T
10–2 10–2 BA.2.12.1 R A ANKR PNS K VGGNYNQ I A F SGP T
10–3 10–3
BA.4/BA.5 R A ANKR PNS K VGGNYNR I A F SGP T
Delta R A ANKR PNS KV GGNYNR I E F SGP T
RaTG13 T A ANKR P K A K EGGN F N L I T YRGP T
G
.1
.1
.3
.2
/5
.1
.1
.3
.2
/5
.1
2.
.1
2.
14
14
BA
.1
BA
BA
BA
.1
BA
BA
.4
.4
T A ANKR PNS K VGGNYN L I E F SGP T
.2
.1
.2
.1 Pangolin-GD
BA
BA
BA
BA
D6
D6
BA
.2
BA
.2
BA
BA
SARS-CoV-1
c Epitope group D1 Epitope group D2 Epitope group E2.1 Epitope group E2.2
Urbani KP A E KKPR A T S T GNYNK I PWDGT T
variants
*** *** ***
Clade 1a
97.6× NS 13.4× 16.4×
PC4-127 KP A E KR PR A T S T GNYNK I PWGGT T
***
15.0× NS ***
4.25× ***
7.4×
Pseudovirus IC50 (μg ml–1)
Clade 3
BM48-31 S P A EMR PNS SNE F L E K SGQS
100 100 100 100 BtKY72 NP A E L R PNS KSGNN Y I S E SGP T
10–1 10–1 10–1 10 YN2013 RPAE T KPAVG S F L R T DPN
10–2 10–2 10–2 SC2018 RPAE I KPA TG S Y L Y T DPN
Clade 2
10
Rs4237 R P A E T KP AQG QY L R T DP T
10–3 10–3 10–3 10
Shaanxi2011 R P A E T KP AQG QY L Y T DP S
Rs4247 RPAE T KPA TG HY L Y T DPN
M
M
G
G
Q
Q
R
R
14
14
14
14
52
52
52
52
52
52
52
52
52
52
52
52
D6
D6
D6
D6
L4
L4
L4
L4
L4
L4
L4
L4
L4
L4
L4
L4
Fig. 4 | Spike L452 mutants can evade cross-reactive NAbs elicited by BA.1 representative potent NAbs in group D1 (n = 24), D2 (n = 12), E2.1 (n = 23) and
infection. a, Epitope of representative antibodies in group D1 (C110; PDB: E2.2 (n = 23) against SARS-CoV-2 spike L452 mutants. Geometric mean of the
7K8V), D2 (LY-CoV1404; PDB: 7MMO), E2.1 (BD-744; PDB: 7EY0) and E2.2 fold change in IC50 relative to D614G is shown above each plot. Two-tailed
(FC08; PDB: 7DX4). Residues highlighted in red indicate sites that are mutated Wilcoxon signed-rank test of paired samples. d, Average escape maps at escape
in Omicron variants. b, Neutralizing activity of NAbs in group D1 (n = 95), hotspots of antibodies in epitope groups D1, D2, E2.1 and E2.2, and the
D2 (n = 53), E2.1 (n = 90) and E2.2 (n = 161) against spike-pseudotyped SARS- corresponding multiple sequence alignment of various sarbecovirus RBDs.
CoV-2 variants. The geometric mean of the fold change in IC50 relative to BA.2 is The height of each amino acid in the escape map represents its mutation
shown above each plot. Two-tailed Wilcoxon signed-rank test of paired escape score. Sites that are mutated in Omicron subvariants are marked in
samples, in comparison to IC50 values versus BA.2. c, Neutralizing activity of bold. All neutralization assays were performed as biological duplicates.
339D
101 E1
100
K440 L441
10–1 T345 R346
F375 N343 glycan
10–2 N343 S371F
N343 glycan
S371L F2 408S
10–3 E340
N343 glycan
/5
.1
.1
.3
.2
.1
-1
.1
14
BA
.1
BA
BA
.4
oV
.2
.1
BA.1 RBD
BA
BA
BA.1 RBD
D6
-C
BA
.2
BA
BA.2 RBD
RS
SA
G104 F111
101 Y380
C379 D109 T376 Y508
334
337
339
340
345
346
365
374
375
376
378
383
384
404
405
408
439
440
441
448
503
504
505
508
K378 V503
100 D405 G504 G107 F110
R408
V503 V382 S106F377
10–1 T108F375 WT NPGE T R Y F S T KS PGDRNN L NVGY Y
Y508 P384 V54 BA.1 NPDE T R Y F F T KS PGDRNK L NVGHY
V58
K378 T376
SARS-CoV-2
10–2 F375 Y369 BA.1.1 NPDE T KY F F T KS PGDRNK L NVGHY
Y380
variants
Y93
Y34 BA.3 NPDE T R Y F F T KS PGNRNK L NVGHY
Y32
Clade 1b
10–3 Y369 BA.2 NPDE T R Y F F A KS PGNSNK L NVGHY
BD55-1239L BA.2.12.1 NPDE T R Y F F A KS PGNSNK L NVGHY
P384
BA.4/BA.5 NPDE T R Y F F A KS PGNSNK L NVGHY
/5
.1
.1
.3
.2
.1
-1
D405
.1
14
BA
.1
BA
BA
.4
oV
.1
BA
BA
D6
G504
-C
BA
.2
BA
G404 R408
SA
SARS-CoV-1
*** 0.02×
*** 0.18×
*** 0.16×
*** NS BD55-3372 BD55-3372H Urbani NPGE T KY F S T KS AGDRRN I N I GY Y
variants
0.002× NS NS NS
Pseudovirus IC50 (μg ml–1)
T71
101 D123 F76 BJ01 NPGE T KY F S T KS AGDRRN I N I GY Y
Clade 1a
G72 Sin852 NPGE T KY F S T KS AGDRRN I N I GY Y
Y122 G502 S73
100 Y505Q498 R408 WIV1 NPGE T T Y F S T KS AGDRRN I N I GY Y
T500 V503
D405
N501 D405 LYRa11 NPGE T T Y F S T KS AGDRRN I N I GY Y
D121 D120N50
10–1 Y505 Y508V407 Rs4231 NPGE T T Y F S T KS AGDRNS KNVGHY
Clade 3
V 503 Q498 BM48-31 QPNE T S Y F S T QS PGDRNS L E I GF Y
R408
10–2 Y508 BtKY72 NPGQSNY F S T KS PGDRNS V NVGY Y
Y52
Y112 ZXC21 NPHK T R Y F S T KS P F SR A KQ L E Y T
10–3
BD55-3372L YN2013 NPDS SR Y F S T KS P F SR ANQ L DY T
S114
Clade 2
SC2018 NPDKSR Y F S T KS P S SR A KQ V A Y T
Rs4237 NPDKSR Y F S T KS P S SR A KQ I E Y T
/5
.1
.1
.3
.2
1
-1
.1
2.
14
BA
.1
BA
BA
.4
oV
N439
.2
.1
BA
D6
-C
BA
.2
BA
RS
Fig. 5 | BA.2 subvariants can escape most broad-specificity BD55-3372 (group F3) (g). Antibody residues are shown in blue, and RBD
sarbecovirus-neutralizing antibodies. a–c, Neutralizing activity against residues are in black or red. Residues highlighted in red indicate sites that are
SARS-CoV-1 and SARS-CoV-2 subvariants by NAbs in group E1 (a; n = 70), mutated in Omicron variants. h, Average escape maps of antibodies in epitope
F2 (b; n = 171) and F3 (c; n = 69). The geometric mean of the fold change in IC50 groups E1, F2 and F3, and the corresponding multiple sequence alignment of
relative to BA.2 is shown above each plot. P-values were calculated using a various sarbecovirus RBDs. The height of each amino acid in the escape map
two-tailed Wilcoxon signed-rank test of paired samples, compared with the IC50 represents its mutation escape score. Sites that are mutated in Omicron
for BA.2. d, The epitope of Group E1 antibody BD55-3152 on the BA.1 RBD. subvariants are marked in bold. All neutralization assays were performed as
e, Overlay of BD55-5840 in the complex with BA.1 or BA.2 RBD. f,g, The epitope biological duplicates.
and interactions on the binding interface of BD55-1239 (group F2) (f) and
require investigation, as they may prove to be crucial for developing glycan moiety attached to N343, which in turn shifts the heavy chain of
broad-spectrum sarbecovirus vaccines and antibody therapies. BD55-5840 upward. This may explain the decreased binding between
To study how BA.2 subvariants, BA.3 and BA.4/BA.5 could system- BD55-5840 and S309, rationalizing their reduced neutralizing activity
atically reduce the neutralization efficacy of group E1 antibodies, we (Fig. 5a and Extended Data Fig. 9e). The N343 glycan is critically recog-
solved the cryo-EM structures of two group E1 BA.1-neutralizing anti- nized by almost all group E1 NAbs, including S309. Thus, this group of
bodies, BD55-3152 and BD55-5840, in complex with BA.1 spike proteins broad and potent NAbs is probably affected by the S371F mutation in
using cryo-EM (Fig. 5d and Extended Data Fig. 9a,b). Similar to S309, a systematic manner through displacement of the N343 glycan.
the epitope of group E1 antibodies includes an N-linked glycan on N343 The epitopes of group F2 and F3 antibodies cover a continuous
(Fig. 5d). Group E1 antibodies are also generally sensitive to mutation surface on the back of the RBD and can only bind to RBDs in the up
of G339, E340, T345 and especially R346, as indicated by their escaping configuration (Fig. 2b). To probe how BA.2 escapes group F2 and F3
mutation profiles (Fig. 5h). Notably, the newly acquired mutations of antibodies, we solved the cryo-EM structure of two representative
BA.2 do not overlap with the shared epitope of E1 antibodies, suggesting BA.1-neutralizing antibodies—BD55-1239 from group F2 and BD55-
that the systematic reduction in neutralization is not caused by amino 3372 from group F3—in complex with BA.1 and Delta spike protein,
acid substitution and is potentially owing to structural alteration. To respectively (Fig. 5f,g and Extended Data Fig. 9a). RBD mutations
explore this hypothesis, we further determined the cryo-EM structure on T376, K378 and R408 can lead to escape from neutralization by
of the prefusion-stabilized BA.2 spike in complex with the BD55-5840 group F2 antibodies (Fig. 5h). Indeed, these residues are centred on
Fab (Fig. 5e). A structural comparison with the BA.1 RBD binding to the core of the BD55-1239 epitope and are fairly conserved across sar-
BD55-5840 described above suggests that the 366–377 hairpin loop becoviruses (Fig. 5h). Notably, D405N and R408S, which are present in
displays significant conformational differences due to S371F and T376A Omicron BA.2 sublineages, may alter the antigenic surface, disrupting
mutations (Fig. 5e and Extended Data Fig. 9d). The overall positions of the binding of F2 antibodies (Fig. 5f) and completely abolishing the
residues 375 and 376 are displaced by more than 3 Å, which probably neutralizing capacity of F2 antibodies (Fig. 5b). Similarly, the D405N
further decreases the binding of group F2 and F3 NAbs in addition to and R408S mutations harboured by BA.2 subvariants could interrupt
the T376A side-chain substitution. As a result, the bulky phenylalanine the heavy chain binding of F3 antibodies, causing large-scale escapes
resulting from the S371F mutation interferes with the positioning of the of BA.1-neutralizing group F3 NAbs (Fig. 5c). These observations were
A Omi
512 cross-reactive
t-SNE1 Non-competing Competing
0.001 0.1 10
d Epitope group AOmi Epitope group BOmi DOmi F3Omi
*** 31×
NS 35× *** 36×
*** 36×
*** 260×
*** * NS 1.9×
NS 1.8× * 1.8×
* 490×
*** R498
101 101 K444 Y501
100 100
10–1 10–1
10–2 10–2 K440
Pseudovirus IC50 (μg ml–1)
10–3 10–3
/5
/5
ta
ta
a
.1
.1
.3
.2
.1
.1
.3
.2
1
-1
lta
-1
lta
G
G
ph
ph
.1
2.
.1
2.
14
14
Be
BA
.1
BA
BA
.4
Be
BA
.1
BA
BA
.4
oV
oV
De
De
.2
.1
.2
.1
BA
BA
BA
BA
Al
Al
D6
D6
-C
-C
BA
.2
BA
.2
BA
BA
RS
RS
f
SA
SA
Epitope group DOmi Epitope group F3Omi
NS *** 6.9×
NS 6.9× *** 7.7×
*** 8.4×
*** NS *** 110×
59× *** 120×
*** 130×>500×
*** *** 501N
101 101
A Omi 417K
505Y
100 100
10–1 10–1
10–2 10–2 484E
10–3 10–3 478T 486V
B Omi
/5
/5
ta
ta
a
.1
.1
.3
.2
.1
.1
.3
.2
1
-1
lta
-1
lta
G
G
ph
ph
.1
2.
.1
2.
14
14
Be
BA
.1
BA
BA
.4
Be
BA
.1
BA
BA
.4
oV
oV
De
De
.2
.1
.2
.1
BA
BA
BA
BA
Al
Al
D6
D6
-C
-C
BA
.2
BA
.2
BA
BA
RS
RS
**
SA
SA
10–3 10–3
498Q
F3 Omi
7K
0N
7K
0N
.1
.2
.1
1N
5Y
5D
.2
1N
5Y
5D
6G
8T
6G
8T
8S
8Q
8S
8Q
4E
4E
BA
BA
BA
BA
47
47
41
50
41
50
48
48
40
40
44
40
44
40
50
50
44
44
49
49
+K
+K
+N
+H
+N
+H
+A
+A
+R
+R
+K
+N
+K
+N
+Y
+Y
+S
+S
+R
+R
.1
.1
.1
.1
.1
.1
.1
.1
.1
.1
.1
.1
.1
.1
.1
.2
.1
.2
.1
.1
BA
BA
BA
BA
BA
BA
BA
BA
BA
BA
BA
BA
BA
BA
BA
BA
BA
BA
BA
BA
405
417
439
440
444
445
447
455
456
472
473
475
478
484
485
486
487
489
490
498
501
502
504
505
NS 0.12×
** 200×
NS 2.2× *** 3.1×
** NS 1.5×
* 8.5×
*** *** 6.4×
NS NS 6.9× *** *** NS
NS 2.8× NS NS *** 73×
NS 320× *** 6.9×
*** 110×
*** 16×
***
101 101 WT DKNNKVGL F I YA T EGF NY FQNGGY
100 100 Alpha DKNNKVGL F I YA T EGF NY FQYGGY
10–1 10–1 Beta DNNNKVGL F I YA T KGF NY FQY GGY
10–2 10–2 Delta DKNNKVGL F I YAKEGF NY FQNGGY
10–3 10–3 BA.1 DNNKKVGL F I YAKAGF NY F RYGGH
BA.2 NNNKKVGL F I YAKAGF NY F RYGGH
7K
0N
7K
0N
.1
.2
.1
.2
1N
5Y
5D
1N
5Y
5D
6G
8T
6G
8T
8S
8Q
8S
8Q
4E
4E
BA
BA
BA
47
47
41
50
41
50
48
48
40
40
44
40
44
40
50
50
44
44
49
49
+K
+K
+N
+H
+N
+H
+A
+A
+R
+R
+K
+N
+K
+N
+Y
+S
+S
+R
+R
.1
.1
.1
.1
.1
.1
.1
.1
.1
.1
.1
.1
.1
.1
.1
.2
.1
.2
.1
.1
BA
BA
BA
BA
BA
BA
BA
BA
BA
BA
BA
BA
BA
BA
BA
BA
BA
BA
BA
Fig. 6 | BA.1-specific antibodies elicited by BA.1 infection exhibit narrow groups, and corresponding multiple sequence alignment of various
specificity. a, Four epitope groups were identified among 102 BA.1-specific sarbecovirus RBDs. The height of each amino acid in the escape map
NAbs via k-means clustering and t-SNE of BA.1 RBD-based DMS profiles. represents its mutation escape score. Sites that are mutated in Omicron
b,c, Distribution of ACE2 competition level (b) and neutralizing activities (c) variants are marked in bold. WT-related escaping mutations are highlighted.
against BA.1. d, Neutralizing activities of BA.1-specific antibodies against g, Neutralizing activities of BA.1-specific NAbs against BA.1- or BA.2-based
pseudovirus with SARS-CoV-1 and SARS-CoV-2 spike variants (AOmi, n = 18; BOmi, pseudoviruses carrying single substitutions (AOmi, n = 18; BOmi, n = 30; DOmi,
n = 30; DOmi, n = 22; F3Omi, n = 32). The geometric mean of the fold change in IC50 n = 22; F3Omi, n = 32). The geometric mean of the fold change in IC50 relative to
relative to BA.1 is shown above each plot. e, Average mutational escape score BA.1 is shown above each plot. Wilcoxon signed-rank test of paired samples,
projection of each BA.1-specific epitope group on SARS-CoV-2 RBD (PDB: 7WPB). compared with IC50 for BA.1. All neutralization assays were performed as
f, Averaged escape maps at escape hotspots of the 102 NAbs in the four epitope biological duplicates.
further validated by neutralizing activity against spike-pseudotyped antibodies. By integrating the analysis of the entire dataset of 1,640
vesicular stomatitis virus (VSV) harbouring D614G/D405N and D614G/ SARS-CoV-2 RBD antibodies, we derived the embedded features of
R408S. As expected, group E1 antibodies were not affected, whereas the BA.1-specific NAbs and performed clustering and t-SNE analysis
group F2 and F3 antibodies displayed significantly decreased activity, (Fig. 6a). The 102 NAbs were clustered into four BA.1-specific epitope
following D405N or R408S single substitutions (Extended Data Fig. 9c). groups, which we designated AOmi, BOmi, DOmi and F3Omi, since these
Nevertheless, several group F3 antibodies, such as BD55-5514, are not groups are closely related to the corresponding WT epitope groups
sensitive to the D405N and R408S mutations of BA.2, making them good (Fig. 6a,e). These antibodies all compete for binding with ACE2 and
therapeutic drug candidates (Fig. 2e). In sum, S371F, D405N and R408S potently neutralize BA.1 but do not neutralize SARS-CoV-2 D614G or
mutations harboured by BA.2 and emerging Omicron variants may SARS-CoV-1 (Fig. 6b–d) because of the differences in the spike protein:
induce large-scale escape of NAbs with broad sarbecovirus specificity, N417K/Y501N/H505Y for AOmi, A484E/K478T for BOmi, K440N for DOmi
which are critical for the development of broad-specificity sarbecovirus and R498Q/Y501N for F3Omi, as indicated by average escape maps of
antibody therapies and vaccines. each group (Fig. 6e,f). Some of the previously circulating variants
also harbour these same mutations—such as N501Y in Alpha (B.1.1.7),
K417N/E484K/N501Y in Beta (B.1.351) and T478K in Delta—and only
BA.1-specific NAbs exhibit narrow breadths a small subset of the antibodies exhibit neutralizing activity against
In addition to the WT–BA.1 cross-reactive NAbs, we also investigated these variants (Fig. 6e). Moreover, nearly all of the BA.1-specific NAbs
the epitope distribution of BA.1-specific NAbs that do not react with showed poor cross-reactivity against other Omicron subvariants
WT RBD. We built a yeast display variants library based on the BA.1 (Fig. 6d). Specifically, most antibodies in the F3Omi and AOmi groups
RBD, and determined the escape mutation maps of 102 BA.1-specific are evaded by BA.2 subvariants and BA.3, possibly because of D405N,
the reference for V(D)J alignment, which can be obtained from https:// Author contributions Y. Cao and X.S.X. designed the study. Y. Cao, F.J. and X.S.X. wrote
support.10xgenomics.com/single-cell-vdj/software/downloads/latest. the manuscript with inputs from all authors. Y. Cao and F.S. coordinated the expression and
IMGT/DomainGapAlign is based on the built-in lastest IMGT antibody characterization of the NAbs. J.W. (BIOPIC), F.J., L. Zhao and H.S. performed and analysed
the yeast display screening experiments. Y.Y., T.X., P.W., J.W. (Changping Laboratory),
database, and we set the ‘Species’ parameter as ‘Homo sapiens’ and kept R.A., Yao Wang, J.Z., N.Z., R.W., X.N., L.Y., C.L., X.S., L. Zheng and F.S. performed the NAb
the others at the default settings. Public DMS datasets involved in the expression and characterization, including pseudovirus neutralization and ELISA. Y.Y., W.H.,
study from literature can be downloaded from https://media.githu- Q.L. and Youchun Wang prepared the VSV-based SARS-CoV-2 pseudovirus. A.Y., Yao Wang,
S.Y., R.A. and W.S. performed and analysed the antigen-specific single B cell V(D)J sequencing.
busercontent.com/media/jbloomlab/ SARS2_RBD_Ab_escape_maps/ S.D., P.L., Z.Z., L.W., R.F., Z.L., X.W. and J.X. performed the structural analyses. X.H., W.Z., D.Z.
main/ processed_data/escape_data.csv. Cryo-EM density maps have and R.J. recruited the SARS-convalescent donors and vaccinated individuals. X.C. and Z.S.
been deposited in the Electron Microscopy Data Bank with accession recruited the Omicron BA.1-convalescent donors. X.C., Y. Chai, Y.H. and Y.S. isolated PBMCs
from BA.1-convalescent donors. Q.G. proofread the manuscript.
codes EMD-33210, EMD-33211, EMD-33212, EMD-33213, EMD-33323,
EMD-33324, EMD-33325, EMD-32732, EMD-32738, EMD-32734, EMD- Competing interests X.S.X. and Y. Cao are inventors on the provisional patent applications for
32718 and EMD-33019. Structural coordinates have been deposited BD series antibodies, which includes BD30-604 (DXP-604), BD55-5840 (SA58) and BD55-
5514 (SA55). X.S.X. and Y. Cao are founders of Singlomics Biopharmaceuticals. The other
in the Protein Data Bank with accession codes 7XIW, 7XIX, 7XIY, 7XIZ, authors declare no competing interests.
7XNQ, 7XNR, 7XNS, 7WRL, 7WRZ, 7WRO, 7WR8 and 7X6A.
Additional information
Supplementary information The online version contains supplementary material available at
https://doi.org/10.1038/s41586-022-04980-y.
Code availability Correspondence and requests for materials should be addressed to Yunlong Cao,
Python and R scripts for analysing escaping mutation profile data and Zhongyang Shen, Youchun Wang, Xiangxi Wang, Junyu Xiao or Xiaoliang Sunney Xie.
Peer review information Nature thanks Patrick Wilson and the other, anonymous, reviewer(s)
reproducing figures in this manuscript are available at https://github. for their contribution to the peer review of this work.
com/jianfcpku/SARS-CoV-2-RBD-DMS-broad. Reprints and permissions information is available at http://www.nature.com/reprints.
a
BA.1
BA.1.1
BA.3
BA.2
BA.2.13
BA.2.12.1
BA.4/5
N764K
N856K
Q954H
N969K
D614G
N679K
D796Y
R408S
K417N
N440K
G446S
L452M
L452Q
L452R
S477N
T478K
E484A
F486V
Q493R
G496S
Q498R
N501Y
Y505H
T547K
H655Y
P681H
S704L
R346K
S371L
S371F
S373P
S375F
T376A
D405N
T19I
L24S
del25-27
A67V
del69-70
T95I
G142D
del143-145
N211I
del212
V213G
G339D
L981F
hACE2 + BA.1 spike
b c
40
Response (RU
BA.2_S6P BA.3_S6P 32
24
3,675 movies 3,651 movies
16
Template Picker Template Picker
8
1,449,818 particles 953,672 particles KD=3.05 nM
0
Class2D Class2D 0 100 200 300 400 500 600
Time (s)
636,412 particles 422,822 particles hACE2 + BA.2 spike
Heterogeneous Refine
Heterogeneous Refine 60
Response (RU)
40
20
Homogeneous
Refine GSFSC Resolution 3.07 Å KD=1.56 nM
1.0
0
No Mask (3.9 Å)
0
0.8
8 Loose (3.4Å)
0 100 200 300 400 500 600
GSFSC Resolution 3.25 Å
GSFSC Resolution 3.62 Å 1.0
No Mask (3.6 Å) 0.6
6 Tight (3 Å) Time (s)
1.0
No Mask (4.9 Å)
0.8 Loose (3.4 Å)
3.4 0.4
4
Corrected (3.1 Å) hACE2 + BA.3 spike
0.8 Loose (4.1 Å)
Tight (3.2 Å)
0.6 Tight (3.6 Å) 0.6
Corrected (3.2 Å)
3.2 0.2
2 100
Response (RU)
Corrected (3.6 Å) 3.0
4.0 3.8 0.4
0.4 3.07 Å 0.0
0 80
3.8 3.4 DC 17Å 8.3Å 5.5Å 4.2Å 3.3Å 2.8Å 2.4Å
0.2 0.2 185,916 particles
3.6 3.0 60
0.0 3.62 Å 3.25 Å 0.0 Block-based Classification
assification
DC 17Å 8.6Å 5.7Å 4.3Å 3.4Å 2.9Å 2.4Å 167,642 particles 182,332 particles DC 17Å 8.6Å 5.7Å 4.3Å 3.4Å 2.9Å 2.4Å Local Refinement
efinement 1.0
GSFSC Resolution 3.72 Å 40
No Mask (7.6 Å)
0.8 Loose (4.5 Å) 20
Tight (4.0 Å) KD=2.62 nM
4.5 0.6
0
BA.2.13_S6P BA.2.12.1_S6P 4.1
Corrected (3.7 Å)
0.4 0 100 200 300 400 500 600
3,752 movies 4,683 movies 3.7 0.2 Time (s)
Template Picker Template Picker 3.72
3 72 Å 0.0 hACE2 + BA.2.13 spike
85,941 particles DC 17Å 8.6Å 5.7Å 4.3Å 3.4Å 2.9Å 2.4Å
60
1,827,318 particles 1,554,387 particles
Response (RU)
48
Class2D Class2D BA.4/5_S6P (with N658S)
36
3,625 movies
474,728 particles 783,568 particles 24
Heterogeneous Refine Heterogeneous Refine Template Picker
12
625,843 particles KD=1.61 nM
0
Class2D 0 100 200 300 400 500 600
Time (s)
186,987 particles hACE2 + BA.2.12.1 spike
Heterogeneous Refine 100
Response (RU)
1.0
GSFSC Resolution 3.48 Å 80
GSFSC Resolution 3.49 Å No Mask (4 Å)
1.0
No Mask (3.9 Å) 0.8 Spherical (3.8 Å) 60
0.8 Loose (3.6Å) Loose (3.6Å)
0.6
0.6 Tight (3.4 Å) Tight (3.4 Å) 40
Corrected (3.5 Å) 0.4 Corrected (3.5 Å)
3.9
9 8
3.8
0.4 20
3.7
7 0.2
6
3.6 0.2
KD=1.77 nM
3.5
5 3.4
4 0.0 0
0.0
DC 17Å 8.6Å 5.7Å 4.3Å 3.4Å 2.9Å 2.4Å
3.48 Å DC 17Å 8.6Å 5.7Å 4.3Å 3.4Å 2.9Å 2.4Å
1.0
GSFSC Resolution 3.52 Å 0 100 200 300 400 500 600
3.49 Å 364,517 particles No Mask (3.9 Å)
0.8 Spherical (3.8 Å)
Time (s)
346,286 particles
0.6
Loose (3.6Å) hACE2 + BA.4/5 spike
Tight (3.5 Å)
3.9 0.4 Corrected (3.5 Å) 60 (with N658S)
Response (RU)
3.7 0.2
3.5
3.52 Å 0.0
DC 17Å 8.3Å 5.5Å 4.2Å 3.3Å 2.8Å 2.4Å
40
129,691 particles
20
KD=16.7 nM
0
0 100 200 300 400 500 600
d Time (s)
Extended Data Fig. 1 | Structures and ACE2 binding of emerging Omicron with S6P and R683A, R685A substitutions. c, Binding affinities of Omicron
subvariants spike glycoprotein. a, Mutations on the spike glycoprotein of variants spike trimers to hACE2 measured by SPR. SPR analyses were
SARS-CoV-2 Omicron subvariants. Residues that are not identical among conducted in biological duplicates. d, MD simulated interactions between
Omicron subvariants are colored red. b, Workflow to generate cryo-EM hACE2 and RBD of Omicron variants. Structures of the RBD from Omicron
structure of BA.2, BA.3, BA.2.13, BA.2.12.1, BA.4/5 spike glycoprotein trimer variants and hACE2 are shown as ribbons.
Article
a CoronaVac × 3
p<0.0001 p<0.0001 p<0.0001 p<0.0001 p<0.0001 p<0.0001 p=0.0002 p=0.053 p=0.003
2.3x 6.8x 7.0x 4.8x 3.7x 3.6x 2.2x n.s. 1.9x
104 1484
830 784 447
Plasma NT50
632 504 418
223
103 122 117 104 79
122 112 105
100
75
2
41
10
101
D614G BA.1 BA.1.1 BA.3 BA.2 BA.2.13 BA.2.12.1 BA.4/5 SARS-CoV-1
b CoronaVac × 2 + ZF2001
p=0.82 p=0.72 p=0.5 p=0.0005 p<0.0001 p<0.0001 p<0.0001 p<0.0001 p<0.0001
n.s. n.s. n.s. 0.38x 0.16x 0.18x 0.19x 0.28x 15x
101
D614G BA.1 BA.1.1 BA.3 BA.2 BA.2.13 BA.2.12.1 BA.4/5 SARS-CoV-1
102
101
D614G BA.1 BA.1.1 BA.3 BA.2 BA.2.13 BA.2.12.1 BA.4/5 SARS-CoV-1
Extended Data Fig. 2 | Different immunity backgrounds lead to distinct (n = 38) previous SARS-CoV-1 infection; c, individuals who received 3 doses
humoral immunity against Omicron subvariants. NT50 against SARS-CoV-2, CoronaVac (n = 40) or 2 doses CoronaVac with ZF2001 booster (n = 38). P-values
SARS-CoV-1 D614G and Omicron subvariants spike-pseudotyped VSV by were calculated using two-tailed Wilcoxon rank-sum tests and labeled above
plasma samples from a, individuals who received 3 doses CoronaVac with the bars. n.s., not significant, p > 0.05. All neutralization assays were conducted
(n = 50) or without (n = 40) BA.1 breakthrough infection; b, individuals who in biological duplicates. Geometric means are labeled. Error bars refer to
received 2 doses CoronaVac and ZF2001 booster with (n = 28) or without geometric standard deviations.
a
Antigen barcode
RBD+
RBD-APC
V(D)J library
b
BA.1 RBD-APC
WT+ in BA.1+
BA.1 RBD+ 75.5
BA.1 RBD-PE
BA.1 RBD-APC
BA.1 RBD-APC
BA.1 RBD-APC
BA.1 RBD-APC
Extended Data Fig. 3 | Workflow for the isolation and characterization of high-throughput deep mutational scanning. b, FACS strategy to enrich
SARS-CoV-2 RBD antibodies. a, Overall schematic of antibody identification BA.1/WT cross-reactive memory B cells or BA.1-specific memory B cells.
by single cell VDJ sequencing with feature barcodes and epitope analysis by
Article
SARS-CoV-1
Clade 1a
variants
SARS-CoV-1-related
sarbecovirus
Clade 1a
SARS-CoV-1-related
sarbecovirus
Clade 1b
BatCoV clade (3)
Africa/Europe
Asian non-ACE2-utilizing
BatCoV clade (2)
ELISA OD450
0 4
Extended Data Fig. 4 | ELISA reactivity against 22 sarbecovirus RBD. Shades of red indicate ELISA OD450 for each antibody against various sarbecoviruses
from different clades.
Extended Data Fig. 5 | Neutralizing activities of antibodies elicited by E2.1, n = 26; E2.2, n = 39; E3, n = 68; F1, n = 97; F2, n = 158; F3, n = 67). Geometric
SARS-CoV-2 BA.1 or wildtype. Neutralizing activity against SARS-CoV-2 mean titers (GMT) are annotated above each group of points, and error bars
D614G and Omicron subvariants pseudovirus by antibodies of each epitope indicate geometric standard deviation. P-values were calculated using
group from BA.1 convalescents (BA.1-stimulated. A, n = 30; B, n = 41; C, n = 20; two-tailed Wilcoxon rank-sum tests and labeled above the bars. n.s., not
D1, n = 49; D2, n = 17; E1, n = 11; E2.1, n = 64; E2.2, n = 122; E3, n = 57; F1, n = 80; significant, p > 0.05. NAbs in the boxed epitope groups showed substantial
F2, n = 13; F3, n = 2), and from wildtype convalescents or vaccinees neutralization potency changes against BA.2.12.1 or BA.4/5 compared to BA.1.
(WT-stimulated. A, n = 98; B, n = 55; C, n = 88; D1, n = 46; D2, n = 36; E1, n = 59; All neutralization assays were conducted in biological duplicates.
Article
a WT-stimulated b BA.1-stimulated
A, n=91 B, n=56 A, n=30 B, n=40
IGH
J2
IG IG
HJ HJ
1 3
3
IG
HJ
J4
IG
HJ
J3
H
IG
HJ
3
IG
IGH
4
C, n=75 D1, n=51 C, n=13 D1, n=57
IGH IGH
J1 J1
IG
HJ
IG 2 3
HJ HJ
IG
IG
3
4
HJ
HJ
IGH
IGHJ4
IGHJ3
IG
2
J3
D2, n=42 E1, n=56 D2, n=17 E1, n=11
IGH
J3 IG
HJ
3
IG
H
4
HJ
J1
IGH
4
IG
IGHJ
J3
E2.1, n=26 E2.2, n=37 E2.1, n=64 E2.2, n=120
IG 3
HJ HJ IG IG
IG
HJ
3 IG HJ 4 HJ
2 HJ 1
IG
IG 3
4
HJ
HJ
3
IG
IGH IG IGH
J3 IG HJ3 J3
3 HJ
HJ 1
IG
4
4
HJ
HJ
IGH
IG
IG
J2
IIG
GHHJ2
J1
IG
4
HJ
HJ
IGH
4
IGHJ
IGHJ2
IGHJ3
IGHJ1
IG
J3
Extended Data Fig. 6 | Heavy chain V-J genes of BA.1-stimulated and each epitope group. The number of NAbs is annotated above the chord plot.
WT-stimulated antibodies in each epitope group. Heavy chain V-J genes IGHV genes are annotated only if the corresponding number of antibodies is
combination of a, WT-stimulated antibodies. b, BA.1-stimulated antibodies or greater than one.
a Epitope group A (BA.1-stimulated) Epitope group A (WT-stimulated) b
*
0.45x n.s. n.s. n.s. n.s. n.s. ***
4.5x ***
0.0076x n.s. n.s. n.s. n.s. n.s. **
1 1 BA.1
10 10 stimulated
A
0 0
10 10
10 10 WT
stimulated
A
10 10
10 10 BA.1
stimulated
4G
.1
.1
/5
4G
.1
.1
/5
B
BA
.1
.4
BA
.1
.4
61
61
BA
BA
BA
BA
D
D
WT
Epitope group B (BA.1-stimulated) Epitope group B (WT-stimulated) stimulated
B
n.s. n.s. n.s. n.s. n.s. n.s. *** *** n.s. n.s. n.s. n.s. n.s. ***
17x
1 1
10 10 BA.1
0 0
stimulated
10 10 C
10 10
WT
Pseudovir
10 10 stimulated
C
10 10
352
417
420
449
452
455
456
460
470
472
473
475
476
480
483
484
485
486
487
488
489
490
493
494
502
4G
.1
.1
/5
4G
.1
.1
/5
BA
.1
.4
BA
.1
.4
WT AKDY L L FNT I YAGCVEGFNCYFQSG
61
61
SARS-CoV-2
BA
BA
BA
BA
BA.1 ANDY L L FNT I YAGCVAGFNCYFRSG
D
variants
BA.3 ANDY L L FNT I YAGCVAGFNCYFRSG
BA.2 ANDY L L FNT I YAGCVAGFNCYFRSG
Clade 1b
Epitope group C (BA.1-stimulated) Epitope group C (WT-stimulated) BA.2.13 ANDYML FNT I YAGCVAGFNCYFRSG
BA.2.12.1 ANDYQL FNT I YAGCVAGFNCYFRSG
** n.s. n.s. n.s. n.s. *
5.8x ***
100x ***
0.0047x n.s. n.s. n.s. n.s. n.s. * BA.4/BA.5 ANDYRL FNT I YAGCVAGVNCYFQSG
1 1
Delta AKDYRL FNT I YAGCVEGFNCYFQSG
10 10 RaTG13 AKDF L L FNT I YAGCQTGLNCYYYRG
Pangolin−GD ARDY L L FNT I YAGCVEGFNCYFQSG
SARS-CoV-1
0 0 Pangolin−GX AVD− L L FKT I YAGCQVGLNCYYERG
10 10
Clade 1a
Urbani AVDYKY L KNPFPDC− PA LNCYWNDG
variants
PC4−127 AVDYKY L KNPFPDC− PAPNCYWRGG
10 10 WIV1 AVDYKS L KNPFPDC− PAFNCYWNDG
Rs4231 AVDY LWVKN I YPGC− I GPNCYNRPG
Clade 3
BM48−31 AVD− FRFKNL FPSC− EGLNCYKASG
10 10 BtKY72 AVD− Y L FKNL YSAC I SQLGCYEKSG
Clade 2
YN2013 AVD− FSHKS − − − − − − − − ENGVRSTP
10 10 SC2018 AVD− YSHKSD− − − − − − −GNGVYSTP
Shaanxi2011 AVD− YSSKS − − − − − − − − ENGVYSTP
4G
.1
.1
/5
4G
.1
.1
/5
BA
.1
.4
BA
.1
.4
61
61
BA
BA
BA
BA
D
Extended Data Fig. 7 | Comparison of BA.1-stimulated and WT-stimulated *, p < 0.05; **, p < 0.01; ***, p < 0.001; n.s., not significant, p > 0.05. All
antibodies in group A, B and C. a, Neutralizing activity against SARS-CoV-2 neutralization assays were conducted in biological duplicates. b, Averaged
D614G and Omicron subvariants by BA.1-stimulated (A, n = 30; B, n = 41; C, escape maps at escape hotspots of BA.1-stimulated and WT-stimulated
n = 20) and WT-stimulated (A, n = 98; B, n = 55; C, n = 88) antibodies in Group A, B antibodies in group A, B and C, and corresponding MSA of various sarbecovirus
and C. Geometric mean of IC50 fold changes compared to IC50 against BA.2 are RBDs. Height of each amino acid in the escape maps represents its mutation
annotated above the bars. P-values were calculated using a two-tailed Wilcoxon escape score. Mutated sites in Omicron variants are marked in bold.
signed-rank test of paired samples, in comparison to IC50 against BA.2.
Article
a roup E3 b c
***
0.36x *
0.85x n.s. n.s. *
0.95x n.s. n.s.
1 RBD E3
10
0
10 Epitope
Group
10 R466 K462
P463
W353 F1
F464 P426
10 D427
R355 Y396
F429 D428
10 R357 S514
E516
357
361
365
369
374
378
381
383
384
385
386
390
391
392
393
394
396
428
462
464
465
468
514
516
518
519
N394 L518
4G
.1
.1
.3
/5
T393
SARS-CoV-2
BA
.1
BA
.4
H519 WT RC Y Y F KGS P T K L C F T N Y D K F E I S E L H
61
BA
BA
A520 BA.1 RC Y Y F KGS P T K L C F T N Y D K F E I S E L H
variants
D
Clade 1b
roup F1 S2H97 BA.2 RC Y Y F KGS P T K L C F T N Y D K F E I S E L H
BA.4/BA.5 RC Y Y F KGS P T K L C F T N Y D K F E I S E L H
*** n.s. n.s. n.s. n.s. n.s. n.s. Delta RC Y Y F KGS P T K L C F T N Y D K F E I S E L H
1 RBD RaTG13 RC Y Y F KGS P T K L C F T N Y D K F E I S E L N
10 Pangolin−GD RC Y Y F KGS P T K L C F T N Y D K F E I S E L N
Pangolin−GX RC Y Y F KGS P T K L C F T N Y D K F E I S E L N
SARS-CoV-1
0 Urbani K C Y Y F KGS A T K L C F S N Y DR F E I S E L N
variants
Clade 3 Clade 1a
10 BJ01 K C Y Y F KGS A T K L C F S N Y DR F E I S E L N
Sin852 K C Y Y F KGS A T K L C F S N Y DR F E I S E L N
WIV1 RC Y Y F KGS A T K L C F S N Y DR F E I S E L N
10 LYRa11 RC Y Y F KGS A I K L C F S N Y DR F E I S E L N
Q414 Rs4231 RC Y Y F KGS A T K L C F S N Y DN Y E L S E L N
G413 T376 S375 BM48−31 RC Y Y F QGS P T K L C F S S Y D K YG L S E L N
10 BtKY72 RC Y Y F KGS P T K L C F S S Y D K Y E I S E L N
D427 K378
F374 ZXC21 K C Y Y F KGS P S K L C F T S Y D K F E L S E L N
10 F429 C379 F377 Anlong112 K C Y Y F KGS P S K L C F T S Y D K F E L S E L N
P384 YN2013 K C Y Y F KGS P S K L C F T S Y D K F E L S E L N
G381 Y369 SC2018 K C Y Y F KGS P S K L C F T S Y D K F E L S E L N
Clade 2
S383
F392 T385 Rs4237 K C Y Y F KGS P S K L C F T S Y D K F E L S E L N
4G
.1
.1
.3
/5
K386N388 Rp3 K C Y Y F KGS P S K L C F T S Y D K F E L S E L N
BA
.1
BA
.4
61
Shaanxi2011 K C Y Y F KGS P S K L C F T S Y D K F E L S E L N
BA
BA
K528
D
Rs4247 K C Y Y F KGS P S K L C F T S Y D K F E L S E L N
S304 HKU3−1 K C Y Y F KGS P S K L C F T S Y D K F E L S E L N
Extended Data Fig. 8 | Antibodies of group E3 and F1 exhibit weak but were conducted in biological duplicates. b, Epitope of representative
broad-spectrum neutralization. a, Neutralizing activity against SARS-CoV-2 antibodies in group E3 (S2H97, PDB: 7M7W) and F1 (S304, PDB: 7JW0). Residues
D614G and Omicron subvariants by antibodies in group E3 (n = 125) and F1 highlighted in red indicate mutated sites in Omicron variants. c, Averaged
(n = 177). Geometric mean of IC50 fold changes compared to BA.2 are annotated escape maps at escape hotspots of antibodies in group E3 and F1, and
above the bars. P-values were calculated using a two-tailed Wilcoxon signed- corresponding MSA of various sarbecovirus RBDs. Height of each amino acid in
rank test of paired samples, in comparison to IC50 against BA.2. *, p < 0.05; the escape maps represents its mutation escape score. Mutated sites in
**, p < 0.01; ***, p < 0.001; n.s., not significant, p > 0.05. All neutralization assays Omicron variants are marked in bold.
a c
d
b
Extended Data Fig. 9 | RBD-binding structures and affinity of broad annotated above the bars. P-values were calculated using a two-tailed Wilcoxon
Sarbecovirus antibodies. a, Cartoon models of Cryo-EM structures of BD55- signed-rank test of paired samples. *, p < 0.05; **, p < 0.01; ***, p < 0.001; n.s., not
3152 in complex of BA.1 RBD, BD55-1239 in complex of BA.1 RBD, and BD55-3372 significant, p > 0.05. All neutralization assays were conducted in biological
in complex of Delta RBD. b, Workflow to generate refined structural model of duplicates. d, Conformational comparison between BA.1 and BA.2 RBD
BD55-3152 and BD55-1239 in complex of BA.1 RBD, BD55-3372 in complex of regarding the 366-377 hairpin. e, Biolayer interferometry analysis of Group E1
Delta RBD, and BD55-5840 in complex of BA.2 RBD. c, Neutralizing activity of antibodies S309 and BD55-5840 binding to Omicron BA.1 and BA.2 Spike
representative NAbs in group E1 (n = 68), F2 (n = 139) and F3 (n = 61) against trimer. Biolayer interferometry analyses were conducted in biological
SARS-CoV-2 D614G, in addition to D614G+D405N and D614G+R408S. duplicates.
Geometric mean of IC50 fold changes compared to IC50 against D614G are
Article
AOmi, n=18 BOmi, n=30 DOmi, n=22 F3Omi, n=32
IGHV3−23
IGHV4−39
IGHV5−51
43
IGHV1 8
IGHV3−
IGHV
1−69
IGH
IGH
−23
IGHV3−
IGH −11
58
IGH
IGH 3−1 1
IGH
IG
V3−
70
V 1−
V1−
V3
24
IGHV
IG
1−18
V1−
HV
2−
−24
V
V3
V1
HV 3−2 −2
1−
70
IGH
IG
IG V 4 −
IG
30
5−
30−
9
IGH
IG
5
HV
3−
−5
IG
2−
HV
HV 3
HV −30
−6
HV
HV
51
IG IG IG
V4
49
IG
HV
3
9
53
3−
HV
1− −3
HV
IG
IG HV
3
5
H
IG 3−
IG
66
4−
8
3−
IG
H H 3
HV 30 69 IG V4− −9
0
V4 9
−3 IG IG −4 1− IG H 3
IGH 0− HV
4− HV HV
4− IGH V4−3 1
V4 4 4 IG 59 V4 4
−31 5 IGH −39
IGH V2− IGH IGH
V3− V4
V4− IGH V5−
51 48 IGHV −4
61 7−4−
1
J6
6
IG
H
IG
HJ
IG
HJ
4
2
4
HJ
HJ
6
IGH
IGH
J3
IGH
IG
IGHJ4
IG
IGH
J3
J6
J3
Extended Data Fig. 10 | HV-HJ gene combination of BA.1-specific antibodies. Heavy chain V-J gene combination of BA.1-specific neutralizing antibodies in
BA.1-specific epitope groups AOmi, BOmi, DOmi and F3Omi.