Evolving antibody evasion and receptor affinity of the Omicron BA.2.75 sublineage of SARS-CoV-2

Summary SARS-CoV-2 Omicron BA.2.75 has diversified into multiple subvariants with additional spike mutations and several are expanding in prevalence, particularly CH.1.1 and BN.1. Here, we investigated the viral receptor affinities and neutralization evasion properties of major BA.2.75 subvariants actively circulating in different regions worldwide. We found two distinct evolutionary pathways and three newly identified mutations that shaped the virological features of these subvariants. One phenotypic group exhibited a discernible decrease in viral receptor affinities, but a noteworthy increase in resistance to antibody neutralization, as exemplified by CH.1.1, which is apparently as resistant as XBB.1.5. In contrast, a second group demonstrated a substantial increase in viral receptor affinity but only a moderate increase in antibody evasion, as exemplified by BN.1. We also observed that all prevalent SARS-CoV-2 variants in the circulation presently, except for BN.1, exhibit profound levels of antibody evasion, suggesting this is the dominant determinant of virus transmissibility today.

The raw IC 50 (the 50% inhibitory concentration) values for each mAb against each pseudovirus are summarized in Table S1, and the fold changes in IC 50 values compared with that of BA.2.75 are presented in Figure 3C.Overall, the non-RBD mAbs and class 4 RBD mAb (C1520, C1717, S3H3, and 10-40) generally did not have impaired neutralization activity against the BA.2.75 subvariants, with the exception of C1717, which had a 3.6-fold drop in neutralizing CH. demonstrated substantial resistance to some of the RBD class 3 mAbs after acquiring L452R, K356T, and R346T, respectively.The single mutation D1199N found in BA.2.75.2 did not alter the neutralization profile of BA.2.75, whereas the mutations K444M/T and F490S both impaired the neutralizing activity of some class 3 RBD mAbs.

Structural modeling of mAb-binding impairment in BA.2.75 subvariants
We conducted structural modeling to further investigate how mutations in the circulating BA.2.75 subvariants confer resistance or sensitization to mAbs against different epitopes (Figure 3D).One of these mutations, F486S, disrupted a common cation-p interaction with R97 of Brii-196 from VH3-53 gene class, 29 as well as the interactions with Y92 and W98 for S2E12.The other RBD mutations, R346T, K356T, K444M/T, L452R, and F490S, are located on the outer surface of the RBD and within the epitope cluster of class 3 mAbs, which likely explains their loss of neutralizing activity.Specifically, the K444M and K444T mutations abolished two salt bridges interacting with D56 and D58 in CDRH2 of LY-CoV1404, 6 and the K356T mutation weakened S309 by breaking the salt bridge and cation-p interaction.Additionally, the K356T mutation may introduce an N-glycan at N354 in RBD, which would reduce the accessibility of the surrounding residues due to steric hindrance, thereby conferring a degree of resistance to the RBD class 3 mAbs.R346T, L452R, and F490S mutations have previously been observed in BA.4.6, 2,30 BA.4/5, 3,21 and Lambda, 31,32 respectively.The R346T mutation removed the salt bridge and several hydrogen bonds with D95 and R96 in SP1-77, the L452R mutation created steric hindrance with I54 in XGv282, and F490S disturbed the cationp interaction with R74 in XGv282 6 (Figure 3D).R403K mutation, a novel substitution sensitizing BA.2.75 to some RBD-directed mAbs, could retain the interaction with G93 and form an extra hydrogen bond with Y33 in Omi-3 (Figure 3D).

Evolution of BA.2.75 subvariant phenotypes
Lastly, we sought to examine whether there were common trends in receptor-binding and antibody evasion phenotypes across the BA.2.75 sublineage, by plotting the fold increases in hACE2-binding affinity versus the fold increases in antibody evasion to sera from the bivalent boosted cohort (Figure 4C).Two general phenotypic combinations were evident: one group including CH.1.1 and DS.1 with substantially higher neutralization resistance but decreased ACE2 affinity and another group including BN.1 with substantially higher ACE2 affinity but only moderately increased neutralization resistance.Placed in the context of other Omicron subvariants (e.g., XBB.1.5 and BQ.1.1),it is quite  S2.
apparent that viral strains that are currently dominant in the circulation are primarily those with highest antibody evasion properties, again indicating this particular phenotype as the principal determinant of transmissibility in the population today (Figure 4C).

DISCUSSION
To gain a deeper understanding of the evolution of the emerging Omicron BA.2.75 sublineage, we systematically evaluated antigenic and viral receptor-binding properties of major subvariants within this sublineage.Our experimental and in silico analyses revealed several critical mutations, including three not previously described (K356T, R403K, and D574V) (Figure S1), which conferred different degrees of antibody resistance and receptor-binding affinity (Figures 2, 3, and 4).Five RBD mutations, R346T, K356T, K444 M/T, and F490S, impaired the neutralization activities of some class 3 RBD mAbs, F486S led to a large loss of neutralizing activities of classes 1 and 2 RBD mAbs, and D574V conferred resistance to all of the class 1 RBD mAbs tested.Our results indicate that the K356T mutation could potentially introduce an additional N-linked glycan at the N354 site, thereby enhancing the ability of BA.2.75.5 to evade RBD class 3 antibodies.Additionally, the D574V mutation may influence the ''up'' and ''down'' dynamics of the RBD.This process might be modulated by changes in the local conformation of SD1. 34,35Notably, the BA.2.75 subvariants with K444 M/T and R346T paired with F486S evaded authorized antibodies bebtelovimab and Evusheld, respectively, which poses a new threat to individuals who need them therapeutically or prophylactically.Interestingly, we made the novel observation that R403K sensitizes BA.2.75 to neutralization by some class 1, 2, and 3 RBD mAbs (Figure 3), while it substantially increases the receptor affinity.This affinity increase could be a compensatory mechanism to regain the fitness loss in receptor binding caused by mutations at F486 in the DS.1 subvariant, mechanistically similar to the R493Q mutation observed in BA.4/5. 3 In addition to R403K, our study shows that R346T, K356T, and D574V not only contributed to neutralization profile changes but also enhanced receptor-binding affinity, which sheds light on the continued co-evolution of immune evasion and factors affecting transmissibility of the virus (Figures 2, 3, and 4).This higher receptor-binding affinity could potentially compensate for lower antibody evasion properties and allow for expansion, as exemplified by BN.1.However, other transmissibility-related factors, such as cell-type tropism, syncytial formation, and viral load/titer in epithelial cells, need further investigation.These factors may co-evolve with immune evasion and receptor-binding affinity, potentially affecting viral transmissibility in the human population.
Overall, our investigations have shown that the evolutionary trajectory of BA.2.75 subvariants has diverged in two different directions: substantially higher neutralization resistance but slightly reduced ACE2 affinity, as seen in CH.1.1 and DS.1, and substantially higher ACE2 affinity but only moderately increased neutralization resistance, as seen in BN.1 (Figure 4C).Globally, and particularly in the US, UK, Ireland, and Singapore, BN.1 is being outcompeted by CH.1.1 (Figure 1C), although it remains unclear why the same has yet to occur in some Asian countries such as Japan.

Limitations of the study
Instead of live viruses, we used VSV-based pseudoviruses for the neutralization test.Although pseudovirus neutralization tests offer advantages, such as increased efficiency and enhanced biosafety, it could still be beneficial to confirm our findings with authentic virus evaluations.However, past studies have shown strong associations between pseudovirus and live virus neutralization tests in assessing antibody responses to SARS-CoV-2. 3,36,37Additionally, we may need to consider non-spike genes that can potentially influence biological functions related to virus transmissibility.

STAR+METHODS
Detailed methods are provided in the online version of this paper and include the following:

Sample collection
The sera samples were all collected at Columbia University Irving Medical Center or at the University of Michigan through the Immunity-Associated with SARS-CoV-2 Study (IASO), and the collections were conducted under protocols reviewed and approved by the Institutional Review Board of Columbia University or the Institutional Review Board of the University of Michigan Medical School. 39All subjects provided written informed consent.Sera from individuals who received three doses of either the mRNA-1273 or BNT162b2 vaccines are described in the text as ''3 shots WT''.Sera from individuals who received three doses of either the mRNA-1273 or BNT162b2 vaccines followed by a

Figure 1 .
Figure 1.Spike alterations and prevalence of BA.2.75 subvariants (A) Phylogenetic tree of selected BA.2.75 subvariants and current variants of concern (VOCs).The mutations on the branches showed the spike amino acid alterations of each variant.The recombination event for XBB from BJ.1 and BA.2.75 is denoted by the asterisk.The BA.2.75 subvariants are highlighted in red, and green boxes indicate current globally dominant variants with a frequency over 2%.(B and C) Proportions of SARS-CoV-2 VOCs (B) and frequencies of BA.2.75 subvariants among BA.2.75 (C) in GISAID from October 2022 to March 2023.The cumulative number of sequences in the denoted time period is displayed at the upper right corner of each graph.See also Figure S1.

Figure 3 .
Figure 3. Resistance of pseudotyped BA.2.75 subvariants to neutralization by monoclonal antibodies (mAbs) (A) Key spike mutations found in BA.2.75 subvariants.Mutations are highlighted in magenta.(B) Footprints of NTD-, NTD-SD2-, and SD1-directed neutralizing mAbs on spike, and RBD class 1 to class 4-directed neutralizing mAbs on RBD.Mutations found in BA.2.75 subvariants are highlighted in magenta.(C) Fold change in IC 50 values of BA.2.75 subvariants relative to BA.2.75, with resistance to neutralization highlighted in red and sensitization in green.Spike mutations found in each of the indicated subvariants in addition to BA.2.75 are highlighted in blue.(D) Structural modeling of the impact on mAbs for the F486S, K444M, K346T, K356T, L452R, and R403K mutations.Clash is shown as the red asterisk; the interactions are shown as yellow dashed lines.See also TableS1.

Figure 4 .
Figure 4. Neutralization of pseudotyped BA.2.75 subvariants by polyclonal sera from four clinical cohorts (A) Neutralization of pseudotyped D614G and Omicron subvariants by sera from four different clinical cohorts.''3 shots WT'' refers to individuals who received three doses of a COVID-19 WT mRNA vaccine, ''3 shots WT + bivalent'' refers to individuals vaccinated with three doses of the WT mRNA vaccine and subsequently one dose of a WA1/BA.5 bivalent mRNA vaccine, and breakthrough refers to individuals who received COVID-19 vaccines and were infected.The results are representative of those obtained in two independent experiments and shown as dots with geometric mean (red line).Values above the dots denote the raw geometric mean ID 50 values, and the sample size (n) for each group is shown on the lower left.The limit of detection is 100 (dotted line).Comparisons were made against BA.2.75, and the fold changes in ID 50 values are shown, with resistance to neutralization highlighted in red and sensitization in green.Statistically significant fold changes (p < 0.05, determined by using two-tailed Wilcoxon matched-pairs signed-rank tests) are highlighted in bold.(B) Antigenic map based on the neutralization data of ''3 shots WT + bivalent'' vaccinee sera.SARS-CoV-2 variants are shown as colored circles, and sera are shown as gray squares.The x and y axes represent antigenic units (AU) with one unit corresponding to a 2-fold serum dilution of the neutralization titer.(C) Changes to receptor-binding affinity and antibody evasion of Omicron BA.2.75 subvariants.The x axis illustrates the fold change in ACE2-binding affinity of the Omicron subvariants relative to the D614G strain.The y axis represents the relative immune evasion capability of Omicron subvariants in comparison to the D614G strain (fold change in geometric mean ID 50 over ''3 shots WT + bivalent'' cohort).Black dashed lines correspond to equivalence to BA.2.75.See also TableS2.

The electrostatic surface potential of the RBD in top view, with red and blue corresponding to negative and positive charges, respectively. The red line on the RBD surface indicates the footprint of ACE2. Black arrows indicate the surrounding mutations found in BA.2.75 subvariants. (C) The summarized profile of viral receptor affinities of spike proteins. Spike mutations found in each of the indicated subvariants in addition to BA.2.75 are highlighted in blue. Enhanced ACE2 affinities compared with that of BA.2.75 are highlighted in red, whereas reduced affinities are in green. The results shown are representative of those obtained in two independent experiments.
1.1.Class 1, 2, and 3 RBD mAbs exhibited diverse neutralization profiles.BM.4.1 and BA.2.75.7 (denoted as BM.4.1/BA.2.75.7 as they share an identical spike) partially or completely escaped neutralization by both class 1 and class 2 RBD mAbs because of the F486S mutation.BA.2.75.1 resisted RBD class 1 mAbs, due to the D574V mutation.BA.2.75.4,BA.2.75.5, and BA.2.75.6

TABLE d
B Sample collection B Cell lines d METHOD DETAILS B Plasmids B Protein expression and purification B Surface plasmon resonance (SPR) (Continued on next page)