Characterization of SARS-CoV-2 Omicron BA.2.75 clinical isolates

The prevalence of the Omicron subvariant BA.2.75 rapidly increased in India and Nepal during the summer of 2022, and spread globally. However, the virological features of BA.2.75 are largely unknown. Here, we evaluated the replicative ability and pathogenicity of BA.2.75 clinical isolates in Syrian hamsters. Although we found no substantial differences in weight change among hamsters infected with BA.2, BA.5, or BA.2.75, the replicative ability of BA.2.75 in the lungs is higher than that of BA.2 and BA.5. Of note, BA.2.75 causes focal viral pneumonia in hamsters, characterized by patchy inflammation interspersed in alveolar regions, which is not observed in BA.5-infected hamsters. Moreover, in competition assays, BA.2.75 replicates better than BA.5 in the lungs of hamsters. These results suggest that BA.2.75 can cause more severe respiratory disease than BA.5 and BA.2 in a hamster model and should be closely monitored.

; the BA.2.75 subvariant rapidly became the dominant form in India and increased in prevalence in other regions suggesting that it had a selective advantage relative to BA. 5 Fig. 1a). The newly emergent forms within the BA.2.75 lineage remain globally relevant in January 2023, and global sampling reflects a remarkable diversity of forms descended from BA.2, BA.5, and BA.2.75 co-circulating with distinct regional patterns (Fig. 1, Supplementary Fig. 1a).
Compared with the ancestral strain Wuhan/Hu-1/2019 (Pango lineage A; Wuhan Hu-1), the Omicron BA.1 virus had more than 30 amino acid differences in the spike protein of SARS-CoV-2 including insertions and deletions ( Supplementary Fig. 1a) [by comparison, Delta (Pango lineage B.1.617.2) differed from the ancestral Wuhan Hu-1 by only 11 amino acids in its Spike] 1 . BA.2 differed from BA.1 at 27 Spike positions, and BA.5 differs from BA.2 by 5 amino acids in the S protein ( Supplementary Fig. 1a). We demonstrated that the pathogenicity of BA.1 and BA.2 sublineage viruses is comparable in animal models and attenuated compared with previously circulating variants of concern (VOCs), consistent with clinical data in humans 2,3 . In addition, our recent data suggest that BA.4 and BA. 5 have similar pathogenicity to that of BA.2 in rodent models 4 . SARS-CoV-2 initiates infection through the binding of the receptor-binding domain (RBD) of its spike protein to host cell surface receptors [i.e., human angiotensin-converting enzyme 2 (hACE2)]. BA.2.75 differs from that of BA.2 by nine amino acids in the Spike, including four in the RBD (i.e., G339H, G446S, N460K, and the wild-type amino acid at position Q493). Recent studies reported that the RBD of BA.2.75 has a higher binding affinity for hACE2 than that of BA.2 5,6 , raising the possibility that this property may increase the replicative ability and/or pathogenicity of BA.2.75. Moreover, in addition to the substitutions in the RBD, there are several amino acid differences in the other viral proteins of BA.2.75, which may also alter its replicative capability and pathogenicity ( Supplementary  Fig. 1b). Here, we assessed the replicative capacity and pathogenicity of authentic BA.2.75 subvariants isolated from COVID-19 patients in established COVID-19 animal models.

Transitions in Omicron variant prevalence throughout 2022
SARS-CoV-2 has undergone a series of variant transitions since the Omicron lineage was first observed in November of 2021. The initial global transition from Delta to the Omicron BA.1 lineage was extremely swift and was followed successively by waves of BA.2 and BA.5, with each variant, essentially replacing the previous dominant form (Fig. 1a). This may indicate that each variant has been more transmissible than the prior variant, particularly in settings with histories of prior infection and vaccination resulting in changes in immune status at the population level. BA.2.75, a BA.2 sublineage, was first detected in India in May of 2022, and since then has been increasing in sampling frequency in some parts of the world, with different BA.2.75 sublineages reaching high prevalence in different countries (Fig. 1b, Supplementary Figs. 2 and 3). As of this writing (Jan. 28,2023), there are 95,000 BA.2.75 lineage sequences in GISAID 7,8 , and these have been sampled globally. After BA.2.75 came to high prevalence in India and Nepal, and had begun to spread to other parts of Asia (e.g., Singapore), it peaked in prevalence and was replaced by the recombinant lineage XBB (Fig. 1b) 9 , hCoV-19/Japan/UT-NCD1757-1N/2022 (NCD1757), and hCoV-19/Japan/UT-NCD1759-1N/2022 (NCD1759). We confirmed that the S protein of all three isolates contained the nine additional amino acid changes (i.e., K147E, W152R, F157L, I210V, G257S, D339H, G446S, N460K, and Q493 (reversion)) ( Supplementary Fig. 1b) that distinguish the consensus form of BA.2.75 (https://cov.lanl.gov/components/ sequence/COV/pangocommonforms.comp) from a BA.2 isolate (hCoV-19/Japan/UT-NCD1288-2N/2022; NCD1288), which carries the most common circulating form of BA.2 in Spike. However, two of the isolates (NCD1757 and NCD1759) had a D574V substitution in the subdomain (SD), in addition to the nine mutations; this and several other distinctive mutations found in other proteins are summarized in Supplementary Fig. 1b.
We first evaluated the pathogenicity of the BA.2.75 isolates in wildtype Syrian hamsters, a well-established small animal model for the study of COVID-19 [10][11][12] (Fig. 2a), consistent with our previous observations 3,4 . By contrast, most of the animals infected with any of the three BA.2.75 isolates gained weight over the 6-day experiment, similar to BA.2-, BA.5-, or mock-infected animals. We also examined pulmonary functions in the infected hamsters by measuring Penh and Rpef, which are surrogate markers for bronchoconstriction and airway obstruction, respectively, by using a whole-body plethysmography system. Inoculation of hamsters with the BA.2, BA.5, BA.2.75 (NCD1757), or BA2.75 (NCD1759) isolate did not cause substantial changes in either Penh or Rpef at any timepoint post-infection compared to the mock-infected group. Infection with BA.2.75 (TY41-716) caused a slight increase in Penh at 3 and 5 dpi, although no statistically significant differences in Penh values were observed among BA.2-, BA.5-, and BA.2.75 (TY41-716)-infected animals. Consistent with our previous data, infection with B.1.617.2 caused significant changes in Rpef in comparison with the five Omicron isolates (Fig. 2b).   1, are increasingly sampled in most countries once they are established, and illustrate the analysis used to identify the countries of interest for display here. All data are from GISAID; the illustrations were created by using the "Embers" web-based tool at cov.lanl.gov 26 . We next assessed levels of infection in the respiratory tract of wildtype Syrian hamsters (Fig. 2c). Hamsters were intranasally infected with Consistent with our previous report 4 , the virus titers in the lungs of animals infected with BA.2 or BA.5 were lower than those in animals infected with B.1.617.2 [mean differences in viral titer = 5.0 or 4.2 and 2.2 or 3.4 log 10 (PFU/g) at 3 and 6 dpi, respectively], although the difference was not statistically significant between the BA.2-and B.1.617.2-infected groups at 6 dpi. The lung titers in the BA.2.75 (TY41-716)-infected groups were also lower than those in the B.1.617.2-infected groups [mean difference in viral titer = 2.1 and 1.7 log 10 (PFU/ g) at 3 and 6 dpi, respectively], although these differences did not reach statistical significance. The viral titers in the lungs of another BA.2.75 strain (NCD1757)-infected groups were similarly lower than those in the B.1.617.2-infected group at 3 dpi [mean differences in viral titer = 2.0 log 10 (PFU/g)]; however, animals infected with BA.2.75 or B.1.617.2 had similar titers in the lungs at 6 dpi. The lung titers in the BA.2.75 (TY41-716)-and BA.2.75 (NCD1757)-infected groups were higher than those in BA.5-infected groups [for BA.2.75 (TY41-716), mean differences in viral titer = 2.1 and 1.7 log 10 (PFU/g), at 3 and 6 dpi, respectively; for BA.2.75 (NCD1757), mean differences in viral titer = 2.2 and 2.8 log 10 (PFU/g), at 3 and 6 dpi, respectively]; however, the differences were not statistically significant among the three groups. At 3 dpi, the virus titers in the lungs were significantly higher in the respiratory tract of animals infected with BA.2.75 (TY41-716), compared to animals infected with BA.2 (NCD1288) [mean difference in viral titer = 2.9 log 10 (PFU/g)]; however, at 6 dpi, similar titers were detected in the lungs of animals inoculated with BA.2.75 (TY41-716) or BA.2. The viral titers in the lungs of the BA.2.75 (NCD1757)-infected groups were also higher than those in the BA.2 (NCD1288)-infected groups [mean differences in viral titer = 3.0 and 1.6 log 10 (PFU/g), at 3 and 6 dpi, respectively], although the difference was not statistically Virus titers in the nasal turbinates and lungs were determined by performing plaque assays with Vero E6-TMPRSS2-T2A-ACE2 cells. Vertical bars show the mean ± s.e.m. Points indicate data from individual hamsters. The lower limit of detection is indicated by the horizontal dashed line. Data were analyzed by using a one-way ANOVA with Tukey's multiple comparisons test (titers in the lungs at 3 dpi and nasal turbinates at 3 and 6 dpi) or the Kruskal-Wallis test followed by Dunn's test (titers in the lungs at 6 dpi). P values of <0.05 were considered statistically significant. Data are from one experiment. significant between the BA.2.75 (NCD1757)-and BA.2-infected groups at 6 dpi. Taken together, these results suggest that the replicative ability of BA.2.75 in the lungs of wild-type hamsters is higher than that of previous Omicron variants, including BA.2 and BA.5.
Histopathological findings in the lungs of SARS-CoV-2 BA. 2
Hamsters were intranasally inoculated with BA.2.75 (TY41-716), BA.2.75 (NCD1757), BA.5 (TY41-702), or B.1.617.2 (Delta, UW5250) and euthanized at 3 and 6 dpi for histopathological evaluation; representative images are shown in Fig. 3. This examination revealed that inflammation was not obvious in the lungs of either BA.2.75 (TY41-716)-, BA.2.75 (NCD1757)-, or BA.5-inoculated animals at 3 dpi; however, infiltration of inflammatory cells such as mononuclear cells and neutrophils was observed in peribronchial and peribronchiolar regions in these two groups at 6 dpi (Fig. 3a, b and Supplementary Fig. 4). It is noteworthy that focal pneumonia, characterized by patchy inflammation interspersed in alveolar regions, was observed in the lungs of both BA.2.75 (TY41-716 and NCD1757)-inoculated animals at 6 dpi. However, there was no obvious pneumonia in the lungs of BA.5-inoculated animals at the same timepoint. By contrast, in the lungs of the B.1.617.2-inoculated animals, peribronchial and peribronchiolar inflammation was prominent at 3 dpi, and extensive pneumonia with focal alveolar hemorrhage was observed in the alveolar regions at 6 dpi (Fig. 3a, b and Supplementary Fig. 4). In addition, we detected viral RNA and protein in the lung tissue of BA. BA.2.75 replication in the extrapulmonary organs of wild-type hamsters As described above, BA.2.75 differs from BA.2 by nine amino acids in the Spike; however, two (NCD1757 and NCD1759) of the three BA.2.75 strains we tested have a substitution, in addition to the nine substitutions ( Supplementary Fig. 1b). For the following experiments, we therefore characterized only the TY41-716 strain, as a representative of the most common circulating form of BA.2.75 with respect to Spike.
Previously, we and other groups have reported that ancestral SARS-CoV-2 is detectable in the brain, liver, spleen, heart, and kidneys of wild-type hamsters 10,11 . To examine whether BA.2.75 replicates in the extrapulmonary organs of wild-type hamsters, four hamsters per group were intranasally infected with 10 5 PFU of BA.2.75 (TY41-716), BA.5 (TY41-702), or B.1.617.2 (Delta: UW5250); at 5 dpi, the heart, kidneys, spleen, brain, and lungs of the infected animals were collected to measure infectious virus titers. Subgenomic viral RNAs in these organs also were examined by using quantitative RT-PCR (qRT-PCR). Both viral RNA levels and viral titers in the lungs of BA.2.75-infected hamsters were significantly higher than those in the lungs of BA.5-infected hamsters and significantly lower than those in the lungs of B.1.617.2infected hamsters at 5 dpi, consistent with our observations in wildtype hamsters infected with these viruses at days 3 and 6 after infection ( Supplementary Fig. 5). No infectious virus was detected from any of the extrapulmonary organs of the BA.5-or BA.2.75-infected animals ( Supplementary Fig. 5a). In contrast, virus was recovered from the heart of two of the four B.  Supplementary Fig. 5b). Viral RNA was detected in the heart of all four B.1.617.2-infected animals and the kidney of one of the four animals ( Supplementary Fig. 5b). These results suggest that BA.2.75 mainly replicates in the respiratory organs of wild-type hamsters.

Host responses in BA.2.75-infected wild-type hamsters
To evaluate the host responses after BA.2.75 infection, we measured several proinflammatory and innate cytokine gene expression levels in the lungs of hamsters at 3 dpi by using qRT-PCR. The expression levels of all four genes [i.e., Interleukin (Il)1β, Il6, tumor necrosis factor (Tnf)α, and interferon (Ifn)γ] tested were remarkably lower in the lungs of BA.2, BA.5-, or BA.2.75-infected animals than in those of B.1.617.2-infected animals, although no statistically significant differences in the expression levels of the Tnfα and Ifnγ genes were found between the BA.2.75-infected and B.1.617.2-infected groups. Among the three Omicron variant groups, the expression levels of all four genes in the lungs were slightly higher in the BA.2.75-infected group than in the BA.2-or BA.5-infected group, although these difference did not reach statistical significance (Supplementary Fig. 6) [for Il1β, mean foldchanges relative to Naive were 1.11, 0.97, and 0.88, respectively; for Il6, mean fold-changes relative to Naive were 1.78, 1.47, and 0.92, respectively; for Tnfα, mean fold-changes relative to Naive were 2.19, 1.53, and 1.83, respectively; and for Ifnγ, mean fold-changes relative to Naïve were 3.52, 1.39, and 1.27, respectively]. These results suggest that infection with BA.2.75 may induce higher proinflammatory responses in the lungs of infected hamsters compared with BA.2 or BA.5 infections, but the responses are attenuated compared to those in response to B.1.617.2 infection, which is consistent with the results of histopathologic examination and virus replication in the lungs of wild-type hamsters.

BA.2.75 infection in hACE2-expressing hamsters
We investigated the infectivity of BA.2.75 in respiratory organs by using a more susceptible model, specifically transgenic hamsters expressing hACE2 (Fig. 4). At 5 dpi, the virus titers in the lungs and nasal turbinates of hACE2-expressing hamsters infected with BA.2.75 (TY41-716) were lower than those in animals infected with To further investigate the replicative fitness of BA.2.75, we compared the growth of BA.2.75 in wild-type hamsters with that of BA.5, which is currently the dominant variant circulating globally. Wild-type hamsters were intranasally inoculated with 2 × 10 5 PFU of a mixture of BA.2.75 (TY41-716) and BA.5 (TY41-702) at ratios of 1:1, 1:3, 1:19, or 1:199. At 4 dpi, the proportion of each virus in the nasal turbinates and lungs of the infected hamsters was determined by using next-generation sequencing (NGS). The proportion was calculated on the basis of the differences between these two viruses across 6 regions in the S protein.
NGS analysis revealed that the proportion of BA.2.75 had increased in the nasal turbinates and lungs of all infected animals compared to that in each inoculum for any ratio, except for the lung samples from hamsters 2, 10, and 19 (Fig. 5). For animals inoculated with a 1:1 or 1:3 ratio of BA.2.75:BA.5, the lung and nasal turbinate samples showed a greater proportion of BA.2.75, except for the lung sample from hamsters 2, 9, and 10 ( Fig. 5a, b). Of note, even though the proportion of BA.2.75 in the inoculum was much lower than that of BA.5 (i.e., a 1:19 or 1:199 mixture of BA.2.75:BA.5), BA.2.75 became dominant in the lungs of four (#s 11, 12, 15, and 20) of the ten animals (Fig. 5c, d). Taken together, these results suggest that BA.2.75 may have greater replicative fitness than BA.5, especially in the upper respiratory tract.

Discussion
We previously showed that Omicron sublineage BA.2 and BA.5 variants exhibit similar pathogenicity in rodent models by using several clinical isolates, and showed that both variants are significantly attenuated compared to previous circulating VOCs 2,4 . Here, we evaluated the replication and pathogenicity of Omicron sublineage BA.2.75 variants in hamsters. Our data show that there are no substantial differences in weight change among hamsters infected with BA.2.75, BA.2, or BA.5 (Fig. 2a); however, viral titers in the lungs of BA.2.75-infected hamsters were higher than those in the lungs of BA.2-or BA.5-infected hamsters (Fig. 2c). In addition, in competition assays, BA.2.75 replicated better than BA.5 in the lungs (Fig. 5). Of note, in the lungs of BA.2.75-inoculated hamsters, we observed focal pneumonia, characterized by patchy inflammation interspersed in alveolar regions, indicating that BA.2.75 can cause mild pneumonia ( Fig. 3 and Supplementary Fig. 4). In contrast, BA.5 mainly affected the bronchi, resulting in bronchitis/ bronchiolitis, and did not cause obvious pneumonia (Fig. 3) 5,6 . SARS-CoV-2 enters cells in two distinct ways: by fusion of the viral lipid envelope with the target cell plasma membrane or fusion of the viral envelope with the endosomal membrane after internalization through the endocytic pathway [17][18][19] . The internalization of SARS-CoV-2 via the endocytic pathway is believed to be induced by the binding of the virus to ACE2 20,21 . The Omicron variants have been shown to preferentially utilize the endocytic pathway to enter cells 22,23 . In addition, previous studies have demonstrated that the enhanced binding affinity between ACE2 and the RBD increases the efficiency of SARS-CoV-2 entry 24,25 . Therefore, the higher ACE2 binding affinity of BA.2.75 may enhance its ability to infect the lungs, thereby allowing BA.2.75 to cause viral pneumonia in hamsters. Also, this higher ACE2 affinity of BA.2.75 may increase its competitive fitness compared to BA.5 in the respiratory tracts of hamsters, as observed in our in vivo competition assay (Fig. 5). Further investigations are required to understand the underlying mechanism of the increased fitness of BA.2.75, for example, the efficiency of virus replication including virus entry, viral RNA synthesis, and budding, and the ability to escape from local immune responses.
We note two key limitations in this study: (1) although hamsters are one of the most widely used animals that are known to be susceptible to SARS-CoV-2, including mice and non-human primates [10][11][12] , it is unclear whether the BA.2.75 variant causes more clinically severe respiratory disease than other Omicron variants in humans; and (2) our study was performed in immunologically naïve animals; however, many people have already acquired immunity to SARS-CoV-2 through natural infection and/or vaccination. Therefore, it remains unclear whether our data reflect the clinical outcome in patients with immunity against SARS-CoV-2. Clinical studies are needed to corroborate our findings in the hamster model.
In summary, our data show that, compared to BA.5 and BA.2, BA.2.75 can replicate efficiently in the lungs of hamsters and cause more severe respiratory disease. This higher replicative ability of BA.2.75 in the lower respiratory tract may affect the clinical outcome in infected humans. Figure 1, Supplementary Figs. 2, and 3 show transitions between variant forms, emphasizing the continuing expansion of the BA.2.75 variant sublineages that is indicative of a possible selective advantage over cocirculating forms in some geographic regions. Web-based updates of these figures based on the most recent GISAID data can be generated via the "Embers" and "Isotonic Regression" tools at the Los Alamos National Laboratory SARS-CoV-2 variant analysis website (https://cov. lanl.gov); the methods used here were first developed by Korber et al. 26 . To resolve the lineage groupings shown in Fig. 1, we first aligned all sequences in GISAID, and then defined all distinct forms of the spike protein in the GISAID sampled within the last two months. This provides a contemporary view of currently circulating common forms. These variant forms were then tallied according to their assigned Pango lineage designations (https://cov.lanl.gov/components/ sequence/COV/pangocommonforms.comp); often distinct Pango lineages share the precise form of the Spike protein. We next grouped Pango lineages that share the same baseline Spike for tracking purposes, and this enabled us to resolve five major distinctive sublineages of BA.2.75 ( Supplementary Fig. 1), with a sixth category of "others", which includes both more ancestral and rarer forms. Figure 1 was created using Embers based on those Pango lineage groupings, and displays running weekly counts and proportions of variants at different geographic levels.

Variant tracking strategies
The Isotonic Regression analysis used for Supplementary Figs. 2 and 3 explores the dynamics of the transition towards higher frequencies of BA.2.75 sublineages over time, and can be run through the cov.lanl.gov Isotonic Regression tool (http://cov-dev.lanl.gov/content/ sequence/ISORG/isorg.html). All countries globally that had a given BA.2.75 sublineage that was well enough established to have been sampled ten or more times are included in the analysis. A one-sided resampling statistic is used to evaluate whether increased sampling over time of a given sublineages is significant 7,8,26 . This was used to identify nations where BA.2.75 sublineages are currently common and increasingly sampled for inclusion in Fig. 1b. The full Isotonic Regression analysis of the two most common sublineages of BA.2.75, CH.1, and BN.1, are shown in Supplementary Figs. 2 and 3.
The sequence data sets for these analyses and figures were uploaded from GISAID 2023-01-27 7,8 . We are grateful for the global contributions of the scientific community to GISAID to enable this work, and the following is a summary of the data that was provided: GISAID Identifier: EPI_SET_230129sh (https://doi.org/10.55876/gis8. 230129sh, https://epicov.org/epi3/epi_set/230129sh?main=true). All genome sequences and associated metadata including the contributors of each individual sequence with details such as accession number, virus name, collection date, originating lab, submitting lab, and the list of authors in this dataset are published in GISAID's EpiCoV database. EPI_SET_230129sh comprises 6,947,638 individual genome sequences. The collection dates range from Jan 1, 2022 to Jan 25, 2023 and data were collected in 205 countries and territories, but the sampling after Jan 18, 2023 is sparse so those data are excluded from the analysis.

Animal experiments and approvals
Animal studies were carried out in accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. The protocols were approved by the Animal Experiment Committee of the Institute of Medical Science, the University of Tokyo (approval number PA19-75) and the Institutional Animal Care and Use Committee at the University of Wisconsin, Madison (assurance number V006426). Virus inoculations were performed under isoflurane, and all efforts were made to minimize animal suffering. In vivo studies were not blinded, and animals were randomly assigned to infection groups. No sample-size calculations were performed to power each study. Instead, sample sizes were determined based on prior in vivo virus challenge experiments.

Lung function
Respiratory parameters were measured by using a whole-body plethysmography system (PrimeBioscience) according to the manufacturer's instructions. In brief, infected hamsters were placed in the unrestrained plethysmography chambers and allowed to acclimate for 1 min before data were acquired over a 3-min period by using Fine-Pointe software.

Whole-genome sequencing
Viral RNA was extracted by using a QIAamp Viral RNA Mini Kit (QIA-GEN). The whole genome of SARS-CoV-2 was amplified by using a modified ARTIC network protocol in which some primers were replaced or added 30,31 . Briefly, viral cDNA was synthesized from the extracted RNA by using a LunarScript RT SuperMix Kit (New England BioLabs). The DNA was then amplified by performing a multiplexed PCR in two pools using the ARTIC-N5 primers and the Q5 Hot Start DNA polymerase (New England BioLabs) 32 . The DNA libraries for Illumina NGS were prepared from pooled amplicons by using a QIAseq FX DNA Library Kit (QIAGEN) and were then analyzed by using the iSeq 100 System (Illumina). To determine the sequences of BA.2.75 (NCD1757) and BA.2.75 (NCD1759), the reads were assembled by the CLC Genomics Workbench (version 22, Qiagen) with the Wuhan/Hu-1/ 2019 sequence (GenBank accession no. MN908947) as a reference. The sequences of BA.2.75 (NCD1757) and BA.2.75 (NCD1759) were deposited in Genbank with accession IDs: OQ326841 and OQ326844, respectively. For the analysis of the ratio of BA.5 to BA.2.75 after coinfection, the ratio of BA.2.75 to BA.5 was calculated from the 6 amino acid differences in the S gene between the two viruses. Samples with more than 300 read-depths were analyzed.

Quantitative reverse transcription PCR (qRT-PCR)
For detection of subgenomic viral RNA (sgRNA), RNA extraction was carried out with an QIAamp Viral RNA Mini Kit (Qiagen). Approximately 400-600 ng of total RNA including sgRNA was reversetranscribed into cDNA by using a QuantiTect Reverse Transcription Kit (Qiagen) with oligo dT primer. The cDNA equivalent to 10 ng of input RNA was amplified and analyzed by using PowerUp SYBR Green Master Mix (Life Technologies) on QuantStudio 6 Flex (Applied Biosystems) following the manufacturer's protocol. The primer sequences targeting sgRNA were as follows: 5′-CGGTGAACCAAGACGCAGTA-3′ (reverse primer targeting SARS-CoV-2 Nucleocapsid gene 33 ). The results are presented as the mean cycle threshold (Ct) value technically performed in triplicate.

Statistical analysis
GraphPad Prism software was used to analyze the data. Statistical analysis included the Kruskal-Wallis test followed by Dunn's test, and an ANOVA with post hoc tests. Differences among groups were considered significant for P values <0.05.

Reporting summary
Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Data availability
All data supporting the findings of this study are available in the paper. There are no restrictions in obtaining access to primary data. All data generated in this study are provided in the Supplementary Information/Source data file. The sequences of BA.2.75 (NCD1757) and BA.2.75 (NCD1759) determined in this study have been deposited in Genbank with accession IDs: OQ326841 and OQ326844, respectively. Source data are provided with this paper.

Materials availability
All reagents described in this paper are available through Material Transfer Agreements.