As the COVID-19 pandemic approaches its third year, most countries have undergone major changes in their fight against its etiological agent, coronavirus SARS-CoV-2. In most countries, stringent health care regulations have shifted to a reactive, rather than preemptive, set of measures, whilst major governmental surveillance efforts give way to self-testing and restrictions are lifted [1]. Consequently, as 2023 draws near, an unprecedented variability of Omicron-only subvariants dominates an ever-shifting global landscape [2, 3] that highlights the utmost importance that genomic surveillance plays nowadays. The evolution of the SARS-CoV-2 genome has been described for more than two years by an unprecedented global genomic surveillance effort (13,784,411 sequences have been deposited into GISAID [4] by November 7, 2022). This worldwide cooperation towards open data has allowed scientists from all countries to contribute towards the characterization of major evolutionary paths that the virus has transited in response to natural and vaccine-acquired immunity [5]. Here, we present BW.1 (an alias for BA.5.6.2.1), an Omicron subvariant descending from the BA.5.6.2 lineage, that may have arisen in Mexico in early July and rapidly became dominant in the Yucatan Peninsula in October (Fig. 1). BW.1 carries significant immune escape mutations that are shared with those found in the BQ.1 variant, one of the most rapidly spreading lineages that have been described to date (namely, mutations S:K444T, S:L452R, S:N460K and S:F486V).
Surveillance In Mexico
Within Latin America, Mexico has been one of the countries that have sequenced the most SARS-CoV-2 genomes (79,455, as of the time of writing), second only to Brazil (189,523). This has been in part thanks to an unprecedented collaborative sequencing effort carried out by the government (INMEGEN, InDRE) and multiple academic institutions (CoViGen-Mex). In spite of the uncertain future of this cooperation, the resulting sequences have enabled the study of the variant turnover in the country throughout the three years of the pandemic [6, 7]. As pointed out in previous studies [8, 9], the Yucatan Peninsula (States of Campeche, Quintana Roo, and Yucatan) is one of the primal entry points for SARS-CoV-2 variants in Mexico due to its large influx of tourists and commercial activities in the areas surrounding Merida and Cancun. Currently, Mexico is undergoing a period of low SARS-CoV-2 transmission after the last BA.2/BA.5 driven wave (30,608 average daily cases were registered as cases peaked on July 7, 2022). Still, genomes from samples collected in Yucatan have had a large prevalence of the BW.1 variant in past weeks. Coincidentally, during October, the Yucatan Peninsula registered a rapid increase of cases that may foreshadow the onset of a new epidemiological surge in the country [10]. During this same period and as shown in Fig. 1, the BW.1 variant managed to outcompete every other variant in the region, including its parental variant BA.5.6.2.
Origins Of The Bw.1
The genomic analysis of the BW.1 variant suggests it might have originated in Mexico during its fifth epidemiological wave (June to September, 2022), derived from BA.5.6.2 viruses circulating in the State of Yucatan, which may have been in turn imported from the United States as shown in Fig. 2. All BW.1 sequences are placed in a monophyletic clade in the phylogenetic reconstruction, owing to an additional pair of mutations from those observed in BA.5.6.2 sequences that circulated in Mexico. Even though intermediate evolutionary steps may be missing due to an incrementally smaller number of total genomes sequenced in Mexico as cases dwindled after July, there are different telltale signs supporting this hypothesis. Out of 105 BW.1 genomes in GISAID on November 7, 2022, [4], 95 (90.47%) were collected in Mexico, the earliest of which were collected during epidemiological week 27 (early July) (Supplementary Table 1) In contrast, the earliest samples from the rest of the world were collected between weeks 37 and 39 (September in the Netherlands and the UK, respectively). In Mexico, most samples of BW.1’s parental variant, BA.5.6.2, were collected in the same region of Yucatan (83.33% (n = 20); Fig. 2).
BW.1 carries two synonymous substitutions that can be traced back specifically to Mexican sequences of BA.5.6.2 genomes. In fact, as shown in Fig. 3, these are, precisely, the only two mutations that make Mexican BA.5.6.2 different from the ones sequenced in the rest of the world. More specifically, nucleotide transition T7666C (ORF1a:D2467D), has been reported in 2,458 sequences worldwide (of which 41.98% belong to Omicron and 69% belong to Delta) and is found in 82.48% of all BA.5.6.2 sequences worldwide but is completely absent in Mexican BA.5.6.2 genomes (and consequently, in those from BW.1) as the mutation was first reported in week 24 (mid-June), possibly after the arrival of BA.5.6.2 to Mexico (Supplementary Fig. 1). On the contrary, in Mexico, all BA.5.6.2 and BW.1 sequences share synonymous substitution C14599T (ORF1b:L378L) whereas only 4.00% of BA.5.6.2 from the rest of the world have it. This mutation has been reported in 84,374 genomes worldwide, 50.56% of which were identified as Omicron BA.4 and 22.74% as Delta sequences.
Most genomic sequences of BW.1 differ from their parental BA.5.6.2 by two point mutations (Fig. 3). Nucleotide transition G2144A, coding for ORF1a:V627I, is present in 98.09% of the BW.1 genomes (93 from Mexico and 10 from the rest of the world) and has only been described in 5,419 sequences worldwide, with 45.65% of these sequences belonging to the Omicron and 27.05% to Delta variants. More importantly, none of parental BA.5.6.2 genomes had it. Transversion T22942G (coding for mutation S:N460K) is present in 92.38% of BW.1 sequences (87 from Mexico and 10 from the world). None in the parental BA.5.6.2 group had this mutation but it was first observed in BW.1 genomes from Mexico in early August (week 31 of 2022; Supplementary Fig. 1), becoming fixated later that month (week 34, late August), further supporting the hypothesis of its Mexican origin. The remaining ten sequences from the rest of the world were collected subsequently, and they all have this mutation (Fig. 3).
Important Mutations Found In Bw.1
At the time of writing (November 7, 2022), the most successful variants worldwide descend from variant BQ.1 (including its rapidly spreading subvariant BQ.1.1), derived from Omicron lineage BA.5.3 (BQ.x variants currently account for 25.16% of daily cases, although they are expected to dominate the viral landscape in Europe and the United States in the near future). BW.1, in contrast, (a regional variant, currently dominant in Mexico), descends from Omicron lineage BA.5.6 but their genomes share a common evolutionary history with BQ.x variants as they both descend from the BA.5 lineage. From this parental lineage, they inherited over 54 changes, including key mutations S:L452R (known to confer spike stability, viral fusogenicity and increased infectivity [11]) and S:F486V (shown to decrease the effectivity of multiple monoclonal antibodies [12]), in the receptor binding domain (RBD) of the spike protein (S), a very important region for antibody recognition [13].
In recent months, immune selection has seemingly become one of the major driving forces behind the fixation of additional key genomic mutations through parallel and convergent evolution, particularly of those in the RBD. Due to the pressure of the incremental immunity in the population, both the BQ.1 and BW.1 variants have fixated two additional key RBD mutations: S:K444T and S:N460K. It has been reported that mutation S:K444T enhances viral resistance to bebtelovimab and P2G3 in Delta and hinder antibody recognition in Omicron BA.4 subvariants [14], as well as neutralization resistance, and evasion of Class 3 antibody recognition in BQ.1, and BQ.1.1 [15]. This mutation represents an instance of parallel evolution in BW.1 and BQ.1 (Fig. 3). Mutation S:N460K, present in both variants through convergent evolution, has been suggested as a driver for enhanced fusogenicity [15], syncytia formation, enhancement of S processing in S1 and S2 subunits on BA.4 and BA.5 Omicron variants, as well as enhanced neutralization resistance with subvariants BQ.1, BQ.1.1, BA.2 and BA.2.75 [15, 16]. Immune escape has been shown to be more efficient whenever S:K444T and S:N460K co-ocurr, as in BQ.1 and BW.1,impairing the effectiveness of monoclonal antibodies Evusheld [3] and bebtelovimab [14].
Of note, other phylogenetically distant lineages have also independently acquired these two mutations, such as the highly infectious recombinant XBB, which is derived from BA.2.75.3 and BA.2.10.1 variants, whose genomes includes mutation S:N460K. Similarly, multiple phylogenetically distant subvariants of the BA.2 (such as BR.4 and CH.1), BA.4 (BA.4.6.3, CS.1) and BA.5 (CK.1, DB.1) clades have acquired different substitutions resulting in variations in the S:K444 amino acid. Ultimately, predicting the dynamics of future outbreaks remains nearly impossible as it depends on very specific regional conditions. However, genomic surveillance allows for preemptively tracking important mutations which may contribute towards enhanced viral phenotypes that might pose an actual threat to human health as the current study suggests for BW.1.