A reverse mutation gives Europe's rising BA.3.2.2 branch a firmer hold on ACE2

The R493Q change is the part worth dwelling on, because it goes backwards. Position 493 mutated away from glutamine during earlier omicron evolution; in RE.2.2 it returns, and the researchers report that the restored glutamine forms an extra hydrogen bond with lysine 31 of human ACE2 ^[4]. That single contact is described as the primary determinant of the sublineage's stronger receptor engagement ^[4]. The measurement behind it is surface plasmon resonance on purified receptor-binding domain plus cryo-EM structures ^[3], which tells you about affinity of isolated proteins. Transmission between people is a separate question, and the study as reported does not measure it.

The immune escape result is the less predictable one. RE.2.2 evaded several antibody classes, but two broadly neutralizing antibodies that had gone quiet against earlier omicron lineages, S2K146 and L4.65, recovered potent activity against it ^[5]. Ternary cryo-EM structures point to G446D as the substitution that lets those antibodies bind ^[6]. Escape here works as a reshuffle: some antibodies lose their grip, others get it back. The authors read it as evidence that antigenic evolution is bounded by structural requirements ^[6].

Then the glycan. Glycoproteomic profiling by liquid chromatography-tandem mass spectrometry, combined with cryo-EM, found up to 26 N-linked glycosylation sites on the spike trimer ^[7]. One of them, at N529 on the receptor-binding domain, has not been observed in SARS-CoV-2 variants before ^[8]. It forms hydrogen bonds between protomers that hold the spike in its closed conformation ^[9]. Read alongside the affinity data, that is an odd pairing: a spike that grips the receptor harder while spending more of its time shut. The paper frames this as high affinity balanced against regulated entry ^[9]. A structure cannot settle whether the closed-state bias offsets the affinity gain in a real infection.

For anyone thinking about antigen design, the load falls on N529 and G446D. A new glycan on the RBD changes what an immunogen presents to B cells at exactly the surface most vaccines are built around, and G446D is the residue that brought two dead antibodies back ^[8][6]. Those are checkable claims. They are also checkable with reagents most serology labs already have.

The surveillance side is thinner than it looks. The lineage designation RE.2.2 covers the BA.3.2.2 branch that gained prevalence after a long stretch of low detection in European surveillance datasets ^[1], and BA.3.2.2 itself, derived from omicron BA.3, has spread quickly in several regions ^[2]. There are no prevalence figures, dates or countries in the source. Prolonged low detection followed by a rise fits cryptic circulation, and it also fits a lineage that was rare and then spread.

One reassuring negative: the team tested RE.2.2 against ACE2 proteins from a range of animal species and found its host range broadly comparable to the omicron variants used for comparison ^[10]. No new reservoir signal, on that assay.

The work came out of the Institute of Microbiology of the Chinese Academy of Sciences and was published in PNAS ^[11], with Linjie Li as first author ^[12].