Abstract
Abstract
Background.
TRIM21 is a cytosolic IgG Fc receptor and E3 ubiquitin ligase that mediates antibody-dependent intracellular neutralization. Its evolutionary dynamics across mammals, and the level of biological organization (gene, lineage, residue) at which diversifying selection is detectable, have not been resolved.
Objective.
To test for gene-wide and lineage-level episodic diversifying selection in mammalian TRIM21, to identify residues with the strongest site-level evidence, and to evaluate one structurally positioned residue for recurrent amino-acid change.
Methods.
Ninety-six mammalian TRIM21 orthologs were aligned as a 631-codon alignment and analysed on a JTT+G4 maximum-likelihood tree with HyPhy (BUSTED, FEL, FUBAR, SLAC, aBSREL) and PAML site models, with explicit alignment-column-to-human-residue verification and a matched-taxon cross-framework test. Residue 453 was examined by coordinate analysis of the TRIM21–IgG Fc structure (PDB 2IWG), ancestral-state reconstruction (IQ-TREE2, JTT), and four JTT recurrence null simulations (N = 100,000 each).
Results.
BUSTED supported gene-wide episodic diversifying selection (LRT = 266.78, p < 0.0001), and PAML M7 vs M8 on a taxon subset agreed (2ΔlnL = 80.28, p = 3.69 × 10⁻¹⁸). aBSREL flagged seven of 27 branches spanning several mammalian lineages, although one is a topology-flagged internal branch and the elephant signal is accession-dependent. In the full 96-taxon alignment, FEL identified six residues surviving FDR correction at q ≤ 0.05 (57, 252, 308, 336, 453, 455); V50 was borderline (q = 0.0516). R57 was the strongest statistical candidate (FEL q = 0.0028; FUBAR 0.983; SLAC p = 0.0022). Residue 453 (glycine in human) was a distinct kind of candidate: FEL q = 0.0145 and SLAC p = 0.042 with FUBAR just below threshold (0.894); a backbone contact 3.50 Å from Fc in 2IWG; the highest entropy (2.433 bits) among coordinate-defined interface residues; and two high-confidence G→R events (Bovidae, Ursidae). Observing at least two such events was statistically unusual under the specified JTT nulls (P = 0.00031–0.0329, depending on root conditioning and null type). Site-level support was not reproduced in an 8-taxon matched test, consistent with loss of power.
Conclusions.
Gene-wide episodic diversifying selection in mammalian TRIM21 is well supported, whereas residue-level attribution is graded and taxon-sensitive. R57 is the strongest statistical candidate and residue 453 is the strongest structurally grounded recurrent-evolution hypothesis. Neither is shown to be functionally adaptive; both generate testable predictions for stability, Fc-binding, and antibody-dependent intracellular neutralization assays.