Le Drezen, C. | Search bioRxiv

bioRxiv2026

Distinct atypical chemokine receptor 1 determinants underlie bacterial toxins recognition and pore formation

Atypical chemokine receptor 1 (ACKR1) is one of the most promiscuous receptors in the human chemokine system, engaging structurally diverse chemokines through a conformationally flexible N-terminal tail. This same interface is exploited by pathogens, including Plasmodium vivax and Staphylococcus aureus (SA), via a compact sulfotyrosine code. Among pathogenic proteins recognizing ACKR1, the SA leukocidin pair HlgAB is a notable exception. HlgAB-mediated pore formation is only weakly competed by chemokines, the Duffy binding protein, or antibodies targeting the receptor's N-terminus, leaving open how HlgAB engage ACKR1. Combining structural biology approaches with cell-based assays, we show that HlgA and HlgB engage ACKR1 sulfated N-terminus with distinct affinities and site hierarchies, with a single higher-affinity site for sulfated tyrosine 41 present in HlgA but absent in HlgB. Unexpectedly, this N-terminal engagement is dispensable for pore formation; productive lysis instead requires a separate interaction between the toxins and ACKR1 extracellular vestibule. These results define a two-step recognition mechanism, toxin capture by the sulfotyrosine N-terminus followed by vestibule-dependent pore formation, and extend the view that ACKR1 promiscuity arises from distributed, ligand-specific use of multiple receptor surfaces rather than a single adaptable interface.

biochemistry↗

bioRxiv2026

Machine Learning-Guided Engineering of High-Affinity Cross-Reactive Antibodies with Minimal Mutations

Antibodies raised against human targets often fail to recognize their animal orthologs, limiting preclinical evaluation in relevant models. We developed a Deep Mutational Scanning (DMS)-coupled deep learning strategy to engineer potent cross-reactive antibodies with minimal sequence divergence. Starting from C4, a fully human anti-PD-L1 antibody with weak recognition of murine PD-L1, DMS identified substitutions that improved binding to both human and mouse antigens. Conventional recombination of beneficial mutations generated highly cross-reactive antibodies but required 13 to 15 substitutions. To reduce this mutational burden, a deep learning model trained on DMS-derived sequence-binding data was used to identify minimal mutation combinations predicted to retain high affinity. This approach yielded variants carrying only 4 to 5 substitutions, with in vitro and cellular binding properties comparable to highly mutated antibodies. Epitope mapping, structural modeling and in vivo assessment further confirmed that these engineered antibodies retained PD-1/PD-L1 blockade and demonstrated therapeutic activity in a mouse tumor model.

molecular biology↗