Most HIV drugs act inside an infected cell. A smaller group works at the threshold, stopping the virus before it fuses with a cell at all. The catch is that the fusion machinery the virus uses keeps mutating, and the peptides built to jam it lose their hold. Enfuvirtide ↗, the first fusion inhibitor approved for HIV, is the textbook case. It works, then resistance erodes it. A team at the National Institute of Pathogen Biology in Beijing now reports in Molecular Therapy ↗ that adding a single amino acid to a newer inhibitor restored its grip against the very mutants that had shrugged it off.
The starting point was LP-98, a lipopeptide the group had already engineered to be far more potent than enfuvirtide. Lipopeptide means a short protein chain with a fatty tail that anchors it near the cell surface, keeping it where fusion happens. LP-98 was strong against ordinary HIV but weakened against resistant strains, the same failure mode that limits every drug in the class.
The fix targeted a specific slot. HIV fuses using a viral surface protein called gp41, which folds into a bundle with a deep greasy pocket. Fusion inhibitors that reach into that pocket bind harder. The researchers extended the front end of LP-98, its N-terminus (the leading tip of the peptide chain), to add residues that seat into the pocket.
The result was not what a "more is better" logic would predict. Adding two extra tryptophans, bulky residues meant to fill the pocket, sharply reduced activity. A single residue, an isoleucine at position 124, did the real work. Peptides carrying that one change, named LP-101 and LP-108, held or improved their activity across divergent HIV and SIV (the primate cousin of HIV that researchers use to model infection in monkeys) strains, and their potency against resistant mutants rose sharply instead of falling.
The animal data are the strongest part. LP-101 drove virus below the limit of detection, the point where standard assays can no longer find any, in mice engineered to carry a human immune system and infected with HIV, and in rhesus macaques infected with SIV. Two different host species, two related viruses, the same result.
The mechanism held up under structural work. Binding assays showed LP-101 gripped a lab-made mimic of the pocket, a fragment called N44, more tightly than its parent did. A crystal structure showed why. The modified N-terminus reached deeper into the gp41 pocket than the original. Molecular dynamics simulations, which model how the atoms move over time, showed the flexible tip adjusting its fit rather than sitting rigidly, which helps explain why it tolerates a target that keeps changing shape.
The caveats are the ordinary ones for work at this stage. Everything here is in cells and animals. No person has taken LP-101, and the jump from a humanized mouse or a macaque to a human trial is where most antiviral candidates stall. Resistance is also a moving target. A virus that escaped LP-98 by one route can, in principle, find another against LP-101, and the paper tests mutants that already exist rather than predicting the ones that do not yet. What the study shows cleanly is a design lesson. The gain came from one carefully placed residue reaching into a conserved pocket, not from piling on bulk, and the two-tryptophan misfire is a useful reminder that in peptide engineering the smallest change can be the one that matters.