The researchers broke humanin down into a short piece (amino acids 5‑15) and showed it can still stick to the Alzheimer‑related amyloid‑beta peptide, but not as well as the full‑length humanin. Changing one building block (Leu11) makes it work worse, and a tiny piece of amyloid‑beta itself does little on its own. The work is all in test‑tube experiments, not in people, so it doesn’t give direct dosing or supplement advice.

Amyloid fibrils in Alzheimer's disease are composed of amyloid-β (Aβ) peptides of variant lengths. Humanin (HN), a 24 amino acid residue neuroprotective peptide, is known to interact with the predominant Aβ isoform in the brain, Aβ (1-40). Here, we constructed smaller segments of Aβ and HN and identified residues in HN important for both HN-HN and HN-Aβ interactions. Peptides corresponding to amino acid residues 5- 15 of HN, HN (5-15), HN (5-15, L11S), where Leu11 was replaced with Ser, and residues 17-28 of Aβ, Aβ (17-28), were synthesized and tested for their ability to block formation of the complex between HN and Aβ (1-40). Co-immunoprecipitation and binding kinetics showed that HN (5-15) was more efficient at blocking the complex between HN and Aβ (1-40) than either HN (5-15, L11S) or Aβ (17-28). Binding kinetics of these smaller peptides with either full-length HN or Aβ (1-40) showed that HN (5- 15) was able to bind either Aβ (1-40) or HN more efficiently than HN (5-15, L11S) or Aβ (17-28). Compared to full-length HN, however, HN (5-15) bound Aβ (1-40) with a weaker affinity suggesting that while HN (5-15) binds Aβ, other residues in the full length HN peptide are necessary for maximum interactions. L11 was more important for interactions with Aβ (1-40) than with HN. Aβ (17-28) was relatively ineffective at binding to either Aβ (1-40) or HN. Moreover, HN, and the smaller HN (5-15), HN (5-15 L11S), and Aβ (17-28) peptides, had different effects on regulating Aβ (1-40) aggregation kinetics.