DISTILLATION №83 — DISCARDED

MOTS-c Pro-12 → Aib substitution | AMPK α2 | YPR hinge rigidification

TLDR

This fold was DISCARDED due to a tool-limit failure at the protein-protein interface: Boltz-2 returned an ipTM of 0.179 for the Aib-12 MOTS-c / AMPK α2 complex, indicating the structural predictor could not converge on a stable, specific docking pose. This is not a biological invalidation of the hypothesis. A parallel complication from the literature — that MOTS-c's dominant established mechanism of AMPK activation is indirect (via AICAR accumulation), not through direct binding to the AMPK α2 catalytic subunit — adds biological uncertainty that the structural predictor cannot resolve either way.

What we tried

MOTS-c (MRWQEMGYIFYPRKLR) carries a proline at position 12 that breaks the otherwise potentially helical C-terminal YPRKLR segment. Prior MOTS-c folds in this lab — particularly DISTILLATION №19 (K13R substitution, PROMISING, pLDDT 0.63) — identified the Arg-13/Lys-14/Arg-16 cationic patch as the likely AMPK-engaging face. The hypothesis here was that replacing Pro-12 with α-aminoisobutyric acid (Aib) — a gem-dimethyl Cα residue that locks φ/ψ into helical values and is the canonical helix-nucleating non-natural amino acid in peptide chemistry — would allow the RKLR patch to project as a coherent helical face, increasing electrostatic complementarity with the acidic regulatory surface of AMPK α2.

This represents a distinct strategy from all prior MOTS-c folds: fold #19 optimized cationic patch character; fold #43 targeted proteolytic stability at the GYIF junction; fold #71 extended half-life via PEGylation; fold #25 explored membrane association via N-terminal myristoylation; fold #30 attempted helical pre-organization via an i,i+4 hydrocarbon staple across residues 5–9. The Aib-12 substitution is the first fold in this series focused on conformational rigidification at the YPR hinge specifically to improve AMPK interface geometry.

Why it was discarded

Boltz-2 returned an ipTM of 0.179 — far below any threshold for confident interface prediction (generally ≥ 0.5 for meaningful docking signal). The peptide monomer pLDDT of 0.618 is typical for MOTS-c folds in this pipeline (consistent with folds #19, #43, #71) and does not itself represent a failure, but the interface did not converge. No Chai-1 cross-validation was available, and the Boltz-2 affinity module produced no values.

This mirrors the outcome of DISTILLATION №30, where an all-hydrocarbon i,i+4 stapled MOTS-c variant was similarly discarded with pLDDT 0.60 and a failing ipTM against AMPK α2. The convergent pattern across folds #30 and #83 suggests that MOTS-c–AMPK α2 docking is currently below the resolution threshold of the structural predictors in this pipeline — likely because there is no deposited co-crystal structure to template against, and the peptide is short enough that Boltz-2 cannot confidently resolve a specific binding pose from sequence alone.

Additionally, the literature does not support a direct MOTS-c–AMPK α2 binding interaction: the mechanistic consensus (PMID:25738459, PMID:36677050, PMID:36761202) holds that MOTS-c activates AMPK via intracellular AICAR accumulation, not peptide-receptor contact. Even if Boltz-2 had returned a high ipTM, the biological interpretation of that result would require validation against this indirect-mechanism baseline.

What this doesn't mean

DISCARDED is not disproved. This fold failed because the structural predictor could not converge on a stable docking pose — a tool-limit outcome that says nothing about whether the Aib-12 substitution actually rigidifies the YPR hinge, whether the RKLR patch aligns better as a helical face, or whether MOTS-c binds AMPK α2 at all. The indirect AICAR-mediated mechanism is the dominant published model, but the identification of direct MOTS-c protein-protein interactions (LARS1, PMID:39321430) establishes that direct binding mode is biologically plausible. No published study has directly measured MOTS-c–AMPK α2 binding affinity or excluded a direct interaction. The hypothesis that Pro-12 → Aib improves helical geometry and C-terminal face presentation remains chemically sound and experimentally untested.

What would answer the question

• Biophysical binding assay (SPR or ITC): Surface plasmon resonance or isothermal titration calorimetry with recombinant AMPK α2 catalytic domain would directly determine whether native MOTS-c and the Aib-12 analog bind the catalytic subunit, and at what affinity — answering both the direct-binding question and the Aib modification question simultaneously.
• Solution NMR / CD spectroscopy: Circular dichroism or 2D NMR of native vs. Aib-12 MOTS-c in aqueous buffer (with and without TFE as a helix-promoting co-solvent) would directly measure whether Pro-12 → Aib rigidifies the C-terminal segment as hypothesized — the conformational premise of the fold.
• Cellular AMPK activation assay with direct vs. indirect mechanism dissection: Comparing MOTS-c and Aib-12 MOTS-c in a cell-free AMPK kinase assay (recombinant AMPK, no AICAR production pathway) vs. a cellular assay (with intact folate cycle) would distinguish direct from indirect mechanisms and whether the modification alters either pathway.
• Free-energy perturbation (FEP) or enhanced sampling MD: Classical or alchemical MD with an AMPK α2 structural template (PDB: 2Y94 or equivalent) would provide conformational and binding free-energy estimates for the Pro-12 → Aib substitution that are inaccessible to single-run AlphaFold-style predictors, especially for short peptides without template interfaces.

Raw metrics

All heuristic values are sequence-based estimates, not experimental measurements.