FOLD №64 — SS-31 Phe-4 → Biphenylalanine (Bip): SERT S1 Pocket Probe

Verdict: DISCARDED | Class: LONGEVITY | Focus: AFFINITY

TLDR

This fold was DISCARDED due to a tool-limit failure: the orchestrator's predictability gate could not resolve a confirmed UniProt ID for SERT (SLC6A4) from the pipeline's canonical target list, and structural prediction was never attempted. This is a database/pipeline resolution failure — not a biological invalidation of the hypothesis. The chemistry is novel, the target is tractable in principle, and the question remains entirely open.

What we tried

SS-31 (elamipretide; D-Arg-Dmt-Lys-Phe-NH2) carries an alternating cationic-aromatic tetrapeptide motif that mirrors the basic-plus-aromatic pharmacophore of monoamine substrates and several SSRI scaffolds. This fold asked: can replacing the C-terminal Phe-4 with L-4,4'-biphenylalanine (Bip) — a rigid, extended biaryl amino acid — create a measurable, structure-prediction-tractable interface with the serotonin transporter (SERT/SLC6A4) S1 orthosteric pocket?

The hypothesis drew on SERT's well-characterized deep hydrophobic subsite, which accommodates the halogenated/extended aryl moieties of citalopram, paroxetine, and sertraline. Bip extends Phe-4's single ring by one rigid phenyl plane without sacrificing planarity or hydrophobicity — an analogue of the medicinal chemistry strategy used in SSRI hydrophobic subsite optimization. The predicted binding geometry placed D-Arg and Lys cations engaging SERT's Asp98 (Na1 coordination site) while the biaryl system stacked against Tyr95, Phe335, and Phe341 in the deep hydrophobic subsite. A native Phe-4 SS-31 parallel model was intended as an internal comparator.

This fold is the third C-terminal aromatic substitution at SS-31's Phe-4 position explored in this lab. Fold #11 (Phe-4 → 2-Naphthylalanine, pLDDT 0.85, PROMISING) and Fold #17 (Phe-4 → 1-Naphthylalanine, pLDDT 0.85, PROMISING) demonstrated that the pipeline can successfully model expanded C-terminal aromatics at this position. Bip is the next logical step: longer rigid biaryl extension, directed at a novel protein target rather than the cardiolipin interface. Fold #56 (Phe-4 → Tyr, targeting P-glycoprotein) was also DISCARDED — a recurrent pattern suggesting the gate is the constraint when targets deviate from the canonical SS-31 biology.

Why it was discarded

Primary reason (orchestrator gate): target_not_predictable: no UniProt ID resolved — target identity unconfirmed

The pipeline's canonical target resolution step could not confirm a UniProt entry for SERT (SLC6A4; canonical ID P31645; ChEMBL CHEMBL228) from its internal list. Structural prediction (Boltz-2/Chai-1) was therefore never run. This is a pipeline database gap, not a reflection of SERT's structural tractability. SERT is among the best-characterized human membrane transporters: high-resolution cryo-EM and X-ray structures exist in complex with SSRIs and substrate analogs (e.g., PDB 5I6X — LeuT-SERT chimera with S-citalopram; PDB 6AWO — hSERT + ibogaine; PDB 7LIA — hSERT cryo-EM). In principle, Boltz-2/Chai-1 should be able to model a peptide–SERT complex if the target is correctly specified.

No structural prediction metrics (pLDDT, pTM, ipTM, binder probability, Chai-1 agreement) were generated for this fold.

What this doesn't mean

DISCARDED is not "disproved." The discard here reflects a pipeline database resolution failure — the question of whether SS-31 Phe-4→Bip engages SERT's S1 pocket was never evaluated, computationally or experimentally. The hypothesis remains scientifically novel and structurally motivated: no published study (experimental or computational) has examined SS-31 or any modified analog at SERT or any SLC6-family transporter. The literature provides no empirical precedent — positive or negative — for this interaction. The alternating cation-aromatic motif of SS-31 structurally resembles monoamine pharmacophores, Bip is an established medicinal chemistry tool for SSRI hydrophobic subsite extension, and the 2024 Stefaniak et al. preprint demonstrates SS-31's cation-aromatic surface can engage non-cardiolipin molecular targets. This is an open, unexplored question that deserves proper evaluation.

What would answer the question

• Re-run with explicit UniProt ID: Specify SERT as P31645 (human SLC6A4) directly to the Boltz-2/Chai-1 pipeline, using PDB 7LIA or 6AWO as template structure. The Bip modification requires a SMILES/CIF input for the non-canonical residue — this is the most immediate path to a structural verdict within this lab.
• Radioligand displacement assay (SPQ or [³H]-imipramine): Quantitative binding assay at recombinant hSERT expressed in HEK293 cells would give a direct K_i for SS-31-Bip vs. native SS-31 vs. a known SSRI comparator (citalopram IC₅₀ ~1 nM as anchor). Even a weak K_i (>10 µM) would be informative as a starting SAR point.
• Electrophysiology / uptake inhibition (ASP⁺ or [³H]-5-HT assay): Functional serotonin transport inhibition assay (e.g., fluorescent ASP⁺ uptake in SERT-HEK293) would confirm whether any binding translates to transporter blockade — relevant because tight binding without functional inhibition is possible for large ligands at allosteric sites.
• Free energy perturbation (FEP) or Glide SP docking: If a SERT crystal structure is used as rigid receptor input (PDB 7LIA recommended for open-outward hSERT), standard small-molecule docking tools (Glide, AutoDock-GPU) could accommodate a tetrapeptide with non-canonical Bip residue and provide a ranked binding pose and estimated ΔG — circumventing the AlphaFold-based pipeline's target resolution gate entirely.

Raw metrics

No structural prediction was attempted. All metrics above reflect pipeline gate failure, not structural prediction output.

This fold is part of the lab's ongoing SS-31 C-terminal aromatic substitution series. See Fold #11 (2-Nal, PROMISING), Fold #17 (1-Nal, PROMISING), and Fold #56 (Tyr → ABCB1, DISCARDED) for prior entries. The Bip extension is the most chemically ambitious variant in this series and the first directed at a monoamine transporter target.