FOLD №31 — Tirzepatide Aib-2 Substitution

Verdict: PROMISING | Class: METABOLIC | Modification: Non-canonical amino acid

Mechanism of action

Tirzepatide is a 39-residue synthetic peptide dual agonist of the GIP receptor (GIPR) and glucagon-like peptide-1 receptor (GLP-1R), approved for type 2 diabetes and obesity management. Its N-terminal sequence (Tyr1-Ala2-Glu3-Gly4-Thr5...) is GIP-derived and engages the orthosteric transmembrane binding pockets of both GIPR and GLP-1R to activate Gαs-coupled cAMP signaling, driving insulin secretion, glucagon suppression, and CNS-mediated appetite reduction. The intact N-terminal pharmacophore — particularly Tyr1, Glu3, and Gly4 — is required for receptor activation at both targets. Metabolic stability is primarily conferred by a C20 fatty diacid conjugated to Lys-20, enabling albumin binding and a ~5-day plasma half-life that supports once-weekly dosing. However, the Tyr1-Ala2 bond remains the canonical DPP-4 scissile bond, and proteolytic removal of the N-terminal dipeptide eliminates receptor agonism regardless of lipidation status.

Performance applications

Tirzepatide's clinical efficacy in the SURPASS and SURMOUNT trial series is well established: HbA1c reductions of up to 2.30 percentage points and body weight reductions approaching 20% at 15 mg weekly doses, with emerging data in MASH, sleep apnea, and cardiovascular outcomes. SURMOUNT-4 weight regain data following discontinuation underscore the dependency on continuous dosing and motivate exploration of variants with enhanced pharmacokinetic durability. A tirzepatide analogue with orthogonal DPP-4 resistance at Aib-2 could, in principle, maintain or extend the pharmacological profile while reducing the theoretical contribution of N-terminal proteolysis to dose-to-dose variability — though the marginal PK benefit on top of the existing C20 diacid half-life remains the central unresolved question.

Modification rationale

Aib (α-aminoisobutyric acid) is a Cα,α-disubstituted non-proteinogenic amino acid with gem-dimethyl geometry at the α-carbon. This geometry sterically blocks DPP-4's S1 pocket, which requires a planar Cα at the P1' position, rendering Aib-containing peptides non-substrates for DPP-4 cleavage. This is the same chemical logic that underpins the Aib-2 strategy used in semaglutide and validated across numerous GLP-1 and GIP analogue programs. Critically, Aib also constrains backbone dihedral angles to the α-helical region (φ ≈ −60°, ψ ≈ −30°), acting as a helix-capping residue that may pre-organize the N-terminal turn prior to receptor binding — a property structurally consistent with the requirement for an ordered helical N-terminus for incretin receptor engagement.

The Aib-2 and C20 diacid modifications are structurally orthogonal: the fatty diacid on Lys-20 operates via albumin-mediated half-life extension, while Aib-2 acts directly at the proteolytic cleavage site. They address different inactivation mechanisms and are not expected to sterically interfere with one another.

Cross-fold context: This fold revisits the Aib-2 concept previously tested in Fold №3 (Retatrutide Aib-2, DISCARDED, pLDDT 0.71), where the triple-agonist scaffold's more complex N-terminal hydrogen-bonding network with GCGR appeared less tolerant of the modification, contributing to a weak structural signal. The present fold tests whether tirzepatide's distinct GIPR-biased pharmacology and simpler dual-receptor N-terminal geometry represents a more permissive context for this substitution. Separately, Fold №23 demonstrated that tirzepatide tolerates Cα-methylation at position 24 (αMe-Cys, pLDDT 0.71, PROMISING), providing precedent that the tirzepatide scaffold is generally compatible with backbone-constraining non-canonical residue insertions without gross conformational disruption.

Predicted properties (where signal is moderate)

The pLDDT of 0.706 and pTM of 0.751 are consistent with a structurally plausible, well-folded helical peptide, matching the researcher's predicted range of 0.70–0.78. The N-terminal Tyr1-Aib2-Glu3-Gly4 pharmacophore is resolved within the helical fold, consistent with backbone geometry preservation. The ipTM of 0.128 is below the threshold for confident receptor engagement inference and should not be interpreted as evidence of reduced binding — it reflects the limitations of the single-run complex prediction, not a predicted potency loss.

The DPP-4 resistance prediction is made on chemical first principles (Cα,α-disubstitution abolishes DPP-4 substrate geometry) rather than the structural run; this is the highest-confidence claim in the fold. The practical magnitude of in vivo benefit is uncertain given tirzepatide's existing ~5-day half-life.

What would strengthen this signal

Additional in silico predictions:

• Ensemble prediction (≥5 AlphaFold3 seeds or Boltz-2 affinity module) on the tirzepatide(Aib2)–GLP-1R and tirzepatide(Aib2)–GIPR complexes with full receptor structure included; ipTM from a single run is insufficient to assess dual agonism preservation.
• Chai-1 agreement run (noted as None in this fold) — inter-model agreement on the complex would materially increase confidence in receptor interface geometry.
• Comparative docking of native tirzepatide vs. Aib-2 variant against available GIPR and GLP-1R cryo-EM structures (PDB: 7DTX, 7MHZ, 7SK8) to assess steric accommodation of the Cα-methyl in both binding pockets.
• MD simulation of the N-terminal turn (residues 1–8) comparing native Ala-2 vs. Aib-2 conformational sampling, to quantify the predicted helix pre-organization effect.

Wet-lab experiments to validate:

• DPP-4 cleavage assay (fluorogenic substrate competition or LC-MS peptide stability assay) comparing tirzepatide vs. tirzepatide(Aib2) in human plasma — this is the most direct test of the primary hypothesis.
• cAMP accumulation assay at GLP-1R and GIPR (HEK293 overexpression or primary cells) with full dose-response curves for tirzepatide vs. Aib-2 variant — essential to confirm that dual agonist potency is retained and to quantify any GIPR/GLP-1R selectivity shift introduced by the Cα-methyl.
• SPR or ITC binding affinity measurements at purified GIPR and GLP-1R extracellular domains to isolate affinity effects from functional assay confounds.
• In vivo PK comparison (rodent, sc dosing) of native tirzepatide vs. tirzepatide(Aib2) — specifically, plasma half-life and intact N-terminal peptide quantification by mass spectrometry to determine whether Aib-2 provides measurable additional stability above the C20 diacid baseline.

Key unknowns to resolve:

1. What fraction of tirzepatide inactivation in vivo is attributable to DPP-4 vs. renal/hepatic clearance? If DPP-4 contributes <10% to total clearance, the Aib-2 modification may be pharmacokinetically inert in practice.
2. Does the GIPR binding pocket at position 2 accommodate the Cα-methyl without potency loss? This is the highest-risk structural unknown and the primary differentiator from the GLP-1R precedent.
3. Does combining Aib-2 with the C20 diacid alter the balance of GLP-1R vs. GIPR agonism, and if so, in which direction?