Binder10246 occupies the reference ETGE-site region after conserved-core alignment
Thirty predicted Binder10246–KEAP1 complexes were aligned to the human KEAP1 Kelch-domain structure 5WFV. The comparison used the 5WFV-bound NRF2 peptide LDEETGEFL as a spatial reference.
Interpretation. Every aligned model places Binder10246 in the spatial region occupied by all nine residues of the reference NRF2 peptide. This is geometric co-localization of independently determined poses. It does not demonstrate physical binding, affinity, NRF2 displacement, competition, or NRF2-pathway activation.
Six construct-by-seed groups
| KEAP1 construct | Seed | Models | Fit RMSD, Å median [range] | 4 Å mask overlap of 16, median [range] | 5 Å mask overlap of 22, median [range] | Binder residues near ETGE 4 Å / 5 Å, median [range] | ETGE peptide residues near binder 4 Å / 5 Å |
|---|---|---|---|---|---|---|---|
| Q14145 315–624 | 7 | 5 | 0.279 [0.262–0.321] | 16 [16–16] | 20 [20–21] | 20 [19–22] / 26 [24–28] | 9 / 9 |
| Q14145 315–624 | 42 | 5 | 0.280 [0.269–0.300] | 16 [15–16] | 19 [19–20] | 19 [18–24] / 24 [23–29] | 9 / 9 |
| Q14145 315–624 | 101 | 5 | 0.291 [0.265–0.295] | 15 [14–16] | 19 [18–19] | 20 [18–21] / 23 [21–26] | 9 / 9 |
| Q14145 325–609 | 7 | 5 | 0.367 [0.300–0.431] | 14 [13–16] | 17 [16–19] | 17 [14–21] / 23 [22–24] | 9 / 9 |
| Q14145 325–609 | 42 | 5 | 0.344 [0.340–0.357] | 10 [10–12] | 15 [14–16] | 14 [13–15] / 16 [15–17] | 9 / 9 |
| Q14145 325–609 | 101 | 5 | 0.295 [0.277–0.344] | 16 [13–16] | 19 [18–20] | 18 [14–22] / 24 [17–25] | 9 / 9 |
Exact molecular identities
Binder10246, 118 aa
Targets are human KEAP1 UniProt Q14145, residues 315–624 and 325–609. The reference is PDB 5WFV: KEAP1 receptor label/auth chain A/A and NRF2 peptide label/auth chain C/P, sequence LDEETGEFL.
Relative to canonical Q14145, the resolved 5WFV receptor has six sequence differences: C319S, E540A, E542A, C613S, C622S, and C624S. Only E540A and E542A are explicitly annotated as mutations in the structure record; the other four differences are retained as observed sequence differences. None of these six positions occurs in the reference or model contact masks reported here.
Human and mouse NCALD references are identical across all 193 residues
The reviewed human NCALD reference P61601 (TaxID 9606; sequence version 2; entry version 175) and reviewed mouse NCALD reference Q91X97 (TaxID 10090; sequence version 4; entry version 183) are 193/193 identical in the sequences used here. The terminal residues Q192 and F193 are therefore conserved. Contact projections are reported per design and per model sample.
Terminal-position sample support
| Design | Distance cutoff | Samples contacting Q192 | Samples contacting F193 | Samples contacting Q192 or F193 | Denominator |
|---|---|---|---|---|---|
| design_0154 | 4 Å | 3 | 1 | 4 | 15 |
| design_0154 | 5 Å | 5 | 2 | 6 | 15 |
| design_0154 | 8 Å | 12 | 13 | 13 | 15 |
| design_0026 | 4 Å | 1 | 1 | 2 | 15 |
| design_0026 | 5 Å | 1 | 1 | 2 | 15 |
| design_0026 | 8 Å | 5 | 5 | 7 | 15 |
The counts above are unions recomputed within each predicted sample. A sample contacting both terminal residues contributes once to the “Q192 or F193” count. The separate source-backbone masks contain 0/5/35 unique target positions for design_0154 and 4/9/25 for design_0026 at 4/5/8 Å; these single-structure masks are not sample frequencies.
Exact molecular identities
design_0154, 64 aa
design_0026, 67 aa
NCALD reference sequence, 193 aa
Interpretation. Human–mouse reference identity supports direct residue-number projection for this sequence. It does not establish that experimental reagents have identical processing, post-translational or calcium states. Predicted contact support does not establish affinity, a shared binding mode, function, or a mouse effect.
Open the previously published full-length NCALD model package.
Methods and statement boundaries
KEAP1 reference projection
Each predicted KEAP1 model was rigidly fitted to 5WFV receptor label/auth chain A/A using the 282 resolved Cα atoms at canonical KEAP1 positions 326–609, excluding positions 540 and 542 because 5WFV explicitly annotates E540A and E542A. The fixed NRF2 peptide is label/auth chain C/P. For alternate atom locations, blank identifiers were preferred; otherwise the atom with the highest occupancy was selected. This deterministic rule selects occupancy 0.67 for 5WFV Cys434 alternate A over occupancy 0.33 for alternate B.
The KEAP1 contact-mask intersection counts shared canonical KEAP1 positions in the reference and model contact sets at the stated cutoff. Spatial overlap is calculated separately from heavy-atom distances between the transformed Binder10246 and the fixed 5WFV NRF2 peptide. None of the six 5WFV-versus-canonical sequence-deviation positions occurs in either contact mask.
NCALD projection
Official reviewed human and mouse UniProt reference sequences were compared residue by residue. For each predicted full-length complex, a target position counts as contacted when any heavy-atom pair falls within the stated cutoff. Q192 and F193 support is retained separately, and their union is calculated within each model sample.
All outputs are descriptive structure-model analyses. They do not substitute for biochemical binding, competition, pathway, pharmacology, or in vivo measurements.
Machine-readable data
The complete model-level and grouped values supporting the displayed summaries are provided without filtering by outcome.