L804P
Category 3/4 — Most DruggableLikely pathogenicLumenal · predictedσ-1 candidateEditorialLeucine → Proline at position 804 in wolframin's C-terminal lumenal domain. ClinVar Likely pathogenic, associated with DFNA6 (WFS1-related hearing loss). AlphaMissense 0.999 (deep pathogenic signal), DynaMut2 ΔΔG -0.44 kcal/mol (destabilising). A fold-intact variant with severe local backbone disruption.
Interactive 3D Structure
Bond changes · DynaMut2 interaction analysis
| Interaction type | Wild-type partner | Mutant partner | Status |
|---|---|---|---|
| Hydrogen bond | E776 | E776 | Preserved |
| Hydrogen bond | I777 | I777 | Preserved |
| Hydrogen bond | F840 | F840 | Preserved |
| Polar contact | E776 | E776 | Preserved |
| Polar contact | I777 | I777 | Preserved |
| Polar contact | — | I802 | Gained |
| Polar contact | F840 | F840 | Preserved |
| Carbonyl | E776 | — | Lost |
| Carbonyl | F840 | F840 | Preserved |
| Van der Waals | — | E776 | Gained |
| Van der Waals | — | I802 | Gained |
| Van der Waals | F825 | — | Lost |
| Hydrophobic | I777 | — | Lost |
| Hydrophobic | V779 | V779 | Preserved |
| Hydrophobic | I802 | I802 | Preserved |
| Hydrophobic | F825 | — | Lost |
| Hydrophobic | F840 | — | Lost |
| Hydrophobic | L842 | L842 | Preserved |
Lost / gained / preserved interatomic contacts at the variant residue, from the DynaMut2 (Arpeggio) interaction analysis of the wild-type and energy-minimized mutant structures.
Computational Predictions
This variant is in a water-facing region, which is the regime the stability predictor was built and benchmarked for. Ordinary predictor error still applies.
Clinical Evidence
Not observed in ~730k individuals — consistent with a rare allele (ACMG PM2_supporting).
ClinVar classifies this variant as Likely pathogenic for Nonsyndromic hearing loss (DFNA6/14/38) (autosomal dominant, OMIM 600965). These are dominant WFS1-related disorders and are clinically distinct from classic recessive Wolfram syndrome. This classification is not a statement about Wolfram syndrome severity.
- Nonsyndromic hearing loss (DFNA6/14/38)autosomal dominant · OMIM 600965submitted as: Autosomal dominant nonsyndromic hearing loss 6
- Phenotype scope is interpretation of submitted records. The computed values on this card (AlphaMissense, DynaMut2 ΔΔG, pLDDT) are measurements of the model, not of a patient.
- Presentation cannot be generalised from neighbouring variants: one nucleotide over can present very differently. Structural similarity supports a mechanistic hypothesis, never a phenotype claim.
Review status: 2★ multiple submitters, no conflicts. ClinVar "conditions" is the union of all submissions; 0–1★ entries are submitted conditions, not documented phenotypes.
No population breakdown — the variant is absent from gnomAD v4, so no ancestry group has an observed frequency.
No homozygotes — the variant is absent from gnomAD v4 altogether.
Source: gnomAD r4, retrieved 2026-08-21. Values are allele frequencies. The carrier estimate is derived from the allele frequency (Hardy-Weinberg, ~2·AF) and is an approximation, not a published figure — where a published carrier frequency exists it outranks this column and is cited on the card.
Structural Context
Position 804 sits deep in wolframin's C-terminal lumenal domain (residues 653–869), the protein's largest soluble region and the documented interaction interface with ATF6 and the Na+/K+ ATPase β1 subunit. The AlphaFold model places L804 against immediate sequence neighbors ARG805 (2.4 Å) and VAL803 (2.5 Å), and into a hydrophobic pocket containing PHE840 (3.3 Å) and ILE777 (3.8 Å). The leucine side chain contributes branched hydrophobic packing into that pocket, supporting the local fold geometry of two distant segments of the lumenal domain.
Replacing leucine with proline at this position has a structural cost that DynaMut2's |ΔΔG| of 0.44 understates. Proline is the only amino acid whose backbone is locked into a ring — its phi angle is constrained, it cannot serve as a hydrogen-bond donor in the backbone, and it forces a kink at the position it occupies. When proline is introduced into a region with defined secondary structure, the local geometry has to rearrange to accommodate it, and the disruption propagates a few residues in either direction along the chain.
The ΔΔG value reflects the energetic cost of the global fold absorbing this disruption — modest, because the lumenal domain has flexibility to spare. But the local geometric perturbation around residues 803–805 will be substantial, and the lost packing against PHE840 and ILE777 is unrecoverable without further fold rearrangement. AlphaMissense's 0.999 score reflects this: the model recognizes the deep pathogenic signal of an introduced proline even when global stability is only modestly perturbed.
Druggability Assessment
The therapeutic strategy here is unusual for the Atlas: the lesion is geometric, not interaction-based. A small molecule that stabilizes the local α-helical geometry around residues 803–805 could compensate for the backbone kink introduced by proline. Alternative: a chaperone that biases the fold toward the wild-type local geometry during synthesis would shift the population toward functional protein.
AlphaMissense's near-maximum score (0.999) confirms that the field's clinical observation of pathogenicity matches the structural prediction — this is a real lesion despite the small ΔΔG. The variant illustrates a category of pathogenicity the Atlas captures well: severe local geometric disruption with mild global energetic cost.
Why this matters
Feed this card to Wolfram Intelligence
Download the L804P PDF below and upload it to Wolfram Intelligence to generate therapeutic-strategy proposals — guanidinium mimetics, sigma-1 agonist docking, NAC thiol-capping. NAC is already on the bench-testing list.