# Expanded NTD homolog and control-geometry search

STRUCTURAL HYPOTHESIS / DEPOSITED HOMOLOG GEOMETRY, NOT ESTABLISHED SCN5A ASSEMBLY.

Item47 asks for relevant resolved contacts or testable hypotheses with provenance and limitations. This expanded route identifies a **useful intramolecular homolog geometry** and a **real N-terminal CaM-peptide control**, but does not establish an R104-containing NaV1.5–CaM or NaV1.5–NaV1.5 interface. It is not another exact-accession inventory.

## Executed search beyond the original16 entries

`search_ntd_homolog_geometry.py` submitted the verified NP_000326.2 residues1–130 to the RCSB experimental-polymer sequence service with identity cutoff0.30 and E-value cutoff1. All80 returned entities fit within the100-result page; no search-result truncation occurred. This avoids relying on a SCN5A accession or an NTD title annotation. Seven returned entities carry exact Q14524 mapping; the rest include paralogs, nonhuman channels and chimeras. Search scores are service ranking scores, not assumed percent identities.

`NTD_HOMOLOG_SEARCH.json`, its raw response, `NTD_HOMOLOG_METADATA.json` and `NTD_HOMOLOG_CANDIDATE_RESULTS.json` preserve the request, reference hash, all80 entity descriptions/reference mappings and an explicit local NTD alignment. Alignment uses match2/mismatch−1/gap−2 against the first220 deposited residues. It is a candidate-mapping aid, not a validated phylogeny or proof that an entire homologous domain is resolved. The full metadata file is about20MB. No coordinate ensembles, MD or arbitrary docking were generated.

The80-entity metadata inventory is complete for this bounded query. Detailed coordinate/contact acceptance below is deliberately restricted to the strongest source-linked homolog and the identified CaM controls. It is not a coordinate-level exclusion of all80 entries or proof that no relevant structure exists anywhere. Very remote homologs, matches below the chosen threshold and unpublished structures remain outside scope.

## Candidate1: resolved NaV1.7 NTD contacts, not an intermolecular assembly

[Zhang et al.,2022](https://www.nature.com/articles/s41594-022-00860-1) explicitly describes the NaV1.7 NTD and its connection to voltage-sensor domainI. Its deposited inhibitor-bound structures are7XM9,7XMF and7XMG. The report identifies conserved R116 in the NTD/VSDI interaction region. This is a direct structural lead, not a title-only inference.

The independently aligned core is:

```text
NaV1.5 80–130: LEDLDPFYSTQKTFIVLNKGKTIFRFSATNALYVLSPFHPIRRAAVKILVH
NaV1.7 75–125: LEDLDPYYADKKTFIVLNKGKTIFRFNATPALYMLSPFSPLRRISIKILVH
```

39/51 positions are identical, with no gap in this core. The exact mappings are NaV1.5 Y87→NaV1.7 Y82, R104→R99 and R121→R116. Thus the two relevant arginines are **different residues** in the homolog; do not confuse R104 with the publication's conserved R116.7XVF is another returned NaV1.7 candidate, but its deposited entity carries a34-position offset: the corresponding entity positions are116/133/150. Entity numbering must not silently replace author/reference numbering.

RCSB unobserved-residue metadata for7XM9 chainA marks positions1–6 and31–48 missing near the N terminus; residues75–125 are not in those missing intervals. The three target residues have complete standard heavy-atom records in the downloaded coordinates. `ntd_homolog_contact_check.py` inspected12,289 deposited ATOM heavy atoms and computed minimum distances to residues outside the operational NTD cutoff130. `NTD_HOMOLOG_CONTACT_RESULTS.json` preserves the atom-pair witnesses:

| SCN5A site |7XM9 homolog site | Observed outside-NTD contacts within4Å |
|---|---|---|
| Y87 | Y82 | Same-chain F181,3.636Å |
| R104 | R99 | Same-chain R174, backbone-O/NH1 contact2.991Å; F181, NH1/CZ3.420Å |
| R121 | R116 | Same-chain L172,3.480Å; A173,3.651Å |

These are static heavy-atom distances, not bond assignments, interaction energies or persistence estimates. The R99/R174 pair involves the **R99 backbone oxygen**, not evidence of favorable arginine–arginine side-chain electrostatics. The operational cutoff is stated to make the contact computation reproducible; it is not a universal domain boundary.

Every reported contact is intrachain.7XM9's entry contains one sodium-channel polymer entity plus beta1/beta2, not a second alpha channel or CaM. Its deposited alpha entity includes a GFP fusion in its construct description, with much of the C-terminal entity unmodeled. An R104Q homology hypothesis can therefore concern altered local NTD/VSD coupling, but these coordinates cannot be relabeled as the missing intermolecular dominant-negative interface. Side-chain and construct-context differences also preclude direct transfer of a mutation energy or trafficking conclusion.

## Candidate2: true N-terminal CaM-peptide control, with a useful false-positive lesson

[Liu and Vogel,2012](https://www.frontiersin.org/journals/molecular-neuroscience/articles/10.3389/fnmol.2012.00038/full) investigates CaV1.2 NSCaTE and CaM. It supplies a related assay-class control, not a homologous SCN5A NTD pose.

**2LQC** contains two polymer entities: CaM and the deposited peptide `GTGAALSWQAAIDAARQAKLMGSA`. SIFTS maps peptide entity residues3–24 to CACNA1C/Q13936 residues47–68; the initial GT is not silently assigned as native channel sequence. The CaM entity maps to P0DP23 residues2–78. The saved metadata and entry record identify a20-model solution-NMR ensemble. It is a candidate positive control for a CaM N-lobe/short alpha-helical N-terminal peptide interaction, not a positive control for recognition of the folded SCN5A NTD or a quantitative cross-channel affinity standard.

**2LQP is not equivalent.** Despite its complex-sounding title, its deposited entry has only one polymer entity: CaM residues79–149. The other entity is calcium, not the NSCaTE peptide. A request for polymer entity2 correctly returned404. It can supply the ligand-bound CaM conformation, but **cannot supply a deposited peptide-interface coordinate control**. This is why candidate qualification must inspect actual entities rather than trust the word “complex” in a title. The literature additionally uses data-driven docking; its inferred poses must be distinguished from deposited joint-coordinate evidence.

`NTD_CONTROL_METADATA.json` preserves exact2LQC entity data and source hashes. The initial metadata script aborted on the2LQP nonexistent-polymer request; the corrected bounded run explicitly retained that404 rather than converting it into a missing whole structure.

## Other routes and disposition

- CaM–NaV C-terminal structures, including4OVN/6MUD, are domain-mismatched controls already covered by the earlier exact-accession review. They are not counted as new R104 geometry.
- Primary-source literature search also returned bacterial NavMs structures. Their pore-module organization and short termini do not establish a corresponding mammalian residues80–130 assembly interface. They are not accepted by the presence of “N terminus” in the article alone.
- Name-based searches for sodium-channel/14-3-3 structures did not yield an accepted atomic NTD complex in the inspected results. Functional interaction literature is not converted into a coordinate model.
- R104Q and R104W mutation comparisons could be proposed against the retained homolog core, but no new mutant structure, energetic calculation, docking or MD was run here. A control must test the intended endpoint, not merely produce a visually plausible complex.

## Acceptance boundary

The expanded search deliverable is complete as a **bounded candidate qualification**: accepted homolog intramolecular geometry, a qualified related-class control, explicit rejected domain/title substitutions, exact sequence mappings and evidence-linked limits. No direct target NTD partner geometry is accepted from the detailed candidates. The safe next use is hypothesis formulation about NTD/VSD coupling and design of endpoint-matched controls; it is not permission to compute an intermolecular binding free energy from these unrelated interfaces.

Any broader exclusion claim would require contact/assembly qualification of additional returned entries and explicit alternative-search coverage. This report does not claim that a finite source search has proven universal absence. The evidence is saved in new program artifacts; no canonical source or private material was changed or read.