G437R

G437 sits inside TM4's helical core, with Thr436 (2.46 Angstrom) and Phe438 (2.46 Angstrom) as covalent neighbors. The through-space cluster is dense and hydrophobic: Ala433 (3.68 Angstrom), Ser544 (3.80 Angstrom) — note that Ser544 is on a neighboring helix or loop, indicating an inter-helix contact — Val434 (3.83 Angstrom), Thr440 (4.13 Angstrom), Ile435 (4.41 Angstrom), Phe439 (4.42 Angstrom). The wild-type glycine occupies the position where TM4 meets its partner helix; only glycine is small enough to allow that close approach.

Substituting arginine here is a near-worst-case lipid-bilayer mutation. Arginine's guanidinium group carries a delocalized positive charge and requires solvation — typically by water or by phosphate headgroups at the membrane interface. Burying it in the middle of TM4 forces one of two outcomes: either the guanidinium snorkels upward toward the lipid headgroups, dragging the helix backbone out of register, or it remains buried at a substantial desolvation cost (estimated 8-12 kcal/mol for an ionized arginine in low dielectric). Both outcomes destabilize the helix-helix interface that requires glycine's small footprint.

DynaMut2's DeltaDeltaG of -0.59 kcal/mol substantially underestimates the cost, because the energy function does not fully capture transmembrane desolvation. AlphaMissense at 0.935 reads the true severity: G437R is essentially incompatible with normal TM4 packing. The structural prediction is that TM4 either ejects from its native register against TM5 or that the protein fails to insert into the membrane co-translationally — exactly the kind of misfolding that triggers ER retention and ER-associated degradation.

ClinVar records this variant for Wolfram syndrome 1 — consistent with a classical loss-of-function mechanism via misfolding and degradation.