Rhodopsin

Homo sapiens · seed P08100 · 348 aa · family defined as ≥30% identity to that seed · compiled 07 October 2026

Open in CODSWALLOP UniProt P08100 RCSB by accession PDBe-KB AlphaFold DB InterPro CATH 1.10.530.40 CATH 1.20.1070.10 SCOP 8033861 SCOP 8042778 SCOP 8070800 SCOP 8057879 RCSB 4ZWJ PDBe

CATH and SCOP identifiers come from the RCSB's own structure annotations, which the Domains panel already reads, so these are looked up rather than guessed at.

Every figure here is counted over the whole family rather than quoted from one entry.

The reference structure

4ZWJ, the structure every other member of this family is superposed onto. Rendered by the RCSB and embedded here: the live app shows an interactive viewport, which a document that fetches nothing cannot.

Which residues anyone has ever seen

How many of this family's constructs contain each residue of the seed. A trough is a stretch nobody has put in a construct, which is a construct-design answer rather than a disorder one.

Constructs, most-used first

38 distinct constructs across 103 entries. 57 polymer entities differ from the UniProt canonical sequence in some way, 14 carry a recognised expression tag and 8 carry a fusion partner.

"Differs from canonical" is not the same as "engineered". The canonical sequence is the full gene product, so a secreted protein whose structures all start after its signal peptide counts every one of them as different: lysozyme's most-used construct, residues 19–147 on 1,239 entities, is simply the mature protein. Read the construct column below for what was actually done, rather than this count.

Showing the 25 most-used of 38.

Positions people deliberately mutate

Columns where the wild-type residue still dominates but a real minority carries something else, which is a different question from "what varies across species".

What it assembles into

44 entries have the depositor's assembly corroborated by PISA, 52 carry the depositor's word alone and 7 were assigned by PISA where the depositor gave none. The middle figure is not a disagreement: PISA may have returned nothing or never run. 7 entries carry more than one assembly with different chain counts, so they have no single answer to quote: 1F88, 1HZX, 1L9H, 3OAX, 3PQR, 5TE3, 6PH7.

Domain architecture

Every source's own domains on the seed axis, one row each. They are not merged: Pfam, CATH, SCOP and InterPro disagree about boundaries, and a merged track would state a consensus none of them gave.

What binds it

RET47 entries NAG10 entries HTO7 entries ACE6 entries DAO5 entries SGV4 entries HSM3 entries BGL1 entries 4E61 entries ID31 entries 7AB1 entries DOK1 entries

How it crystallises

Parsed from the free text 52 depositors typed into _exptl_crystal_grow.pdbx_details, out of 52 entries that recorded anything at all. Median pH 6.0 (range 4.5 to 9.0).

Precipitants

Buffers

Which entries to trust

103 entries carry a wwPDB validation report: 39 clean, 31 worth a check and 33 with something to explain. Median clashscore 6.43, median RSRZ outliers 8.59%, median R-free minus R-work 0.029. 95 have released structure factors.

Across species

Seed sequence

348 residues, numbered every ten. Every identity figure in this document is measured against this sequence.

active or binding site modified residue or glycosylation disulphide cysteine transmembrane or signal the 15 most-substituted positions

Sites are UniProt's curated features where the seed is a UniProt accession; the substituted positions are measured from this family's own alignment rather than annotated, and only the fifteen most substituted are marked: every position carrying a minority substitution would be most of the protein, because the family holds orthologues. A residue can carry more than one and is drawn with the first that applies, in the order of the key above.

Primary citations

One record per paper, not per entry.