The Antarctic octopus has a potassium channel gene that codes for isoleucine at position 321. But the Antarctic octopus does not have isoleucine at position 321. It has valine.
The gene says one thing. The protein does another. The distance between them is not a mutation — the gene hasn't changed. It is a rewrite: an enzyme called ADAR finds the relevant spot in the RNA transcript, chemically modifies adenosine to inosine, and the ribosome reads inosine as guanosine, incorporating valine where the template called for isoleucine. The gene remains intact. The transcript is corrected before it is translated.
This happens at one specific site — I321V — in the S5 helix of the channel's pore region. It matters: a channel with valine at 321 closes more than twice as fast as a channel with isoleucine there. In cold water, nerve signals depend on precise timing of channel opening and closing; if channels stay open too long, the signal fails. Warm-water octopuses, mostly unedited at this site, have channels tuned for warmth. Antarctic and Arctic octopuses, extensively edited, have channels tuned for cold. The genes are nearly identical — four positions differ across the entire gene between Antarctic and tropical species. The functional difference lives not in the genome but in what happens to the RNA in transit.
The warm-water octopus tells a larger version of the same story. The California two-spot experiences water temperature swings with tides and seasons. When it gets cold, editing increases at more than 13,000 sites across the nervous system. Kinesin-1, a motor protein that hauls cargo along axons, gets edited in a way that adjusts its walking speed. Synaptotagmin, the protein that triggers neurotransmitter release in response to calcium, gets edited in a way that alters how it binds calcium.
When the water warms back up, the editing reverses at those sites. The genome didn't change. The editing tracks temperature, responds within hours, reaches a new steady state within days, and undoes itself when conditions reverse. A stable genome, a variable proteome.
One asymmetry: cold triggers editing at 13,285 sites; warmth triggers editing at only 550. The default is warm-tuned. Cold requires extensive recalibration; warmth mostly doesn't. Whatever the editing system holds to as its baseline, it is a warm baseline.
There is a question about what words mean here.
When biologists say "the gene for X," they usually mean: the gene that, when expressed, produces the protein that does X. The Antarctic octopus has a potassium channel gene. That gene, in polar conditions, produces a channel that closes twice as fast. Is the gene "for" the fast channel or the slow one?
If you read the gene, you'd predict a slow channel. If you watch the organism, you'd see a fast one. The prediction from the sequence is wrong for any animal that was doing the editing — which, in Antarctica, is all of them, always.
The gene isn't wrong. The edit isn't a correction of an error. The genome encodes an amino acid that the organism consistently doesn't use, in an animal that has apparently survived Antarctic temperatures for as long as the editing has been in place. The "unedited" version of the Antarctic octopus K+ channel is not a real thing that exists — it's an artifact of reading the DNA without accounting for what happens to the RNA in transit.
What you inherit from an octopus parent is not a channel that closes at a certain rate. You inherit the gene — the draft — plus the editing machinery that will rewrite it in context. The channel's properties are the interaction of both, and both are heritable. The edited sequence itself isn't.
The polar species differs from the California two-spot in an important way. The polar octopus lives with extensive editing as its permanent state; the California two-spot holds the rewrite in reserve for when the water cools. One has fixed the rewrite into its routine; the other keeps it conditional. The machinery is the same — ADAR enzymes acting on RNA before translation. What differs is how often it runs, and at which sites, and whether that setting is itself adjustable or locked in.
The genome contains not just the first draft but the editing rules. Both are heritable. The edited sequence itself is not, and cannot be — the edit doesn't reach the germ line. What each generation inherits is the capacity to rewrite, under the right conditions. The rewrite has to happen again fresh, in each animal, in each context.
Read the Antarctic octopus genome and you find an isoleucine gene. Watch the Antarctic octopus function and you find a valine channel. The gene doesn't describe the organism. It describes the organism's starting point, and the rules the organism will use to depart from it.