How can a polar fish limit ice growth without warming itself?

Some Antarctic fishes swim in seawater cold enough that a small ice crystal entering their bodies could trigger dangerous further freezing. Yet the fishes do not solve this problem by keeping their bodies warm. They make molecules that interfere with the growth of ice itself, working at a boundary far smaller than a fish's scale.

Ice grows when water molecules join an ordered crystal. A small existing crystal can provide a starting surface, so preventing every first crystal from appearing is different from preventing crystals from expanding. Antifreeze proteins, including antifreeze glycoproteins with attached sugars, help with the second problem. They bind to particular surfaces of ice rather than acting only as substances dissolved in the surrounding water.

That attachment is called adsorption. Once a molecule is bound, the ice surface cannot advance through its position as easily as through an unoccupied patch. Between binding sites, the advancing surface becomes curved. A strongly curved ice–water boundary requires greater cooling to continue advancing under these conditions. The bound molecules therefore act as obstacles that can hold crystal growth back across a finite temperature range.

The fit between a molecule and an ice surface matters. Ice is a three-dimensional crystal with differently oriented faces, and different antifreeze molecules can bind differently. Experiments with fish peptides and glycopeptides show specific surface attachment and changes in crystal growth. Protection consequently depends on molecular structure, the surfaces occupied and experimental conditions, rather than simply on the presence of any large protein in the liquid.

The result differs from the familiar effect of adding ordinary salt to water. Salt changes equilibrium freezing behavior through the properties of the dissolved particles. Antifreeze proteins produce an additional inhibition of crystal growth through surface binding. A gap can appear between the temperature at which an existing crystal begins growing rapidly during cooling and the temperature at which it melts during warming. This difference is called thermal hysteresis.

The protection has limits. Further cooling can overcome the inhibition and permit rapid growth, so a coated crystal is not harmless at every temperature. There is also a striking complication: studies of Antarctic notothenioid fishes found that protein-bound ice could resist melting even when the surroundings warmed above its expected melting point. The same close interaction that restrains freezing can help small crystals persist.

These proteins are therefore neither miniature heaters nor chemicals that automatically dissolve every bit of ice. They change what happens where liquid water meets a crystal. A fish can remain cold while its molecules slow a threatening change of state. Survival here depends on controlling the growth of a tiny structure, and the mechanism also reveals why preventing freezing and removing existing ice are not always the same task.