Credit: © 2008 ACS

Antifreeze proteins (AFPs) that occur in creatures such as Antarctic fish, moths and snow fleas — and are added to ice cream — are known to inhibit the growth of ice crystals. They work through a non-colligative method, unlike the ethylene-glycol-based antifreezes commonly used to clear car windscreens on a cold winter morning. In contrast, they function by binding to ice-crystal surfaces at key sites where they block crystal growth. This forces any lateral crystal growth to occur with a curved radius, which lowers the local freezing temperature of the ice. The AFP binds through its 'ice-binding face' but the specifics of the preliminary stages are unclear.

Now, using atomistic molecular dynamics simulations, David Nutt and Jeremy Smith, of the University of Heidelberg, have shown1 that the structural ordering of solvation water around the ice-binding face of an AFP mediates the binding of the protein to the ice surface. They simulated a triangular-prism-shaped AFP from a spruce budworm and then calculated the solvent density–distribution functions and local molecular arrangement of water around its three separate faces. This showed that the ice-binding face orders the adjacent solvation water into a quasi-ice-like structure. Thus, the structural similarity between the protein–water interface and the ice–water interface is predicted to reduce the entropy loss on binding, making it more favourable.

Nutt and Smith also show that the solvation water around the other non-ice-binding faces of the AFP is disordered because of the strong hydrophilic interactions of the water molecules with these faces. This interfacial water disruption works to prevent the protein being covered in further ice growth, which would prevent the AFP from functioning.