Living with ice: how do polar fishes survive freezing?

In a classic case of convergent evolution, unrelated Arctic and Antarctic polar fishes have evolved almost identical mechanisms to deal with the shared problem of living with ice.

HFSP Program Grant holders Margaret Brimble and Arthur DeVries and colleagues
authored on Mon, 25 June 2012

Polar fishes, whose body temperatures are close to that of their environment, survive freezing of their body fluids through the secretion of a range of antifreeze molecules. One class of fish antifreezes, the antifreeze glycoproteins (whose activity is critically dependent on the presence of specific sugars), is particularly remarkable in that essentially identical molecules are synthesized in both Arctic and Antarctic polar fishes, despite their distant evolutionary relationship.  Like other fish antifreezes, these molecules function by binding to ice crystals to inhibit their growth. They are found in high levels in the blood where they presumably function to inhibit the potentially lethal growth of ice crystals internalized through injury.

Figure 1: Juvenile Antarctic fish (Order Notothenioidei, Family Nototheniidae) inhabit an icy domain (dark background) that may provide a source of food and offer protection from predators, but also puts them at risk of freezing.

A number of paradoxes are raised in relation to antifreeze glycoprotein synthesis and activity. First, they are synthesized in a part of the pancreas (the exocrine pancreas) that discharges into the gut from where they are likely to be egested along with the feces. For fishes that truly “live on the edge” this is undoubtedly energetically wasteful. Second, since the exocrine pancreas discharges into the gut the question arises as to how high levels of antifreeze glycoproteins accumulate in the blood. This is of particular relevance to Antarctic notothenioids (the dominant group of Antarctic fishes) since, unlike the Arctic gadids (cods), they do not also synthesize antifreeze glycoproteins in the liver (from which direct entry into the circulation might be expected). Third, since fish antifreeze glycoproteins work by inhibiting ice crystal growth, rather than by inhibiting ice crystal formation, then potentially lethal ice must be present in the fish’s body if the antifreezes are to function internally. How can a fish whose body temperature is close to that of its freezing environment deal with this internalized ice?

Here we demonstrate that the phylogenetically distant Antarctic notothenioids and Arctic gadids, which show molecular convergence through the synthesis of essentially identical antifreeze glycoproteins, also display convergence in their physiological deployment. In order to investigate the associated phenomena we purified antifreeze glycoproteins from Antarctic fish serum and prepared specific antibodies for immunodetection in histological sections. We also prepared two labeled antifreeze glycoprotein probes for in vivo analysis. One was prepared by fluorescently labeling natural fish antifreeze glycoproteins, while the other was synthesized chemically, optimizing the synthesis of the required glycosylated amino acid building block in the process. The synthetic antifreeze glycoprotein was isotopically labeled by using a heavier (15N) form of alanine, the most common amino acid in the protein backbone of the antifreeze molecule.  This probe will prove useful in further studies investigating in vivo distribution since (unlike the fluorescent probe) it should not elicit any host response.


Figure 2: The exocrine pancreas of a larval Antarctic fish is charged with antifreeze glycoproteins, colored red in this immunostained histological section. The pancreas is to the left (dorsal) of the junction between the stomach (upper right bulbous structure) and the intestine. The head of the larval fish is towards the top. The blue structures are nuclei of individual cells reacting with a specific nuclear stain. Their disposition is an aid to identifying anatomical features.

Both the Antarctic notothenioids and Arctic gadids synthesize antifreeze glycoproteins in the exocrine pancreas (supplemented by the liver in Arctic gadids) from where they are discharged into the gut to inhibit the growth of ingested ice. This is a critical function in these polar fishes, since both groups ingest ice through eating and drinking ice-laden food and seawater. Antifreeze glycoproteins that bind to ingested ice are purged from the gut ridding the body of a potential problem, whereas unbound antifreeze glycoproteins are able to be resorbed via the rectal epithelium. Following resorption they are transported to the blood, still structurally and functionally intact. From here, they are secreted into the gall bladder (via the liver) to re-enter the gastrointestinal tract in the bile. This unique recirculation pathway maintains high levels of antifreeze glycoproteins in the gut (where the risk from ingested ice is significant), conserves energy expenditure (through recycling unbound antifreeze molecules), and explains how they reach high levels in the blood in Antarctic notothenioids (through rectal resorption).

We also show that in both groups of polar fishes antifreeze glycoproteins accumulate within macrophages of the spleen where, being removed from the circulation, they are likely at less risk. Since the spleen is not a site of antifreeze synthesis, the inevitable conclusion is that these molecules have been engulfed by splenic macrophages following adsorption to circulating ice crystals in a process reflecting the phagocytosis of foreign particulate material, one of the major functions of the spleen. This conclusion is supported by independent evidence showing that the spleen is the only major organ of polar fishes identified to date that accumulates measurable quantities of ice crystals. As to the fate of the accumulated ice, it would seem that in the absence of any known endogenous physiological process to induce melting of ice in these fishes, they must await an environmental warming event. For Antarctic notothenioids that are highly territorial, such environmental warming may never happen in their lifetime, and indeed the burden of internalized ice may inevitably prove fatal. The accumulated ice may be less of a long-term problem for Arctic gadids, since in the summer these fishes are often exposed to seawater temperatures warm enough to melt the entrapped ice.  

What is remarkable about these observations is the extent to which they are shared by Antarctic notothenioids and Arctic gadids, illustrating that these two divergent groups of polar fishes display unparalleled elements of convergent evolution, not only in terms of virtually identical antifreeze glycoprotein structure (with evolutionary origins in unrelated genes), but also in terms of a common site of synthesis (the exocrine pancreas) and a similar mode of physiological deployment (through identical recycling pathways and phagocytic uptake in the spleen).


(1) Synthesis and recycling of antifreeze glycoproteins in polar fishes. Clive W Evans, Linn Hellman, Martin Middleditch, Joanna M Wojnar, Margaret A Brimble and Arthur L DeVries. Antarctic Science 24: 259 – 268 (2012).

(2) How do Antarctic notothenioid fishes cope with internal ice? A novel function for antifreeze glycoproteins. Clive W Evans, Vladimir Gubala, Robert Nooney, David E Williams, Margaret A Brimble and Arthur L DeVries.Antarctic Science23: 57–64 (2011).

(3) Synthesis of an isotopically-labelled Antarctic fish antifreeze glycoprotein probe. Joanna M Wojnar, Clive W Evans, Arthur L. DeVries and Margaret A Brimble. Australian Journal of Chemistry 64: 723-731 (2011).

Other References

(4) Convergent evolution of antifreeze glycoproteins in Antarctic notothenioid fish and Arctic cod. Liangbiao Chen, Arthur L DeVries, and Chi-Hing C Cheng. Proceedings of the National Academy of Sciences USA 94: 3817–3822 (1997).

Antarctic Science link for ref (1)

Antarctic Science link for ref (2)

Faculty 1000 link for ref (2)