In July of 2000, during the long, unbroken night of the Antarctic winter, graduate student Jack Gilbert found himself careening through the darkness on a quad-bike. Driving over rocky hills along the continent’s eastern shore, he finally arrived at Ace Lake—a salty body of water that freezes over for several months of the year.

It is a truly inhospitable environment, but it still harbors life—microbes, thriving in its frigid waters. These extreme survivors were the organisms Gilbert was there to study. He reached them by drilling through 1.4 meters of ice and pulling up samples of water. Then, he retired to a bare-bones hut for the night.

Gilbert had to wrap up warm to avoid freezing to death. But microbes can survive in the planet’s coldest places through other means. Some produce antifreeze proteins—molecules that latch onto small ice crystals and stop them from growing bigger. By releasing these proteins into the space around them, the microbes can lower the freezing point of water and create a network of liquid in which they swim and feed.

Among the hundreds of cold-loving microbes that Gilbert scooped up from Ace Lake, he found that one produced an “uncharacteristically powerful” antifreeze. It was called Marinomonas primoryensis. It had only been discovered the year before, and it was clearly doing something unexpected. For a start, its antifreeze was a titan of a molecule—between 30 and 50 times bigger than the average protein, and almost half as long as the microbe that makes it.

Gilbert moved on to other things, but his supervisor Peter Davies kept studying the giant protein. His team realized that only one tip of the long molecule can latch onto ice; the other stays stuck to the bacterium’s body. It seemed that M. primoryensis was using the protein like a grappling hook, to grab onto ice rather than stop it from forming.

Davies proved this idea by teaming up with Maya Bar Dolev and Ido Braslavsky at The Hebrew University of Jerusalem. They introduced M. primoryensis into microscopically narrow canals full of pure water, and then seeded a single ice crystal by dipping in a chilly finger of copper. “We saw that they swim around, make contact with the ice, bind to it, and form a little colony,” says Davies.

It’s not entirely clear how the bacteria stick, but they certainly use the tip of that long antifreeze protein. When the team blocked the tip using an antibody—essentially, sticking a cushion over the head of the grappling hook—the microbes could no longer adhere to ice.

Davies suspects that the microscopic grappling hook is shaped so that it organizes the water molecules around it. These line up in an orderly lattice—a bit like they would in ice, but slightly more fluid. They’re now halfway between liquid and solid, between water and ice. The water at the very surface of ice sheets is in a similar state, and when these two liquid-ish layers meet, they mingle and freeze. The tip of the protein is then embedded within the ice, and the bacterium is firmly fastened. “It’s as if you had put your tongue onto a cold metal railing and got stuck,” says Davies.

So what happens when the ice grows, as ice is wont to do? How does M. primoryensis avoid falling off or becoming entombed? The key to understanding what it does, Davies says, is to realize that ice doesn’t grow by just spreading outwards. If you watch it under a microscope, you see that it forms little ridges at its surface that then spread sideways, like moving cliff faces, or perhaps mini-glaciers. If a microbe is in the way, it might get pushed off. But fortunately, M. primoryensis forms little colonies, where each member holds onto the ice, and they all hold on to each other. If one is dislodged, it can just swing round and re-attach itself.

Why would a bacterium evolve to do this at all? Why adhere to a surface that, as Braslavsky notes, “can just melt and disappear?” Davies thinks it’s because the water immediately below Ace Lake’s ice is the only place where M. primoryensis can thrive. They can use the nutrients and oxygen produced by other microbes that live in this zone.

And they’ve done so by converting the proteins that other microbes use to stop ice from forming into tools for living on ice. “It shows how evolution shapes the same tools to adapt to new niches,” says Gilbert. “It’s extremely exciting to see work I started 17 years ago continually being advanced.”

“Bacteria are known to attach to virtually every other mineral surface, except—until now—ice,” says Jody Deming from the University of Washington, who studies cold-loving bacteria. She adds that understanding how microbes stick to cold surfaces “carries significance for many human activities, from Arctic Inuit storage of harvested meats, to the efficient operation of marine vessels and oceanographic instruments that spend months submerged in subzero waters.”

Deming also thinks that these extreme Earth microbes might provide hints about life on other worlds. “In the foreseeable future, our best chance of finding microbial life on an icy moon like Europa or Enceladus will be finding it associated with ice,” she says. “The chances are discovering it increase if we can better design missions with this new information in hand.”