When Thitipan Meemongkolkiat arrived at the University of Sydney, Australia, from Chulalongkorn University, Thailand, bringing with him almost 9000 bees, Benjamin Oldroyd suspected that the visiting graduate student could help him to answer a question that had bugged him for 25 years. Fascinated by the genetics behind honeybee (Apis mellifera) behaviour, Oldroyd had discovered that one form (allele) of a key enzyme – cytosolic malate dehydrogenase – is remarkably vulnerable when the temperatures rise: it stops working. Honeybee populations can carry three different forms of the enzyme – known as slow, medium and fast – and the three forms are found at different frequencies in warm and cool populations, with the medium form occurring rarely in tropical bee populations. The fact that the medium form isn't stable in hot environments fits nicely with its low prevalence in hotter locations. But it didn't explain why populations from colder climes hang onto the medium form of the enzyme. ‘I've always wanted to know what the selective benefit of the medium allele is in cold climates’, says Oldroyd. So when Meemongkolkiat arrived in the lab, Oldroyd hoped that his chance had come to discover why the allele holds on in chillier climes.But first Meemongkolkiat needed to find out whether any of the Thai bees carried the elusive medium form of the gene. ‘I expected that the population would have no medium alleles’, says Oldroyd. However, after months of painstakingly isolating the different enzyme forms from more than a thousand Thai bees, it was apparent that many of the insects carried at least one copy of the medium form. The Thai bees had not lost that form of the gene even though Meemongkolkiat and Frank Seebacher, also from the University of Sydney, found that forms of the enzyme including the medium allele disintegrate faster at high temperatures and are less efficient than the slow and fast forms.Wondering whether the medium allele might contribute to the insects’ ability to recover from a cold shock, Meemongkolkiat gave over 500 bees a 2 h icy bath and waited to see which recovered from the chill. Remarkably, the bees that carried two copies of the medium form of the enzyme were up and about within 15 min of warmth returning, while the bees that only had one copy of the medium enzyme resumed activity after 19 min. In contrast, the bees that had fast and slow versions of the enzyme only recovered after 21–23 min. The bees that carry the medium form of the enzyme are able to recover better from an icy blast, but why?Meemongkolkiat teamed up with molecular modeller Jane Allison from the University of Auckland, New Zealand, to find out why bees with the medium form of the enzyme recover better from a chill. Searching the three proteins' predicted gene sequences, the pair found several differences that could alter the proteins' stabilities at different temperatures. Next they built molecular models of the enzymes and ‘warmed’ them in a computer simulation to 15, 30 and 50°C, to find out which held together best at the higher temperatures and which locked up at 15°C. After hours of intensive computer calculations, it was clear that the medium form was too floppy to function at 50°C; the protein couldn't function at the higher temperatures, explaining why many honeybee populations in tropical locations have lost it. However, the flexibility of the medium form allowed it to continue functioning at 10°C – when the slow and fast forms lock solid – providing an advantage for bees that retain the medium form when winter temperatures plummet.