ABOVE: The distinctive smell that soil gives off after rain comes from microbe-made compounds. © ISTOCK.COM, BJÖRN FORENIUS

There are few surer heralds of spring than petrichor. It’s the earthy aroma that wafts up after a good rain and comes primarily from chemical compounds called terpenes. Geosmin, for example, is a terpene most commonly associated with Streptomyces bacteria, although other bacteria and fungi also make it, and it’s found in soils and bodies of fresh water the world over. Its ubiquity has long fascinated scientists—not least because it hasn’t been clear why its various producers make it. 

Liana Zaroubi, a PhD candidate at Simon Fraser University in British Columbia, Canada, first came to this mystery as an undergrad at Concordia University in Montreal in 2015. She took a class with Brandon Findlay on chemical ecology and recalls being intrigued by the bonus question he posed on the final exam: “What do you think geosmin does, and how would you test that?” She ended up joining his lab as a master’s student so she could look into that question. “I thought it was super interesting,” she says. 

After months spent reviewing the literature and testing hypothesis after hypothesis, she and Findlay considered whether geosmin and another terpene that contributes to petrichor, 2-methyl-isoborneol (2-MIB), might indirectly deter predators. Although these chemicals aren’t themselves toxic to animals, other compounds the bacteria make are, so geosmin and 2-MIB could be an aposematic signal, like the coloration of many poisonous insects that tells hungry birds to dine elsewhere. 

An initial round of experiments with bacteria-eating amoebae went poorly, Zaroubi says. The organisms are very slow predators, and geosmin is highly unstable, she explains. The amoebae would need weeks to get to the bacteria in the researchers’ experimental setup, but the geosmin would degrade in days, or even hours. “So we thought of faster predators, like nematodes.”

First, the researchers tested whether C. elegans would react to the presence of geosmin. They found that while the chemical didn’t appear to affect the nematodes’ health, it drastically affected their movements, causing them to move much faster and to make more frequent changes in direction. Mutant worms with deficiencies in detecting soluble and volatile odorants showed no such behavioral changes, suggesting the wildtype animals were smelling or tasting the compound. 

Next, the researchers plopped C. elegans and Streptomyces coelicolor, a bacterium that produces both geosmin and 2-MIB, into a petri dish. On the whole, the worms avoided the bacteria, the team found. But if the researchers engineered either the bacteria not to produce the chemicals, or the worms to be deficient in detecting those chemicals, the nematodes more frequently consumed the bacteria—and became ill from the toxic metabolites also produced by the microbes. “Geosmin thus acts as an aposematic signal,” the authors write in their paper, “honestly and reliably advertising the unpalatability of its producers and providing a mutual benefit to predator and prey.”

While the chemical didn’t appear to affect the nematodes’ health, it drastically affected their movements.

It’s the first time aposematic signaling has been documented in bacteria, says Findlay. He adds that it’s unsurprising that geosmin and 2-MIB should make good aposematic signals: composed of hydrocarbons arranged into rings or chains, the compounds are very good at fitting into cellular receptors. But because they also degrade so rapidly, they can’t accumulate in the environment or travel very far, meaning that they accurately reveal the organism producing them right here, right now. “As a chemical messenger, that makes [them] very, very valuable,” Findlay says.

The study is just one of a handful of recent papers identifying possible functions for geosmin and 2-MIB. For example, research by scientists at the Swedish University of Agricultural Sciences demonstrated that the two chemicals actually attract springtails, which feed on the bacteria producing them with no ill effects from the toxins. In turn, the springtails distribute bacterial spores in their fecal pellets and by carrying them on their bodies, helping the microbes move to new environments. 

Some fly species have also found ways to interpret the smelly signal of petrichor. While working at the Max Planck Institute for Chemical Ecology in Germany in 2012, sensory neuroecologist Marcus Stensmyr published a paper showing that fruit flies are repelled by geosmin, even when it’s added to vinegar. “Flies absolutely love vinegar,” says Stensmyr, now at Lund University in Sweden. “Anything that can make it less attractive must be important.” This aversion, the team showed, is governed by a single receptor, tuned specifically to geosmin and capable of detecting the chemical at concentrations as low as 1 part per 100 million, he says. 

It’s not clear why flies don’t like geosmin. It’s possible fly larvae are sensitive to toxins produced by various geosmin makers, Stensmyr suggests. It could also have to do with competition for food. Some molds, such as Penicillium, that produce geosmin eat the yeast that grows on rotting fruit. Since fruit fly larvae also eat yeast, the presence of mold, as signaled by geosmin, means that larvae laid on a particular piece of fruit could starve.

Follow-up research from Stensmyr found that female Aedes aegypti mosquitoes, which possess a very similar geosmin-specific receptor, react completely differently. “They loved it,” he says. This makes sense, given that mosquitoes are insensitive to the toxins the bacteria produce and, in fact, mosquito larvae eat geosmin-making bacteria. Stensmyr notes that female mosquitoes in his study prioritized egg-laying sites where geosmin was present. “If you just look at mosquitoes and flies, which are not too distantly related, this compound seems to be very important,” he says. “But it has different meanings.”

Stensmyr says it’s likely that a huge number of animals are capable of detecting geosmin. Even humans are highly sensitive to it, being able to smell geosmin at concentrations as low as 400 parts per trillion. “We have the example of nematodes, we have it from insects, we have it from humans; we have a whole range of animal phyla in between that possibly also react to this chemical, or can use it in one way or another.” Indeed, some animals respond to the compound in ways that appear totally unrelated to the bacteria. Research from the 1990s suggests geosmin might help European glass eels find freshwater, a function that Stensmyr speculates may have been used by humans’ ancestors as well. And low, but not high, concentrations of the chemical appear to suppress stinging behavior in honey bees.

Geosmin may hold yet more secrets. Zaroubi notes, for example, that fungal strains that produce the chemical don’t appear to use the same gene pathway as bacteria to make it, meaning perhaps that geosmin production has evolved multiple times independently. Findlay adds that the research can help scientists view aposematism in a new way: from the perspective of the predator, rather than just the prey producing the don’t-eat-me signal. Aposematic signals depend “on both the sender and the receiver of the signal,” he says. “In our case, we have pretty full control over the genetics of these nematodes. So we can interrogate the evolution from both sides, from more than one angle. I’m super excited about that.”