One day in October 2010, a volunteer firefighter named Thomas Fritz cut down a crab apple tree outside his house, and impaled his hand on one of the branches. He dressed the wound, but it still became infected. His doctor gave him a course of antibiotics and sent fluid from the wound to the University of Utah for analysis. The technicians there tested the microbes in the fluid, and found that their DNA was a close match to a bacterium called Sodalis.

By coincidence, the man who discovered SodalisColin Dale—was working at the university. Dale didn’t buy the results. He had only ever seen Sodalis in the cells of insects, and he assumed that it was permanently dependent on these animal hosts. After all, it lacked many of the genes that it would need to live independently. It couldn't possibly survive on its own, much less lurk on a tree branch, or successfully infect a firefighter’s hand.

But the DNA wasn’t lying. The bacterium that had infected Fritz was indeed a version of Sodalis, but a free-living one with a bigger, self-sufficient genome. Dale called it Sodalis praecaptivus—“Sodalis before captivity.” It represents what these microbes might have looked like before, over millions of years, they evolve to be permanent parts of an insect’s body. And it has an odd knack for getting into clumsy people. A few years after analyzing Fritz’s sample, Dale found Sodalis praecaptivus again—this time, in a kid who had impaled himself on a branch while climbing a tree.

Clearly, this microbe is a common part of the environment. And it provides scientists like Dale with a unique look at something that’s very hard to study—how the relationships between animals and microbes first form.

Our world is built on such partnerships. Humans rely on our microbes to digest our food, train our immune system, and sculpt our organs. Many insects are even more dependent, relying on microbes within their cells to provide them with essential nutrients that are missing from their diets. But since these relationships are typically millions of years old, it’s hard to work out how and why the partners first started dancing together.  

Sodalis provides some clues. It’s out there, and it gets in animals. Sometimes, its intrusions lead to temporary infections, as in Fritz, the tree-climbing kid, and insects like stinkbugs and chestnut weevils. Sometimes, it leads to permanent bonds, as per many tsetse flies, cicadas, aphids, and more. As John McCutcheon from the University of Montana once told me, “It’s really good at getting into insect cells. Sodalis is special, and I don’t know why.”

Dale now has an idea about why. His team member Shinichiro Enomoto began deleting genes from Sodalis praecaptivus to see if that changed its ability to colonize insects. One gene called ypeI had a particularly dramatic effect. It controls an ability called quorum sensing, where bacteria sense molecules made by their neighbors to coordinate their growth and behavior. And without this ability, Sodalis turned into a killer. The bacterium normally colonizes grain weevils without harming them. But if Enomoto injected the insects with the mutant strains, they became the lesser of two weevils: They grew lethargic, and died early. “In ten to fourteen days, the insects are on their backs,” says Dale. “We were kind of shocked.”

The team showed that Sodalis contains several genes that are lethal to its hosts. Some make insecticidal poisons. Others create enzymes that break down the chitin in an insect’s external skeleton. But in a normal infection, Sodalis uses quorum sensing to deactivate these genes. “It usually has this very quiet, benign association with its insect host,” says Dale. “But when you knock out that one gene and mess up the quorum sensing, it becomes a killer. It’s like it’s really trying hard to be nice, but it’s one mutation away from being a bad guy.”

So perhaps the thing that makes Sodalis special is its capacity for restraint. “A major problem for a would-be symbiont is that it needs to get into the host but then immediately become as [harmless] as possible,” says Nancy Moran, from the University of Texas in Austin. And Sodalis can do exactly that. It has all the right tools for breaking into an insect’s body and setting up an infection, but once its population gets big enough, it uses quorum sensing to turn down its destructive side and live at peace. “This trick may allow it to regularly enter insect hosts and persist there, increasing the chances of becoming a long-term symbiont,” says Moran.

“It provides a really unique tool to probe microbial features important for establishing associations with a host,” adds Nichole Broderick, from the University of Connecticut. Quorum sensing is probably not its only trick. Sodalis probably also has ways of evading its host’s immune system, or even providing benefits to its host.

Discoveries like this show that among microbes and animals, there’s a very fine line between harmful infections and beneficial alliances. Sodalis might be a useful companion to many insects, but such relationships likely began with a barely concealed capacity for violence and conflict.