Can Disease-Sniffing Dogs Save the World’s Citrus?

Once trained, canines can detect citrus greening disease earlier and more accurately than current diagnostics

citrus detector dog 1
A detector dog named Szaboles, trained to sniff out the bacterial pathogen Candidatus Liberibacter asiaticus in a citrus orchard. Courtesy of Tim R. Gottwald

Tim Gottwald will never forget the sight: the mottled yellow leaves, the withered branches, the small, misshapen fruits, tinged with sickly green. These were the signs he’d learned to associate with huanglongbing, or citrus greening—a devastating and wildly infectious bacterial infection that slashed the United States’ orange juice yields by more than 70 percent in the span of a decade.

“It’s like a cancer,” says Gottwald, a plant pathologist with the United States Department of Agriculture. “One that’s metastasized, and can’t be eradicated or cured.”

Once they’ve begun to sport splotchy foliage and stunted fruit, trees can be diagnosed with a single glance. A symptomatic plant, Gottwald says, is a diseased one. Unfortunately, the converse isn’t true: Infected trees can appear normal for months, sometimes years, before visibly deteriorating, leaving researchers with few reliable ways to suss out sick citrus early on—and giving the deadly bacteria ample opportunity to spread unnoticed.

Now, Gottwald and his colleagues may have a creative new strategy to fill this diagnostic gap—one that relies not on vision, but smell. They’ve taught dogs to recognize the telltale scent of a huanglongbing infection—an odor that eludes the attention of humans, but consistently tickles the super-sensitive schnozz of a mutt. Once trained up, canines can nose out the disease within weeks of infection, trouncing all other available detection methods in both timing and accuracy, the researchers report today in Proceedings of the National Academy of Sciences.

“This is a major step in the development of what could be a really important early detection tool,” says Monique Rivera, an entomologist and citrus pest expert at the University of California, Riverside who wasn’t involved in the study. “It could give growers information about potential exposure … to the causative bacteria.”

First described in China in the early 1900s, huanglongbing has now crippled orchards in more than 50 countries around the globe. Fifteen years ago, the scourge took hold in Florida, where infected trees are now the norm; the state’s $9 billion citrus industry, once the second largest in the world, is now on the verge of collapse. From oranges to grapefruits to lemons, no variety of citrus is immune.

As the disease continues to creep into new regions, researchers worldwide are scrambling to contain it. But the task has proved difficult: No effective treatments, cures or vaccines exist for huanglongbing, the product of a bacterium called Candidatus Liberibacter asiaticus (or CLas, pronounced “sea lass”) that’s ferried from tree to tree by winged insects. Scientists have also found the microbes to be extraordinarily difficult to grow and study in the lab.

Can Disease-Sniffing Dogs Save the World’s Citrus?
The 20 canines in the study trained to detect CLas-infected citrus trees. Gottwald et al., PNAS, 2020

Currently, the only surefire way to curb citrus greening’s spread is to extract and eliminate infected trees. This strategy depends entirely on early detection—“one of the biggest problems in the field right now,” says Carolyn Slupsky, a plant pathologist at the University of California, Davis who wasn’t involved in the study. Spotting an asymptomatic infection by eye is essentially impossible. And though genetic tests can sometimes pinpoint microbes in apparently healthy trees, their success rates are low and inconsistent, due in part to the patchiness with which CLas distributes itself in plant tissue.

In many ways, huanglongbing is “the perfect storm of a disease,” Slupsky says.

But canines may just be the perfect candidates to lend a helping paw. With a sense of smell that’s 10,000 to 100,000 times more powerful than a human’s, dogs are superstar sniffers, capable of nosing out everything from bombs to drugs. In recent years, they’ve even been deployed to detect pathogenic diseases like malaria. Infections, it turns out, stink—and dogs definitely take notice.

To see if pooches’ powers of perception might extend to huanglongbing, Gottwald and his team taught 20 dogs to pick up on the smell of citrus plants with known infections, rewarding the pups with toys when they identified the correct trees. After just a few weeks of training, the newly-minted citrus sniffers were picking out infected trees with about 99 percent accuracy. Put in pairs to corroborate each other’s results, the dogs got close-to-perfect scores.

Gottwald was floored. “I wasn’t surprised [the dogs] could do it,” he says. “But I was surprised by how well they could do it. It was pretty amazing.”

The team then pitted the pups against a common but expensive laboratory test that’s often used to verify the presence of CLas DNA in suspicious-looking citrus. After spiking the microbes into 30 trees, the researchers mixed the newly-infected plants into rows of healthy ones and allowed the dogs to inspect them on a weekly basis. Within a month, the canines had collectively homed in on every single CLas-positive plant.

The DNA test, on the other hand, had no such luck: Seventeen months into the infection, it was still failing to identify a third of the diseased trees.

If Gottwald’s team sees continued success, “this could be very exciting for [citrus] growers,” who could someday keep dogs around as a fast and relatively inexpensive way to survey their orchards, says Phuc Ha, a microbiologist at Washington State University who wasn’t involved in the study. For now, the most immediate applications lie in disease prevention. But, she adds, should researchers develop treatments for huanglongbing, canines could eventually play a role in curing the condition as well.

Gottwald and his team have already begun to send small teams of citrus-sniffer dogs to inspect vulnerable trees in California and Texas. In both locations, the canines have alerted the researchers to trees that have yet to test positive in the lab.

This, however, evokes the double-edged sword of early detection research: The dogs are so much faster at finding potentially diseased trees that their picks can’t actually be confirmed, Slupsky points out. Maybe they’re more sensitive than the molecular test, and the disease is more widespread than researchers feared. Or maybe the canine’s noses are leading them astray. “Specificity is always an issue,” Slupsky says, “because you’re comparing them to an imperfect test.”

Dogs also come with their own drawbacks. They can tire; they can be distracted. They’re not machines. And while they can make fast work of orchards where infections are rare, their performance will probably plummet in heavily afflicted groves. In an ideal world, Slupsky says, the dogs would serve strictly as a first line of defense, screening trees for further monitoring or testing in the lab. She and her colleagues are hard at work on one such diagnostic, built to detect the unique suite of chemicals infected leaves produce early on.

Many questions remain unanswered. Gottwald still isn’t sure what exactly the dogs are smelling on the plants, though a series of experiments indicate the scent is probably coming from the CLas bacteria themselves. That theory may be tough to test: Though researchers like Washington State’s Ha have now grown CLas in the presence of other microbes, no one has yet managed to isolate the strain in a pure culture, hampering efforts to understand its basic biology and develop precise treatments.

While exciting, the team’s dog-nostic developments ultimately underscore “just how distant we still are from understanding a lot of the mechanistic processes that are going on [with this disease],” Rivera says. But with more collaboration and multidisciplinary work, she adds, “I think we’ll keep heading toward solutions.”