A trail of tears has led a team of Japanese researchers to an unusual aspect of courtship in mice, details of which are published on page 898 of this issue.

The group, led by Kazushige Touhara at the University of Tokyo, had been investigating the chemical signals used by mice to recognize a potential mate. The team's first port of call was a selection of chemicals that had previously been flagged up as potential mouse pheromones. But when they began studying these compounds, the researchers discovered that they didn't really fit the bill. The compounds did not seem to be recognized by the organ in the nose associated with pheromone detection in behaving mice. There was only one conclusion. “There must be other pheromones,” Touhara thought.

And so the hunt began. The first place the scientists looked was in mouse urine, but they found no suitable candidates there. Then Hiroko Kimoto, one of Touhara's graduate students, examined a little-known gland below the ear and got a promising response. Turning to a book on mouse anatomy, she realized that this gland was a type of tear gland.

“It took time to identify the secreting organ, because nobody expected the pheromone to be released from the eyes,” Touhara says. “Apparently, no one has looked at this gland closely before.”

The researchers investigated further and were surprised to find that the active substance secreted by the gland was non-volatile. This is unusual for pheromones, as they are usually volatile compounds that are detected by the nose. If the new compound was the elusive mouse pheromone, mice would have to detect it by touch.

Kazushige Touhara and some of his mice.

To confirm the team's suspicions, Kimoto set about purifying the peptide, now named ESP1. This was an involved process, and meant extracting the peptide from some 80 mice. The team then tested each fraction on mice to see whether it caused gene expression associated with pheromone recognition. This stage took about a week per mouse and required scores of mice to get a statistically significant result.

But even with positive data from this work, the researchers couldn't be certain that ESP1 was a pheromone — the gene-expression work could not prove that the peptide generated an electrical signal in the pheromone-detecting tissue in the mouse nose. “The electrophysiology that we show in the paper was an important experiment to prove that the isolated pheromone was real,” Touhara says. Koji Sato did these experiments, and Sachiko Haga found that ESP1 is indeed recognized by a suspected pheromone receptor.

Although the team is confident it has now found its pheromone, Touhara says that there is a lot of work ahead. He wants to look at the evolution and function of the ESP gene family, and pick apart its neurological mechanisms. Ultimately, he would like to find out more about ESP1's receptor and how the peptide binds to it, and reveal where the signal is integrated in the brain and what effects it has on behaviour.