Primates have trichromatic rather than dichromatic vision, because their genomes encode three rather than two types of cone photopigment. But was the acquisition of a new receptor type in a dichromatic ancestor sufficient for trichromacy to evolve, or was more complex neural rewiring required? A new study shows that mice can acquire new colour vision simply by expressing an additional human photoreceptor.

Most mammals have just two types of photoreceptor, sensitive to short (S) or medium (M) wavelengths of light. However, in most New World monkeys, the X-chromosomal gene that encodes the M receptor has a second allele that is sensitive to long (L) wavelengths; therefore, heterozygous females have S, M and L receptors and are consequently trichromatic. In Old World primates, the X-chromosomal locus has duplicated and diverged, giving all individuals trichromatic vision; however, the New World monkey system is thought to be ancestral.

Gerald Jacobs and colleagues attempted to recreate the New World monkey situation in mice by replacing the M locus with the human L locus and breeding heterozygous females. The process of random X-chromosomal inactivation in clones of cells means that there are separate populations of M- and L-expressing cones in the eyes of such individuals, but the proportions of the two are variable. Initial physiological tests showed that the L cones were functional, so the authors carried out behavioural tests to assess new colour vision.

Mice were trained to discriminate between monochromatic light at two different medium and long wavelengths. As expected, homozygous mice did not respond to this training. Although the first two heterozygous mice that were tested did not respond either, the M:L ratio was particularly skewed in both animals. Other mice with more balanced proportions of M and L receptors were able to discriminate between several different pairs of wavelengths.

These results suggest that an ancestral female primate with a mutant allele that changes the wavelength sensitivity of the M receptor could have had an immediate selective advantage. Subsequent genetic changes might have refined the downstream circuitry for trichromatic vision but the initial step could have been relatively simple. Further work should elucidate these subsequent changes, as well as the relationship between X-inactivation, the M:L receptor ratio and trichromatic vision.