Evolution of color vision: transgenic mice see red
This week’s paper, "Emergence of novel color vision in mice engineered to express human cone pigment", by Gerald Jacobs and colleagues at UC Santa Barbara and Johns Hopkins Medical School (Science 315:1723), is yet another experimental study that increases our understanding of how repeated cycles of natural selection, each producing a fairly small change, can lead to adaptations that may seem irreducibly complex.
Most humans have three different photopigment color sensors, as do our closest relatives. Many other mammals, including mice, have only two. Three-color vision is useful for many purposes, from identifying higher-protein leaves to eat (Nature 410:363) to telling which wire to cut to disarm the nuclear bomb buried under the stadium. But eventual usefulness isn’t enough for a trait to evolve. If a series of steps is required, each step must be beneficial, or at least not lethal. Such a series of steps has been worked out for the evolution of optically sophisticated eyes from light-sensitive spots (Proc. Roy. Soc. B 256:53), but what about color vision?
New World monkeys have a form of color vision that may have been a stepping stone between the two-color vision of other mammals and the three-color vision we apes enjoy. One of their photopigment genes is on the X sex chromosome; another is on a nonsex chromosome. Males have one X chromosome (and one Y sex chromosome), so two different color receptors. Females with the same photopigment on both X chromosomes also have only two different color receptors. But "heterozygous" females (with two different X chromosomes), can have two different photopigments on their X chromosomes, plus one on the nonsex chromosome, for a total of three photopigments, each sensitive to different colors of light.
How could this form of three-color vision evolve from the two-color version in other mammals? Although mutation of the photopigment gene on the X chromosome would yield some heterozygous females with three photopigments, I would have thought that three-color vision would have additional requirements (special neurons, etc.). But apparently I would have been wrong.
Jacobs and colleagues genetically transformed mice, adding a human gene for a photopigment not normally present in mice. They then tested color vision in heterozygous females, which had three different photopigments: the two normal mice versions (X and non-X), plus the human version on their second X chromosome. These mice varied in the proportion of the two X-linked photopigments. Those with roughly equal proportions were usually able to tell the difference between colors that looked the same to normal mice. The tests involved distinguishing between differently-illuminated color panels, presumably in exchange for mouse treats.
So specialized neuronal circuitry for each photopigment apparently isn’t essential for at least simple three-color vision. The authors suggest that inactivation of different X chromosomes in different cones in the mouse eyes may have helped the mouse brains learn to distinguish nerve signals linked to photopigments that respond to different colors of light.
Once this simple color vision system was established, further evolution of the nervous system could lead to further improvements, so it seems likely that monkeys have better color vision than these transgenic mice. But why haven't monkeys evolved an ape-like color vision, which works for males as well as females? The required evolutionary changes in the nervous system may be trickier for photopigment genes on nonsex chromosomes. Or maybe the benefits of three-color vision aren’t great enough for natural selection to favor allocation of more brain cells to processing additional color information, except in species with large brains. Or, for monkeys living in troops, maybe there are always enough heterozygous females around to serve as color consultants.
The same issue of Science has 1) a paper showing that fewer genes have been transferred from mitochondria (respiratory organelles descended from intercellular symbionts, and retaining some of their own genes) to the cell nucleus in plants that out-cross (produce seeds with another plant rather than by self-pollination), perhaps because out-crossing can scramble interactions between nuclear and mitochondrial genes, and 2) a report on a virus-defense system in bacteria in which viral DNA from a previous infection is used against new infections, analogous to the vertebrate immune system but using a very different mechanism. As in previous weeks, I haven't seen any papers on "intelligent design" this week, maybe because I've only been looking in scientific journals.