Nematodes see without eyes


The humble nematode Caenorhabditis elegans is a millimeter-long roundworm which eeks out its existence in the soil and feeds on bacteria.

Because it lives in a dark environment, and lacks specialized light-sensing organs, the nematode has always been assumed to be completely blind.

However, a new study published online in Nature Neuroscience shows that C. elegans they possess neurons which are sensitive to light. As well as showing for the first time that C. elegans has a rudimentary sense of vision, the findings also shed some light on the evolution of the eye.

Despite lacking eyes, nematodes somehow manage to avoid the light and remain in their natural environment. Shawn Xu and his colleagues of the University of Michigan’s Life Sciences Institute therefore reasoned that there must be some mechanism which keeps them in the dark.

Xu’s team first tested whether or not the worms could respond to light in any way, by placing individual worms onto small plates and viewing them under the microscope. They found that pulses of ultraviolet light focused onto any part of the worms’ bodies elicited robust movements away from the light source. This avoidance response occurred within 1 second, and was found to be “dose-dependent”; i.e. the distance moved away from the light source, and the duration of the movement, increased with the intensity of the pulse of light.

Xu’s team found that prolonged exposure to UV light first paralyzes and then kills the worms. The negative phototactic response they observed therefore seems to be essential to the nematode’s survival – it serves as a protective mechanism which keeps the worms away from bright light and within their dark natural environment.

The researchers then sought to determine which cells in the C. elegans nervous system are responsible for this behaviour, by using pulses of laser light to destroy identified cells. This is a relatively simple procedure, as C. elegans has a very simple nervous system consisting of just 302 neurons.By selectively destroying combinations of sensory neurons, the researchers identified 4 pairs of candidate photoreceptors which, when destroyed together, led to a severe deficit in the avoidance response.

Further investigation showed that these neurons respond to light by generating small inward currents. In the vertebrate retina, the light-sensitive cells convert light into electrical impulses by means of proteins called cyclic nucleotide-gated (CNG) ion channels. The C. elegans genome encodes 6 of these proteins, so Xu’s team wondered if they also mediate the phototaxis they had observed.

To test whether or not this is the case, they looked at worms with mutations in the CNG channels. They found that worms carrying mutations in one of the channels, called tax-2, had severe defects in phototaxis. But when the tax-2 gene was re-inserted into these mutants, in such a way that it was expressed specifically in the candidate photoreceptors, the light avoidance resposne was rescued. Further confirmation of these findings was obtained with a chemical called LY83857, which blocks the enzymes that produce cyclic GMP, the nucleotide which binds to and activates the CNG channels.

What does this tell us about the evolution of the eye? From classical anatomical studies, it was believed that the eyes of vertebrates and invertebrates evolved independently of each other, because they are so different in structure. These findings, however, show that cyclic GMP/ CNG phototransduction is an evolutionarily ancient biochemical pathway, which may have been present in the urbilateria, the hypothetical last common ancestors of all bilaterally symmetrical organisms which are believed to have lived some 600 million years ago.

(Image from the Max Planck Institute for Developmental Biology)


Ward, A., et al (2008). Light-sensitive neurons and channels mediate phototaxis in C. elegans. Nature Neurosci. DOI: 10.1038/nn.2155

One thought on “Nematodes see without eyes

  1. Interesting! Read the paper, but confused about the logarithm they used. So in the paper, is log(I/Io) 10, e or 2 based? I just assume it’s 10 based.

Comments are closed.