Friday, June 19, 2009

Eyes abound

Unraveling and disentangling homology and convergence is one of the most fascinating endeavors in biology. Homology indicates common origin and maintenance, and is often taken as evidence for importance: ancient features are thought to be maintained because they are too useful to dispose of during evolution. In contrast, convergence, is the separate invention of similar features or functions during evolution. Convergence is taken as evidence for an element of predictability in evolution. For a simple example, fish and dolphins are highly convergent, and we can use this knowledge to predict that when vertebrates evolve to live in the ocean, that evolution will produce particular features like flippers/fins.

I recently came across a fascinating paper, arguing that structures that interact with light - either by altering or receiving it - are highly convergent, and may even be homologous at some level. Namely, bird feathers that reflect UV light have some striking similarities with eyes! Furthermore, a paper I am a co-author on just came out in PNAS that further supports this general claim. We found that the light producing structure of a bioluminescent squid shares many features with eyes, including the ability to detect ('see') the light it produces!

First, the feathers. Bleiweiss studied the uv/blue feathers of Tanagers and Bluebirds. In nature, short wavelength colors are often produced by structures, as opposed to pigments which produce longer wave colors like orange and red. Structural colors work by differentially interfering/reflecting different wavelengths of light. A familiar example of structural color is a CD/DVD. These disks contain grooves that are spaced very closely together. Because the spacing is similar to the wavelengths of visible light, interference of certain wavelengths occurs, leaving other specific wavelengths that we see as color. These spaced grooves are called diffraction gratings, and they are known in nature, for example on the antennae of some ostracod crustaceans which reflect blue light. Bluebird and tanager feathers do not use diffraction gratings, but instead a different structural mechanism. In the course of studying these feathers, Bleiweiss found some striking similarities with eyes. Perhaps similar to fins/flippers that push water for locomotion, the physical similarities of feathers and eyes may reflect convergence due to shared physical necessities of interactions with light.

An attractive tanager.  Image from

What are these similarities between eyes and structurally colored feathers? First is a wide, domed surface to receive the light. Second is tissue that is transparent to some light but reflective of other wavelengths. In eyes, this is the cornea and lens, which are transparent to much light, but often reflect UV (human retinas are actually sensitive to UV, except the light never gets there because the cornea and lens reflect it.). Tanager feathers have physically similar tissues with similar properties to reflect UV/blue light and allow other light to pass through. Third, there is a large central space in both eyes and ocular feathers: eyes contain humors and feathers a space filled with gas (air). Finally, at the bottom is a reflective layer. In eyes, this is the tapetum lucidum, which produces eye shine in cats, coons and other night-active animals. Again, optical feathers share a similar pigmented structure also designed to reflect light.

These similarities seem to be a perfect case of convergent evolution: two structures that perform physically similar functions (light gathering, or light reflecting) have converged on similar solutions. However, Bleiweiss also raises the intriguing possibility that eyes and feathers actually share some (partial) homology. Complex traits like eyes and feathers are made of many components, each with a potentially different evolutionary history. Amazingly, some of the genetic components, developmental features, and signal transduction cascades of eyes and feathers are also shared, in addition to their functional similarities. These similarities might be evidence of a deep shared ancestry between multiple organs, including eyes, feathers, and even teeth.

I was particularly struck by Bleiweiss' paper because I've been thinking about similar things in the context of a collaboration studying the light-producing organ of a squid that yielded a PNAS paper this week. Not unlike tanager feathers and eyes, the convergence of squid light-producing organs and eyes has long been noted. Many squid, including Euprymna scolopes, the object of our study, are bioluminescent. Euprymna seems to use its bioluminescence for camouflage. In the ocean, most light comes from straight above, so animals would cast a distinct and conspicuous shadow below them. Instead of eliciting the shadow response of a predator or prey, Euprynma matches downwelling light to make itself more cryptic. The light is produced in a light organ that houses symbiotic bacteria. It is the bacteria that actually generate the light. Consistent with Bleiweiss' general hypotheis, this light organ has many similarities with eyes.

Light organs and eyes both have lenses. Eyes focus incoming light for better visual acuity, and light organs focus outgoing light, similar to a flashlight. Eyes and light organ have an open space below the lens, and a pigmented layer opposite to the lens. In addition to these similarities, we found that the light organ responds physiologically to light using the same genes (opsin and its signaling components) that are used in eyes. Just as with optical feathers, squid light organs are functionally convergent, yet also share structural components in common, indicating some elements of homology.

Euprymna scolopes Hawaiian Bobtail Squid.  Picture by Chris Frazee, image from

These findings indicate an interesting new research program using the tools of phylogenetics. By reconstructing the evolutionary history of multiple components of convergent/partially homologous traits, we can see how and when these components came together, illustrating the pathways by which evolution has produced new features. This will allow a richer, more fundamental understanding of the origins of biodiversity and complexity, topics that intrigue everyone.

Bleiweiss, R. (2009). Feathers with Ocular Architecture: Implications for Functional and Evolutionary Similarities of Visual Signals and Receptors Evolutionary Biology, 36 (2), 171-189 DOI: 10.1007/s11692-009-9059-6

Tong, D., Rozas, N., Oakley, T., Mitchell, J., Colley, N., & McFall-Ngai, M. (2009). From the Cover: Evidence for light perception in a bioluminescent organ Proceedings of the National Academy of Sciences, 106 (24), 9836-9841 DOI: 10.1073/pnas.0904571106


gillt said...

Don't you mean orthology, not homology, when discussing cross-species similarities, be it geno or phenotype?

Todd Oakley said...

gillt - no, I don't think I mean orthology per se, when discussing cross-species similarities, although I wouldn't rule it out. Paralogy could also be a very useful word here. I used homology as a word to include either orthology or paralogy and was not explicitly referring to any one level (geno or phenotype).

gillt said...

Oh, I assumed your example of dolphins and fish sharing fins was relevant here.

Todd Oakley said...

gillt "Oh, I assumed your example of dolphins and fish sharing fins was relevant here."

Dolphin and fish fins would be a case of convergence, or homoplasy (perhaps that is what you meant in your original comment instead of orthology).

But the point is that even convergent traits can have some elements of homology. For example, the bones of dolphin fins and a few fish (lobe fin fish) are homologous, even though we consider the overall structure and function of the fins to be convergent in those 2 taxa (lobe fins and dol-fins).

Anonymous said...

Aren't eyes and feathers both made of bird cells containing bird DNA?

Of course they are going to have some similar features suggesting a shared origin.

Bob Miller said...

Hi Todd,
Could this imply that some "eyespots" in moths, fish etc actually could have sensory capability? The evidence for their anti-predator function seems weak.
Bob Miller

Todd Oakley said...

zayzayem - Of course all animal cells share some homologous features. However, what I described goes far beyond that.

Bob - Interesting thought you have ( and an easily testable hypothesis) about butterfly eyespots. Butterflies do have photoreceptors in strange places, like on their genitalia.