This topic is among the most clear-cut examples of “gene-environment interactions” topics…!! During evolution, when vertebrates first left the water and ventured onto land, they encountered a visual world that was radically different from that of their aquatic ancestors. In order to survive successfully, these “new” land species were required to adapt visually, in terms of being able to find food, avoid predators, and sexually reproduce. “The need to see acutely” — represents the environmental pressure; and necessary (and relatively evolutionarily “quick”) changes in visual networks of the eye (and/or brain) — represent the response by genes…
Fish exploit the strong wavelength-dependent interactions of light with water by differentially sending visual image signals — from as many as five spectral photoreceptor types into distinct behavioral programs. However, out of the water — the same spectral rules do not apply, and this evolutionary adaptation required rapid changes in response to this new environmental pressure.
Early tetrapods [e.g., Kenichthys from China ~395 million years ago (MYA), Gogonasus & Panderichthys ~380 MYA, and then salamanders with four legs] soon evolved the double cone, still a poorly understood pair of new photoreceptors that increased the “ancestral terrestrial” complement from five to seven types of photoreceptors. Subsequent non-mammalian lineages differentially adapted this highly parallelized retinal input strategy for their diverse visual ecologies. In contrast, mammals (first appearing ~225 MYA) shed most of their ancestral photoreceptors and converged on an input strategy that is extraordinarily general. In eutherian mammals (i.e., animals born via placenta), including humans, parallelization emerged gradually during evolution, as the visual signal began to traverse the layers of the retina and onwards, into the brain. [“Parallelization” can be defined as “parallel genotypic adaptation,” i.e., the independent evolution of homologous loci (similar genes) to fulfill the same function (trait) in two or more lineages. N.B. — these changes need not be identical, just functionally equivalent.]
Vertebrate vision first evolved in the water, where (for >50 million years) it was consistently based on visual signals from five anatomically and molecularly distinct types of photoreceptor neurons: rods, as well as ancestral red, green, blue, and UV cones (expressing RH, LWS, RH2, SWS2, and SWS1 opsins, respectively). In the water, these five input streams are probably best thought of as parallel feature channels that deliver distinct types of information to distinct downstream circuits. This is because water absorbs and scatters light in a wavelength-dependent manner (see Fig 1A in attached pdf file), which means that “beyond color,” different spectral photoreceptor channels inherently deliver different types of visual information.
Aquatic visual systems have recently been proposed to evolutionarily reach “answers” that exploit these differences. In this view, photoreceptors represent parallel channels that are differentially wired to drive and/or regulate distinct behavioral programs (see Fig 1B in attached pdf file): First, rods and ancestral red cones are the eyes’ primary brightness sensors; they are used for general-purpose vision and to drive circuits for body stabilization and navigation. Second, ancestral UV cones are used as a specialized foreground system, primarily wired into circuits related to predator–prey interactions and general threat detection. Third, ancestral green and blue cones probably represent an auxiliary system, tasked with regulating, rather than driving, the primary red/rod and UV circuits.
This ancestral strategy exploits the specific peculiarities of aquatic visual worlds; however, in air the same rules do not necessarily apply. For example, in water, object vision can be a relatively easy task, because background structure tends to be heavily obscured by an approximately homogeneous aquatic backdrop. At short wavelengths, including in the UV range, this effect can be so extreme that no background is visible at all. Many small fish exploit this fact of physics to find their food. Out of water, this and many other “ancestral visual tricks” no longer work, because in air, contrast tends to be largely independent of viewing distance: everything is visible at high contrast. Accordingly, when early would-be tetrapods started to crawl out of the water, strong selection pressures would have favored a functional reorganization of some of these inherited aquatic circuits; nowhere is this more evident than at the level of the photoreceptors themselves.
One of the earliest and perhaps most important retinal circuit changes was the emergence of the double cone, which took the “aquatic ancestral” photoreceptor complement of five to a “terrestrial ancestral” complement of seven (see Fig 1 in attached pdf file). The visual systems of all extant tetrapods, including humans, directly descend from this early “terrestrialized” retinal blueprint. However, from there, different descendant lineages have taken this highly parallelized retinal input strategy and embarked upon radically different visual paths. Most lineages, including those that led to modern-day amphibians, reptiles, and birds — have retained the terrestrialized ancestral blueprint, modifying upon it to suit their unique visual ecologies.
Mammals, however, have ended up on a very different path. Their early synapsid ancestors gradually shifted some of their visual systems’ “heavy lifting” out of the eye and into the brain. Along this path — whether as cause or consequence — descendant lineages gradually decreased their photoreceptor complements from seven types to six, then five, and eventually to the mere three that we see in eutherians today (see Fig 1C in attached pdf file): Rods (RH), as well as ancestral red (LWS) and UV cones (SWS1).
Primates, including humans, have then taken this eutherian strategy to the extreme: >99.9% of all photoreceptors in our eyes are either rods or ancestral red cones (including both “red-” and “green-shifted LWS variants”), the ancestral “general purpose” system of the eye. The remaining 0.1% is what is left of the ancestral UV system, today expressing a blue-shifted variant of the SWS1 opsin (hence, often called “blue cones,” not to be confused with ancestral blue cones that express SWS2). In concert, the “three” cone variants drive achromatic vision (although with limited contribution from ancestral UV cones), while in opposition they serve color vision.
However, this “textbook strategy” is far removed from the original aquatic circuit design and probably quite unique to our own lineage. Accordingly, for understanding vision in a general sense, and to understand our own visual heritage, it will be critical to respect the vertebrates’ shared evolutionary past. Here, vision is built on a retinal circuit design that begins with major parallelization — right from the original evolutionary input.
DwN
PLoS Biol Jan 2024; 22: e3002422
COMMENTS:
By coincidence, I happen to be re-re-reading the Biology of Spiders by Rainer F. Foelix. It is an incredible book full of pictures, including electron microscope picture of tissue preparations (I like books that have lots of pictures). In terms of vision, diverse animals share the structures of cones and rods, corresponding to select frequencies of the electromagnetic spectrum. This confirms in my mind, once again, that the inherent conditions for life (and reproduction) are fundamental manifestations of energy and matter. That spiders’ eyes have frequency-tuned structures that report the world outside the spider — to the world inside the spider’s brain — boggles my mind. … Every species seems directed intentionally not just to survival, but survival to reproduce life further in that species.
Hence, … again … the one thing I can conclude with certainty about the God of our Universe is “oblivion (or “randomness” or “chaos” or “infinite entropy” is not an option.”
I regret that I missed this January publication, when I was teaching Evolutionary Neurobiology this spring! The topic of “Vision” has always been a great opportunity to talk about selective environmental pressures, convergent versus divergent evolutionary trends, etc.
Thanks for sharing!
Reply:
Chris, while working on summarizing this paper, I came very close to discussing “divergent” versus “convergent” evolution (but I felt my summary was already too long). ☹
“Divergent evolution” is when individuals of one species, or closely related species, acquire enough variations in their traits (due to novel, constant environmental pressures) that it leads to two distinct new species (each of which can usually survive better in two distinctly different niches). A simple example would be a water-living fish evolving into a land-living salamander.
On the other hand, “convergent evolution” occurs when two unrelated species develop similar traits because they experience similar environments; moreover, the DNA responsible for these traits can either be homologous genes in the same gene family, or a completely different genetic network. An example would be “synthesis of antifreeze protein” by fish in Antarctica versus fish in the Arctic Ocean. Authors [see attached article] used the term “parallelization” instead of “convergent evolution.” Clearly, over the past 4 billion years, divergent evolution has occurred far, far more commonly than convergent evolution…
DwN
This is a terrific review, which I have downloaded for my files!
From the engineering point of view, the striking difference between underwater vision and vision through the air is the much greater “clutter” for air vision. An interesting example of this is America’s first US missile early-warning satellites which intentionally looked for 2.7-micron infrared (IR) radiation from burning rocket plumes that had emerged from the cloud deck. Because of the heavy absorption of water vapor at 2.7 microns, however — much of the cloud clutter and ground clutter for clear skies gets suppressed. As described in this excellent essay, fish use short-range, low-clutter ultraviolet (UV) for catching nearby prey for reasons similar to those problems encountered by America’s first US missile early-warning satellites.
Best wishes,
MORE COMMENTS:
Attached is the “Accelerated Article Preview” — which has been peer-reviewed and will appear soon in the complete published form. However, the journal felt these findings were sufficiently worthy of rapid distribution to scientists who might find these data extremely interesting.