Why Sensory Ecology Needs to Become More Evolutionary
25
We would expect particularly strong constraints, if pigments sensitive at two
peak wavelength values are separated by adaptive valleys, so that, for example,
two combined changes are necessary to alter a receptor's sensitivity, but each
single mutation renders the pigment nonfunctional. We would also expect that
pigment adaptations might be compromised by molecular spell-checkers that pit
protein folding requirements against spectral tuning (Nakayama et al. 1998).
Each class of color receptor contains a distinct visual pigment, which consists of
two components. One is the chromophore, retinal (or one of its congeners), which
changes its configuration on absorption of a single photon (Seki and Vogt 1998).
The other component is a protein moiety, the opsin (Fig.l ). Opsins are integrated
into the membrane of photoreceptive organelles of the receptors, and contain
about 370 amino acids (Deeb and Motulsky 1996). They contain seven
transmembrane helices, arranged in a quasicircle so that they form a pocket. This
pocket holds the chromophore. Specific amino acids in the transmembrane helices
oriented towards the center of the pocket (and, thus, interact electrostatically with
the chromophore) are responsible for spectral tuning (Hope et al. 1997).
To understand how visual pigments in insects changed over evolutionary time, it
is informative to evaluate the phylogeny of their opsins. To this end, we compared
the amino acid sequences of the opsins of 54 species of arthropods, as well as
different opsins found within the same animal species. The basis of such an
analysis is that one groups together those proteins that are most similar; the nodes
of the tree represent (hypothetical) ancestral opsins (Goldsmith 1990). It is
immediately apparent that invertebrate opsins fall into distinct functional clades
according to spectral sensitivity (Fig.2).
There is one cluster of UV pigments, a distinct group of blue pigments, a third
group of long wave pigments, which includes pigments with peak sensitivity from
green to red. Most interestingly, chelicerate and crustacean green sensitive
pigments are more similar to insect green pigments than they are to UV and blue
pigments, which confirms the phylogenetic analysis above: opsin clades diverged
from one another before the major groups of arthropods diverged, and it is
therefore likely that ancient arthropods already possessed (at least) UV and green
visual pigments. Somewhat puzzling are the origins of two clades of blue-green
(480 nm) pigments of Drosophila and the crab Hemigrapsus sanguineus. They
might be the result of convergent evolution, or of several independent gene losses.
On the basis of electrophysiology, not many insects are expected to have pigments
that fall in the 480 nm spectral class, which favors the latter hypothesis. We
advocate testing this hypothesis by looking for opsin pseudogenes to members of
this clade.
Another important result of this comparative analysis is that rather large
portions of opsin amino acid sequence can be exchanged while the spectral
function remains surprisingly constant. Bee UV and D. melanogaster Rh3, which
have nearly identical Amax values (353 and 345 nm) differ by 36.6%, while the two
Limulus sequences, which are thought to differ by I 0 nm, differ by only 1 %!
Of course, our argument is dependent on the assumption that differences
between opsins do not represent cases of convergent evolution, possibly as a
response to similar selective pressures.
25
We would expect particularly strong constraints, if pigments sensitive at two
peak wavelength values are separated by adaptive valleys, so that, for example,
two combined changes are necessary to alter a receptor's sensitivity, but each
single mutation renders the pigment nonfunctional. We would also expect that
pigment adaptations might be compromised by molecular spell-checkers that pit
protein folding requirements against spectral tuning (Nakayama et al. 1998).
Each class of color receptor contains a distinct visual pigment, which consists of
two components. One is the chromophore, retinal (or one of its congeners), which
changes its configuration on absorption of a single photon (Seki and Vogt 1998).
The other component is a protein moiety, the opsin (Fig.l ). Opsins are integrated
into the membrane of photoreceptive organelles of the receptors, and contain
about 370 amino acids (Deeb and Motulsky 1996). They contain seven
transmembrane helices, arranged in a quasicircle so that they form a pocket. This
pocket holds the chromophore. Specific amino acids in the transmembrane helices
oriented towards the center of the pocket (and, thus, interact electrostatically with
the chromophore) are responsible for spectral tuning (Hope et al. 1997).
To understand how visual pigments in insects changed over evolutionary time, it
is informative to evaluate the phylogeny of their opsins. To this end, we compared
the amino acid sequences of the opsins of 54 species of arthropods, as well as
different opsins found within the same animal species. The basis of such an
analysis is that one groups together those proteins that are most similar; the nodes
of the tree represent (hypothetical) ancestral opsins (Goldsmith 1990). It is
immediately apparent that invertebrate opsins fall into distinct functional clades
according to spectral sensitivity (Fig.2).
There is one cluster of UV pigments, a distinct group of blue pigments, a third
group of long wave pigments, which includes pigments with peak sensitivity from
green to red. Most interestingly, chelicerate and crustacean green sensitive
pigments are more similar to insect green pigments than they are to UV and blue
pigments, which confirms the phylogenetic analysis above: opsin clades diverged
from one another before the major groups of arthropods diverged, and it is
therefore likely that ancient arthropods already possessed (at least) UV and green
visual pigments. Somewhat puzzling are the origins of two clades of blue-green
(480 nm) pigments of Drosophila and the crab Hemigrapsus sanguineus. They
might be the result of convergent evolution, or of several independent gene losses.
On the basis of electrophysiology, not many insects are expected to have pigments
that fall in the 480 nm spectral class, which favors the latter hypothesis. We
advocate testing this hypothesis by looking for opsin pseudogenes to members of
this clade.
Another important result of this comparative analysis is that rather large
portions of opsin amino acid sequence can be exchanged while the spectral
function remains surprisingly constant. Bee UV and D. melanogaster Rh3, which
have nearly identical Amax values (353 and 345 nm) differ by 36.6%, while the two
Limulus sequences, which are thought to differ by I 0 nm, differ by only 1 %!
Of course, our argument is dependent on the assumption that differences
between opsins do not represent cases of convergent evolution, possibly as a
response to similar selective pressures.