Light can be detected in many ways, but light detection is not exactly what we mean when we think of vision. Vision is more complex because it implies the processing of intricate spatial patterns of incoming light along with its intensity and aspects of its spectral characteristics. In higher organisms, vision involves interpretation of complex signals by the central nervous system and translation of these into decision-making that usually includes a directed response, such as by muscular instructions directing movement. Before the response, the patterned aspect of the signal must be detected.
Because light travels in straight lines, an organism receives spatially organized light energy that is a map of the objects from which the energy comes.To see pattern, the organism must have some corresponding form of neurological spatial "map". This is in contrast to the much less spatially organized nature of olfaction, sound, or (sometimes) heat, the essentially nonspatial nature of immunological information, and the temporal arrangement of daylight or air temperature. For spatial processing, there must be a receptor map. Interestingly, as we will discuss in Chapter 16, even senses that do not depend on spatial perception use spatial maps in the brain to decode signal. This is probably a function of the fairly uniform histological, laminar, and synaptic organization of the segments of the brain that decode sensory input.
The hundreds of ommatidia that comprise the compound eyes of insects provide a fixed receptor array, each with its own neuronal connection to the brain. In camera-type eyes, such as those in mammals, the rods and cones in the retina are fixed in location and their corresponding neurons can be connected to the brain in a way that directly conserves spatial relationships, or at least the brain can reconstruct them because the signals from the same matrix position in the retina (or compound eye) are fixed. In mammals, unified stereoscopic vision further integrates the slightly different images from right and left eyes. An important part of this is that there need be no specific a priori aspect of each ommatidium or retinal cell, relative to images that are going to be detected: the organism is set up to detect any spatial light pattern. It does not need, for example, built-in diagrams of its predators, food, or landscape panoramas.
There are many ways in which eyes that are used for spatial inference can be constructed. Figure 14-8 shows some of them. The arrangement and number of eyes vary extensively, even when considering just bilateral cerebral eyes (e.g., see Arendt and Wittbrodt 2001). These include relatively simple larval eye spots in many branches of animal taxa, to more elaborate eyes formed by an array, usually a cup, of retinal cells. In some eyes, this has deepened and been closed except for a small opening, which like a pinhole camera allows a single image to be received in an array of photoreceptor cells on the inside. More precise (in human terms, at least), focused images are formed by eyes with lenses in this aperture, as found in mammals. Each ommatidium in an insect compound eye is a simple eye in itself, with spatially distributed neurons from its individual receptor cells receiving similar signals that are sent to similar regions of the brain.
Comparative anatomy and taxonomy have shown that eyes have evolved independently many times (a classic analysis suggests 40-65) (Arendt and Wittbrodt 2001; Salvini-Plawin and Mayr 1961). For that reason, eyes have long been used as exemplars of parallel evolution whose distribution is in the phylogeny of animals; that is, eyes are examples of analogy rather than homology because similar eyes appeared intermingled among taxa that do not appear to have shared one type of eye in their common ancestor. The evolutionary complexity of eyes perplexed Darwin (Origin of Species), who confessed that he could hardly imagine something as structurally complex as an eye evolving and reevolving independently so many times. A simpler explanation would be that animal eyes were homologous rather than analogous, but to rescue such a notion required a common ancestral eye that could have evolved into the diverse descendant forms seen today. Darwin hypothesized a primitive animal eye consisting of two cell types, a photoreceptor and a nonphotosensitive pigment cell with a substance that shades the light so that its direction could be detected, all perhaps encased in some kind of translucent covering. Primitive eyes resembling this structure exist in various taxa, but it was the remarkable discovery that genes fundamental to vision were widely shared among animals that made the idea credible.
Pax6 and Conserved homology Among Eyes
The transcription factor (TF) Pax6 was one of these genes. Expressed early in development, Pax6 appears to serve as a selector for the initiation of differentiation cascades involved in eye development in most metazoans tested. Evidence showing this has come from expression experiments, as well as from natural or artificial muta-tional studies in mouse, humans, flies, and other animal species (e.g., Gehring 2002; Gehring and Ikeo 1999; Punzo et al. 2001). Rhodopsins and intracellular lightsensitive mechanisms exist in bacteria, suggesting that light sensitivity evolved before multicellular life. Furthermore, Pax6 is expressed in several other later-stage eye structures and even regulates the expression of at least some opsin genes, like insect rhodopsin, expressed in mature photoreceptors.
The phylogenetically deep sharing of Pax6 and rhodopsins can be seen as devel-opmentally bracketing the morphological and sensory aspects of light reception among animals. This might seem to provide Darwin with the elements of his suggested common Precambrian origin for animal eyes (Figure 14-8), (Gehring 2002; Gehring and Ikeo 1999). But the sharing of a couple of genes does not provide obvious ease for his concern as to how similar types of complex eye morphologies have evolved several times independently or how similar eyes can evolve through different developmental pathways even in related species and from very different tissue contexts (e.g., Hall 1999). It is these morphologies that account for how light is used in various taxa. One explanation is that developmental patterning mechanisms were recruited, or intercalated, between the initiation of visual systems by Pax6 and the later induction of opsin gene expression (e.g., Gehring 2002; Gehring and Ikeo 1999).
Different types of eyes involve different morphogenic processes, such as placode formation, invagination (e.g., of an optic cup), and periodic patterning (e.g., omma-tidia). As reviewed in Chapter 9, these are standard parts of the developmental repertoire of complex organisms, which during evolution could have been invoked in new contexts in between the first expression of Pax6 and the later expression of photoreceptors. The various forms of eye could have evolved sometimes inserting shared but other times different morphogenic mechanisms. Because the developmental mechanisms are resident parts of the animal developmental toolkit, such evolution could be relatively rapid (Pichaud et al. 2001).
Relevant to this is that eyelike structures can be induced experimentally in insects and vertebrates by ectopic application of Pax6 protein (Gehring and Ikeo 1999), in locations where relevant signaling molecules are already expressed (e.g., Kumar and Moses 2001; Pichaud, Treisman et al. 2001), but in contexts that normally do not develop into eyes (e.g., other imaginal disks in insects). Pax6 from one species can even induce such development in another.
These results suggest that different eyes share genetic homologies and that they did not entirely evolve independently as analogous organs. This is interesting because the intercalated mechanisms are also used in other structures, so that eyes would in part be homologous to antennae, feathers, and teeth (which also use Hedgehog, Bmps, and the like). The basic elements of the logic of the rapid evolution of diverse eyes are also found in other systems, including olfaction and various aspects of developmental polarity, repeated structures, and the like that we have seen are so characteristic of evolution.
Crystallins and the Nonconserved "Homology"
Another interesting aspect of eye evolution concerns the crystallins, major proteins found in vertebrate lens and corneal tissue.An important characteristic of crystallins is their light transparency properties when desiccated and compacted. However, genes coding crystallins did not evolve to serve a visual purpose. Instead, different species have opportunistically recruited different, often wholly unrelated proteins to use as their lens or corneal proteins (Tomarev and Piatigorsky 1996). These proteins typically also serve other functions in the same organism, such as heat shock response or housekeeping enzymes. Lens and cornea develop embryologically from the same tissue, but in a given species the corneal crystallins are different from those used in the lens. Many crystallins include Pax6 REs in their 5' flanking regulatory regions, and experiments show that the enhancers drive appropriate eye-specific expression, often across species (e.g., the enhancer from a chick can induce lens expression in a mouse). Sox2 and retinoic acid receptor pathways may be alternative or additional parts of the shared regulatory mechanism (see below). However, unlike the conservation of the Pax6 regulatory pathway itself, in the case of crys-tallins, the function rather than the specific gene is conserved. Phenogenetic drift has replaced one gene with another in different lineages that have shared the form of their eyes since a common ancestor.
Was this article helpful?