INTRODUCTION
Discussing the evolution of melatonin function is a particularly difficult task because of the lack of data on the subject. In truth, very few reported studies have dealt with the evolutionary aspects of melatonin function in the Vertebrata. Although a careful accounting of these actions has not been made, it is probable that only prolactin has more ascribed functions (see Bern and Nicoll, 1968). Because of this, it would be difficult indeed to discuss the evoultion of more than a few of these functions. Therefore, this paper will examine known information about melatonin and the evolution of the cells which synthesize melatonin and present a hypothesis concerning the evolution of only one aspect of melatonin function, its role in timing various biological rhythms.
In many of the discussions concerning melatonin function, it was either demonstrated or implied that melatonin may be serving to either entrain or reinforce rhythmical physiological or behavioral phenomena (see Binkley, 1980). By examining the large amount of literature on melatonin and reproductive cyclicity, it becomes apparent that melatonin is somehow involved in the temporal arrangement of various reproductive events (see Reiter, 1978). Melatonin has been shown to be involved in rhythmical thermoregulatory events in some ectotherms (see Ralph et al., 1979) and in the nightly pallor response in amphibian larvae (Bagnara and Hadley, 1973) and some fish. Finally melatonin has been shown to affect the locomotor activity rhythms in sparrows (Turek et al. 1976).
Much of the research on melatonin and biological rhythms stems from the fact that melatonin synthesis is an extremely regular event in nearly all vertebrates. Many reports have now demonstrated that melatonin content of the pineal gland (see Ralph, 1976) and in circulation increases several fold at night. It is also known that the pineal gland is the chief secretory source of melatonin. The cell responsible for melatonin synthesis and secretion within the mammalian pineal gland is the pinealocyte. Work primarily from the laboratory of Collin (see Collin, 1971) has established that the pinealocyte was derived through the evolutionary process of rudimentation from secretory photoreceptor cells in the line of mammalian progenitors. Therefore in lower vertebrates (cyclostomes, fish, amphibians and some reptiles), pineal photoreceptor cells are both sensory and secretory. The transition from a sensory-secretory photoreceptor cell to a secretory rudimentary photoreceptor cell is quite evident in some reptiles and birds. Finally, only the secretory pinealocyte is present in adult mammals. Although it has not been unequivocally demonstrated, it is thought that the photoreceptor cells in the pineal organs of fish (Gern et al., 1978b) and amphibians (Gern and Norris, 1979) and the rudimentary secretory photoreceptor cells of reptiles (Firth et al., 1979) and birds (Ralph, 1976) synthesize melatonin. Retinal photoreceptor cells probably also synthesize melatonin (Bubenik et al., 1974) which indicates that melatonin synthesis is a biosynthetic pathway shared by the ciliate photoreceptor cells in general. Some support for this may be found in the very similar embryonic origins of the pineal organ and associated structures and the retina and pigmented-epithelium of the lateral eye (Eakin, 1973).
What function melatonin may have in photoreceptive tissue is not clearly understood. It is assumed that melatonin synthesis in general is a nightly event in ciliate photoreceptor cells or their derivatives. Furthermore, this nocturnal melatonin synthesis is assumed to be an evolutionarily conserved feature of these cells. Melatonin is detoxified or converted to other molecules within the liver, where it is hydroxylated at the 6 position. Most (70%) of the 6-hydroxymelatonin is then conjugated to sulfate while a much smaller fraction is conjugated with glucuronic acid (Kopin, et al. 1961). Both conjugates are released from the body in urine or feces. Another pathway for melatonin degradation is by deacetylation by the hepatic aryl acylamidase system (Rogawski et al., 1979), although only a small percentage of the circulating melatonin is converted to 5-methoxytrypamine and eventually to 5-methoxyindoleacetic acid by this system. As noted earlier, both systems are primarily present in the liver, with no reports as to their presence in photoreceptive tissue available. Neither have been found in nervous tissue (Kopin et al., 1961; Rogawski, et al., 1979).
With this background the following hypothesis is advanced as well as possible avenues to test the hypothesis. It is hypothesized that melatonin is a synthetic product of ciliate photoreceptor cells found in vertebrates. This melatonin has an action either within the photoreceptor cell or on cells closely associated with the photoreceptor cells such as pigmented-epithelium. Such actions may be to cause the photomechanical displacement of the photoreceptor outer segments or aggregation of melanosomes within the pigmented-epithelium. Both actions are consistent with the temporal occurrence of melatonin during any 24-h day, where melatonin would act to unmask the sensitive photoreceptor outer segment for greater light gathering activity at night. Whatever the action of melatonin is in photoreceptive tissue, it is assumed that it occurs primarily at night and that this action is the most primitive. It is probable that the enzymes necessary for conversion or detoxification of melatonin are not present at the site of melatonin synthesis and that this too is reflective of the primitive condition. Therefore, detoxification or conversion of this indoleamide must be done by the liver, requiring that melatonin be placed into circulation. In the most primitive vertebrates, melatonin synthesized by various photoreceptor cells was placed into circulation for detoxification, creating a pulse of melatonin in the plasma which could be used to determine either the onset or duration of the scotophase, an event which occurs in most modern vertebrates. Because of the regularity of this pulse, melatonin could be used to cue other rhythmical phenomena if the proper cellular receptor for the molcule was present. Furthermore, if melatonin was acting in some way on photoreceptive tissue, in primitive vertebrates, it implies that a cellular receptor for the molecule existed. Therefore, the genetic information necessary for melatonin receptor formation was present but repressed in most cells. It is hypothesized that either through mutation or through the derepression of that portion of the genome which contains the genetic information coding for the production of melatonin receptors were formed in cells at a new site. Animals possessing this adaptation functioned better in their environment because of their ability to use the melatonin pulse in a new way, as a biologica...