Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Frontiers in Neuroendocrinology 20, 224–240 (1999) Article ID frne.1999.0181, available online at http://www.idealibrary.com on Gonadotropin-Releasing Hormone Genes: Phylogeny, Structure, and Functions Russell D. Fernald and Richard B. White Program in Neuroscience, Stanford University, Stanford, California 94305-2130 Gonadotropin-releasing hormone (GnRH, previously called leutinizing hormonereleasing hormone, LHRH) is the final common signaling molecule used by the brain to regulate reproduction in all vertebrates. Recently, genes encoding two other GnRH forms have been discovered. Here we present a phylogenetic analysis that shows that the GnRH genes fall naturally into three distinct branches, each of which shares not only a molecular signature but also characteristic expression sites in the brain. The GnRH genes appear to have arisen through gene duplication from a single ancestral GnRH whose origin predates vertebrates. Several lines of data support this suggestion, including the fact that all three genes share an identical exonic structure. The existence of three distinct GnRH families suggests a new, natural nomenclature for the genes, and in addition, we present a logical proposal for naming the peptide sequences. The two recently discovered GnRH genes are unusual because they encode decapeptides that are identical in all the species in which they have been found. The control of gene expression also differs among the three gene families as might be expected since they have had separate evolutionary trajectories for perhaps 500 million years. r 1999 Academic Press INTRODUCTION Gonadotropin-releasing hormone (GnRH, previously called luteinizinghormone releasing hormone or LHRH) is known and named for its role as the final common signaling molecule used by the brain to regulate reproduction in all vertebrates. The GnRH decapeptide is synthesized by neurosecretory cells in the hypothalamus and secreted into portal vessels, to be transported to the pituitary gland where it stimulates secretion of luteinizing hormone (LH) and follicle-stimulating hormone (FSH) from pituitary gonadotrophs. Because of its central importance in controlling reproduction in all vertebrates, the action of this GnRH form, called the ‘‘releasing’’ form, is well characterized. Recently, genes encoding decapeptides slightly different from this GnRH form have been found in many vertebrates (see 32 for review), including humans (68). Moreover, these additional genes and the original releasing form of GnRH are expressed in the body as well as in the brain (e.g., 68). Here we will review the history of GnRH discoveries and assess what is known about the genes encoding different GnRH forms. In addition, we will show the phylogenetic relationships, structural similarities, ontogeny, sites of expression, and 0091-3022/99 $30.00 Copyright r 1999 by Academic Press All rights of reproduction in any form reserved. 224 frne 0181 GnRH GENES: PHYLOGENY, STRUCTURE, AND CONTROL 225 regulation of the GnRH genes. Finally, we will discuss the new insights these discoveries provide about GnRH. DISCOVERY AND NOMENCLATURE Discovery The releasing form of GnRH was predicted to exist in 1950 (24) but since the hypothalamus contains minuscule amounts of GnRH, and the peptide has modified N- and C-termini, it was difficult to isolate. Consequently, its structure was not determined until 20 years later (10, 41). Twelve years after this proof of its existence, the gene encoding GnRH and its cDNA were described for human (53) and rat (1). Interestingly, cloning this gene enabled one of the first demonstrations of the prospect for gene therapy when a GnRH transgene injected into hypogonadal mice (hpg; 11) embryos was shown to rescue the phenotype (54). Subsequently, it was discovered that both strands of the DNA within the GnRH gene are transcribed (2, 8). In 1991, the first cDNA encoding GnRH in a nonmammalian species was identified in a teleost fish, Haplochromis burtoni, by Bond et al. (9), and shortly thereafter the cDNA encoding a second form of GnRH was found in the same species (66). Discovery of multiple GnRH genes was not completely surprising because indirect evidence had been accumulating since 1987 about multiple GnRH peptide forms in a single species. These suggestions came from studies in which GnRH forms were separated using high-pressure liquid chromatography (HPLC) and were distinguished using peptide antisera (reviewed in 34; 55). However, this evidence was not always easy to interpret, as evidenced by attempts to characterize the structure of guinea pig GnRH (cf. 33 vs 29). As noted above, a second cDNA encoding GnRH was identified in a cichild fish in 1994 (66) followed by the discovery of a third GnRH cDNA in the same species a year later (67). Using PCR, several groups were subsequently able to identify more than one cDNA or gene encoding GnRH peptides in several fish species (4, 7, 14, 20, 37, 51). Since the original discovery in fish, a cDNA encoding a second GnRH has been found in several placental mammals (tree shrew, 32; humans, 68; rhesus monkey, 63). The only information to date regarding the spatial distribution in the genome of different GnRH genes is in humans, and at least in this one species the genes reside on different chromosomes (68). Nomenclature An unfortunate outcome of the relatively rapid discovery of new GnRH peptide sequences and additional GnRH genes is a cumbersome terminology to identify them. Each new GnRH peptide form was named according to the first species in which that particular peptide sequence was found. Though initially convenient and now familiar to the cognescenti, this nomenclature provides frne 0181 226 FERNALD AND WHITE neither useful structural nor functional information about the growing number of known peptides. Recent phylogenetic analysis has provided a more rational terminology for identifying both GnRH gene and peptide forms. As more GnRH encoding genes are cloned and their associated peptide sequences known, there needs to be a common, universal, informative nomenclature for GnRH genes as well as for GnRH peptides. Because two GnRH genes have now been found in humans, nomenclature recommended by the Genome Database Nomenclature Committee of March 5, 1997 has been adopted perforce. Accordingly, the human gene encoding the releasing form of GnRH (pGlu-His-Trp-Ser-Tyr-Gly-Leu-Arg-Pro-Gly) is now termed GNRH1 (formerly GnRH or LHRH). The second form of GnRH (pGlu-His-Trp-Ser-His-Gly-Trp-TyrPro-Gly) is known as GNRH2. Correspondingly, GnRH genes in nonhumans are known as Gnrh1 for the releasing form, Gnrh2 for the second form (previously called ‘‘chicken II GnRH’’), and Gnrh3 for a third form identified in several teleost fish but not (yet?) found in humans (previously called salmon GnRH or sometimes sGnRH). The identities of the different GnRH peptides encoded by the GnRH genes are best described by listing the amino acid differences from the human hypothalamic-releasing form. Thus, the GNRH2 and Gnrh2 genes encode a peptide identified as 5His5Trp7Tyr8 6GnRH (meaning it has the three substitutions in the three positions noted) in all species in which it has been found. Correspondingly, GnRH3 genes encode 5Trp7Leu8 6GnRH in all species found to have a third GnRH form. As described in more detail below, the peptides encoded by the Gnrh1 genes are expressed primarily in the hypothalamus and comprise the originally described releasing hormone. Interestingly, the peptides encoded by the Gnrh1 genes are the only ones to date that have been shown to vary among species. Five of the ten amino acids encoded by Gnrh1 genes are invariant while two other positions show conservative changes (reviewed in 32). PHYLOGENETIC RELATIONSHIPS AND THE ORIGIN OF GnRH The discovery and sequencing of many genes that encode different GnRH peptides among vertebrate species have made it possible to examine their relatedness using molecular phylogenetic techniques. These methods use amino acids or nucleotides as the traits or characters whose comparison form the basis for constructing phylogenetic relationships. When constructing such a tree, more cases in which gene sequences are known for a single species will increase the accuracy of the phylogenetically based predictions. Based on molecular phylogenetic analysis (Fig. 1), the three known GnRH genes appear to have arisen from a common ancestral form through gene duplications prior to the appearance of vertebrates (68). Several methods (maximum parsimony, maximum likelihood, and neighbor-joining) were used to test the robustness of this analysis, and each method showed essentially the same branching patterns. The phylogenetic tree shows the existence of three frne 0181 GnRH GENES: PHYLOGENY, STRUCTURE, AND CONTROL 227 evolutionarily distinct GnRH groups that are expressed in three distinct regions of the brain: releasing forms localized to the hypothalamus (GnRH1), forms previously localized solely to midbrain nuclei (GnRH2), and forms localized to the telencephalon, to date found only in teleost fish (GnRH3). Several lines of evidence support the conclusion that the three GnRH groups identified by this phylogenetic comparison comprise evolutionarily distinct forms. First, the groupings correspond to GnRH forms with distinct expression patterns in the brain and thus presumed common biological roles. As described by Kasten et al. (32), in all species studied the GnRH1 forms appear in the hypothalamus, the GnRH2 forms appear in the midbrain, and the GnRH3 forms are found in the terminal nerve and olfactory system of the forebrain telencephalon. This suggests that the phylogenetic groupings can be used to predict the activities of yet uncharacterized forms once their molecular structure is determined. Second, although only the neighbor-joining tree is shown, the same groupings were found using other distance-based methods as well as maximum parsimony and maximum likelihood (not shown). Third, highreliability (e.g., bootstrap) values were generated for the neighbor-joining and parsimony trees, strongly supporting identification of these groups. Finally, the branch lengths to each group are relatively long, suggesting an ancient divergence. Furthermore, the branching order within each GnRH gene group matches the evolutionary branching order of the species (see in particular GnRH1, for which sequences were available from diverse species). Although these phylogenetic relationships cannot identify the selective forces that generated multiple GnRH forms, the analysis does reveal that different forms of GnRH very likely diverged from one another prior to the divergence of species represented in the tree. Because the tree is unrooted, the exact order of the duplications is unclear. However, GnRH1 and GnRH2 include representatives from fish and mammals, so the separation of these two groups must have occurred prior to the separation of mammalian and fish ancestors. The origin of GnRH3 is less clear, since, at this time, it has been described only in fish. The data suggest two alternative hypotheses for the appearance of GnRH3. The gene duplication leading to GnRH3 may have occurred relatively recently, within the fish lineage, and thus is restricted to this taxon. Alternatively, GnRH3 may have resulted from an ancient duplication and either has been lost in higher vertebrates during evolution or has not yet been discovered. We favor the latter hypothesis because if the duplication were recent, the GnRH3 forms would be expected to cluster with one of the other GnRH forms in fish, which does not occur (68). Perhaps a version of GnRH3 remains to be found in other vertebrates including humans. Supporting this latter notion, several lines of evidence indicate that duplication of most of the genome occurred during the transition from simple chordates to vertebrates (27, 40, 48). Most likely, the multiple GnRH forms are yet another consequence of these ancient duplications. Recent immunocytochemical work suggests that a chordate ancestor, an ascidian, has at least two distinct GnRH forms (62). In retrospect, we might have predicted that there would be several descendants of an ancient GnRH in extant vertebrates. frne 0181 228 FERNALD AND WHITE frne 0181 GnRH GENES: PHYLOGENY, STRUCTURE, AND CONTROL 229 In summary, genes encoding GnRH forms fall into three distinct evolutionary branches. Phylogenetic reconstruction supported by brain distribution of GnRH gene forms suggests that they arose through gene duplications prior to the rise of vertebrates and have been adapted to distinct functions over 500 myr of vertebrate evolution. GENE STRUCTURE All GnRH genes share the same basic structure. Within each gene, a given exon encodes a corresponding preprohormone region as first described for humans (53). In all cases, the GnRH preprohormone mRNA encodes GnRH and the GnRH-associated peptide (GAP), separated by a canonical cleavage site. The preprohormone mRNAs are encoded by four exons (Fig. 2). Exon 1 encodes the 58-UTR exclusively. Exon 2 encodes the signal peptide, GnRH decapeptide, the proteolytic cleavage site, and the N-terminus of GAP. Exon 3 encodes the central portion of GAP and exon 4 encodes the C terminus of GAP along with the 38-UTR. Among all known GnRH genes, there seem to be two generalities: (i) exons 2 and 3 are the most similar in length and (ii) the UTR sizes and intron sizes are the most different. The level of similarity in the coding sequences can be read from the phylogenetic tree distances (Fig. 1). The greatest differences within the preprohormone are within the GAP coding sequences. The striking contrast between conservation of the GnRH coding sequence and lack thereof in the GAP coding sequence is evidence of differential selective pressure within the gene. This is evident in cases where the identity and similarity of GnRH and GAP coding sequences have been compared for mRNAs of different GnRH genes within a species (e.g., 32, 67, 69). Since animals as phylogenetically distant as fish and humans have essentially identical GnRH gene structures, we expect that GnRH genes yet to be described will also fit this pattern. The molecular divergence of the 58- and 38-flanking sequences in the different GnRH gene forms, as well as in their introns, suggests that each of the three branches has had a long period of independent evolution, consistent with our recent phylogenetic analysis of GnRH evolution based on cDNA sequences (68). FIG. 1. Unrooted neighbor-joining phylogenetic tree of cDNAs encoding GnRH. The tree was generated using the neighbor-joining algorithm and the Hall distance calculation correction procedure. Evolutionary distances are represented only by the length of the branches and not by branch angles. Bootstrap values, indicating the number of times a particular set of sequences groups together when trees are generated from resampled alignments, are indicated for some important nodes on the tree. An unrooted tree is shown because of the lack of an obvious outgroup. Some GnRH3 sequences from different species were identical and are grouped together. The scale bar corresponds to estimated evolutionary distance units as calculated by Protdist in Phylip. Redrawn from (68). frne 0181 230 FERNALD AND WHITE FIG. 2. Schematic diagram showing organization of GnRH genes. The genes encoding GnRH in the cichlid fish, H. burtoni, are shown. Within each gene, a given exon encodes a corresponding preprohormone region. Single hatched bars, signal sequence coding region; cross-hatched bars, GnRH-coding region; open bars, GnRH-associated peptide (GAP) coding region. Horizontal lines adjacent to exons represent introns. Modified from (69). ORIGIN OF GnRH EXPRESSING CELLS AND THEIR SITES OF EXPRESSION The three GnRH genes described above are expressed in cells that have distinctly different origins, locations, and, presumably, functions. Information about these topics is best understood for GnRH expression in the brain. In the adult brain, GnRH1 is expressed in the hypothalamus, GnRH2 in the midbrain– tegmentum, and GnRH3 in the forebrain telencephalon. The hypothalamic and telencephalic gene forms (Gnrh1 and Gnrh3) are expressed by cells that arise in the olfactory placode and migrate into the brain (44, 46, 52, 72). These cells migrate centrally and take up positions along the ventral surface of the adult brain in a rostrocaudal continuum from the terminal nerve area to the hypothalamus. The final distribution of these neurons in fish species divides into two populations. The anterior population of cells expresses the Gnrh3 gene and the posterior population expresses the Gnrh1 gene (70). Although the original observations of migration were performed in mouse tissue, using immunocytochemistry (e.g., 52), they have been confirmed using in situ hybridization with molecular probes (70). The placodal ancestry of these neurons along with their expression in the terminal nerve suggests that Gnrh3 in the telencephalon might coordinate sensory input at the time of reproduction (cf. 67), consistent with the report of Yamamoto et al. (73). GnRH2 expressing neurons of the brain do not arise from the olfactory frne 0181 GnRH GENES: PHYLOGENY, STRUCTURE, AND CONTROL 231 placode. This was first shown indirectly by lesioning the placode prior to development in the salamander (47). Animals with the placode lesioned had neurons with GnRH immunoreactivity in the midbrain–tegmentum but not in the telencephalon or hypothalamus. More recently, using gene probes, White and Fernald (70) demonstrated directly that Gnrh2 is expressed in cells born in the midbrain ventricle in a teleost fish. The expression of this GnRH gene form begins shortly after the cell is born, while it is still within the ventricular germinal zone. All three GnRH genes in this species are expressed by the end of the first third of development (day 4 of 14). This early onset suggests that GnRH might play additional roles during development. At present, very little is known about the ontogeny or type of cells that express any of the three GnRH forms outside the brain (e.g., 5, 6, 64). Expression of GnRH Outside the Brain GnRH1 gene expression has been well studied in the brain of vertebrates, but only recently has investigation begun of its expression elsewhere in the body. For example, GnRH1 mRNA has been reported in splenocytes and peripheral lymphocytes in rats (5, 71) and humans (6, 64). By using reverse transcription– polymerase chain reaction (RT-PCR) on human tissue, GnRH1 mRNA has been found in liver, heart, skeletal muscle, kidney, and placenta (e.g., 31). GnRH1 gene expression has been also been reported in the testis of several primates and rats (15), human ovary (50), and rat prostate (6). Widespread GnRH1 expression was also found in the teleost fish, H. burtoni, by using RT-PCR. Heart, liver, spleen, kidney, testis, and brain of an adult male all contained GnRH1 mRNA (69). Much less is known about the distribution of GnRH2 expression outside the brain. In humans, White et al. (68) reported widespread distribution of GNRH2 mRNA in the prostate, bone marrow, and kidney, in addition to its expression in the brain. In goldfish, Lin and Peter (37) demonstrated expression of GnRH2 mRNA in the ovary, and Yu et al. (74) used RT-PCR to detect it in both testis and ovary. In the cichlid H. burtoni, White and Fernald (69) showed that GnRH2 mRNA is expressed in the testis and brain but not in kidney. Similarly, there are only a few reports about expression of GnRH3. In the teleost Porichthys notatus GnRH3 mRNA was found in the testis and ovary (22). In the cichlid fish H. burtoni, GnRH3 expression was identified in the testis. All three forms of GnRH have been detected in the testis of H. burtoni, which suggests that the different GnRH peptides may play distinct roles in both the brain and the testis. Understanding of the role of multiple GnRH forms within a single organ will ultimately depend on detailed knowledge of their specific sites of action, including the distribution of GnRH receptors and the possible existence of multiple GnRH-receptor types. frne 0181 232 FERNALD AND WHITE CONTROL OF GnRH GENE EXPRESSION Regulation of the GnRH genes is a topic of intense interest, particularly because it has become possible to compare control of gene expression of several GnRH genes within a single organism. Such analyses should provide information and insight about how GnRH is regulated at its many sites of action. Previously most studies of gene regulation were directed at GnRH1 both because of its importance and because no other genes were known. Gore and Roberts (19) recently provided an excellent review of hypothalamic GnRH1 gene regulation in mammals. In contrast, little is known about how the other two GnRH gene forms are regulated, and comparative analysis of gene regulation among the three forms of GnRH is in an early stage. Thus the broad and interesting issues such as differences between in vitro and in vivo GnRH regulation or between transcriptional and posttranscriptional regulation mechanisms are not yet known except for the GnRH1 gene. We will focus here on what is known about regulation of the other GnRH genes. Similar Regulation in Different GnRH Genes? The discovery that different GnRH peptides are expressed in a variety of tissues suggests that, as a group, they serve multiple functions, probably requiring complex gene regulation. Analysis of the 58-flanking sequence in the three GnRH genes in H. burtoni suggests that this may be the case. Because each decapeptide is encoded by a different gene, each with a unique 58upstream regulatory sequence, transcriptional regulation of the three GnRH genes is likely to be independent. Even though some short regulatory motifs are conserved among the three H. burtoni genes, the proximate 500 bp of upstream sequence are quite dissimilar. Indeed, the promoter regions of the three genes are not significantly more similar to one another than are their introns (69). Promoter sequence comparisons of multiple GnRH genes in closely related species has only recently become possible with the sequencing of GnRH1 and GnRH2 genes in two teleosts, H. burtoni (69) and Morone saxatilis (14). Preliminary comparison of the upstream regions of the Gnrh2 genes in H. burtoni, an African cichlid, and the striped bass, M. saxatilis, an anadromous North American species, reveals several regions with remarkable conservation, in addition to the region near the most proximate promoter. For example, at a distance more than 2000 bp upstream of the mRNA coding sequence in each GnRH2 gene, there is a 74% nucleotide identity in a 750-nucleotide overlap. Apparently there has been a strong selective pressure to maintain this alternative promoter, thus suggesting that it may play an important role in regulation of Gnrh2 gene expression in these species and perhaps others. Multiple promoter usage has previously been suggested for the GnRH2 gene in humans (68) and M. saxatilis (14), and this may prove to be widespread within GnRH2 genes. frne 0181 GnRH GENES: PHYLOGENY, STRUCTURE, AND CONTROL 233 Comparison of Gnrh1 gene upstream sequences in H. burtoni and M. saxatilis shows no region of sequence conservation comparable to that in the GnRH2 genes, although within the proximate 500 bp of the promoter region, there is a 77% nucleotide identity in a 178-nucleotide overlap. It is tempting to speculate that these gross differences in sequence conservation between the two gene forms may mirror the relative conservation of the peptide forms encoded by the genes. Direct Regulation of GnRH Genes by Steroids? cDNAs encoding two or even three of the GnRH forms in a single species have now been characterized in two primates (human, 68; rhesus monkey, 63), one rodent (tree shrew, 32), and six fish species (African cichlid, 69; African catfish, 7; two species of seabream, 20; striped bass, 14; and goldfish, 38). However, genomic sequencing that includes the upstream regulatory structures of multiple GnRH gene forms in a single species has been described for only three species. Two complete GnRH genes have been described for humans (68). In the striped bass, M. saxatilis, two of the three known GnRH gene forms have been described (14). Complete genes encoding three distinct GnRH forms have been characterized in only one species, H. burtoni (69), so this review will focus on the GnRH genes in this species. Several putative steroid receptor binding sites are evident in the proximate 500 bp of 58-flanking sequence of each GnRH gene in H. burtoni (69). Specifically, putative glucocorticoid receptor (GR) binding sites (6 in GnRH1 and 10 in GnRH3) hint at a possible direct role for GR and its ligand cortisol in regulating these genes. In contrast, GnRH2 has only 2 putative GR sites in this region. Behavioral, endocrine, and morphological data in H. burtoni are consistent with GnRH gene regulation by cortisol in hypothalamic GnRH neurons (17). Though still unknown in fish, a subset of GnRH neurons in the rat hypothalamus contain GR (3). Furthermore, Chandran et al. (12) provided strong evidence for the direct role of GR in GnRH gene regulation in GT1-7 cells. In contrast to putative GR binding sites, estrogen receptor binding sites were not found in the 500 bp upstream of any of the three GnRH genes, suggesting that estradiol may not have a direct effect on GnRH gene regulation in H. burtoni. Consistent with this, forebrain GnRH neurons in the rainbow trout (Oncorhynchus mykiss) do not contain immunoreactive estrogen receptor (ER) (45). In another salmonid, the Atlantic salmon, ER binding sites were found 1.5 kb upstream of the 5Trp7Leu8 6GnRH gene (GnRH3—see 68), and human ER protein was shown to bind to this sequence (36). However, ER itself has not been localized to GnRH neurons in that species, so a direct role for estradiol in the transcriptional regulation of GnRH remains uncertain. Moreover, there is scant evidence for the presence of ER within GnRH neurons in other vertebrates for which GnRH and ER colocalization has been attempted (e.g., negative results in rat, 25, 56; rhesus macaque, 60; mink, 65; sheep, 26). The frne 0181 234 FERNALD AND WHITE well-described role of estrogen in the regulation of GnRH is thus likely to be indirect, mediated by other cells. A single putative androgen receptor binding site is upstream of the GnRH1 gene in H. burtoni, consistent with the idea that androgen acts directly on GnRH1 neurons in this species. Gonadal steroid feedback experiments are also consistent with this hypothesis (59), although these effects could also be indirect. The possible role of the two progesterone binding sites upstream of GnRH1 and GnRH3 is unclear, although direct transcriptional regulation of the GnRH gene by progesterone receptor (PR) has been proposed in other animals, and Cho et al. (13) have shown that translation effects in vivo appear to be stimulated by PR. However, in the tilipia Oreochromis niloticus, preliminary studies indicate that progesterone does not influence any of the three populations of GnRH neurons (49). By using semiquantitative in situ hybridization, they found that progesterone treatment had no effect on mRNA levels for any of the three GnRH forms. However, castration significantly elevated GnRH3 mRNA while GnRH1 and GnRH2 mRNA levels remained unchanged, suggesting that in this species, the terminal nerve GnRH neurons are under negative gonadal steroid feedback, while the other two GnRH neuron populations are not. In contrast, in H. burtoni, a closely related cichlid fish, cells expressing GnRH1 are clearly under gonadal steroid regulation: these cells hypertrophy after castration (59). In other teleosts, hypothalamic GnRH neurons are also affected by gonadal steroids (e.g., 21, 23). GnRH FUNCTIONS Among the GnRH peptides, the function of GnRH1 is best understood because the peptide encoded by this gene is essential for reproduction. Consequently, many features of its regulation are now known (see above and 19). Recently, the hpg mouse has been used to tease apart the role of GnRH1 in sexual maturation from its role in reproductive behavior (18). Both the GnRH2 and the GnRH3 peptides are identical in all species in which they have been found. Since GnRH2 has been identified in many more species to date, its complete conservation through 500 myr of evolution suggests that it has been subject to extremely stringent selective pressures. Such selection might have arisen if GnRH2 serves different functions at different loci in the body, which is suggested by a growing body of evidence. First, GnRH2 has been shown to be present in the sympathetic ganglia of amphibia (28), where it has been shown that GnRH can act as a neuromodulator in spinal cord ganglion neurons (30). In fish, neurons previously shown to control sperm duct and oviduct contractility receive input from cells that show GnRH2 immunoreactivity (42). Examples of possible GnRH2 expression outside the midbrain cells include transient GnRH2 immunoreactivity within mast cells in the habenula of ring doves following courtship (57, 58). These data have led to the suggestion that mast cells might be an alternative delivery system for GnRH (57). Has GnRH2 remained unchanged because of selective pressure from mul- frne 0181 GnRH GENES: PHYLOGENY, STRUCTURE, AND CONTROL 235 tiple functions? We have speculated elsewhere (68) that GnRH2 may have originated in the immune system and only later acquired a neuromodulatory function in the brain. Further characterization of GnRH2 including decisive functional tests should help clarify whether it serves distinct functions in the body. This in turn might shed light on what selective forces have been responsible for its remarkable conservation over evolutionary time. The function of GnRH3 also remains unclear. Lesions of the terminal nerve Gnrh3 expressing neurons in a fish (Colisa lalia, dwarf gourami; 73) resulted in a threshold change in courtship initiation. However, virtually no other data are available about possible functions of this peptide. Interestingly, it has been known for some time that a subpopulation of axons from the terminal nerve contain GnRH3 (67) and project to the retina (43), although no function has yet been ascribed to GnRH3 in the retina. We have previously noted (67) that each of the sensory, motor, and humoral systems in teleost fish might utilize distinct but related forms of GnRH peptide: the terminal nerve form perhaps coordinates the sensory and motivational systems, the mesencephalic form may modify motor action in the service of reproductive behavior, and the hypothalamic form causes the release of gonadotropins to facilitate reproduction (67). The report that lesioning the terminal nerve affected reproductive readiness (73) suggests that GnRH3 expression might indeed influence reproductive motivation. The distribution of GnRH peptide types hints at possible functions. However, the distribution of GnRH receptors, which mediate any effect of GnRH, is essential for understanding what GnRH forms actually do. However, it is not yet known whether different forms of GnRH might have different receptor types. GnRH receptors have been found throughout the body, including the kidney (31) and prostate (16). Because previous studies have shown that GnRH receptors bind GnRH2 peptide up to 100 times more effectively than GnRH1 (35), GnRH2 could act through these receptors outside the brain. Recently, three putative GnRH receptor subtypes were identified, suggesting that there may be a specific cognate receptor type for each GnRH type (61). These authors suggest that GnRH receptor types coevolved with the GnRH gene forms, resulting in a cognate receptor for each GnRH type. SUMMARY AND FUTURE PREDICTIONS GnRH was once thought to be a single peptide, acting solely as a brain signal to the pituitary. However, new discoveries about GnRH have changed our views of GnRH. Clearly the original gonadotropin-releasing hormone was just the beginning of the GnRH story. As shown above, there are three GnRH gene families based on molecular phylogenetic analysis. Each family shares not only a molecular signature but also a characteristic expression site in the brain. Surprisingly, the two newly discovered gene families encode decapeptides that are identical in all species in which they have been found. This is in contrast to GnRH1, which has substantial variation among species. Although a great deal frne 0181 236 FERNALD AND WHITE is known about the hypothalamic GnRH form, exactly where and how the two recently discovered GnRH forms act remain unknown. As might be suspected, past predictions about GnRH have had mixed success. Most notably, GnRH2 was hypothesized to have been lost in placental mammals (55) and to be the oldest GnRH (39), neither of which now appears to be true. Still, it seems likely that there are some predictions that could be made based on the data presented here. First, it seems probable that a third form of GnRH as described in teleost fish will be found in other vertebrate classes. Since this form is expressed in the telencephalon and the neurons that express it share an ontogenetic origin with the releasing form, it may play a role in reproductive readiness in other species as it appears to do in teleost fish. Second, the two gene duplications that are proposed to have led to vertebrates probably produced the multiple GnRH forms. If so, it seems likely that there will be multiple GnRH receptor types that are also a product of these duplications. Finally, the GnRH genes are closely related to one another but are evidently not a part of a larger gene family, possibly making them unique among peptide encoding genes. Does this apparent isolation have any biological meaning? Perhaps the GnRH genes encode peptides that are unique, making more urgent the discovery of the roles of GnRH2 and GnRH3. ACKNOWLEDGMENTS We thank J. Eisen for collaboration on the phylogeny, A. Greenwood and K. Hoke for helpful comments on the manuscript, and E. Bennett for invaluable help with the manuscript. This research was supported by NIH NS 34950 to R.D.F. REFERENCES 1. Adelman JP, Mason AJ, Hayflick JS, Seeburg PH. Isolation of the gene and hypothalamic cDNA for the common precursor of gonadotropin-releasing hormone and prolactin releaseinhibiting factor in human and rat. Proc Natl Acad Sci USA 1986; 83: 179–183. 2. Adelman JP, Bond CT, Douglass J, Herbert E. Two mammalian genes transcribed from opposite strands of the same DNA locus. Science 1987; 235: 1514–7. 3. Ahima RS, Harlan RE. Glucocorticoid receptors in LHRH neurons. Neuroendocrinology 1992; 56: 845–850. 4. Ashihara M, Suzuki M, Kubokawa K, Yoshiura Y, Kobayashi M, Urano A, Aida K. Two differing precursor genes for the salmon-type gonadotropin-releasing hormone exist in salmonids. J Mol Endocrinol 1995; 15: 1–9. 5. Azad N, Emanuele NV, Halloran MM, Tentler J, Kelly MR. Presence of LHRH messenger RNA in rat spleen lymphocytes. Endocrinology 1991; 128: 1679–1681. 6. Azad N, Lapaglia N, Abel K, Jurgens KAJ, Kirsteins L, Emanuele NV, Kelley MR, Lawrence AM, Mohagheghpour N. Immunoactivation enhances the concentration of luteinizing hormonereleasing hormone peptide and its gene expression in human peripheral T-lymphocytes. Endocrinology 1993; 133: 215–223. 7. Bogerd J, Zandbergen T, Andersson E, Goos H. Isolation, characterization and expression of cDNAs encoding the catfish-type and chicken-II type gonadotropin-releasing hormone precursors in the African catfish. Eur J Biochem 1994; 222: 541–549. frne 0181 GnRH GENES: PHYLOGENY, STRUCTURE, AND CONTROL 237 8. Bond CT, Hayflick JS, Seeburg PH, Adelman JP. The rat gonadotropin-releasing hormone SH locus: Structure and hypothalamic expression. Mol Endocrinol 1989; 3: 1257–1262. 9. Bond C, Francis R, Fernald R, Adelman J. Characterization of complementary DNA encoding the precursor for gonadotropin-releasing hormone and its associated peptide from a teleost fish. Mol Endocrinol 1991; 5: 931–937. 10. Burgus R, Butcher M, Amoss M, Ling N, Monahan M, Rivier J, Fellows R, Blackwell R, Vale W, Guillemin R. Primary structure of ovine hypothalamic lutenizing hormone-releasing factor (LRF). Proc Natl Acad Sci USA 1972; 69: 278–282. 11. Cattanach BM, Iddon CA, Charlton HM, Chiappa SA, Fink G. Gonadotropin-releasing hormone deficiency in a mutant mouse with hypogonadism. Nature 1977; 269: 338–340. 12. Chandran UR, Attardi B, Friedman R, Zheng ZW, Roberts JL, DeFranco DB. Glucocorticoid repression of the mouse gonadotropin-releasing hormone gene is mediated by promoter elements that are recognized by heteromeric complexes containing glucocorticoid receptor. J Biol Chem 1996; 271: 20412–20420. 13. Cho BN, Seong JY, Cho H, Kim K. Progesterone stimulates GnRH gene expression in the hypothalamus of ovariectomized, estrogen treated adult rats. Brain Res 1994; 652: 177–180. Chow MM, Kight KE, Gothilf Y, Alok D, Stubblefield J, Zohar Y. Multiple GnRHs present in a teleost species are encoded by separate genes: Analysis of the sbGnRH and GnRH-II genes from the striped bass, Morone saxatilis. J Mol Endocrinol 1998; 21: 277–289. Dong KW, Duval P, Zeng Z, Gordon K, Williams RF, Hodgen GD. Multiple transcription start sites for the GnRH gene in rhesus and cynomolgus monkeys: A non-human primate model for studying GnRH gene regulation. Mol Cell Endocrinol 1996; 117: 121–130. Fekete M, Redding TW, Comraru-Schally AM, Pontes JE, Connelly RW, Srkalovic G, Schally AV. Receptors for luteinizing-hormone-releasing hormone, somatostatin, prolactin, and epidermal growth-factor in rat and human-prostate cancers and in benign prostate hyperplasia. Prostate 1989; 14: 191–208. Fox HE, White SA, Kao MHF, Fernald RD. Stress and dominance in a social fish. J Neurosci 1997; 17: 6463–6469. Gibson MJ, Wu TJ, Miller GM, Silverman AJ. What nature’s knockout teaches us about GnRH activity: Hypogonadal mice and neuronal grafts. Horm Behav 1997; 31: 212–20. Gore AC, Roberts JL. Regulation of gonadotropin-releasing hormone gene expression in vivo and in vitro. Front Neuroendocrinol 1997; 18: 209–245. Gothilf Y, Munoz-Cueto JA, Sagrillo CA, Chen TT, Kah O, Elizur A, Zohar Y. Three forms of gonadotropin-releasing hormone in a perciform fish (Sparus aurata): Complementary deoxyribonucleic-acid characterization and brain localization. Biol Reprod 1996; 55: 636–645. Grober MS, Jackson IMD, Bass A. Gonadal steroids affect LHRH preoptic cell number in sex-role changing fish. J Neurobiol 1991; 22: 734–741. Grober MS, Myers TR, Marchaterre MA, Bass AH, Myers DA. Structure, localization, and molecular phylogeny of a GnRH cDNA from a paracanthopterygian fish, the plainfin midshipman (Porichthys notatus). Gen Comp Endocrinol 1995; 99: 85–99. Halpern-Sebold LR, Schreibman MP, Margolis-Nunno H. Differences between early-and-latematuring genotypes of the platyfish Xiphophorus maculatis in the morphometry of their immunoreactive LHRH containing cells: A developmental study. J Exp Zool 1986; 240: 245–258. Harris GW. Oestrous rhythm: Pseudopregnancy and the pituitary stalk in the rat. J Physiol 1950; 111: 347–360. Herbison AE, Theodosis DT. Localization of oestrogen receptors in preoptic area neurons containing neurotensin but not tyrosine hydroxylase, cholecystokinin or luteinizing hormonereleasing hormone in the male and female rat. Neuroscience 1992; 50: 283–298. Herbison AE, Robinson JE, Skinner DC. Distribution of estrogen receptor-immunoreactive cells in the preoptic area of the ewe: Co-localization with glutamic acid decarboxylase but not luteinizing hormone-releasing hormone. Neuroendocrinology 1993; 57: 751–759. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. frne 0181 238 FERNALD AND WHITE 27. Holland PWH, Garcia-Fernandez J, Williams NA, Sidow A. Gene duplications and the origins of vertebrate development. Dev Suppl 1994; 125–133. 28. Jan YN, Jan LY, Brownfield MS. Peptidergic transmitters in synaptic boutons of sympathetic ganglia. Nature 1980; 288: 380–382. 29. Jimenez-Liñan M, Rubin BS, King JC. Examination of guinea pig luteinizing hormonereleasing hormone gene reveals a unique decapeptide and existence of two transcripts in the brain. Endocrinology 1997; 138: 4123–30. 30. Jones SW, Adams PR, Brownstein MJ, Rivier JE. Teleost luteinizing hormone-releasing hormone: Action on bullfrog sympathetic ganglia is consistent with role as neurotransmitter. J Neurosci 1984; 4: 420–429. 31. Kakar SS, Jennes L. Expression of gonadotropin-releasing hormone and gonadotropinreleasing hormone receptor mRNAs in various non-reproductive human tissues. Cancer Lett 1995; 98: 57–62. 32. Kasten TL, White SA, Norton TT, Bond CT, Adelman JP, Fernald RD. Characterization of two new preproGnRH mRNAs in the tree shrew: First direct evidence for mesencephalic GnRH gene expression in a placental mammal. Gen Comp Endocrinol 1996; 104: 7–19. 33. Kelsall R, Coe IR, Sherwood NM. Phylogeny and ontogeny of gonadotropin-releasing hormone: Comparison of guinea pig, rat, and a protochordate. Gen Comp Endocrinol 1990; 78: 479–94. 34. King J, Millar R. Genealogy of the GnRH family. Prog Comp Endocrinol 1990; 54–59. 35. King JA, Millar RP. Gonadotropin-releasing hormones. In: Pang PK, Schriebmen MP, Eds. Vertebrate Endocrinology: Fundamental and Biomedical Implications. Vol 4, Part B. New York: Academic Press, 1991: 1–31. 36. Klungland H, Andersen O, Kisen G, Alestrom P, Tora L. Estrogen receptor binds to the salmon GnRH gene in a region with long palindromic sequences. Mol Cell Endocrinol 1993; 95: 147–154. 37. Lin XW, Peter RE. Expression of salmon gonadotropin-releasing hormone (GnRH) and chicken GnRH-II precursor messenger ribonucleic acids in the brain and ovary of goldfish. Gen Comp Endocrinol 1996; 101: 282–296. 38. Lin XW, Peter RE. Cloning and expression pattern of a second [His5Trp7Tyr8]gonadotropinreleasing hormone (chicken GnRH-H-II) mRNA in goldfish: Evidence for two distinct genes. Gen Comp Endocrinol 1997; 107: 262–72. 39. Lovejoy DA, Fischer WH, Ngamvongchon S, Craig AG, Nahorniak CS, Peter RE, Rivier JE, Sherwood NM. Distinct sequence of gonadotropin-releasing hormone (GnRH) in dogfish brain provides insight into GnRH evolution. Proc Natl Acad Sci USA 1992; 89: 6373–7. 40. Lundin LG. Evolution of the vertebrate genome as reflected i nparalogous chromosomal regions in man and the house mouse. Genomics 1993; 16: 1–19. 41. Matsuo H, Baba Y, Nair RMG, Arimura A, Schally AV. Structure of the porcine LH- and FSH-releasing hormone. I. The proposed amino acid sequence. Biochem Biophys Res Commun 1971; 43: 1334–1339. 42. Miller KE, Kriebel RM. Peptidergic innervation of caudal neurosecretory neurons. Gen Comp Endocrinol 1986; 64: 396–400. 43. Munz H, Stumpf W, Jennes L. LHRH systems in the brain of platyfish. Brain Res 1981; 221: 1–13. 44. Muske L. Evolution of gonadotropin-releasing hormone (GnRH) neuronal systems. Brain Behav Evol 1993; 42: 215–230. 45. Navas JM, Anglade I, Bailhache T, Pakdel F, Breton B, Jego P, Kah O. Do gonadotrophinreleasing hormone neurons express estrogen receptors in the rainbow trout? A double immunohistochemical study. J Comp Neurol 1995; 363: 461–474. 46. Norgren RB, Gao C. LHRH neuronal subtypes have multiple origins in chickens. Dev Brain Res 1994; 165: 735–738. frne 0181 GnRH GENES: PHYLOGENY, STRUCTURE, AND CONTROL 239 47. Northcutt RG, Muske LE. Multiple embryonic origins of gonadotropin-releasing hormone (GnRH) immunoreactive neurons. Dev Brain Res 1994; 78: 279–290. 48. Ohno S. Patterns in genome evolution. Curr Opin Genet Dev 1993; 3: 911–914. 49. Parhar IS, Soga T, Ishikawa Y, Nagahama Y, Sakuma Y. Neurons synthesizing gonadotropinreleasing hormone mRNA subtypes have multiple developmental origins in the medaka. J Comp Neurol 1998; 401: 217–226. 50. Peng C, Fan NC, Ligier M, Vaananen J, Leung PC. Expression and regulation of gonadotropinreleasing hormone (GnRH) and GnRH receptor messenger ribonucleic acids in human granulosa-luteal cells. Endocrinology 1994; 135: 1740–1746. 51. Penlington MC, Williams MA, Sumpter JP, Rand-Weaver M, Hoole D, Arme C. Isolation and characterization of mRNA encoding the salmon- and chicken-II type gonadotropin-releasing hormones in the teleost fish Rutilus rutilus (Cyprinidae). J Mol Endocrinol 1997; 19: 337–346. 52. Schwanzel-Fukuda M, Pfaff D. Origin of luteinizing hormone-releasing hormone neurons. Nature 1989; 338: 161–163. 53. Seeburg PH, Adelman JP. Characterization of cDNA for precursor of human luteinizing hormone releasing hormone. Nature 1984; 311: 666–668. 54. Seeburg P, Mason A, Stewart T, Nikolics K. The mammalian GnRH gene and its pivotal role in reproduction. Rec Prog Horm Res 1987; 43: 69–98. 55. Sherwood N, Lovejoy D, Coe I. Origin of mammalian gonadotropin-releasing hormones. Endocrinology 1993; 14: 241–254. 56. Shivers BD, Harlan HE, Morell JI, Pfaff DW. Absence of oestradiol concentration in cell nuclei of LHRH immunoreactive neurons. Nature 1983; 304: 345–347. 57. Silver R, Ramos CL, Silverman AJ. Sexual behavior triggers the appearance of non-neuronal cells containing gonadotropin-releasing hormone-like immunoreactivity. J Neuroendocrinol 1992; 4: 207–210. 58. Silverman AJ, Millar RP, King JA, Zhuang X, Silver R. Mast-cells with gonadotropinreleasing hormone-like immunoreactivity in the brain of doves. Proc Natl Acad Sci USA 1994; 91: 3694–3699. 59. Soma KK, Francis RC, Wingfield JC, Fernald RD. Androgen regulation of hypothalamic neurons containing gonadotropin-releasing hormone in a cichlid fish: Integration with social cues. Horm Behav 1996; 30: 216–226. 60. Sullivan KA, Witkin JW, Ferin M, Silverman AJ. Gonadotropin-releasing hormone neurons in the rhesus macaque are not immunoreactive for the estrogen receptor. Brain Res 1995; 685: 198–200. 61. Troskie B, Illing N, Rumbak E, Sun Y-M, Hapgood J, Conklin D, Millar R. Identification of three putative GnRH receptor subtypes in vertebrates. Gen Comp Endocrinol 1998; 112: 296–302. 62. Tsutsui H, Yamamamoto N, Ito H, Oka Y. GnRH-immunoreactive neuronal system in the presumptive ancestral chordate, Ciona intestinalis (Ascidian). Gen Comp Endocrinol 1998; 112: 426–432. 63. Urbanski HF, White RB, Fernald RD, Kohama SG, Densmore VS. Regional expression of mRNA encoding a second form of gonadotropin-releasing hormone in the macaque brain. Endocrinology 1999; 140: 1945–1948. 64. Varma S, Sabharwal P, Malarkey WB. Human splenocytes secrete LHRH, which inhibits lymphocyte proliferation. Prog Neuroendocrinimmunol 1992; 5: 187–191. 65. Warembourg M, Leroy D, Peytevin J, Martinet L. Estrogen receptor and progesterone receptor-immunoreactive cells are not colocalized with gonadotropin-releasing hormone in the brain of the female mink (Mustela vison). Cell Tissue Res 1998; 291: 33–41. 66. White SA, Bond CT, Francis RC, Kasten TL, Fernald RD, Adelman JP. A second gene for GnRH: cDNA and patterns of expression. Proc Natl Acad Sci USA 1994; 91: 1423–1427. frne 0181 240 FERNALD AND WHITE 67. White SA, Kasten T, Bond C, Adelman J, Fernald R. Three gonadotropin-releasing hormone genes in one organism suggest novel roles for an ancient peptide. Proc Natl Acad Sci USA 1995; 92: 8363–8367. 68. White RB, Eisen JA, Kasten TL, Fernald RD. Second gene for gonadotropin-releasing hormone in humans. Proc Natl Acad Sci USA 1998; 95: 305–309. 69. White R, Fernald R. Genomic structure and expression sites of three gonadotropin-releasing hormone genes in one species. Gen Comp Endocrinol 1998; 112: 17–25. 70. White R, Fernald R. Ontogeny of gonadotropin-releasing hormone (GnRH) gene expression reveals distinct origin for GnRH-containing neurons in the midbrain. Gen Comp Endocrinol 1998; 112: 322–329. 71. Wilson TM, Yu-Lee L-Y, Kelley MR. Coordinate gene expression of luteinizing hormonereleasing hormone (LHRH) and the LHRH-receptor after prolactin stimulation in the rat Nb2 T-cell line: Implications for a role in immunomodulation and cell cycle gene expression. Mol Endocrinol 1995; 9: 44–53. 72. Wray S, Grant P, Gainer H. Evidence that cells expressing lutienizing hormone-releasing hormone mRNA in the mouse are derived from progenitor cells in the olfactory placode. Proc Natl Acad Sci USA 1989; 86: 8132–8136. 73. Yamamoto N, Oka Y, Kawashima S. Lesions of gonadotropin-releasing hormone-immunoreactive terminal nerve cells: Effects on the reproductive behavior of male dwarf gouramis. Neuroendocrinology 1997; 65: 403–412. 74. Yu K-L, He M-L, Chik C-C, Lin X-W, Chang JP, Peter RE. mRNA expression of gonadotropinreleasing hormones (GnRHs) and GnRH receptor in goldfish. Gen Comp Endocrinol 1998; 112: 303–311. frne 0181