Molecular cloning, localization and circadian expression of chicken melanopsin (Opn4): differential regulation of expression in pineal and retinal cell types.

The avian retina and pineal gland contain autonomous circadian oscillators and photo-entrainment pathways, but the photopigment(s) that mediate entrainment have not been definitively identified. Melanopsin (Opn4) is a novel opsin involved in entrainment of circadian rhythms in mammals. Here, we report the cDNA cloning of chicken melanopsin and show its expression in retina, brain and pineal gland. Like the melanopsins identified in amphibians and mammals, chicken melanopsin is more similar to the invertebrate retinaldehyde-based photopigments than the retinaldehyde-based photopigments typically found in vertebrates. In retina, melanopsin mRNA is expressed in cells of all retinal layers. In pineal gland, expression was strong throughout the parenchyma of the gland. In brain, expression was observed in a few discrete nuclei, including the lateral septal area and medial preoptic nucleus. The retina and pineal gland showed distinct diurnal expression patterns. In pineal gland, melanopsin mRNA levels were highest at night at Zeitgeber time (ZT) 16. In contrast, transcript levels in the whole retina reached their highest levels in the early morning (ZT 0-4). Further analysis of melanopsin mRNA expression in retinal layers isolated by laser capture microdissection revealed different patterns in different layers. There was diurnal expression in all retinal layers except the ganglion cell layer, where heavy expression was localized to a small number of cells. Expression of melanopsin mRNA peaked during the daytime in the retinal pigment epithelium and inner nuclear layer but, like in the pineal, at night in the photoreceptors. Localization and regulation of melanopsin mRNA in the retina and pineal gland is consistent with the hypothesis that this novel photopigment plays a role in photic regulation of circadian function in these tissues.

Characterization of the rat GRIK5 kainate receptor subunit gene promoter and its intragenic regions involved in neural cell specificity.

The GRIK5 (glutamate receptor ionotropic kainate-5) gene encodes the kainate-preferring glutamate receptor subunit KA2. The GRIK5 promoter is TATA-less and GC-rich, with multiple consensus initiator sequences. Transgenic mouse lines carrying 4 kilobases of the GRIK5 5'-flanking sequence showed lacZ reporter expression predominantly in the nervous system. Reporter assays in central glial (CG-4) and non-neural cells indicated that a 1200-base pair (bp) 5'-flanking region could sustain neural cell-specific promoter activity. Transcriptional activity was associated with the formation of a transcription factor IID-containing complex on an initiator sequence located 1100 bp upstream of the first intron. In transfection studies, deletion of exonic sequences downstream of the promoter resulted in reporter gene activity that was no longer neural cell-specific. When placed downstream of the GRIK5 promoter, a 77-bp sequence from the deleted fragment completely silenced reporter expression in NIH3T3 fibroblasts while attenuating activity in CG-4 cells. Analysis of the 77-bp sequence revealed a functional SP1-binding site and a sequence resembling a neuron-restrictive silencer element. The latter sequence, however, did not display cell-specific binding of REST-like proteins. Our studies thus provide evidence for intragenic control of GRIK5 promoter activity and suggest that elements contributing to tissue-specific expression are contained within the first exon.

Expression and regulation of the LIM-class homeobox gene rlim-1 in neuronal progenitors of the rat cerebellum.

To investigate LIM gene function in the rat cerebellar system, we analyzed expression and regulation of the rat homologue of frog Xlim-1 (rlim-1) in vivo and in cultured cells. In developing cerebellum, peak levels of rlim-1 mRNA at postnatal day 8 (p8) are coincident with the peak period of granule cell proliferation. Analysis of rlim-1 protein with a specific antibody showed that expression was also maximal at p8. In situ hybridization showed that at p8 rlim-1 mRNA was expressed in Purkinje and granule cells. Both the proliferative and the premigratory granule cells in the external germinal zone displayed high levels of rlim-1 mRNA expression. Immunocytochemical staining demonstrated that at p8 rlim-1 protein was also present in proliferative and premigratory granule cells. In adult cerebellum (p30), rlim-1 mRNA and protein expression in granule cells was strongly attenuated. The down-regulation of rlim-1 mRNA occurred in granule cells just after the time of final division, coinciding with the onset of their migration. rlim-1 protein was detected in migratory granule neurons. The developmental decrease in rlim-1 mRNA and protein found in vivo was reproduced in pure cerebellar granule cell cultures. In these cultures, granule neurons were postmitotic 1 day after plating but still displayed high levels of rlim-1 protein expression up to 3 days in vitro. Our findings indicate that 1) rlim-1 is likely to act in concert with other genes to specify granule cell fate, 2) rlim-1 expression in granule neurons is regulated autonomously, and 3) rlim-1 protein may also play an important role in granule neuron differentiation and survival. Published 2001 Wiley-Liss, Inc.

A novel human opsin in the inner retina.

Here we report the identification of a novel human opsin, melanopsin, that is expressed in cells of the mammalian inner retina. The human melanopsin gene consists of 10 exons and is mapped to chromosome 10q22. This chromosomal localization and gene structure differs significantly from that of other human opsins that typically have four to seven exons. A survey of 26 anatomical sites indicates that, in humans, melanopsin is expressed only in the eye. In situ hybridization histochemistry shows that melanopsin expression is restricted to cells within the ganglion and amacrine cell layers of the primate and murine retinas. Notably, expression is not observed in retinal photoreceptor cells, the opsin-containing cells of the outer retina that initiate vision. The unique inner retinal localization of melanopsin suggests that it is not involved in image formation but rather may mediate nonvisual photoreceptive tasks, such as the regulation of circadian rhythms and the acute suppression of pineal melatonin. The anatomical distribution of melanopsin-positive retinal cells is similar to the pattern of cells known to project from the retina to the suprachiasmatic nuclei of the hypothalamus, a primary circadian pacemaker.

Melanopsin: An opsin in melanophores, brain, and eye.

We have identified an opsin, melanopsin, in photosensitive dermal melanophores of Xenopus laevis. Its deduced amino acid sequence shares greatest homology with cephalopod opsins. The predicted secondary structure of melanopsin indicates the presence of a long cytoplasmic tail with multiple putative phosphorylation sites, suggesting that this opsin's function may be finely regulated. Melanopsin mRNA is expressed in hypothalamic sites thought to contain deep brain photoreceptors and in the iris, a structure known to be directly photosensitive in amphibians. Melanopsin message is also localized in retinal cells residing in the outermost lamina of the inner nuclear layer where horizontal cells are typically found. Its expression in retinal and nonretinal tissues suggests a role in vision and nonvisual photoreceptive tasks, such as photic control of skin pigmentation, pupillary aperture, and circadian and photoperiodic physiology.

Frog prohormone convertase PC2 mRNA has a mammalian-like expression pattern in the central nervous system and is colocalized with a subset of thyrotropin-releasing hormone-expressing neurons.

The prohormone convertase (PC2) is expressed in the mammalian central nervous system (CNS) and has been shown to play an important role in the processing of certain neuropeptide precursors and prohormones at paired basic residues. Amphibian PC2 cDNA was recently cloned for the frog Xenopus laevis, and both its sequence and its pituitary expression pattern were shown to be very similar to those of mammalian PC2. To investigate further the function of PC2 in the vertebrate CNS, we used in situ hybridization histochemistry to localize the distribution of cells expressing PC2 mRNA in the frog brain and the spinal cord. The distribution of PC2-expressing cells was also compared with that of cells expressing thyrotropin-releasing hormone (TRH) mRNA or peptide. PC2-expressing cells were detected in specific nuclei that were widely distributed in the frog CNS. In forebrain, telencephalic PC2 mRNA was found in the olfactory bulb, pallium, striatum, amygdala, and septum, and diencephalic PC2 mRNA was seen in the preoptic area, thalamus, and hypothalamus. More posteriorly, PC2 cells were localized to midbrain tegmentum, the torus semicircularis, and the optic tectum, as well as the cerebellum, brainstem, and spinal cord. Despite this wide distribution steady-state levels of PC2 mRNA were clearly different in various brain nuclei. Regions with higher levels showed good correspondence to areas shown by others in frog to contain large numbers of neuropeptide-expressing cells, including TRH cells. On the other hand, not all brain areas with high levels of TRH mRNA had high levels of PC2 mRNA. Localization studies combining in situ hybridization and immunocytochemistry showed that, at least in optic tectum and brainstem, PC2 mRNA and pro-TRH peptide coexist. These findings suggest that pro-TRH is processed by PC2 in some, but possibly not all, brain regions. Thus, different converting enzymes may be involved in pro-TRH processing in different brain regions.

Transcriptional activation of the matrix metalloproteinase gene stromelysin-3 coincides with thyroid hormone-induced cell death during frog metamorphosis.

A full-length cDNA was isolated for a thyroid hormone response gene in the metamorphosing frog intestine and shown by sequence analysis to be the frog homolog of the mammalian extracellular matrix metalloproteinase stromelysin-3 (ST3). Northern hybridization indicated that ST3 gene expression is differentially activated in tadpole tissues during metamorphosis. In the small intestine, in situ hybridization localized high levels of ST3 mRNA to fibroblast-like cells during thyroid hormone-induced metamorphosis. ST3mRNA was undetectable in the intestine prior to metamorphosis, while high levels were present at the metamorphic climax. At this time, primary intestinal epithelial cells are known to undergo cell death and replacement by secondary epithelial cells, arguing that ST3 is involved in the modification of the extracellular matrix during apoptosis. ST3mRNA was also expressed at high levels during tadpole tail resorption, but not in premetamorphic tail or developing hindlimb, further supporting a role for ST3 when tissue remodeling is accompanied by large-scale cell death. Premetamorphic tadpoles treated with thyroid hormone showed a similar but compressed time course of ST3 gene regulation, suggesting that thyroid hormone controls ST3 gene expression during metamorphosis. In contrast, during embryogenesis, ST3 was expressed before endogenous thyroid hormone is detectable, indicating that ST3 can also be regulated independently of thyroid hormone. These findings implicate that ST3 participates in the modification of the extracellular matrix during matamorphic apoptosis, but Northern analyses using heterologous probes raise the possibility that additional matrix metalloproteinases may also be involved.

The TRH neuronal phenotype forms embryonic cell clusters that go on to establish a regionalized cell fate in forebrain.

How neurons diversify in developing brain to produce discrete cell fates in their appropriate regions remains a fundamental question. Embryonic Xenopus was previously used to identify juxtaposed embryonic cells that first express proopiomelanocortin mRNA in forebrain and pituitary, supporting the idea that this neuropeptide phenotype is induced locally (Hayes and Loh, 1990, Development 110: 747-757). To begin to examine how a more widespread population of forebrain cells is set up, the present focus is on the thyrotropin-releasing hormone (TRH) phenotype. Serial section in situ hybridization histochemistry produced the unexpected finding that the adult-like TRH system spanning forebrain and comprising over six different telencephalic and diencephalic nuclei, is preceded by an embryonic TRH cell population that is initially localized and then highly regionalized in the area from which the adult pattern develops. Thus, the first TRH cells, detected in vivo after 35h (stage 29/30), were confined to discrete anterior or posterior bilateral clusters in embryonic forebrain or hindbrain. Thereafter, the TRH cell clusters in diencephalon, but not hindbrain, expanded to form rows, extending anteriorly into telencephalon and bifurcating posteriorly around the infundibulum. By 80 h (stage 42), after extensive brain morphogenesis, these forebrain rows showed regional differences in levels of TRH mRNA corresponding to the specific brain nuclei that have been shown to contain TRH cells in adult. These findings show that subsets of phenotype-specific forebrain cells first form a regionalized neuronal cell fate before distinct brain nuclei form. This in turn points to the testable hypothesis in Xenopus that certain neuronal cell fates in forebrain may be dictated by cell lineage or local induction.

The frog gonadotropin-releasing hormone-I (GnRH-I) gene has a mammalian-like expression pattern and conserved domains in GnRH-associated peptide, but brain onset is delayed until metamorphosis.

Recent evidence indicates a localized origin in the olfactory placode for the mammalian forebrain neurons that express GnRH. To identify the cellular and molecular signals that induce the GnRH phenotype, we cloned and characterized a cDNA encoding the GnRH prohormone, the precursor for both GnRH-I and GnRH-associated peptide in the frog, Xenopus laevis, an embryonic model accessible to experimental manipulation. The 396-base cDNA represented a single mRNA species encoding an 89-amino acid prepro-GnRH that, unlike a recently cloned fish GnRH gene, was identical to both the mammalian GnRH decapeptide as well as multiple domains within GnRH-associated peptide. Serial section in situ hybridization histochemistry and immunocytochemistry in adult frog localized a forebrain system comprising 250-350 cell bodies whose overall neuroanatomy, including fiber projections, was very similar to that described for mammals. However, neither Northern nor in situ hybridization detected GnRH expression in midbrain, arguing that another frog gene encodes the midbrain GnRH-II expression pattern described by many others using antisera directed against the fish GnRH-I or chicken GnRH-II decapeptides. In contrast to mammals and birds, in which GnRH-expressing cells migrate into embryonic forebrain, frog GnRH cells were first detected after they reached their final position in the preoptic area during the late larval period. Thus, although previous studies proposed a complex organization for the GnRH system in the frog, our findings show that similar to mammals, there is a single gene that can account for the continuum of GnRH-I cells spanning frog forebrain. However, unlike mammals, in frogs, for unknown reasons, GnRH-I gene expression is suppressed until metamorphic climax.

Thyroid hormone-dependent regulation of the intestinal fatty acid-binding protein gene during amphibian metamorphosis.

To investigate, at the molecular level, the remodeling of small intestine during amphibian metamorphosis, a subtractive hybridization approach was used to identify genes that are differentially regulated by thyroid hormone. A frog cDNA was isolated from Xenopus laevis and determined to be the gene encoding the intestinal fatty acid-binding protein (IFABP) based on its high sequence homology to the previously cloned mammalian IFABP gene. Northern blot analyses and in situ hybridization histochemistry also showed that, like the mammalian IFABP genes, frog IFABP gene expression is restricted to the intestinal epithelium. Xenopus embryos express detectable IFABP mRNA at stage 33/34, suggesting that intestinal epithelial cells differentiate well before feeding begins at stage 45. Moreover, during metamorphosis, levels of IFABP mRNA were gradually down-regulated over a period of about 20 days between stages 54 and 62, reaching a minimum at metamorphic climax, after which they were reelevated as the secondary epithelium forms. This reduction in IFABP gene expression could be reproduced in only 3 days by treating premetamorphic tadpoles with thyroid hormone. Our findings also show that this effect, while likely to be indirect, takes place before overt morphological changes are evident in primary epithelial cells. Thus, the down-regulation of IFABP mRNA is one of the early molecular events preceding epithelial cell death during intestinal remodeling.

Upstream sequences of the myogenin gene convey responsiveness to skeletal muscle denervation in transgenic mice.

Myogenin, as well as other MyoD-related skeletal muscle-specific transcription factors, regulate a large number of skeletal muscle genes during myogenic differentiation. During later development, innervation suppresses myogenin expression in the fetal hind limb musculature. Denervation of skeletal muscle reverses the effects of the nerve, and results in the reactivation of myogenin expression, as well as of other embryonic muscle proteins. Here we report that myogenin upstream sequences confer tissue- and developmental-specific expression in transgenic mice harboring a myogenin/chloramphenicol acetyltransferase (CAT) reporter construct. Using in situ hybridization to analyze serial sections of E12.5 embryos, we found colocalization of CAT and endogenous myogenin transcripts in the primordial muscle of the head and limbs, in the intercostal muscle masses, and in the most caudal somites. Later in development, we observed that the expression of the transgene and endogenous myogenin gene continued to be restricted to skeletal muscle but decreased shortly after birth; a period that coincides with the innervation of secondary myotubes. Furthermore, denervation of the mouse hind limbs induced a 10-fold accumulation of CAT and endogenous myogenin transcripts by 1 day after sciatic nerve resection; a 25-fold increase was observed by 4 days after denervation. Interestingly, we observed that the accumulation of CAT enzyme activity lagged considerably with respect to the increase in CAT transcripts. Our results indicate that the cis-acting elements that temporally and spatially confine transcription of the gene during embryonic development, and that mediate the responses to innervation and denervation of muscle, lie within the upstream sequences analyzed in these studies.

Expression of LIM class homeobox gene Xlim-3 in Xenopus development is limited to neural and neuroendocrine tissues.

The Xenopus LIM class homeobox gene Xlim-3 was identified initially as a fragment isolated by polymerase chain reaction cloning with an embryonic cDNA library as template (Taira et al., 1992, Genes Dev. 6, 356-366). cDNA clones representing most of the Xlim-3 mRNA were isolated from an adult brain library. The predicted Xlim-3 protein contains two copies of the LIM domain, a homeodomain, and a C-terminal region rich in proline, glycine, and serine. RNA blot hybridization showed that Xlim-3 mRNA is detected in dorsal regions at neural tube and tailbud stages and in adults predominantly in the pituitary gland and weakly in the eye and brain. Whole mount in situ hybridization revealed that Xlim-3 mRNA is first detectable at the neural plate stage in the stomodeal-hypophyseal (pituitary) anlage and in the neural plate where labeled cells were found adjacent to the forming floor plate. In situ hybridization analysis on serial sections at later stages showed that embryonic Xlim-3 expression persists in the pituitary and pineal, as well as in some cells of the retina, hindbrain, and spinal cord. In the retina, Xlim-3 mRNA was only detected in a distinct sublamina of the inner nuclear layer, but not in dividing cells of ciliary margin. This discrete manner of Xlim-3 expression, especially persistent expression in the pituitary (before morphogenesis of the gland to adult), supports a role in the specification and maintenance of differentiation of distinct neuronal and neuroendocrine tissues.

Molecular cloning and development analysis of a new glutamate receptor subunit isoform in cerebellum.

The glutamate receptor gene GluR-4 is proposed to generate two spliced isoforms (Sommer et al., 1990). Screening a rat cerebellar cDNA library, we have now identified a third type of transcript derived from GluR-4 gene by differential RNA processing. This transcript encodes a protein with a "flop" module between transmembrane regions 3 and 4, but with a C-terminus segment of 36 amino acids different from the previously described GluR-4 flip/flop cDNAs. This subunit was therefore designated as GluR-4c flop. Transcripts synthesized in vitro from GluR-4c cDNA form kainate/AMPA-activated channels when expressed in Xenopus oocytes. The current-voltage relationship for kainate-evoked responses in oocytes injected with GluR-4c showed strong inward rectification. The different transcripts derived from the GluR-4 gene were studied on Northern blots hybridized with either a cDNA probe or oligonucleotides specific for the GluR-4 flip/flop and C-terminal domains. Three transcripts of 6.2, 4.2, and 3.0 kilobases (kb) derived from the GluR-4 gene were identified on Northern blots containing total RNA prepared from different brain regions, using a cDNA probe or an oligonucleotide corresponding to the N-terminal region common to all transcripts. These transcripts were much more abundant in the cerebellum than in other brain areas, and their levels increased during cerebellar development. The maximal increase was observed between postnatal days 1 and 20, an age corresponding to the division and maturation of granule neurons. The flip/flop and the C-terminal oligonucleotides hybridized to the two higher molecular weight transcripts but did not hybridize to the small RNA. Interestingly, using cerebellar cells that were cultured for up to 12 d, we observed that the three transcripts are present in granule neurons, but that astrocytes only express the 6.2 and the 4.2 kb transcripts. The 3.0 kb transcript accumulates in cerebellar granule cells during development in vitro. Furthermore, in situ hybridization histochemistry revealed that the GluR-4c transcripts are preferentially expressed in cerebellar granule cells and Bergmann glial cells, whereas the expression of GluR-4 flip mRNAs is restricted to Bergmann glial cells. Interestingly, we also show that granule cells already express GluR-4c in the premigratory zone of the external granular layer, indicating that intrinsic or highly localized cues induce GluR-4c expression before these cells reach their final position.

Correlated onset and patterning of proopiomelanocortin gene expression in embryonic Xenopus brain and pituitary.

To identify cellular interactions that underlie the spatially appropriate transcription of neural genes, we characterized the embryonic development of proopiomelanocortin (POMC) gene expression in Xenopus laevis using in situ hybridization histochemistry. This has led to the establishment of a unique model system for studying how a neuropeptide gene program in four distinct cell groups is set up in pituitary and forebrain. The embryonic onset and patterning of POMC expression was found to be spatially and temporally correlated inside and outside the brain. The first POMC cells in the pituitary primordium and diencephalon were juxtaposed near the infundibulum at stage 29/30, indicating they undergo molecular differentiation much earlier than previously reported for this system. By stage 31/32, many more POMC cells appeared in the morphologically undifferentiated pituitary primordium and brain. In fact, these cells were seen throughout the presumptive anterior and intermediate lobes of the pituitary and posterior diencephalon at the same time that the pituitary primordium is translocating ventral to diencephalon. By stage 39/40, coordinated morphogenesis produced the adult pattern of POMC cells localized in distinct anterior and intermediate pituitary lobes and two diencephalic nuclei. We propose in light of these findings that embryonic cells in the pituitary primordium and brain are simultaneously induced to transcribe the POMC gene, possibly as a result of reciprocal brain-pituitary interactions.

Normal numbers of retinotectal synapses during the activity-sensitive period of optic regeneration in goldfish: HRP-EM evidence implicating synapse rearrangement and collateral elimination during map refinement.

Optic and nonoptic fibers and synapses were counted in the primary optic innervation layer (S-SO-SFGS) in anteromedial tectum in normal goldfish and in fish 30, 60, and 240 d after the optic nerve was crushed. A newly developed "cold-fill" HRP-labeling protocol was used to label optic afferents for electron microscopy, and counts were then made on EM photomontages of columns through the HRP-labeled S-SO-SFGS. Normal numbers of retinotectal synapses were present at 30 d regeneration, at a time when activity-dependent refinement of the optic projection is incomplete. Normal numbers were also found at 60 and 240 d, when refinement is largely completed. In contrast to this constancy in optic synapse numbers, there was nearly 10 times the normal number of optic fibers in the SFGS at 30 d, and these were reduced by 50% at 60 d, remaining over 4 times normal at 240 d. These findings imply extensive rearrangement of optic synapses during map refinement. They also indicate that synapse rearrangement is associated with the elimination of optic collaterals.

Impulse blockade by intraocular tetrodotoxin during optic regeneration in goldfish: HRP-EM evidence that the formation of normal numbers of optic synapses and the elimination of exuberant optic fibers is activity independent.

Optic fibers and synapses labeled with HRP were counted in the primary optic innervation layer of tectum after continuously blocking visual impulse activity with TTX during regeneration. Normal numbers of optic and nonoptic fibers and synapses were found at both 30 and 60 d, and key ultrastructural features of optic afferents such as fiber fasciculation, myelination, terminal clustering, synaptogenesis onto different classes of postsynaptic targets and general morphology were not notably affected by impulse blockade. These findings indicate that during regeneration the normal proliferation and elimination of optic fibers and the formation of normal numbers of optic synapses are not regulated by activity and are consistent with a pattern formation role for impulse activity rather than a trophic one.

Normal and regenerating optic fibers in goldfish tectum: HRP-EM evidence for rapid synaptogenesis and optic fiber-fiber affinity.

The distribution of normal and regenerating retinal fibers and synapses was studied on tectum in goldfish by light (LM) and electron microscopy (EM). Since labeling of the early regenerating fibers was previously reported to be difficult, a new 'cold-fill' HRP labeling protocol was developed, which labeled regenerating optic fibers and terminals on tectum as early as 14 days after nerve crush when they first arrive on tectum. In order to characterize the laminar distribution of optic afferents in normal fish and in fish regenerating for 14-240 days, EM photomontages of areas 14 microns wide by 160 microns deep through the HRP-labeled primary optic innervation layer (S-SO-SFGS) were constructed. The time points in regeneration that were examined spanned the period in which others have shown that an initially diffuse retinotopic map becomes spatially restricted. At the LM level regenerating optic fibers were restricted to the optic lamina. They reinnervated tectum in an anterior to posterior sequence as previously seen with autoradiography. In addition, at 14 days, some "pioneer" optic fascicles were found to have already grown to posterior tectum where they gave rise to branches with boutonlike terminations and growth-cone-like processes. Form the ultrastructural analysis it was clear that optic fibers and terminals observed strict laminar boundaries as they partitioned themselves in the optic laminae (S, SO and SFGS) in both normal and regenerating fish. The behavior of optic fibers was lamina specific with respect to synapse formation and the orientation of fiber outgrowth. As early as 14 days regeneration, optic fibers made synapses onto the four types of postsynaptic profiles observed in normal fish. Numerous optic terminals were labeled at 14 days, and there appeared to be no waiting period between fiber ingrowth to the SO and synapse formation in the S and SFGS. At 14-60 days, atypical synaptic contacts which appear to be nascent synapses were made by labeled optic fibers in fascicles and by growth-cone-like processes. By 21-30 days, the density of optic terminals was high and there were many more fasciculated optic fibers in the SFGS than normal as late as 350 days. These findings suggest that optic fiber lamination is highly constrained by tectal cues, that fibers rapidly regenerate many synaptic terminals before retinotopic map refinement is complete, and that fibers have a strong affinity for each other.

Optic synapse number but not density is constrained during regeneration onto surgically halved tectum in goldfish: HRP-EM evidence that optic fibers compete for fixed numbers of postsynaptic sites on the tectum.

The number of optic synapses in the half tectum of goldfish was counted by using an improved HRP-labeling protocol and a columnar sampling method that spanned the entire optic innervation layer, S-SO-SFGS. It was previously found by using this procedure in intact tectum that the normal number of optic synapses was regenerated by 30 days and maintained thereafter even in the absence of impulse activity. This suggested that the number of synapses in this system was intrinsically fixed. In order to examine whether this limit was imposed by optic fibers or by target cells, optic synapses were counted in surgically halved tecta which received compressed optic projections consisting of regenerating optic fibers from the entire retina. We reasoned that if synapse number is a function of the number of afferents, then there should be twice the normal number of optic synapses per column; on the other hand, if their number is fixed by target, then their number per column should be normal. We found that the number of optic (labeled) synapses was normal in sample columns from fish at 70 days and 160 days after optic nerve crush. Thus, retinal ganglion cells, on average, formed half as many synapses on the half tectum compared to intact tectum, indicating the number of optic synapses was limited by the tectum. The number of nonoptic (unlabeled) synapses was also found to be normal. By contrast, the S-SO-SFGS was found to be 88-103% thicker compared to normal fish, apparently because of a 20-fold increase in the number of optic fibers. As a result, the density of synapses was about half normal in half tecta, and so, in contrast to synapse number, synaptic density is not constrained during regeneration. We infer from these data that optic fibers compete for limited numbers of postsynaptic sites during regeneration and suggest that this competition promotes neural map refinement and the various plasticities described for this projection.

Retinotopically inappropriate synapses of subnormal density formed by surgically misdirected optic fibers in goldfish tectum.

Selected optic fibers were surgically deflected from one tectum onto the opposite host tectum which was denervated by eye enucleation. At 6-8 months, deflected fibers were labeled with horseradish peroxidase and the retinotopically inappropriate part of tectum was examined using electron microscopy. Numerous (labeled) optic synapses were found in the primary optic innervation layer of the 'wrong' part of tectum but they were about half the normal density. The number and density of non-optic synapses was not found to be affected. These findings indicate optic fibers compete with each other but not with non-optic fibers for synaptic sites in tectum.