Interommatidial cells build a tensile collagen network during Drosophila retinal morphogenesis

Drosophila compound eye morphogenesis transforms a simple epithelium into an approximate hollow hemisphere comprised of ∼700 ommatidia, packed as tapering hexagonal prisms between a rigid external array of cuticular lenses and a parallel, rigid internal floor, the fenestrated membrane (FM). Critical to vision, photosensory rhabdomeres are sprung between these two surfaces, grading their length and shape accurately across the eye and aligning them to the optical axis. Using fluorescently tagged collagen and laminin, we show that that the FM assembles sequentially, emerging in the larval eye disc in the wake of the morphogenetic furrow as the original collagen-containing basement membrane (BM) separates from the epithelial floor and is replaced by a new, laminin-rich BM, which advances around axon bundles of newly differentiated photoreceptors as they exit the retina, forming fenestrae in this new, laminin-rich BM. In mid-pupal development, the interommatidial cells (IOCs) autonomously deposit collagen at fenestrae, forming rigid, tension-resisting grommets. In turn, stress fibers assemble in the IOC basal endfeet, where they contact grommets at anchorages mediated by integrin linked kinase (ILK). The hexagonal network of IOC endfeet tiling the retinal floor couples nearest-neighbor grommets into a supracellular tri-axial tension network. Late in pupal development, IOC stress fiber contraction folds pliable BM into a hexagonal grid of collagen-stiffened ridges, concomitantly decreasing the area of convex FM and applying essential morphogenetic longitudinal tension to rapidly growing rhabdomeres. Together, our results reveal an orderly program of sequential assembly and activation of a supramolecular tensile network that governs Drosophila retinal morphogenesis.

Calcium waves facilitate and coordinate the contraction of endfeet actin stress fibers in Drosophila interommatidial cells

Actomyosin contraction shapes the Drosophila eye's panoramic view. The convex curvature of the retinal epithelium, organized in ∼800 close-packed ommatidia, depends upon a fourfold condensation of the retinal floor mediated by contraction of actin stress fibers in the endfeet of interommatidial cells (IOCs). How these tensile forces are coordinated is not known. Here, we discover a novel phenomenon: Ca2+ waves regularly propagate across the IOC network in pupal and adult eyes. Genetic evidence demonstrates that IOC waves are independent of phototransduction, but require inositol 1,4,5-triphosphate receptor (IP3R), suggesting these waves are mediated by Ca2+ releases from ER stores. Removal of IP3R disrupts stress fibers in IOC endfeet and increases the basal retinal surface by ∼40%, linking IOC waves to facilitating stress fiber contraction and floor morphogenesis. Further, IP3R loss disrupts the organization of a collagen IV network underneath the IOC endfeet, implicating ECM and its interaction with stress fibers in eye morphogenesis. We propose that coordinated cytosolic Ca2+ increases in IOC waves promote stress fiber contractions, ensuring an organized application of the planar tensile forces that condense the retinal floor.

Cytochrome b5 protects photoreceptors from light stress-induced lipid peroxidation and retinal degeneration

Lipid peroxides are generated by oxidative stress in cells, and contribute to ageing and neurodegenerative disease. The eye is at special risk for lipid peroxidation because photoreceptors possess amplified sensory membranes rich in peroxidation-susceptible polyunsaturated fatty acids. Light-induced lipid peroxidation in the retina contributes to retinal degeneration, and lipid peroxidation has been implicated in the progression of age-associated ocular diseases such as age-related macular degeneration (AMD). Here, we show that exposing Drosophila melanogaster to strong blue light induces oxidative stress including lipid peroxidation that results in retinal degeneration. Surprisingly, very young flies are resilient to this acute light stress, suggesting they possess endogenous neuroprotective mechanisms. While lipophilic antioxidants partially suppressed blue light-induced retinal degeneration in older flies, we find that overexpression of cytochrome b5 (Cyt-b5) completely suppressed both blue light-induced lipid peroxidation and retinal degeneration. Our data identify Cyt-b5 as a neuroprotective factor that targets light-induced oxidative damage, particularly lipid peroxidation. Cyt-b5 may function via supporting antioxidant recycling, thereby providing a strategy to prevent oxidative stress in ageing photoreceptors that would be synergistic with dietary antioxidant supplementation.

Paradoxically, light is essential for vision, yet it also induces stress that damages the sensitive cells in the eye. Vikki Weake and her team at Purdue University examined how exposure to blue light causes damage to the retina in fruit flies. Blue light causes death of photoreceptors, the light-sensing neurons. Surprisingly, very young flies are resistant to blue light. Increasing levels of a single protein, Cytochrome-b5, mimicked youthful resilience in older flies. Cytochrome-b5 is central to an ancient cellular defense system that protects membranes from oxidative damage. With expansive sensory membranes containing specialized lipids, photoreceptors are especially sensitive to membrane lipid peroxidation, an emerging final common pathway for cell death in aging and disease. Research into preventing lipid peroxidation might help to develop therapies for age-related diseases such as age-related macular degeneration.

A Programmable Optical Stimulator for the Drosophila Eye

A programmable optical stimulator for Drosophila eyes is presented. The target application of the stimulator is to induce retinal degeneration in fly photoreceptor cells by exposing them to light in a controlled manner. The goal of this work is to obtain a reproducible system for studying age-related changes in susceptibility to environmental ocular stress. The stimulator uses light emitting diodes and an embedded computer to control illuminance, color (blue or red) and duration in two independent chambers. Further, the stimulator is equipped with per-chamber light and temperature sensors and a fan to monitor light intensity and to control temperature. An ON/OFF temperature control implemented on the embedded computer keeps the temperature from reaching levels that will induce the heat shock stress response in the flies. A custom enclosure was fabricated to house the electronic components of the stimulator. The enclosure provides a light-impermeable environment that allows air flow and lets users easily load and unload fly vials. Characterization results show that the fabricated stimulator can produce light at illuminances ranging from 0 to 16000 lux and power density levels from 0 to 7.2 mW/cm2 for blue light. For red light the maximum illuminance is 8000 lux which corresponds to a power density of 3.54 mW/cm2. The fans and the ON/OFF temperature control are able to keep the temperature inside the chambers below 28.17°C. Experiments with white-eye male flies were performed to assess the ability of the fabricated simulator to induce blue light-dependent retinal degeneration. Retinal degeneration is observed in flies exposed to 8 hours of blue light at 7949 lux. Flies in a control experiment with no light exposure show no retinal degeneration. Flies exposed to red light for the similar duration and light intensity (8 hours and 7994 lux) do not show retinal degeneration either. Hence, the fabricated stimulator can be used to create environmental ocular stress using blue light.

Ire1 supports normal ER differentiation in developing Drosophila photoreceptors

The endoplasmic reticulum (ER) serves virtually all aspects of cell physiology and, by pathways that are incompletely understood, is dynamically remodeled to meet changing cell needs. Inositol-requiring enzyme 1 (Ire1), a conserved core protein of the unfolded protein response (UPR), participates in ER remodeling and is particularly required during the differentiation of cells devoted to intense secretory activity, so-called 'professional' secretory cells. Here, we characterize the role of Ire1 in ER differentiation in the developing Drosophila compound eye photoreceptors (R cells). As part of normal development, R cells take a turn as professional secretory cells with a massive secretory effort that builds the photosensitive membrane organelle, the rhabdomere. We find rough ER sheets proliferate as rhabdomere biogenesis culminates, and Ire1 is required for normal ER differentiation. Ire1 is active early in R cell development and is required in anticipation of peak biosynthesis. Without Ire1, the amount of rough ER sheets is strongly reduced and the extensive cortical ER network at the rhabdomere base, the subrhabdomere cisterna (SRC), fails. Instead, ER proliferates in persistent and ribosome-poor tubular tangles. A phase of Ire1 activity early in R cell development thus shapes dynamic ER.

Depletion of PtdIns(4,5)P₂ underlies retinal degeneration in Drosophila trp mutants

The prototypical transient receptor potential (TRP) channel is the major light-sensitive, and Ca(2+)-permeable channel in the microvillar photoreceptors of Drosophila. TRP channels are activated following hydrolysis of phosphatidylinositol 4,5-bisphosphate [PtdIns(4,5)P₂] by the key effector enzyme phospholipase C (PLC). Mutants lacking TRP channels undergo light-dependent retinal degeneration, as a consequence of the reduced Ca(2+) influx. It has been proposed that degeneration is caused by defects in the Ca(2+)-dependent visual pigment cycle, which result in accumulation of toxic phosphorylated metarhodopsin-arrestin complexes (MPP-Arr2). Here we show that two interventions, which prevent accumulation of MPP-Arr2, namely rearing under red light or eliminating the C-terminal rhodopsin phosphorylation sites, failed to rescue degeneration in trp mutants. Instead, degeneration in trp mutants reared under red light was rescued by mutation of PLC. Degeneration correlated closely with the light-induced depletion of PtdIns(4,5)P₂ that occurs in trp mutants due to failure of Ca(2+)-dependent inhibition of PLC. Severe retinal degeneration was also induced in the dark in otherwise wild-type flies by overexpression of a bacterial PtdInsPn phosphatase (SigD) to deplete PtdIns(4,5)P₂. In degenerating trp photoreceptors, phosphorylated Moesin, a PtdIns(4,5)P₂-regulated membrane-cytoskeleton linker essential for normal microvillar morphology, was found to delocalize from the rhabdomere and there was extensive microvillar actin depolymerisation. The results suggest that compromised light-induced Ca(2+) influx, due to loss of TRP channels, leads to PtdIns(4,5)P₂ depletion, resulting in dephosphorylation of Moesin, actin depolymerisation and disintegration of photoreceptor structure.

Ectoplasm, ghost in the R cell machine?

Drosophila photoreceptors (R cells) are an extreme instance of sensory membrane amplification via apical microvilli, a widely deployed and deeply conserved operation of polarized epithelial cells. Developmental rotation of R cell apices aligns rhabdomere microvilli across the optical axis and enables enormous membrane expansion in a new, proximal distal dimension. R cell ectoplasm, the specialized cortical cytoplasm abutting the rhabdomere is likewise enormously amplified. Ectoplasm is dominated by the actin-rich terminal web, a conserved operational domain of the ancient vesicle-transport motor, Myosin V. R cells harness Myosin V to move two distinct cargoes, the biosynthetic traffic that builds the rhabdomere during development, and the migration of pigment granules that mediates the adaptive "longitudinal pupil" in adults, using two distinct Rab proteins. Ectoplasm further shapes a distinct cortical endosome compartment, the subrhabdomeral cisterna (SRC), vital to normal cell function. Reticulon, a protein that promotes endomembrane curvature, marks the SRC. R cell visual arrestin 2 (Arr2) is predominantly cytoplasmic in dark-adapted photoreceptors but on illumination it translocates to the rhabdomere, where it quenches ongoing photosignaling by binding to activated metarhodopsin. Arr2 translocation is "powered" by diffusion; a motor is not required to move Arr2 and ectoplasm does not obstruct its rapid diffusion to the rhabdomere.

Arrestin translocation is stoichiometric to rhodopsin isomerization and accelerated by phototransduction in Drosophila photoreceptors

Upon illumination, visual arrestin translocates from photoreceptor cell bodies to rhodopsin and membrane-rich photosensory compartments, vertebrate outer segments or invertebrate rhabdomeres, where it quenches activated rhodopsin. Both the mechanism and function of arrestin translocation are unresolved and controversial. In dark-adapted photoreceptors of the fruitfly Drosophila, confocal immunocytochemistry shows arrestin (Arr2) associated with distributed photoreceptor endomembranes. Immunocytochemistry and live imaging of GFP-tagged Arr2 demonstrate rapid reversible translocation to stimulated rhabdomeres in stoichiometric proportion to rhodopsin photoisomerization. Translocation is very rapid in normal photoreceptors (time constant <10 s) and can also be resolved in the time course of electroretinogram recordings. Genetic elimination of key phototransduction proteins, including phospholipase C (PLC), Gq, and the light-sensitive Ca2+-permeable TRP channels, slows translocation by 10- to 100-fold. Our results indicate that Arr2 translocation in Drosophila photoreceptors is driven by diffusion, but profoundly accelerated by phototransduction and Ca2+ influx.

Ca2+-dependent metarhodopsin inactivation mediated by calmodulin and NINAC myosin III

Phototransduction in flies is the fastest known G protein-coupled signaling cascade, but how this performance is achieved remains unclear. Here, we investigate the mechanism and role of rhodopsin inactivation. We determined the lifetime of activated rhodopsin (metarhodopsin = M( *)) in whole-cell recordings from Drosophila photoreceptors by measuring the time window within which inactivating M( *) by photoreisomerization to rhodopsin could suppress responses to prior illumination. M( *) was inactivated rapidly (tau approximately 20 ms) under control conditions, but approximately 10-fold more slowly in Ca2+-free solutions. This pronounced Ca2+ dependence of M( *) inactivation was unaffected by mutations affecting phosphorylation of rhodopsin or arrestin but was abolished in mutants of calmodulin (CaM) or the CaM-binding myosin III, NINAC. This suggests a mechanism whereby Ca2+ influx acting via CaM and NINAC accelerates the binding of arrestin to M( *). Our results indicate that this strategy promotes quantum efficiency, temporal resolution, and fidelity of visual signaling.

Neuronal differentiation in Drosophila ommatidium

Using monoclonal and polyclonal antibodies as differentiation markers, we have found that the eight photoreceptors of the Drosophila ommatidium differentiate in a fixed sequence. The foundation photoreceptor, R8, expresses neural antigens first. The paired photoreceptors R2/5 are next to express, followed by the pair R3/4, followed by the pair R1/6; R7 is the final photoreceptor to differentiate. From previous studies it is known that Drosophila photoreceptors use local, positional cues to select their identities. Together with the morphological picture of ommatidial development, the sequential order of photoreceptor differentiation demonstrated here suggests that these cues may be encoded in the particular combination of cells an undetermined cell finds itself in contact with.

Myosin V, Rab11, and dRip11 direct apical secretion and cellular morphogenesis in developing Drosophila photoreceptors

Sensory neuron terminal differentiation tasks apical secretory transport with delivery of abundant biosynthetic traffic to the growing sensory membrane. We recently showed Drosophila Rab11 is essential for rhodopsin transport in developing photoreceptors and asked here if myosin V and the Drosophila Rab11 interacting protein, dRip11, also participate in secretory transport. Reduction of either protein impaired rhodopsin transport, stunting rhabdomere growth and promoting accumulation of cytoplasmic rhodopsin. MyoV-reduced photoreceptors also developed ectopic rhabdomeres inappropriately located in basolateral membrane, indicating a role for MyoV in photoreceptor polarity. Binary yeast two hybrids and in vitro protein–protein interaction predict a ternary complex assembled by independent dRip11 and MyoV binding to Rab11. We propose this complex delivers morphogenic secretory traffic along polarized actin filaments of the subcortical terminal web to the exocytic plasma membrane target, the rhabdomere base. A protein trio conserved across eukaryotes thus mediates normal, in vivo sensory neuron morphogenesis.

Arrestin1 mediates light-dependent rhodopsin endocytosis and cell survival

BACKGROUND

Arrestins are pivotal, multifunctional organizers of cell responses to GPCR stimulation, including cell survival and cell death. In Drosophila norpA and rdgC mutants, endocytosis of abnormally stable complexes of rhodopsin (Rh1) and fly photoreceptor Arrestin2 (Arr2) triggers cell death, implicating Rh1/Arr2-bearing endosomes in pro-cell death signaling, potentially via arrestin-mediated GPCR activation of effector kinase pathways. In order to further investigate arrestin function in photoreceptor physiology and survival, we studied Arr2's partner photoreceptor arrestin, Arr1, in developing and adult Drosophila compound eyes.

RESULTS

We report that Arr1, but not Arr2, is essential for normal, light-induced rhodopsin endocytosis. Also distinct from Arr2, Arr1 is essential for light-independent photoreceptor survival. Photoreceptor cell death caused by loss of Arr1 is strongly suppressed by coordinate loss of Arr2. We further find that Rh1 C-terminal phosphorylation is essential for light-induced endocytosis and also for translocation of Arr1, but not Arr2, from dark-adapted photoreceptor cytoplasm to photosensory membrane rhabdomeres. In contrast to a previous report, we do not find a requirement for photoreceptor myosin kinase NINAC in Arr1 or Arr2 translocation.

CONCLUSIONS

The two Drosophila photoreceptor arrestins mediate distinct and essential cell pathways downstream of rhodopsin activation. We propose that Arr1 mediates an endocytotic cell-survival activity, scavenging phosphorylated rhodopsin and thereby countering toxic Arr2/Rh1 accumulation; elimination of toxic Arr2/Rh1 in double mutants could thus rescue arr1 mutant photoreceptor degeneration.

Rab11 mediates post-Golgi trafficking of rhodopsin to the photosensitive apical membrane of Drosophila photoreceptors

In developing Drosophila photoreceptors, rhodopsin is trafficked to the rhabdomere, a specialized domain within the apical membrane surface. Rab11, a small GTPase implicated in membrane traffic, immunolocalizes to the trans-Golgi network, cytoplasmic vesicles and tubules, and the base of rhabdomeres. One hour after release from the endoplasmic reticulum, rhodopsin colocalizes with Rab11 in vesicles at the base of the rhabdomere. When Rab11 activity is reduced by three different genetic procedures, rhabdomere morphogenesis is inhibited and rhodopsin-bearing vesicles proliferate within the cytosol. Rab11 activity is also essential for development of MVB endosomal compartments; this is probably a secondary consequence of impaired rhabdomere development. Furthermore, Rab11 is required for transport of TRP, another rhabdomeric protein, and for development of specialized membrane structures within Garland cells. These results establish a role for Rab11 in the post-Golgi transport of rhodopsin and of other proteins to the rhabdomeric membranes of photoreceptors, and in analogous transport processes in other cells.

Moesin contributes an essential structural role in Drosophila photoreceptor morphogenesis

Ezrin-Radixin-Moesin (ERM) family proteins organize heterogeneous sub-plasma membrane protein scaffolds that shape membranes and their physiology. In Drosophila oocytes and imaginal discs, epithelial organization, fundamental to development and physiology, is devastated by the loss of Moesin. Here, we show that Moesin is crucial for Drosophila photoreceptor morphogenesis. Beyond its requirement for retinal epithelium integrity, Moesin is essential for the proper assembly of the apical membrane skeleton that builds the photosensitive membrane, the rhabdomere. Moesin localizes to the rhabdomere base, a dynamic locus of cytoskeletal reorganization and membrane traffic. Downregulation of Moesin through RNAi or genetic loss of function profoundly disrupts the membrane cytoskeleton and apical membrane organization. We find normal levels and distribution of Moesin in photoreceptors of a Moesin mutant previously regarded as protein null, suggesting alternative interpretations for studies using this allele. Our results show an essential structural role for Moesin in photoreceptor morphology.

Crumbs, the Drosophila homologue of human CRB1/RP12, is essential for photoreceptor morphogenesis

The apical transmembrane protein Crumbs is a central regulator of epithelial apical-basal polarity in Drosophila. Loss-of-function mutations in the human homologue of Crumbs, CRB1 (RP12), cause recessive retinal dystrophies, including retinitis pigmentosa. Here we show that Crumbs and CRB1 localize to corresponding subdomains of the photoreceptor apical plasma membrane: the stalk of the Drosophila photoreceptor and the inner segment of mammalian photoreceptors. These subdomains support the morphogenesis and orientation of the photosensitive membrane organelles: rhabdomeres and outer segments, respectively. Drosophila Crumbs is required to maintain zonula adherens integrity during the rapid apical membrane expansion that builds the rhabdomere. Crumbs also regulates stalk development by stabilizing the membrane-associated spectrin cytoskeleton, a function mechanistically distinct from its role in epithelial apical-basal polarity. We propose that Crumbs is a central component of a molecular scaffold that controls zonula adherens assembly and defines the stalk as an apical membrane subdomain. Defects in such scaffolds may contribute to human CRB1-related retinal dystrophies.

Eyes closed, aDrosophilap47 homolog, is essential for photoreceptor morphogenesis

Starting with a mutation impacting photoreceptor morphogenesis, we identify here a Drosophila gene, eyes closed (eyc), as a fly homolog of p47, a protein co-factor of the p97 ATPase implicated in membrane fusion. Temporal misexpression of Eyc during rhabdomere extension early in pupal life results in inappropriate retention of normally transient adhesions between developing rhabdomeres. Later Eyc misexpression results in endoplasmic reticulum proliferation and inhibits rhodopsin transport to the developing photosensitive membrane. Loss of Eyc function results in a lethal failure of nuclear envelope assembly in early zygotic divisions. Phenotypes resulting from eyc mutations provide the first in vivo evidence for a role for p47 in membrane biogenesis.

Movies available on-line

Rescue of photoreceptor degeneration in rhodopsin-null Drosophila mutants by activated Rac1

Rhodopsin is essential for photoreceptor morphogenesis; photoreceptors lacking rhodopsin degenerate in humans, mice, and Drosophila. Here we report that transgenic expression of a dominant-active Drosophila Rho guanosine triphosphatase, Drac1, rescued photoreceptor morphogenesis in rhodopsin-null mutants; expression of dominant-negative Drac1 resulted in a phenotype similar to that seen in rhodopsin-null mutants. Drac1 was localized in a specialization of the photoreceptor cortical actin cytoskeleton, which was lost in rhodopsin-null mutants. Thus, rhodopsin appears to organize the actin cytoskeleton through Drac1, contributing a structural support essential for photoreceptor morphogenesis.

Rhodopsin replacement rescues photoreceptor structure during a critical developmental window

Rhodopsin is essential for normal photoreceptor development in Drosophila (O'Tousa et al., 1989; Leonard et al., 1992; Kumar and Ready, 1995) and in mice (Humphries et al., 1997). Here we report studies in which a rhodopsin transgene is expressed at restricted stages during the development of Drosophila photoreceptors otherwise lacking rhodopsin. Substantial rescue of normal photoreceptor structure and physiology is effected by rhodopsin expression during the time of the normal onset of rhodopsin synthesis. Expression shortly before or after this critical period does not rescue these deficits. There is a critical developmental period in which rhodopsin plays its key role in photoreceptor morphogenesis.

Glued participates in distinct microtubule-based activities in Drosophila eye development

A C-terminal truncation of Glued, the Drosophila homolog of the cytoplasmic dynein activating protein, dynactin, results in a severe and complex retinal phenotype, including a roughening of the facet array, malformation of the photosensitive rhabdomeres, and a general deficit and disorder of retinal cells. We have characterized the developmental phenotype in Glued1 and found defects in multiple stages of eye development, including mitosis, nuclear migration, cell fate determination, rhabdomere morphogenesis and cell death. Transgenic flies that express dominant negative Glued under heat-shock control reproduce distinct features of the original Glued1 phenotype depending on the stage of development. The multiple phenotypes effected by truncated Glued point to the multiple roles served by dynactin/dynein during eye development.

Rhodopsin plays an essential structural role in Drosophila photoreceptor development

Null mutations of the Drosophila Rh1 rhodopsin gene, ninaE, result in developmental defects in the photosensitive membranes, the rhabdomeres, of compound eye photoreceptors R1-R6. In normal flies, Rh1 expression begins at about 78% of pupal life. At approximately 90% of pupal life, a specialized catacomb-like membrane architecture develops at the base of normal rhabdomeres. In ninaE null mutants, these catacombs do not form and developing rhabdomere membrane involutes into the cell as curtains of apposed plasma membrane. A filamentous cytoskeletal complex that includes F-actin and the unconventional myosin, NINAC, decorates the cytoplasmic surface of these curtains.

Integrins and the development of three-dimensional structure in the Drosophila compound eye

A stereotyped, three-dimensional network of cell-cell contacts mediated by adherens junctions and cell-extracellular matrix contacts mediated by focal adhesions defines the architecture of the Drosophila ommatidium. Developmental reconstruction shows that this network is built in an incremental and generally conservative sequence; contacts established early in eye development typically persist into adulthood. Reconstructions show that photoreceptor apical surfaces are involuted into the retinal epithelium and are subsequently elaborated to form the photosensitive rhabdomeres. Rhabdomeres become aligned to the ommatidial optical axis via their anchorage to the retinal floor at the cone cell plate, a specialized nexus of cell-cell and cell-extracellular matrix contacts. Parallel reconstructions of retinal development in integrin mutants show that several eye phenotypes trace their origin to the structural failure of the cone cell plate.

Degeneration of photoreceptors in rhodopsin mutants of Drosophila

Five different, well-characterized mutants of the R1-6 rhodopsin gene (ninaE), which corresponds to the rod opsin gene of vertebrates, have been examined morphologically as a function of age (up to 9 weeks) to determine whether or not the photoreceptors degenerate and to assess the pattern of degeneration. Structural deterioration of R1-6 photoreceptors with age has been found in all five mutants. The structural pattern of degeneration is similar in the five mutants, but the time course of degeneration is allele dependent and varies greatly among the five, with the strongest alleles causing the fastest degeneration. The degeneration appears to be independent of either the illumination cycle to which the animals are exposed or the presence of screening pigments in the eye. Although the degeneration first appears in R1-6 photoreceptors, eventually R7/8 photoreceptors, which correspond to cones of vertebrates, are also affected. In many of these mutants, striking proliferations of membrane processes have been observed in the subrhabdomeric region of R1-6 photoreceptors. It is hypothesized that (1) this accumulation of membranes may be caused by the failure of newly synthesized membranes that are inserted into the base of microvilli to be assembled into R1-6 rhabdomeres and (2) this failure may be caused by the extremely low concentration of normal R1-6 rhodopsin in the ninaE mutants.

Cell death in normal and rough eye mutants of Drosophila

The regular, reiterated cellular pattern of the Drosophila compound eye makes it a sensitive amplifier of defects in cell death. Quantitative and histological methods reveal a phase of cell death between 35 and 50 h of development which removes between 2 and 3 surplus cells per ommatidium. The timing of this epoch is consistent with cell death as the last fate to be specified in the progressive sequence of cell fates that build the ommatidium. An ultrastructural survey of cell death suggests dying cells in the fly eye have similarities as well as differences with standard descriptions of programmed cell death. A failure of cell death to remove surplus cells disorganizes the retinal lattice. A screen of rough eye mutants identifies two genes, roughest and echinus, required for the normal elimination of cells from the retinal epithelium. The use of an enhancer trap as a cell lineage marker shows that the cone cells, like other retinal cells, are not clonally related to each other or to their neighbors.

The beginning of pattern formation in the Drosophila compound eye: the morphogenetic furrow and the second mitotic wave

Events in the morphogenetic furrow set the stage for all subsequent compound eye development in Drosophila. The periodic pattern of the adult eye begins in the furrow with the spaced initiation of ommatidial rudiments, the preclusters. A wave of mitosis closely follows the furrow. A cell-by-cell analysis reveals details of these events. Early stages of ommatidial assembly can be resolved using a lead sulfide stain. Overt ommatidial organization begins in the morphogenetic furrow as cells gather into periodically spaced concentric aggregates. A stereotyped sequence of cell rearrangements converts these aggregates into preclusters. In the furrow, new rows of ommatidia are initiated at the equator and grow as new clusters are added to the peripheral ends. Mitotic labeling using BrdU feeds shows that all cells not incorporated into a precluster divide. BrdU injections show that cells divide roughly simultaneously between two adjacent rows of ommatidia.

The emergence of order in the Drosophila pupal retina

During pupation, long-range order is imposed on the autonomously developing ommatidia which compose the Drosophila eye. To accomplish this, eight additional cell types arise: the primary, secondary, and tertiary pigment cells, and the four cells that form the bristle. These cells form an interweaving lattice between ommatidia. The lattice is refined when excess cells are removed to bring neighboring ommatidia into register. Recent evidence suggests that in larval development, local contacts direct cell fate. The same appears to be true during pupal development: the contacts a cell makes predict the cell type it will become. Cells which contact the anterior or posterior cone cells in an ommatidium invariably become primary pigment cells. Cells which contact primary pigment cells from different ommatidia become secondary and tertiary pigment cells. Bristle development is in several ways distinct from ommatidial development. The four cells of each bristle group appear to be immediate descendents of a single founder cell. During their early differentiation, they do not make stereotyped contacts with surrounding ommatidial cells, but do make particular contacts within the bristle group. And unlike the surrounding ommatidia, differentiation of the bristles radiates from the center of the eye to the edges. As cells are removed during two stages of programmed cell death, the bristles are brought into their final position. When all cells in the lattice have achieved their final position, a second stage of retinal development begins as structures specific to each cell type are produced. This paper follows these various stages of pupal development, and suggests how local cell-cell contacts may produce the cells needed for a functional retina.

Notch is required for successive cell decisions in the developing Drosophila retina

Mutations in the Notch locus affect a variety of developmental decisions in Drosophila. In this paper, we examine the role of Notch in the developing retina. We reduced Notch activity at successive intervals during development of the retina, and then examined the effect on individual cells. When Notch activity was reduced, cells responded by selecting inappropriate developmental pathways. We found that all cell types appear to require Notch when establishing their fate. To examine further Notch's role in eye development, we examined two alleles of Notch--split and facet-glossy. split flies show defects in the initial clustering of photoreceptors, whereas the defects in facet-glossy flies are due to the misrouting of presumptive primary pigment cells into the secondary pigment cell pathway. Our results suggest that Notch plays a permissive role in the cell-cell interactions used to assemble the eye.

A multifaceted approach to neural development

The developing compound eye of the fruit fly, Drosophila offers notable advantages for a study of neural differentiation. It is a sensitive amplifier of a simple neural module; each eye is an approximately 700-fold repetition of the unit eye or ommatidium, which is a precise, stereotyped assembly of photoreceptors and accessory cells. The eye develops in a monolayer epithelium, which greatly reduces the complexities of cell-cell interactions often encountered in CNS development, and has permitted a detailed, cell-by-cell description of cell behavior during ommatidial development. Finally, the foundation of fly genetics permits a mutational analysis of eye development and the advanced molecular genetics of Drosophila allows close scrutiny of genes of interest. A recent convergence of cellular, genetic and molecular studies of ommatidial development suggests a model for neural differentiation in the fly eye.

Associative learning modifies two behaviors in the leech, Hirudo medicinalis

We report that 2 behaviors, stepping and shortening, are modified by associative learning in the leech, Hirudo medicinalis. Experiment 1 explored conditioning of the "stepping" response. Paired presentations of touch to the medial dorsal surface of the leech and shock to the tail of the leech resulted in the development of stepping to the touch. Leeches in control groups experiencing the CS alone, US alone, or explicitly unpaired presentations of the CS and US did not. In experiments 2-4, classical conditioning explored conditioning of the touch-elicited shortening reflex. We found that the reflex was enhanced following paired CS-US presentations but not following CS alone, US alone, or explicitly unpaired presentations of the stimuli. Moreover, the learning was extinguished following 15 unreinforced presentations of the CS but was retained for at least 24 hr without extinction training. Moreover, the associative effect was not evident when the CS and US were presented in a backward relationship. That is, no learning was observed when the US preceded the CS. Lastly, the hand-held stimuli were replaced with implanted electrodes. Using a 3 V pulse that mimicked the touch stimulus (CS), we found that paired CS-US presentations produced a significant enhancement in the shortening reflex. Again, no enhancement was observed following unpaired CS, US presentations.

Ommatidial development in Drosophila eye disc fragments

We have tested the hypothesis that the leading edge of the growing Drosophila compound eye acts as a template that organizes unpatterned cells of the retinal epithelium into the accurate cellular mosaic of the eye. Unpatterned fragments of the epithelium, not containing the leading edge of the growing field, were transplanted into larval hosts. After hosts pupated, the implants were recovered; most contained ommatidia, demonstrating that the leading edge of the growing eye pattern is not required for its propagation. In a second set of experiments, implants were recovered before hosts pupated and examined for ommatidia using a monoclonal antibody. These implants likewise differentiated ommatidia and the temporal progress of retinal development in the implants mirrored that of normal development. A schedule of ommatidial development thus appears to be mapped onto the retinal epithelium in advance of the leading edge.

Sevenless: A Cell-Specific Homeotic Mutation of the Drosophila Eye

Each ommatidium in the compound eye of the Drosophila mutant sevenless lacks photoreceptor number seven (R7) from the normal ommatidial complement of eight photoreceptors. A comparison of mutant and normal development reveals that this deficit is caused by the cell-specific transformation of the cell normally fated to produce R7 into a lens-secreting accessory cell, a cone cell.

Membrane properties and selective connexions of identified leech neurones in culture

1. Individual, identified neurones, dissected from the central nervous system of the leech and maintained in culture for several weeks, sprouted processes and formed synaptic connexions.2. The action potentials of isolated touch (T), pressure (P), nociceptive (N) cells and Retzius cells resembled those of their counterparts in situ, enabling them to be recognized unambiguously. Their input resistances were approximately 4 times greater than those of corresponding cells within the animal. In T, P and N cells trains of impulses were followed by a pronounced after-hyperpolarization, as in the animal.3. In certain cells, notably the L motoneurones, membrane properties became altered in culture. The current-voltage relation showed novel rectification and action potentials became much larger.4. Numerous neurites often extended for hundreds of micrometres from isolated neurones and ended in typical growth cones. Electron micrographs revealed that many fine axons were braided together to form thicker fascicles. Frequently, the processes were orientated between two neighbouring cells rather than at random. The fine structure of the cytoplasm, nucleus and organelles in cultured cells resembled those of their counterparts in situ. The glial cell that normally surrounds the neurones was, however, absent.5. Pairs of Retzius cells in culture usually became coupled electrically after about 6 days. Similarly L motoneurones became coupled in vitro. These junctions allowed current to pass in both directions and resembled those seen in the animal.6. Selective connexions were made by certain types of cells. Thus, P sensory neurones did not become coupled with Retzius cells but did develop electrical connexions with L motoneurones, as in the animal.7. Novel synaptic interactions not obvious in the animal could appear in culture. Retzius and L cells became electrically coupled and, in some instances where electrical coupling between Retzius cells failed to develop, chemically mediated inhibitory potentials became apparent.8. Isolated, identified leech neurones not only survive but regenerate processes and are capable of forming selective connexions in culture. The ability to define interactions between isolated pairs of cells offers the opportunity to explore in detail problems relating to synapse formation and cell-cell recognition.

Identified neurones isolated from leech CNS make selective connections in culture

Neurones cultured in vitro offer distinct advantages for studying how processes grow towards their targets and form synaptic connections. In contrast to the complex events occurring during the development of the nervous system, synapse formation in culture can be analysed in a few neurones at a time and under controlled conditions. We have now dissected out and cultured single identified neurones from the central nervous system (CNS) of the adult leech. Various types of sensory cells, motor cells, and interneurones can be identified in leech ganglia--each with a stereotyped set of properties, including: (1) the electrical characteristics of its membrane, (2) the arborisation of its branches and the morphology of its terminals and (3) the pattern of connections it makes with other identified neurones, skin or muscle. Thus, cultured cells can be compared in detail with their counterparts in situ. We have found that isolated cells survive for several weeks, maintain their membrane properties, sprout and form selective connections.