Presenilin

DEVELOPMENTAL BIOLOGY

Embryonic

High levels of Presenilin mRNA are present during oogenesis in the nurse cells and the developing oocyte. Psn is also uniformly distributed in blastoderm stage embryos; this probably represents maternal mRNAs which have accumulated during oogenesis. When the segmentation of the embryo is apparent, the Psn hybridization signal is detected in a striped pattern. In third instar larval imaginal discs, Psn expression is restricted to the eye-antennae and the leg discs. In the eye-antennal disc, Psn mRNA accumulates in the precursor of the postoccipital sensilla of the antennae and in areas of the eye disc corresponding to progenitors of the ocellus and the ocellar bristles. In the leg discs Psn is restricted to the posterior compartment, specifically in areas that give rise to the adult tibia and femur. This dynamic expression pattern for Psn suggests that it may have multiple roles during development (Boulianne, 1997 and Marfany, 1998).

Drosophila Presenilin is widely expressed throughout oogenesis, embryogenesis, and imaginal development, and generally accumulates at comparable levels in neuronal and nonneuronal tissues. During cellularization, Psn is present in all cells of the embryo, including pole cells, and appears to be enriched near cell membranes. During gastrulation, Psn expression is fairly ubiquitous but higher levels are apparent in ectoderm relative to mesoderm. Double immunolabeling with Notch antibodies reveals that Presenilin and Notch are coexpressed in many tissues throughout Drosophila development and display partially overlapping subcellular localizations, supporting a possible functional link between Presenilin and Notch (Ye, 1998).

To examine the processing as well as the cellular and subcellular distribution of Psn during development, polyclonal antibodies were raised to the N- and C-terminal domains of Psn. Psn is found to be ubiquitously expressed throughout development. The high levels of expression observed in 0-2 hr embryos appear to result from large maternal contributions. In wild-type extracts the majority of endogenous Psn protein is proteolytically cleaved to an ~25 kDa N-terminal fragment and an ~35 kDa C-terminal fragment. However, full-length Psn protein (~60 kDa) can be detected in transgenic flies that overexpress Psn from a heatshock promoter or by the GAL4/UAS system.. The fact that presenilins are processed in worms, flies, and vertebrates suggests that the cellular machinery involved in this processing must be highly conserved between species. Using immunocytochemical techniques, Psn is widely distributed during development, although some tissues, including the larval CNS, express higher levels of Psn. Within the CNS, Psn is mainly found in the axons and cell bodies of neurons and primarily within the cytoplasm (Guo, 1999).

Larval and Adult

During oogenesis, Psn is expressed mainly in nurse cells and follicle cells throughout oogenesis. Vesicular staining is seen throughout the anterior germarium region of each ovariole and is limited to nurse cells and follicle cells of late-stage egg chambers. Psn protein is hardly detected within the oocyte in a stage 11 egg chamber: this may simply reflect poor antibody penetration of this tissue. High levels of Psn are seen near membranes of the follicle cells and nurse cells, whereas lower levels are detected in the cytoplasm of the nurse cells. In nurse cells, membrane staining appears to be confined to the apical membranes and pronounced Psn accumulation is seen around the nuclei (Ye, 1998).

Psn is primarily localized in the cytoplasm of immature photoreceptor neurons and is preferentially localized near the apical membranes of the imaginal columnar epithelial cells. Psn localization at the apical membranes of imaginal wing discs results in prominently stained disc surface folds in histological preparations. Psn is expressed at high levels in cells that give rise to the interommatidial bristles of the adult eye. Psn is expressed at a very high level in certain regions of the larval and pupal brain. Psn expression is enhanced in part of the optic lobe anlagen, presumably in undifferentiated neurons that give rise to optic lobe structures of the adult brain. Intense Psn expression in the developing medulla, an optic lobe structure, is readily visible as an intensely labeled crescent-shaped structure proximal to the pupal eye disc. In most cases, Psn colocalizes with Notch, but there exceptions. For example, although both proteins are found near the cell membranes of most tissues, the Psn level in the cytoplasm is relatively higher than Notch, especially in larval discs and the pupal retina. Largely nonoverlapping vesicular staining for both proteins is seen in the larval eye disc. In general, Notch is more highly restricted to the apical cell region and more tightly associated with cell membranes, clearly outlining the apical profiles of the cells when visualized by confocal optical sectioning. In contrast, Psn is detected at more equivalent levels both in the cytoplasm and at the cell membrane, such that Psn immunostaining appears to be more diffuse and graded than Notch in the same cell types (Ye, 1998).

Effects of Mutation or Deletion

Clones of Psn minus cells in the wing cause scalloping and vein thickening, reflecting failures in well defined signaling events that depend on Notch, such as the specification of Wingless-secreting margin cells along the dorso-ventral compartment boundary and the lateral inhibition of vein differentiation (Struhl, 1999).

Genetic characterization of cytological region 77A-D harboring the Presenilin gene of Drosophila

A systematic lethal mutagenesis of the genomic region uncovered by Df(3L)rdgC-co2 (cytological interval 77A-D) has been carried out to isolate mutations in the single known Presenilin (Psn) gene of Drosophila. Because this segment of chromosome III has not previously been systematically characterized, inter se complementation testing of newly recovered mutants was carried out. A total of 79 lethal mutations were isolated, representing at least 17 lethal complementation groups, including one corresponding to the Psn gene. Fine structure mapping of the genomic region surrounding the Psn transcription unit by transgenic rescue experiments allowed the localization of two of the essential loci together with Psn within an ~12-kb genomic DNA region. One of these loci, located 3' to Psn, encodes a Drosophila protein related to the yeast 60S ribosomal protein L10 precursor. It was also determined which of the newly recovered lethal mutant groups corresponded to previously isolated lethal P-element insertions, lethal inversion breakpoints, and/or lethal polo gene mutants. Point mutations were identified in all five recovered Psn alleles, one of which results in a single amino acid substitution G-E at a conserved residue in the C-terminal cytoplasmic tail of the protein, suggesting an important functional role for this Presenilin C-terminal domain. In addition, some viable mutations were recovered in the screen, including new alleles of the clipped and inturned loci (Lukinova, 1999).

Sequence analysis confirms that all five Psn mutants bear lesions in the Psn gene as deduced from the genomic DNA rescue experiments. Psn^C4, Psn^S3, Psn^I2, and Psn^K2 are predicted to encode prematurely truncated Presenilins, in agreement with genetic studies suggesting that these mutant Psn alleles are likely to represent strong or complete loss-of-function mutants. Psn^B3 is predicted to cause an amino acid substitution at a conserved residue of the C-terminal region of Presenilin, directing attention to this portion of the molecule. Most Alzheimer's disease-associated mutations in the human Presenilins occur in either transmembrane domain 2 (TM2) or in an N-terminal stretch of the large hydrophilic loop between TM6 and TM7; only one recently discovered mutation maps to the C-terminal tail of Psn1. In the case of Psn^B3, a single amino acid substitution (G516E) at a conserved position 26 residues before the C-terminus results in a complete or nearly complete loss of Presenilin function. The affected glycine residue presumably plays an important role in normal Presenilin function, perhaps as part of a protein-protein interaction site or by contributing to the conformation of a functionally important C-terminal Presenilin domain. The significance of the C-terminal tail of Presenilin family members for the normal biological functions of these proteins will require further genetic and molecular analysis (Lukinova, 1999 and references therein).

Apoptotic activities of wild-type and Alzheimer's disease-related mutant Presenilins in Drosophila melanogaster

Larvae mutant for Psn secrete a pupal case and complete the last stages of larval development but do not form any adult structures: instead, they collapse into a homogeneous oily mass within the pupal case. This zygotic lethal null phenotype closely resembles that observed for Suppressor of Hairless [Su(H)], which encodes an effector of the Notch signaling pathway. (Ye, 1999).

Formation of the dorsal/ventral (D/V) boundary along the presumptive wing margin requires Notch activity, and may be visualized by the expression of specific reporter genes in the margin zone of the larval wing disc. Expression of cut-lacZ along the wing margin is abolished in Psn mutant wing discs. Similarly, margin zone expression of a wingless-lacZ construct and a vestigial (vg) D/V enhancer-lacZ transgene is absent in the Psn mutant wing disc, whereas lacZ expression outside the wing pouch persists in the mutants. Expression of a vg quandrant enhancer-lacZ transgene, which is specific to the non-margin zones of the wing pouch, is also completely absent in the Psn mutant wing disc. Expression of these wingless and vestigial gene reporters along the wing margin and in the pouch region is directly dependent on Notch signaling activity. The vg D/V enhancer contains a critical binding site for the Su(H) protein. The absence of margin structures in the Psn mutant wing disc, together with the overall reduction in size of the mutant wing pouch region, is reminscent of conditional Notch mutant and Su(H) null mutant phenotypes (Ye, 1999).

Histochemical characterization of Psn mutant wing discs using the early SOP markers achaete-lacZ and scabrous-lacZ reveals clusters of supernumerary SOP cells arising at certain normal SOP locations, as seen in typical neurogenic mutants. Lateral inhibition within these proneural cell clusters is severely impaired in Psn mutant discs, resulting in enlarged SOP territories and increased proneural achaete-lacZ and scabrous-lacZ activity. Embryos derived from females that lack both maternal and zygotic Psn activity display a Notch-like lethal hyperplasia of the embryonic nervous system, and the maternal phenotype is only weakly rescued by wild-type paternal zygotic Psn function (Ye, 1999).

Mutant human presenilins cause early-onset familial Alzheimer's disease and render cells susceptible to apoptosis in cultured cell models. Loss of presenilin function in Drosophila increases levels of apoptosis in developing tissues. Moreover, overexpression of presenilin causes apoptotic and neurogenic phenotypes resembling those of Presenilin loss-of-function mutants, suggesting that presenilin exerts a dominant negative effect when expressed at high levels. In Drosophila S2 cells, Psn overexpression leads to reduced Notch receptor synthesis affecting levels of the intact ~300-kD precursor and its ~120-kD processed COOH-terminal derivatives. Presenilin-induced apoptosis is cell autonomous and can be blocked by constitutive Notch activation, suggesting that the increased cell death is due to a developmental mechanism that eliminates improperly specified cell types. A genetic model is described in which the apoptotic activities of wild-type and mutant presenilins can be assessed, and it is found that Alzheimer's disease-linked mutant presenilins are less effective at inducing apoptosis than wild-type presenilin (Ye, 1999).

Although the Alzheimer's disease-linked Psn mutations (Psn referring to human presenilin) are not complete loss-of-function alleles, it is not clear if they are partial loss-of-function or gain-of-function mutations. All Alzheimer's disease-linked mutations occur at amino acid residues that are conserved in the Drosophila Psn protein, allowing these mutations to be introduced into the fly protein and their apoptotic effects assessed in transgenic animals. Four missense mutations, N141I, M146V, L235P, and E280A were tested in this manner. The N141I mutant of PS2 has been shown to induce more apoptosis in certain cell types, compared with wild-type PS2, and may thus represent a gain-of-function mutation. Both the M146V and E280A mutants of PS1 increase the ratio of Abeta42/Abeta40 by increasing the level of the more neurotoxic and amyloidogenic Abeta42 cleavage product of amyloid precursor protein. The L235P mutation has been identified in a family with an onset of Alzheimer's disease as early as age 29, and may therefore represent a particularly severe mutant form of presenilin. In addition, a truncated Psn protein named D-ALG3, consisting of the most COOH-terminal 100 amino acids, two loop constructs consisting of either the long or short variable hydrophilic loop, and transmembrane domain 7 (TM7) following the loop were also included in this analysis. These constructs were chosen because D-ALG3 resembles truncated mammalian PS proteins that confer resistance to cell death in PC12 cells, and the loop may be a cytoplasmic domain required for protein-protein interactions (Ye, 1999).

The apoptotic activities of mutant presenilins were assessed by the following criteria: their ability to generate a rough eye phenotype in flies bearing two copies of each mutant transgene and GMR-GAL4; their ability to modify the eye phenotype of GMR-GAL4, 2X UAS-Psn+14 flies; and their ability to modify the eye phenotype of flies expressing the Drosophila death-domain protein Reaper under GMR promoter control. At least five independent transgenic lines were analyzed for each construct. Only the M146V substitution produced a rough eye phenotype in two independent transgenic lines, out of the five analyzed. The remaining mutations and Psn fragments fail to produce rough eye phenotypes when expressed under GMR-GAL4 control, but all four Alzheimer's disease-associated mutants enhance the rough eye phenotype of GMR-GAL4, 2X UAS-Psn+14 flies. M146V and N141I also enhance the phenotype of flies bearing GMR-reaper. Although the ALG-3 fragment of PS2 inhibits programmed cell death in PC12 cells, the equivalent segment of Drosophila Psn, D-ALG3, instead weakly enhances the GMR-GAL4, 2X UAS-Psn+14 phenotype, indicating that it may possess weak apoptotic activity. The two loop variant fragments possess no modifying activity in these transgenic assays (Ye, 1999).

To gain further insight into the mechanism of Psn-induced cell death, an attempt was made to suppress this apoptosis by coexpression of Drosophila cell death inhibitors (DIAP1 or DIAP2) or the baculoviral survival factor p35. DIAP1 and DIAP2 are Drosophila homologs of baculoviral inhibitor of apoptosis proteins (IAPs), and can block programmed cell death induced by proapoptotic factors or mutations. Baculoviral p35 protein inhibits caspases and thus blocks apoptosis in many species. In the presence of any of these antiapoptotic proteins, the Psn-induced rough eye phenotype is largely suppressed, as revealed by the more regular external eye surface, the more normal trapezoidal pattern and orientation of R1-6 photoreceptor cells, and a restoration of the pigment cell lattice between ommatidia. The observation that Psn-induced rough eye phenotypes are suppressed by coexpressing either DIAP1, DIAP2, or p35 confirms that increased cell death is the main cause of the rough eye phenotype. Psn-induced programmed cell death may thus be mediated by the conserved caspase pathway of apoptosis or, alternatively, it may be circumvented by inhibition of this pathway (Ye, 1999).

The ability of wild-type and, to a lesser extent, mutant forms of Drosophila Psn to induce low levels of apoptosis similar to the apoptotic levels seen in developing imaginal tissues of Psn loss-of-function mutants suggests that the apoptotic effects of Psn may be a secondary consequence of reduced or dominant-negative Psn activity during developmental patterning. In C. elegans and mice, presenilin proteins have been shown to facilitate Notch signaling, and worms or mice lacking presenilin activity display typical Notch or lin-12/glp-1 loss-of-function phenotypes. Similarly, flies lacking functional Psn gene activity exhibit embryonic neurogenic phenotypes and imaginal disc phenotypes that are characteristic of impaired Notch signaling. Moreover, elevated levels of apoptosis have been noted previously in wing imaginal discs of flies having the partial loss-of-function heteroallelic Notch genotype N^ts/N^55e11; neur^A101/+. These observations raise the possibility that the apoptotic effects of Psn overexpression may be due to a primary interference with Notch signaling, followed by elimination of cells that have not adopted their proper cell fate by a normal corrective mechanism of developmentally controlled apoptosis (Ye, 1999).

The in vivo genetic model for Psn-mediated apoptosis allowed for an examination of the potential involvement of Notch signaling in the apoptotic response, an important issue that has not been possible to assess in the widely used mammalian cell culture assays for Psn-induced apoptosis. First, it was determined if the UAS-Psn constructs that cause apoptosis in the Drosophila eye when driven by GMR-GAL4 are able to produce Notch pathway phenotypes in other tissues when expressed using suitable GAL4 driver constructs. Several GAL4 driver lines that are active in the wing and cuticle anlagen are indeed capable of producing adult Notch-like phenotypes in the wing blade and thorax, including wing margin notching, vein thickening, ectopic wing margin bristles, ectopic wing vein campaniform sensilla, ectopic thoracic macrochaetae, and missing thoracic microchaetae. These phenotypes are consistent with the notion that Psn overexpression leads to dominant-negative effects, since Psn loss-of-function mutants exhibit similar Notch-like phenotypes. To determine if Psn-induced apoptosis might be an indirect effect of reduced Notch activity, apoptosis was analyzed in imaginal wing discs of the conditional temperature-sensitive Notch mutant N^ts1. Increased levels of programmed cell death are spatiotemporally correlated with progressive loss of Notch activity as visualized by reduced wing-pouch-specific expression of the Notch target gene reporter vg(quadrant enhancer)-lacZ and expansion of proneural cell clusters positive for ac-lacZ expression in the presumptive notum region, suggesting that developmental patterning defects caused by reducing Notch activity directly lead to elimination of affected cells by apoptosis. The next step involved testing whether reduction in the dosage of the wild-type Notch gene or coexpression of constitutively activated Notch is able to suppress or enhance Psn-related apoptotic phenotypes. The rough eye phenotype of GMR-GAL4, 2X UAS-Psn+14 is strongly enhanced in an N^54l9 mutant background bearing only one functional copy of the Notch gene. Apoptosis caused by either Psn overexpression or removal of Psn gene function is also dramatically suppressed by coexpression of constitutively activated Notch in the retina. These studies show that genetic removal or overexpression of Psn in developing Drosophila tissues is able to induce Notch-like phenotypes, as well as apoptosis, and that when genetic methods are used to compensate for effects on Notch signaling, the levels of apoptosis are dramatically reduced. These results, together with the observed correlation between impaired Notch signaling and high levels of developmental apoptosis, offer a potential explanation of Psn-mediated apoptosis as a developmental response to a primary failure in cellular patterning events requiring Psn activity for proper Notch synthesis or signaling (Ye, 1999).

To further elucidate the molecular mechanism underlying the Psn overexpression phenotypes, Notch processing and trafficking was examined in S2 cells cotransfected with Psn and Notch. When expressed before Notch induction, wild-type Psn and the various Psn mutants lead to reduced Notch protein levels, affecting both the full-length and the processed COOH-terminal fragments of Notch. In agreement with genetic results, the biochemical effect appears to be more pronounced for wild-type Psn than for the mutant forms. The expression of a nonmembrane-bound control protein, Suppressor of Hairless, is not affected by either the wild-type or the mutant Psn proteins. The effect on Notch synthesis is unlikely to be due to increased protein degradation, because ectopic expression of the same Psn construct in S2 cells after Notch induction has no detectable effect on Notch protein levels. In addition, live cell surface immunostainings reveal that Notch protein is trafficked and inserted into the cell membrane normally in spite of the reduced protein levels caused by Psn overexpression (Ye, 1999).

Thus, the proapoptotic activity of different Alzheimer's disease-linked mutant variants has been assessed and it has been found that these variants possess less apoptotic activity than wild-type presenilin, consistent with the idea that this class of presenilin mutations represents partial loss-of-function mutations in nematodes and transgenic mice. In contrast to previous studies showing that Alzheimer's disease-associated missense mutations enhance the ability of PS1 and PS2 to induce cell death and that the Psn COOH-terminal ALG-3 fragment blocks the proapoptotic activity of PS2 , the corresponding missense mutants of fly Psn induce less cell death than wild-type Psn, and the Drosophila Psn COOH-terminal D-ALG3 fragment appears to induce apoptosis only weakly in vivo. In view of the dominant-negative role that is suggested for overexpressed Psn, the reduced apoptotic activity of these mutant variants may reflect a decreased ability to interfere with endogenous presenilin function or a more rapid clearance of misfolded mutant proteins from the ER/Golgi compartment. Analogous partial loss-of-function effects on Abeta secretion of human mutant presenilins may be relatively weak under physiological conditions, but may contribute to gradual deterioration of brain tissue during the aging process and eventually trigger disease onset when neuronal loss reaches a certain threshold (Ye, 1999).

Drosophila presenilin is required for neuronal differentiation and affects Notch subcellular localization and signaling

Recent studies have suggested a role for presenilins in the Notch-signaling pathway, but their specific function within this pathway remains unclear. The Drosophila Presenilin gene and protein have been characterized and their interaction with Notch has been studied in both mutants and transgenics. The Drosophila Psn protein is proteolytically cleaved and broadly expressed during development with the highest levels in neurons within the larval CNS. Mutations in Drosophila Psn genetically interact with Notch and result in an early pupal-lethal phenotype characterized by defects in eye and wing development and incomplete neuronal differentiation within the larval CNS. Moreover, processing of Notch in the Golgi by the furin protease is unaffected in Psn mutants and Notch is present and may even accumulate on the plasma membrane of neuroblasts in the larval CNS of Psn mutants. In contrast, overexpression of Psn in transgenics causes Notch to accumulate in the cytoplasm. Taken together, these results indicate that Drosophila Psn is required for proper neuronal differentiation and may regulate the subcellular localization of Notch proteins within cells, necessary for Notch accumulation and subsequent signaling capabilities (Guo, 1999).

To examine the function of Psn, a series of overlapping deletions was generated by imprecise P-element excision of P4, a homozygous viable P-element insertion located ~1 kb 3' from the Psn gene. Six deletions were generated, three of which remove portions of the Psn coding region (PS^W20, Dps^W6, and PS^W11). All of the deletions fail to complement one another, and none produce any detectable Psn protein as determined by Western blot analysis on extracts from mutant larvae. The largest deletion (PS^W20) deletes the Psn gene as well as portions of the lipoic acid synthase and the 50 S ribosomal protein L15 genes and is embryonic lethal. PS^W11 and PS^W6 that delete Psn and the 50S ribosomal protein L15 genes are lethal during the second larval instar, and these larvae are small and grow at a much slower rate (Guo, 1999).

In addition to the deletion mutants in Psn two EMS alleles, PS³⁰ and PS⁴⁶ have been characterized. These EMS alleles were originally identified in a lethal screen over the deficiency Df(3L)ri-79c/TM3 that uncovers the Psn locus and has break points at 77B-C;77F-78A. Both EMS lines fail to complement each other as well as the two Psn deletions tested, PS^W6 and PS^W11. To determine the molecular basis of these mutations, both PS³⁰ and PS⁴⁶ were sequenced. To date, no mutation within the coding region of PS³⁰ could be detected and alterations in surrounding regulatory sequences are being sought. A single missense mutation was found in PS⁴⁶, consisting of a proline to leucine substitution at amino acid 507 within the highly conserved C-terminal domain of Psn (Guo, 1999).

To address the function of PS during development, a test was made of the phenotype in flies of the genotype PS⁴⁶/PS^W6 or PS⁴⁶/PS^W11, that is, flies that are mutant for Psn but wild-type for lipoic acid synthase and hemizygous for the 50 S ribosomal protein L15 gene. These mutants are early pupal lethal; the third instar larvae form a pupal case, but no adult structures develop. When the phenotype of the mutants is examined during the late third larval instar stage, the mutant larvae have underdeveloped eye and wing imaginal discs. Although the eye imaginal disc forms, it fails to undergo proper neuronal differentiation. The wing imaginal disc is also smaller with the most severe defects observed in the region that will give rise to the wing blade. Furthermore, Wingless expression, which is normally detected in two domains within the developing disc, is disrupted in Psn mutant discs. In contrast, both the antennae and leg imaginal discs appeared normal. Both the lethality as well as the wing and eye phenotypes could be rescued using a Psn transgene driven by a heatshock promoter. To determine whether there was any evidence of neuronal differentiation within the eyes and optic lobes of these pupal lethal mutants, the expression of two neuronal markers, Elav and Cut, that are broadly expressed in neurons during postembryonic development, were examined. In wild-type eye-antennae imaginal discs of third larval instars, Elav is expressed in photoreceptor neurons. Cut is absent from the eye disc at this stage but is broadly expressed in the antennae disc. At later stages, Cut can be detected in cone cells and sensory organ precursors that give rise to interommatidial bristles. Wild-type discs stain positively for Elav (eye) and Cut (antennae), whereas mutant eye imaginal discs appear small and underdeveloped and fail to express Elav. Furthermore, although both Cut and Elav could be detected in the developing optic lobes, the highly organized, laminar pattern of expression that is normally found in wild-type optic lobes fails to develop in Psn mutants. This suggests that Psn is required for eye and wing imaginal disc development, and although it is not required for the development of the optic lobes, they fail to differentiate properly without it (Guo, 1999).

To examine the link between Psn and Notch, the ability of Psn mutants to interact genetically with Notch was examined. Specifically, each Psn mutant was crossed to several alleles of Notch and dosage-sensitive interactions were sought in transheterozygotes. An examination was made of the ability of the point mutant PS⁴⁶ and the deletions PS^W6, PS^W11, and PS^W20 to interact with the deficiency Df(1)N-8/In(1)dl-49, y1 Hw1 m2 g4. These interactions give rise to dominant notching of the wing blade. Transheterozygotes containing one copy of the Notch deficiency and one copy of the PS⁴⁶ mutation or a Psn deletion all exhibit an enhanced wing phenotype. In addition, both Psn alleles enhance the phenotype of N^nd-3, a loss-of-function allele of Notch that gives rise to a thickened wing vein phenotype at 29°C. Specifically, transheterozygotes between N^nd-3 and Psn alleles have enhanced thickening of wing veins and notches at the tips of the wing similar to what is observed in stronger Notch alleles. This suggests that Psn mutations reduce Notch signaling within cells. In agreement with this, Psn alleles suppress the interrupted wing vein phenotype observed in the gain-of-function N^Ax-1 allele. Finally, Psn mutants also enhance the phenotype of the Delta allele Dl[7] that is characterized by thickened wing veins and small deltas located near cross-veins. Specifically, transheterozygotes between Dl[7] and Psn mutants exhibit a greater extent of thickening and deltas at the wing veins. Taken together, these results indicate that Psn mutations reduce Notch signaling and support a role for Psn in the Notch-signaling pathway (Guo, 1999).

To gain insight into the mechanism underlying Psn and Notch interactions, immunocytochemical studies were performed to examine the subcellular distribution of Notch within the larval CNS of Psn mutants. Using an antibody that recognizes the Notch intracellular domain, it has been found that Notch is expressed at high levels within neuroblasts throughout the proliferative centers in the developing optic lobes and at somewhat lower levels in neuroblasts within the thoracic ganglia of third larval instar CNS. Within wild-type neuroblasts, Notch is distributed throughout the cytoplasm and on the plasma membrane. In contrast, the overall distribution of Notch within the optic lobes is disrupted in Psn mutants, and the protein levels are reduced in the cytoplasm and appear to be preferentially retained on the plasma membrane of specific neuroblasts. To distinguish between primary effects of Psn on Notch localization versus secondary defects caused by altered differentiation of Psn mutants, the distribution of Notch protein within neuroblasts of second instar larvae was examined. At this stage, Notch is expressed in small groups of cells that consist of a large neuroblast surrounded by smaller ganglia mother cells. Within these groups of cells Notch appears throughout the cytoplasm and at high levels at the plasma membrane in regions of contact between the neuroblast and its progeny. There is little, if any, accumulation of Notch at the plasma membrane in regions that contact the remaining surrounding cells. In contrast, Notch distribution is altered with less in the cytoplasm and more staining seen over the entire plasma membrane of neuroblasts in Psn mutants. In agreement with this observation, processing of Notch, found to give rise to a functional heterodimeric receptor on the cell surface, is unaffected in Psn mutants (Guo, 1999).

The reciprocal experiment was performed to determine whether overexpression of Psn also affects the subcellular distribution of Notch. Specifically, the GAL4/UAS system was used to target expression of Psn to specific subsets of cells within a developing tissue. Specifically, a pannier-GAL4 line was used to drive expression of UAS-Dps in regions of the wing imaginal disc that will give rise to the adult notum and subsets of neurons within the eye imaginal disc or a cut-GAL4 line that drives expression in the presumptive wing margin. Notch protein is found to specifically accumulate within Psn-expressing cells and appears to be primarily localized within the cytoplasm. In contrast, neighboring cells that do not overexpress Psn show no change in Notch expression or distribution. Notch specifically accumulates in Psn-overexpressing cells but not in the adjacent cells (internal control) that only express basal levels of Psn. Taken together, these results show Psn does not affect the ability of Notch to undergo furin-dependent cleavage within the Golgi and to accumulate on the plasma membrane. Rather, Psn may be affecting the subcellular distribution of Notch and later cleavage events leading to defects in cell signaling (Guo, 1999).

In the absence of ligand interaction, proteolytic processing of Notch does not occur. Alternatively, Psn could be required for internalization of the receptor-ligand complex and subsequent proteolytic processing because it is not known where within the cell these cleavage events occur. Although presenilin in mammals is thought to reside predominantly in the endoplasmic reticulum (ER) and Golgi, consistent with possible roles in early stages of protein processing, some evidence suggests that it may transiently reach the plasma membrane where it could participate in endocytic processes. In fact, in Drosophila, Psn has been shown to accumulate within vesicular structures throughout the cytoplasm, including the ER and Golgi, and is also found at the apical regions of cells in some tissues. A role for presenilin in protein processing or trafficking has been suggested previously on the basis of its subcellular localization and its ability to affect the processing of APP in vertebrates. APP, like Notch, encodes a single-pass transmembrane protein found at the plasma membrane that is known to undergo several cleavage steps giving rise to two secreted peptides, Abeta(40) and Abeta(42-43). Recent studies in transgenic mice have shown that presenilins can affect the processing of both Notch and APP by mechanisms that are unknown. Current models suggest that presenilins may directly cause cleavage of Notch and APP or, alternatively, affect the ability of secretases to cleave. Alternatively, presenilins may play a more general role in regulating the subcellular distribution of transmembrane proteins like APP and Notch within the cell, thereby physically altering their ability to interact with proteases necessary for their proper signaling functions (Guo, 1999 and references therein).

A screen for mutations that suppress the phenotype of Drosophila armadillo, the ß-catenin homolog identifies Pen

During development, signaling pathways coordinate cell fates and regulate the choice between cell survival or programmed cell death. The well-conserved Wingless/Wnt pathway is required for many developmental decisions in all animals. One transducer of the Wingless/Wnt signal is Armadillo/ß-catenin. Drosophila Armadillo not only transduces Wingless signal, but also acts in cell-cell adhesion via its role in the epithelial adherens junction. While many components of both the Wingless/Wnt signaling pathway and adherens junctions are known, both processes are complex, suggesting that unknown components influence signaling and junctions. A genetic modifier screen was carried out to identify some of these components by screening for mutations that can suppress the armadillo mutant phenotype. Twelve regions of the genome were identified that have this property. From these regions and from additional candidate genes tested, four genes were identified that suppress arm: dTCF, puckered, head involution defective (hid), and presenilin. The interaction with hid, a known regulator of programmed cell death, was further investigated. The data suggest that Wg signaling modulates Hid activity and that Hid regulates programmed cell death in a dose-sensitive fashion (Cox, 2000).

While evaluating the effectiveness of this screen, a variety of candidate genes, including some that map within noninteracting Deficiencies, were tested. Heterozygosity for one of these, Presenilin, strongly suppresses arm. Both ßcat and other Arm repeat proteins such as delta-catenin associate with Presenilins in vivo. The function of this interaction remains confusing. Wild-type Presenilin stabilizes ßcat; this stabilization is abrogated by missense mutations in presenilin, and presenilin missense mutant cells from mutant patients have less nuclear ßcat. These data support a role for Presenilins as positive regulators of Wnt signaling via Arm/ßcat. In contrast, it has been reported that overexpression of wild-type Presenilin destabilizes ßcat; ßcat is stabilized in both Presenilin1 null fibroblasts or if Presenilin1 mutations are overexpressed, while a Wnt-responsive promoter is downregulated by Presenilin overexpression. These data support a conclusion that is the opposite of the one above, in which wild-type Presenilins negatively regulate Wnt signaling. It has been suggested that the presenilin-ßcat complex includes cadherins, in contravention of most other data. The genetic data are most consistent with a model in which Presenilins negatively regulate Wg signaling either directly or indirectly by binding Arm/ßcat or by influencing adherens junction assembly. Clearly much work remains to differentiate between the different possible mechanisms (Cox, 2000 and references therein).

Drosophila Nicastrin is essential for the intramembranous cleavage of Notch

To determine where Nicastrin (Nct) acts in the Notch signal transduction pathway, advantage was taken of transgenic Notch constructs that bypass some of the processing steps required for signaling. (1) hs-N^FL was used as a control; it expresses full-length Notch protein under the control of a heat shock promoter. Like endogenous Notch, the protein from this construct requires both the ligand-induced S2 cleavage [this involves cleavage of the extracellular domain of Notch to produce a transient form of the receptor called NEXT (Notch extracellular truncation)] and the Psn-dependent S3 cleavage (an intramembranous cleavage to release the intracellular domain of Notch, which translocates to the nucleus, where it acts as a transcriptional activator in association with Suppressor of Hairless protein. (2) hs-N^ECN is a deletion of the extracellular domain of Notch beyond the S2 cleavage site, and therefore mimics NEXT. This protein signals independently of ligand, but still requires the S3 cleavage to release the intracellular domain of Notch from the membrane. (3) hs-N^ICD expresses the intracellular domain of the receptor and requires neither ligand, the S2 nor S3 cleavages, for signaling. The expression of hs-N^FL has no effect on the development of wild-type embryos, whereas hs-N^ECN and hs-N^ICD disrupt germ band retraction because their ligand-independent signaling overactivates the Notch pathway. To test which of these constructs requires Nct for signaling, each was expressed in embryos that lack both maternal and zygotic Nct activity. hs-N^ECN and hs-N^FL have no effect on the neurogenic phenotype of nct null embryos. In contrast, hs-N^ICD expression strongly rescues the nct phenotype: most embryos form patches of normal cuticle, and in some cases, the wild-type cuticular pattern is almost completely restored. Thus, the nct null mutation blocks Notch signaling after the S2 cleavage but before the release of the intracellular domain from the membrane, indicating that Nct is required for the S3 cleavage, as is Psn (López-Schier, 2002).

Presenilins and the human Nct associate with Notch during its passage through the secretory pathway, raising the possibility that they also function in earlier steps in Notch processing. Since the experiments above only reveal the last step at which these proteins are required, the behavior of Notch protein was followed in the follicle cells, where the downregulation of Notch in response to Delta binding can be visualized. Notch accumulates on the apical side of these cells until stage 7 of oogenesis, when Delta signals to trigger its proteolysis. Thus, Notch disappears from the apical membrane of cells that contact wild-type germline cells, whereas high levels of apical Notch persist in follicle cells that contact Delta germline clones, and this can be detected with antibodies against both the extracellular (N^ECD) and intracellular domains (N^ICD) of the protein. This reduction of apical Notch staining in response to Delta still occurs in Su(H) mutant follicle cell clones, in which Notch signaling in the nucleus is blocked. Therefore the disappearance of most Notch from the apical membrane appears to be a direct consequence of ligand-induced processing, and is not due to downregulation of the receptor in response to the activation of the signaling pathway. In nct and psn mutant follicle cells, apical N^ECD staining disappears at stage 7 as it does in wild-type, indicating that the ligand-dependent S2 cleavage and subsequent degradation of N^ECD occur normally. Unlike wild-type cells, however, mutant cells accumulate a processed form of Notch that can be stained with the anti-N^ICD antibody, and this is concentrated at the apical side of the cells and in intracellular clusters that may be endocytic vesicles. Thus, nct or psn mutant cells can transport Notch to the plasma membrane and process it to form a functional receptor that binds Delta and undergoes the S2 cleavage, indicating that both proteins are specifically required for the S3 cleavage, but not for any earlier steps in the pathway (López-Schier, 2002).

The processed form of Notch that accumulates on the apical side of nct and psn mutant follicle cells presumably corresponds to NEXT, which is the membrane-tethered product of the S2 cleavage and the substrate for the S3 cleavage. To test this possibility, Western blots of extracts from wild-type and embryos lacking maternal and zygotic Nct or Psn were probed with the anti-N^ICD antibody. In addition to the 300 kDa band that corresponds to the uncleaved form of Notch, nct and psn null embryos accumulate a 120 kDa species that is barely detectable in the wild-type extracts. This band migrates at the expected size of NEXT and N^ICD, which cannot be distinguished on these gels as they differ by only 3 kDa. This band is unlikely to be N^ICD, however, because the results above demonstrate that nct and psn block the S3 cleavage (López-Schier, 2002).

To investigate whether Nct is required for the localization or stability of Psn, nct^agro clones of varying sizes were generated in the follicular epithelium. In wild-type follicle cells, Psn protein shows a punctate distribution in the cytoplasm that may correspond to the endoplasmic reticulum, and it localizes at the cell cortex. When nct clones are analyzed early in oogenesis, the mutant cells show a normal distribution of Psn, and this is also the case for small clones in late-stage egg chambers. In contrast, very large clones in late-stage egg chambers show a strong reduction in the levels of Psn. Thus, Nct seems to be required for the long-term stability of Psn in these cells. Since small nct clones show a completely penetrant Notch loss-of-function phenotype but have normal levels of Psn, this effect of Nct on Psn stability is unlikely to cause the defect in Notch signaling in these cells (López-Schier, 2002).

During an analysis of follicle cell clones, an additional function of Nct and Psn in the organization of the submembranous spectrin cytoskeleton was discovered. In wild-type cells, ß_Heavy-spectrin (ß_H-spectrin) associates with alpha-spectrin to form tetramers that localize to the apical membrane, while the basolateral spectrin cytoskeleton is composed of ß-spectrin/alpha-spectrin complexes. In psn and nct follicle cell clones, ß_H-spectrin does not localize to the apical membrane. In contrast, ß-spectrin localization to the basolateral membrane is unaffected in mutant clones. The apical localization of ß_H-spectrin requires its association with alpha-spectrin and vice versa, and the distribution of alpha-spectrin was therefore examined in follicle cell clones. Surprisingly, both psn and nct mutant cells show an increase in the amount of cortical alpha-spectrin, with the highest levels on the apical side. alpha-spectrin therefore appears to be recruited to the apical membrane of mutant cells by a novel mechanism that does not require its association with either ß subunit, neither of which localizes apically in these cells (López-Schier, 2002).

To test whether the mislocalization of alpha- and ß_H-spectrin in nct and psn mutants is a consequence of the defect in Notch signaling, the localization of both proteins was examined in follicle cell clones of a null allele of Notch (N^55e11) and in wild-type cells that abut Delta mutant germline clones. ß_H-spectrin is recruited normally to the apical membrane in both cases, while alpha-spectrin localizes uniformly around the cell cortex at the same level as in wild-type cells. These results indicate that Psn and Nct have a novel function independent of their role in Notch signaling that somehow affects the organization of the spectrin cytoskeleton (López-Schier, 2002).

The absence of apical ß_H-spectrin in nct and psn mutant cells suggested that their apical-basal polarity might be disrupted, and the localization of a variety of other polarity markers was therefore examined. The overall polarity of the cells is unaffected in either nct or psn mutant clones, since Coracle, Neurotactin, DE-Cadherin, Armadillo (ß-catenin), alpha-catenin, and Notch itself are localized to the proper membrane domains. However, DE-Cadherin, Armadillo, and alpha-catenin accumulate to much higher levels in mutant cells than in wild-type. These three proteins are components of the Cadherin adhesion complex, and are enriched at the sites of adherens junction formation at the apical margins of the cell. Their overaccumulation in mutant cells may therefore be linked to the loss of apical ß_H-spectrin and the apical enrichment of alpha-spectrin, although the nature of this link remains unclear (López-Schier, 2002).

Previous studies have shown that almost all Psn is associated with the ER and intermediate compartment and that there is little or no protein at the plasma membrane, where APP and Notch reside. This discrepancy, which has been called the 'spatial paradox,' raises the question of where in the cell the S3 cleavage occurs. One possible solution to this paradox is suggested by the observation that small amounts of Psn can be coimmunoprecipitated with Notch at the cell surface. Both Nct and Psn have been shown to interact with Notch in the secretory pathway, and a minor fraction of the S3 protease complex could therefore be transported to the plasma membrane through binding to its future substrate. A second possibility is that the active protease resides in an intracellular compartment, and that the products of S2 cleavage of Notch are internalized and transported to this site. This analysis provides strong evidence for the first model. Since Nct and Psn are required for the S3 cleavage of Notch, the substrate for this cleavage, NEXT, accumulates in mutant cells. Most NEXT remains closely associated with the apical membrane, arguing against the existence of a major transport pathway to an intracellular compartment. Although a small amount of NEXT is found in intracellular clusters, these do not correspond to the major sites of Psn localization, and may be endocytic vesicles. Furthermore, recent data indicate that endocytosis is not required for the S3 cleavage of NEXT, because a membrane-tethered Notch derivative lacking the extracellular domain can signal normally in shibire mutant embryos, in which Dynamin-dependent endocytosis is blocked. Taken together, these results strongly suggest that S3 cleavage occurs predominantly at the plasma membrane (López-Schier, 2002).

Follicle cell clones mutant for either Nct or Psn have a more severe phenotype than that seen in Notch or Delta mutants, indicating that both proteins must have at least one additional function in these cells that is independent of their role in Notch signaling. One aspect of this phenotype is the overaccumulation of the components of the adherens junctions, DE-Cadherin, Armadillo, and alpha-catenin, and this is probably related to the fact that both alpha-catenin and the Armadillo ortholog ß-catenin associate with Psn in mammalian cells. Although neither is required for the activity of the S3 protease or gamma-secretase, loss of Psn leads to an overaccumulation of ß-catenin in Drosophila embryos and mouse epithelial cells. The precise function of Psn in ß-catenin regulation is unknown, but the overexpressed protein in Drosophila psn mutant embryos is associated with polyubiquitin-positive cytoplasmic inclusions, suggesting that Psn is required in some way to regulate Armadillo degradation. Psn also regulates the turnover of DE-Cadherin and alpha-catenin. Furthermore, Nct is necessary for this function, suggesting that it requires the formation of the high molecular weight protease complex. Since Psn is thought to mediate the proteolysis of membrane proteins, one possibility is that Psn is recruited to DE-Cadherin by binding to the catenins, and cleaves DE-Cadherin to trigger degradation. Alternatively, Psn could regulate the turnover of the catenins in some other way, and their overaccumulation in psn and nct mutants might then lead to the stabilization of Cadherin complexes at the membrane (López-Schier, 2002).

nct and psn mutants also disrupt the organization of the spectrin cytoskeleton in follicle cell clones. This phenotype is particularly puzzling, because the mutants have opposite effects on the two subunits of apical spectrin: ß_H-spectrin disappears from the apical membrane, whereas alpha-spectrin accumulates all round the cortex of the cell, but particularly at the apical side. In wild-type cells, the apical localization of alpha-spectrin normally requires its association with ß_H-spectrin to form (alphaß_H)₂ tetramers. Its apical localization in mutant cells must therefore occur by a different mechanism from normal, and this may sequester it in a form that cannot bind ß_H-spectrin, which would account for the failure of the latter to localize. However, alpha-spectrin must still tetramerize with ß-spectrin, because it shows wild-type localization to the basolateral domain. It is unclear whether the defect in the spectrin cytoskeleton is related to the overaccumulation of Cadherin complexes, but it is intriguing that alpha-spectrin accumulates at sites where DE-Cadherin, Armadillo, and alpha-catenin are most enriched in nct and psn mutant cells. Thus, this localization and stabilization of alpha-spectrin could be mediated through binding to some component of these abnormal adhesion complexes. It is also possible, however, that the Nct/Psn complexes have multiple functions in the follicle cells that independently regulate Notch signaling and the spectrin cytoskeleton and Cadherin adhesion complexes (López-Schier, 2002).

The tumor suppressor gene l(2)giant discs is required to restrict the activity of Notch to the dorsoventral boundary during Drosophila wing development

During the development of the Drosophila wing, the activity of the Notch signalling pathway is required to establish and maintain the organizing activity at the dorsoventral boundary (D/V boundary). At early stages, the activity of the pathway is restricted to a small stripe straddling the D/V boundary, and the establishment of this activity domain requires the secreted molecule Fringe (Fng). The activity domain will be established symmetrically at each side of the boundary between Fng-expressing and non-expressing cells. Evidence is presented that the Drosophila tumor-suppressor gene lethal giant discs (lgd), a gene whose coding region has yet to be identified, is required to restrict the activity of Notch to the D/V boundary. In the absence of lgd function, the activity of Notch expands from its initial domain at the D/V boundary. This expansion requires the presence of at least one of the Notch ligands, which can activate Notch more efficiently in the mutants. The results further suggest that Lgd appears to act as a general repressor of Notch activity, because it also affects vein, eye, and bristle development (Klein, 2003).

It has also been observed that wingless (wg) is expressed ectopically in the pouch of lgd mutants during wing development. Similar phenotypes are observed, if the Notch pathway is ectopically activated during wing development, raising the possibility that the lgd mutant phenotype could stem from the ectopic activation of the Notch pathway. The Notch pathway is indeed ectopically active in lgd mutants, and hyperactivation as well as ectopic activation of the pathway accounts for the lgd phenotype during wing development. In lgd mutants, the expression of Notch target genes along the D/V boundary is expanded, indicating that Lgd is required for the restriction of Notch activity to the D/V boundary. Furthermore, the mutant phenotype of lgd is suppressed by concomitant loss of Presenilin or Suppressor of Hairless function, indicating that the mutant phenotype is caused by the activation of the Notch pathway. Evidence is provided that the activity of fng and Serrate seem to be dispensable in lgd mutant wing disc and that Delta can activate Notch efficiently enough to maintain its activity during wing development. The presented results indicate that the negative regulation of Notch by Lgd is not restricted to wing development and occurs during several other developmental processes, such as vein, eye, and bristle development, suggesting that Lgd suppresses the activity of the Notch pathway in a variety of developmental processes (Klein, 2003).

Loss of lgd function leads to an overgrowth of the imaginal discs, clearly noticeable in the wing region of the wing disc, which becomes enlarged and flat (Bryant, 1971). wg expression is normally restricted to the D/V boundary of the wing pouch. In lgd mutants, wg is activated ectopically in a much broader domain that extends into the wing pouch. In addition, lgd mutant wing discs often develop a second wing pouch in the region of the anlage of the scutellum. Similar phenotypes are caused by gain-of-function alleles of N (for example, Abruptex) and are also observed upon expression of the activated intracellular form of Notch, Nintra, or expression of Notch ligands, such as Dl. The ectopic activation of wg can already be observed in early third instar wing discs and precedes the visible morphological changes that occur at later stages. The deficiency Df(2L) FCK-20 deletes the lgd locus, allowing the classification of the relative strength of the available alleles. The phenotype is always variable, but the overall phenotype of lgd^d7 and lgd^d10 in homozygotes and in trans over Df(2L)FCK-20 is very similar, indicating that these two alleles are strong, probably amorphic alleles. lgd^d4 and lgd^d1 are weaker alleles. All alleles display a qualitatively similar phenotype over the deficiency as in homozygotes, indicating that the observed phenotype is probably caused by the loss-of-function of the lgd gene (Klein, 2003).

The similarity between the loss of lgd function and ectopic N activation suggests that the phenotype of lgd could be caused by ectopic activation of the Notch pathway. To examine this possibility, the expression of E(spl)m8, cut, Dl, and Ser was monitored as well as the activity of the vg-boundary enhancer (vgBE) in mutant wing discs. The expression of all these markers is initiated in cells at the D/V boundary in a Notch-dependent manner. The vgBE is initially expressed along the D/V boundary of the wing, but late in the third instar, it is activated in an additional stripe along the anteroposterior compartment boundary (A/P boundary), which is also dependent on Notch activity. Both domains depend on the presence of a single Su(H) binding site in the enhancer. Similarly, the expression of cut and E(spl)m8 is initiated in cells at the boundary by the Notch-pathway, and E(spl)m8 is also dependent on the presence of Su(H) binding sites in its promoter. As described above, the expression of Dl and Ser is more complex but always dependent on the activity of Notch in cells at the D/V boundary. In lgd mutant wing discs, the vgBE as well as cut, Dl, Ser, and E(spl)m8 are activated ectopically within the wing pouch. The activation of the vgBE is dependent on the presence of the Su(H) binding site in the enhancer, since a version lacking it shows no ectopic activity in the mutants. As in the case of wg, the expression of the vgBE is already expanded in early third larval wing discs. Altogether, these results show that the loss of lgd function leads to the ectopic expression of Notch target genes. This suggests that the Notch pathway is ectopically activated in lgd mutants (Klein, 2003).

All tested Notch-target genes are ectopically activated in lgd mutant wing discs or lgd mutant cell clones. The ectopic activation of Notch target genes as well as the observed overproliferation of lgd mutants is abolished in lgd;Psn double mutants. In addition, Notch target gene expression is also abolished in Psn or Su(H) mutant clones generated in lgd mutant wing imaginal discs. These data suggest that the Notch pathway becomes ectopically active in the absence of lgd function. Furthermore, the fact that Delta alone seems to provide sufficient Notch activity to sustain wing development in lgd mutants indicates that the pathway can be activated more efficiently in the mutant background. The activation of Notch is a consequence of loss of lgd function also in other developmental processes, such as bristle, leg, and wing vein development. Thus, the presented data make lgd a good candidate gene that regulates activity of the Notch pathway during adult development of Drosophila (Klein, 2003).

Identification of genes involved in Drosophila melanogaster geotaxis, a complex behavioral trait

Identifying the genes involved in polygenic traits has been difficult. In the 1950s and 1960s, laboratory selection experiments for extreme geotaxic behavior in fruit flies established for the first time that a complex behavioral trait has a genetic basis. But the specific genes responsible for the behavior have never been identified using this classical model. To identify the individual genes involved in geotaxic response, cDNA microarrays were used to identify candidate genes and fly lines mutant in these genes were assessed for behavioral confirmation. The identities of several genes that contribute to the complex, polygenic behavior of geotaxis have thus been determined (Toma, 2002).

Pioneering experiments on Drosophila melanogaster and Drosophila pseudoobscura investigated the nature of the genetic basis for extreme, selected geotaxic behavior. These experiments constituted the first attempt at the genetic analysis of a behavior. Selection and chromosomal substitution experiments successfully showed that there is a genetic basis for extreme geotaxic response in flies and, by implication, for behavior in general. These experiments also added to understanding of the role of variation in phenotypic evolution and selection. Despite their seminal contributions in behavioral genetics, population genetics and the study of selection, by their nature these experiments could not identify specific genes (Toma, 2002 and references therein).

These results highlight both the success and the limitation of behavioral selection experiments. Although selection results tend to be representative of the natural interactions of genes that produce behavior and can demonstrate that a trait has a genetic basis, they do not pinpoint specific genes that influence the trait. This is partly due to the involvement of many genes and the relatively minor role of each in complex polygenic phenotypes -- a problem that is especially acute for the intrinsically more variable phenotypes that are associated with behavior. The advent of cDNA microarray technology offers an easily generalized strategy for detecting gene expression differences and can complement other means of identifying the genes that underlie complex traits. An expression difference may occur in a gene that is not itself polymorphic, but that gene may contribute to the realization of the phenotypic difference (Toma, 2002).

As a starting point for identifying genes that affect a complex trait, the selected, established Hi5 and Lo extreme geotaxic lines were examined for changes in gene expression between strains of Drosophila melanogaster subjected to long-term selection and isolation. A two-step approach was used: (1) the differential expression levels of mRNAs isolated from the heads of Hi5 and Lo flies was determined using cDNA microarrays and real-time quantitative PCR (qPCR); (2) a subset of the differentially expressed genes was independently tested for their influence on geotaxis behavior by running mutants for these genes through a geotaxis maze. It was reasoned that some of the differences in gene expression between strains might be related to phenotypic differences and that it should therefore be possible, at least in part, to reconstruct the phenotype with independently derived mutations in some of the differentially expressed genes (Toma, 2002).

The findings indicate that differences in gene expression can be used to identify phenotypically relevant genes, even when no large, single-gene effects are detectable by classical, quantitative genetic analysis. Three of the four genes implicated by microarray and qPCR measurements caused differences in geotaxis, whereas none of the six control genes had an effect. Only those genes that had larger differences in expression according to the microarrays, or that were significantly different according to qPCR results (cry, Pdf and Pen), significantly changed geotaxis scores. The converse was not true, because altered geotaxis behavior did not always accompany larger differences in mRNA levels, as shown by pros, although this might reflect the sensitivity of pros to aspects of the genetic context. All of the genes tested for which there was little or no difference in mRNA levels between the selected Hi5 and Lo lines also showed no influence on geotaxis behavior (Toma, 2002).

The directionality of behavioral and mRNA differences proved to be consistent with predictions that were based on expression levels. Homozygous null mutants of Pdf and cry showed a significant increase in geotaxis score, which is consistent with a lower level of expression of these genes in Hi5 relative to Lo. Similarly, the heterozygous Pen mutant showed a significant downward shift in geotaxis score, which is consistent with a lower level of Pen expression in Lo relative to Hi5. Thus, the change in behavior of the tested mutants corresponds to the direction predicted by differences in transcript level in the selected Hi5 and Lo lines (Toma, 2002).

Whereas the cry, Pen and female Pdf mutants produced the anticipated effect on behavior, the magnitude of behavioral effect was smaller than in the original selected lines. This probably reflects the difference between the aggregate effect of an ensemble of genes in the selected lines as opposed to the individual effect of a single mutant gene in a neutral background. In addition, their relatively small effects are exactly the results that one would predict in a polygenic system such as geotaxis behavior, in which many genes have small contributions to the overall phenotype. The three genes identified in this study would not have been predicted on the basis of their previously defined functions (Toma, 2002).

These results show that the two separate approaches to behavioral genetics -- the classical Hirschian quantitative analysis of genetic architecture and the modern Benzerian approach of single-gene mutant analysis -- are complementary and can be unified. This study used the results of a Hirschian approach of laboratory selection for natural variants to identify single gene differences, such as one would find in a Benzerian approach. The results are consistent with the suggestion that naturally occurring variants in behavior correspond to mild lesions in pleiotropic genes (Toma, 2002).

Finally, the results show that differences in gene expression identified by cDNA microarray analysis can be used as a starting point for narrowing down the numbers of candidate genes involved in complex genetic processes. Such an approach is analogous, as well as complementary, to the current method of mapping quantitative trait loci to large chromosomal intervals and then making educated guesses about which genes within those intervals may be involved in the trait (Toma, 2002).

The combination of selection, with its ability to exaggerate natural phenotypic variation, and global analysis of differences in gene expression by cDNA microarray analysis offers a promising approach to previously intractable molecular analyses of behavior. The geotaxis genes that were identified might have been the direct targets of selection, or they might be downstream of the direct targets. Additional studies using the Hi5 and Lo selected lines will be required to distinguish between these possibilities and to determine the causal role that these genes have in the context of the selected lines (Toma, 2002).

This study has gone from the selection of a 'laboratory-evolved' behavioral phenotype, to screening for mRNA differences, to partially reconstituting the phenotype using mutants. This shows the feasibility of combining genomic and classical genetic approaches for the breakdown and partial reassembly of an artificially selected behavioral trait (Toma, 2002).

Modeling clinically heterogeneous Presenilin mutations with transgenic Drosophila

To assess the potential of Drosophila to analyze clinically graded aspects of human disease, a transgenic fly model was developed to characterize Presenilin (PS) gene mutations that cause early-onset familial Alzheimer's disease (FAD). FAD exhibits a wide range in severity defined by ages of onset from 24 to 65 years. PS FAD mutants have been analyzed in mammalian cell culture, but conflicting data emerged concerning correlations between age of onset and PS biochemical activity. Choosing from over 130 FAD mutations in Presenilin-1, 14 corresponding mutations at conserved residues were introduced in Drosophila Presenilin (Psn) and their biological activity in transgenic flies was assessed by using genetic, molecular, and statistical methods. Psn FAD mutant activities were tightly linked to their age-of-onset values, providing evidence that disease severity in humans primarily reflects differences in PS mutant lesions rather than contributions from unlinked genetic or environmental modifiers. This study establishes a precedent for using transgenic Drosophila to study clinical heterogeneity in human disease (Seidner, 2006).

Presenilin is an evolutionarily conserved polytopic membrane protein that is part of the multisubunit γ-secretase complex responsible for intramembranous cleavage of several transmembrane proteins, including APP, Notch, Delta, DCC, ErbB4, N-Cadherin, and E-Cadherin. Familial Alzheimer's disease (FAD) mutations in Presenilin-1 (PS1) and PS2 alter proteolytic processing of APP to generate more toxic Aβ42 peptides that accelerate amyloid plaque formation in brain tissues. PS mutations also contribute to neurodegeneration and cognitive decline through amyloid-independent mechanisms, involving altered regulation of receptor signaling and intracellular kinase pathways (Seidner, 2006).

The Drosophila APP ortholog, APPL, lacks homology to APP within the Aβ peptide region, and loss of APPL produces subtle behavioral deficits that are difficult to measure quantitatively. A more suitable substrate to monitor Presenilin (Psn) FAD mutant activity in Drosophila is the Notch receptor, the most extensively characterized fly γ-secretase substrate. Psn is required throughout Drosophila development for Notch signaling, and a wide variety of Notch-related phenotypes exist that range from severe embryonic lethality to specific defects in adult tissues. This phenotypic range is amenable to characterizing variable degrees of function among different Psn FAD mutants. In addition, genetic and molecular reagents are available to characterize Notch biochemical cleavage and subsequent target-gene activation in Drosophila (Seidner, 2006).

Previous studies have successfully used transgenic mice and C. elegans to assess the genetic properties of human PS FAD mutants, and one study suggested a potential difference between two FAD variants. However, quantitative comparison between FAD mutants was not practical because of heterologous expression methods and difficulties controlling transgene copy numbers. Two studies in mammalian cell culture showed little or no relationship between average ages of onset and Aβ42 secretion levels, contrary to one study reporting a strong correlation (Seidner, 2006).

This study assessed PS FAD mutant function in Drosophila by expressing 14 FAD-linked mutant psn transgenes in animals lacking endogenous psn function. The mutations were selected on the basis of the following criteria: (1) they span the entire protein, (2) relatively large numbers of families and affected individuals have been identified for most mutations, and (3) their ages of onset range from 24 to 65 years, or are possibly asymptomatic (E318G and F175S). Thirteen of the mutations affect residues that are conserved in fly Psn, and the remaining one alters a conservatively substituted residue. To achieve physiological expression of the transgenes, a 1.5 kb promoter of the endogenous psn gene was defined, termed PEPC (Presenilin endogenous promoter cassette), . At the larval-pupal transition, loss of psn function causes highly uniform lethality that is fully rescued by a wild-type PEPC-psn transgene but not by wild-type human PS1 or PS2 transgenes driven by PEPC. To circumvent this technical problem with human PS transgenes, each FAD mutation was engineered into the fly psn gene (Seidner, 2006).

Examining transgenic lines expressing PEPC-psn FAD mutants in a psn null genetic background, eight distinct phenotypic categories were defined that represent a graded series in order of increasing rescue ability, as follows: (1) prepupal lethal, (2) late prepupal lethal, (3) pupal lethal, (4) pharate lethal, (5) severe neurogenic/adult lethal, (6) moderate neurogenic/adult viable, (7) weak neurogenic/adult viable, and (8) morphologically normal. Quantitative analysis of survival rates, incidence of dorsoscutellar bristle duplications, wing notching, wing vein defects, and other morphological features was performed to assign each transgenic line to the appropriate rescue category (Seidner, 2006).

The phenotypic categories are consistent with an ordered series attributable to progressive increases in Psn-dependent Notch activation, but it was important to verify that they do not instead represent the effects of Psn on other substrates. In the canonical Notch pathway, ligand binding to the Notch receptor leads to ectodomain removal and subsequent γ-secretase-mediated intramembranous cleavage of Notch. The liberated intracellular domain, termed NICD, translocates to the nucleus and participates directly in transcriptional regulation of target genes, including the Enhancer of split m7 (E(spl)m7) gene in Drosophila. Activation of Notch signaling within proneural cell clusters destined to give rise to adult sensory organs results in restricted expression of the proneural marker scabrous (sca). Visualization of sca expression in the larval imaginal wing disc with a sca-lacZ transgene provides evidence for a progressive range of Notch signaling with different PEPC-psn transgenes (Seidner, 2006).

Biochemical evidence for progressively increasing Psn function was obtained by immunoblot analysis of Notch cleavage products across the spectrum of PEPC-psn FAD phenotypes. Representative PEPC-FAD transgenic lines of the different categories exhibit partial-to-complete failure in γ-secretase-mediated NICD production. Interestingly, although transgenic lines belonging to categories 2–4 exhibit significant levels of biological rescue activity, they do not produce levels of NICD detectable by immunoblot analysis. These results demonstrate that genetic assays such as these are valuable for studying low-level or tissue-specific aspects of γ-secretase substrate cleavage that might not be amenable to biochemical analysis (Seidner, 2006).

To determine whether the PEPC-psn FAD phenotypic series is similarly correlated with Notch target-gene activation, semiquantitative RT-PCR was used to monitor activation of the E(spl)m7 gene. Normalizing transcriptional activation to two control genes, a progressive increase in E(spl)m7 transcript levels was observed across categories 1-8, confirming that they represent graded increases in γ-secretase-dependent Notch activation. To verify this relationship, a temperature-sensitive Notch allele termed N^ts1 was used. Incremental levels of Notch activity obtained by raising N^ts1 flies at different temperatures produced phenotypes matching those seen in the Psn FAD mutant series, but were independent of manipulations involving endogenous or transgenic Psn. Levels of E(spl)m7 transcriptional activation seen in these N^ts1 phenotypic classes resembled closely those observed in the corresponding PEPC-psn FAD classes, confirming the progressive increase in Notch target-gene activation across this phenotypic spectrum. Taken together, these results validate the use of the phenotypic criteria to characterize varying degrees of PEPC-psn FAD transgene activity toward the endogenous γ-secretase substrate Notch (Seidner, 2006).

A limitation of transgenic analysis in Drosophila is that the transgenes are inserted at essentially random genomic locations, leading to variations in expression as a result of local position effects at different sites. These effects are a confounding factor in comparisons of relative degrees of rescue function, which reflects properties of the primary lesion in each Psn mutant protein as well as the expression level of each transgenic insertion. Therefore, from four to 16 independent insertions were generated for each transgene, and each insertion was scored with multiple quantitative morphological criteria and assigned to the appropriate phenotypic category. Visual inspection of the results suggests a positive overall trend between increasing age of onset in human FAD pedigrees and increasing biological activity for these FAD mutants in transgenic Drosophila (Seidner, 2006).

Statistical tests demonstrate an overall correlation across the dataset as a whole, but the correlation is not smooth. Most FAD mutants appear to cluster in groups separated by discontinuities, with a few mutants that are noticeably different from nearby mutants. Mutants having similar ages of onset are, in general, not statistically different from one another. Nevertheless, the mutants can be statistically grouped into three distinct classes: strong (L166P, L173W, P436Q, V272A, and L235P), intermediate (M146L, M139V, H163R, E280A, A246E, and G206A), and weak (A79V, F175S, and E318G). Strong mutants differ most clearly from the weak ones in terms of function, whereas the intermediate group shows more modest differences compared to either of these two flanking groups. Pairwise comparisons of aggregated mutant groups confirmed these classes. The strong mutant group is significantly different from both the intermediate and weak groups. Despite its age of onset of ~60, the A79V mutant is more similar to the weak mutants than the intermediate ones, as confirmed by pairwise comparisons of aggregated intermediate and weak groups with A79V assigned to either the intermediate or the weak group (Seidner, 2006).

Similarly, two mutants, L235P and P436Q, are less functional in transgenic flies than might be predicted on the basis of their ages of onset, whereas another mutant, M139V, is somewhat more functional than expected. Whether these disparities reflect inadequate age-of-onset data, modifier effects in the corresponding human pedigrees, or an unknown feature of the fly assay system is not clear. Finally, the E318G and F175S mutants, proposed to be either very weak pathogenic mutants or functionally normal polymorphisms, are statistically indistinguishable from the wild-type transgene in the genetic assay, consistent with the idea that they are nonpathogenic polymorphisms. Overall, the results support the assertion that disease severity in early-onset Alzheimer's disease is primarily determined by PS mutant lesion type as opposed to unlinked genetic or environmental modifiers, as was also deduced from a study of Aβ secretion levels in PS FAD mutant transfected cells (Seidner, 2006).

A few other interesting patterns emerge from these comparisons. Mutants for which relatively few independent lines were obtained, most notably H163R, fail to show significant differences when compared to other mutant groups. On the basis of the complete dataset, it is estimated that ~10 independent lines of a given mutant are required to obtain statistically useful data. This problem might be circumvented by employing a 'knock in' strategy to precisely replace the endogenous psn gene with FAD variants, an approach that is not yet reliable in Drosophila. Additionally, one source of experimental noise is that transgenes occasionally insert into locations where they are poorly expressed or damaged during insertion, as is evident from a few instances involving the 'asymptomatic' F175S and E318G mutant and wild-type transgenes. Normalization of functional read-outs for each transgene insertion relative to its mRNA or protein expression level in the appropriate transcript null or protein null psn mutant background should reduce this noise and lead to further refinement of the statistical data (Seidner, 2006).

These findings establish the validity of using transgenic Drosophila or other heterologous organisms to evaluate clinically heterogeneous aspects of human diseases with a clearly defined genetic etiology. Transgenic Drosophila offer several advantages to augment more traditional clinical assessments as well as transgenic mouse models. Transgenic flies are relatively inexpensive and rapid to produce, large numbers of independent lines can be easily generated, and limitless numbers of progeny for each line can be examined under controlled genetic conditions. These features of the assay might make it useful for obtaining a rapid estimate of approximate disease severity for new PS1 and PS2 mutations, especially for those with limited pedigree data, small numbers of affected individuals, or suspected environmental or genetic confounding factors. Although the primary goal of this study was to assess an array of genetically diverse FAD mutant variants in a more standardized genetic background, the transgenic flies characterized could be used to study the effects of suspected modifiers or search for new modifiers of PS function. The transgenic lines could also be combined with APP-expressing transgenes to investigate more directly the role of PS FAD mutants in APP cleavage, amyloid-peptide accumulation, and neurotoxicity in a fly model. The correlation observed between the effects of different FAD mutations on Drosophila Notch signaling and human disease onset underscores recent proposals that in addition to APP processing, more global perturbations in pathways involving other γ-secretase substrates should be considered in early-onset Alzheimer's disease. Finally, the results offer encouragement that additional transgenic Drosophila models might be developed to investigate clinical heterogeneity in other human diseases (Seidner, 2006).

Home page: The Interactive Fly © 1997 Thomas B. Brody, Ph.D.

The Interactive Fly resides on the
Society for Developmental Biology's Web server.