Mouse Disabled 1 Regulates the Nuclear Position of Neurons in a Drosophila Eye Model

Abstract

Nucleokinesis has recently been suggested as a critical regulator of neuronal migration. Here we show that Disabled 1 (Dab1), which is required for neuronal positioning in mammals, regulates the nuclear position of postmitotic neurons in a phosphorylation-site dependent manner. Dab1 expression in the Drosophila visual system partially rescues nuclear position defects caused by a mutation in the Dynactin subunit Glued. Furthermore, we observed that a loss-of-function allele of amyloid precursor protein (APP)-like, a kinesin cargo receptor, enhanced the severity of a Dab1 overexpression phenotype characterized by misplaced nuclei in the adult retina. In mammalian neurons, overexpression of APP reduced the ability of Reelin to induce Dab1 tyrosine phosphorylation, suggesting an antagonistic relationship between APP family members and Dab1 function. This is the first evidence that signaling which regulates Dab1 tyrosine phosphorylation determines nuclear positioning through Dab1-mediated influences on microtubule motor proteins in a subset of neurons.

Precise neuronal migration determines the structural organization of the mammalian nervous system, which is intimately linked to its function. Neuronal migration is known to be under genetic control; however, the mechanism that various gene products employ to coordinate neuronal migration is poorly defined (16, 36, 51). Recently, nucleokinesis has been suggested as a key mediator of neuronal migration, though the signaling pathways that regulate it are not known (44, 51). The Reelin signaling pathway determines the positions of several classes of neurons in the central nervous systems of mammals. One essential downstream consequence of Reelin signaling is the posttranslational modification of the Disabled 1 (Dab1) protein (21, 22). Defining the cellular machinery that is regulated by Dab1 modification is central to understanding the control of neuronal positioning in several areas of the mammalian nervous system.

Mutations that inactivate the Reelin signaling pathway cause anomalies in the organization of several brain regions and the spinal cord (48). Previous genetic and biochemical studies reveal that Reelin function requires the action of two partially redundant receptors, ApoER2 and VLDLR, and the cytoplasmic adaptor protein Dab1 (9, 18). Reelin induces rapid receptor-dependent tyrosine phosphorylation of Dab1 within minutes and the slow degradation of Dab1 over several hours (1, 21). The Src family kinases Src, Fyn, and Yes have been implicated in this process, since genetic disruption of these kinases or specific inhibitors against them dramatically reduce Reelin-induced Dab1 phosphorylation and cause defects in neuronal migration in slice culture assays (2, 5, 27). Recently it has been demonstrated that mice lacking Src and Fyn have phenotypes consistent with loss of Reelin signaling (30). In addition, Dab1 tyrosine phosphorylation has been suggested to be a critical mediator of the signaling cascade. Mice that express only a tyrosine-to-phenylalanine-substituted Dab1 molecule phenotypically resemble the Reelin- and the Dab1-null mutants (22). Dab1 tyrosine phosphorylation has been shown to promote protein interactions between Dab1 and other signaling molecules, such as Nckβ, Crk, CrkL, phosphatidylinositol 3-kinase, and LIS1 (3, 4, 6, 8, 26, 40). It is thought that phosphorylation-dependent interactions likely mediate Reelin-regulated Dab1 functions.

A subset of genes that control neuronal migration encode proteins that regulate microtubule motor proteins. The proteins Ndel1, CDK5, Fak, Dcx, and Lis1 have been implicated in the control of nuclear placement, suggesting a significant role for this process in neuronal migration (37, 42, 43, 46, 55). The Lis1 (or Pafah1b1) gene, which was identified as the gene mutated in human lissencephaly type I, encodes a protein that interacts with the dynein motor complex (41, 47). Interestingly, Lis1 homologs in Aspergillus nidulans and Drosophila melanogaster have been shown to regulate nuclear position in a microtubule-dependent manner (10, 45, 54). It remains to be elucidated how these proteins are regulated to control nuclear movements and neuronal positioning.

The Drosophila eye is the only in vivo system in which it has been demonstrated that the action of dynein and kinesin motor proteins regulates nuclear position in postmitotic neurons (52). In addition, requirements for the nuclear lamin Lam Dm(0), the microtubule regulatory protein Klar, the serine threonine kinase Msn, and its substrate protein BicD have been demonstrated (13, 19, 39, 45, 52). Here, we address whether mouse Dab1 controls the molecular machinery that regulates nuclear placement, using the Drosophila visual system as a model. We found that expression of a tyrosine-phosphorylated Dab1 fusion protein partially rescued nuclear positioning defects caused by dominant mutations in Glued, which encodes a subunit of dynactin. We also identified a role for the amyloid precursor protein (APP)-like (APPL) in the compensation for Dab1 overexpression. Together our findings support a model in which Dab1 functions to regulate neuronal positioning through influences on nucleokinesis within postmitotic neurons.

MATERIALS AND METHODS

Drosophila stocks and culture.

All fly culture, genetics, and production of the transgenic lines were done according to standard protocols. The w¹¹¹⁸ strain was used to generate the transgenic lines UAS-RFP (expressing the red fluorescent protein [RFP]), UAS-Dab1RFP (expressing a wild-type mouse Dab1-RFP fusion protein), UAS-Dab1RFP-5F (encoding a mutant Dab1 with five tyrosines mutated to phenylalanines), and UAS-Dab1RFP158V (encoding a mutant with a defective phosphotyrosine binding domain) as previously described (23, 40). The GMR-Gal4 and Elav-Gal4 strains were kindly provided by Kyung-Tai Min, the Appl-deficient fly line was a gift from Kalpana White, and the Gl¹ mutant was obtained from the Bloomington Stock Center.

Scanning electron microscopy.

Three-day-old adult female flies were fixed in 2.5% glutaraldehyde in 0.1 M cacodylate buffer and processed as previously described (40).

Larval eye disk dissections and immunostaining.

Central nervous system/eye-antennal disk complexes were dissected from third-instar wandering larvae (with 16 to 20 rows of Elav-positive photoreceptor clusters), fixed in 4% paraformaldehyde, and stained using anti-Elav 7E8A10 (1:300; Developmental Studies Hybridoma Bank [DSHB]), anti-Chaoptin 24B10 (1:100; DSHB), and Alexa Fluor-phalloidin (Molecular Probes). Images were collected on a Zeiss LSM 510 confocal microscope.

Fly histology.

The heads from adult female flies were removed using microdissection scissors and fixed in 4% paraformaldehyde (Electron Microscopy Sciences). The tissue was then infiltrated with 20% sucrose overnight and frozen in Tissue-Tek OCT compound (Sakura Finetechnical Co.), and 12-μm horizontal sections were cut. The samples were stained with DAPI (4′,6′-diamidino-2-phenylindole), mounted using Vectashield mounting medium (Vector Laboratories), and examined using an inverted DeltaVision system microscope (Applied Precision).

Reelin stimulations of primary neuronal cultures.

Reelin- and control-conditioned media were prepared as previously described (40). Primary neuronal cultures were established using forebrains from embryonic day 15.5 embryos from wild-type and human APP (hAPP) transgenic littermates (line APPSWE 2576) as previously described (40). The culture medium was changed 3 to 6 hours prior to stimulation. Reelin time course stimulations were performed by replacing the growth medium with Reelin-conditioned or control-conditioned medium and incubating the cells at 37°C as indicated. To test the effects of the Aβ fragment of APP on Reelin stimulation, purified peptide (Biosource) was added to Reelin-conditioned medium at 500 pg/ml, a level that is observed in conditioned media after several hours of culture. The stimulation was stopped on ice by the addition of ice-cold radioimmunoprecipitation assay lysis buffer. Blots were probed with rabbit anti-Dab1 (1:1,000; Biodesign), mouse antiphosphotyrosine 4G10 (1:2,000; Upstate), mouse anti-β-tubulin E7 (1:100; DSHB), or mouse antiactin AC-15 (1:10,000; Sigma). The autoradiographs were scanned, and bands were quantified using the ImageJ software. Values were normalized to actin.

All mice used to generate cultures in this study were handled under the animal care and use guidelines of the NIH.

RESULTS AND DISCUSSION

Using the Drosophila visual system paradigm, we examined whether expression of tyrosine-phosphorylated mouse Dab1 would alter nuclear position in neurons. The photoreceptor neurons form a polarized monolayer epithelium during the third-instar phase of larval life and the nuclei migrate to stereotypic positions during this period (53). Given the importance of Dab1 tyrosine phosphorylation in Reelin signaling, we wished to determine the effects of tyrosine-phosphorylated Dab1 on the positioning of nuclei within neurons. Conveniently, mouse Dab1 expressed as a fusion with the tetrameric red fluorescent protein (Dab1RFP) is tyrosine phosphorylated in cultured cells and when expressed in the Drosophila eye (data available upon request) (40). Comparison of phenotypes caused by Dab1RFP expression and expression of a phenylalanine-substituted mutant Dab1RFP-5F, which is not tyrosine phosphorylated under the same conditions (data available upon request) (40), was used to determine whether the observed phenotypes are dependent upon phosphorylation.

Expression of tyrosine-phosphorylated Dab1 partially rescues the Glued phenotype.

The Drosophila Glued mutant provides an ideal model for the analysis of genetic influences on nuclear position in postmitotic neurons. Dominant mutations in Glued produce readily apparent external eye anomalies, which can be used to assess genetic interactions (52). Furthermore, the photoreceptor nuclei migrate into the optic stalk in the Glued mutant. We investigated the consequences of Dab1 expression on the external eye morphology in the Glued mutant as an initial test. We compared the external eye morphology of Drosophila with a dominant negative allele of Glued, Gl¹, to that of Gl¹ flies expressing Dab1RFP in the photoreceptor cells. As previously demonstrated, Gl¹ flies show reductions in eye volume, irregular organization of the ommatidial array, and numerous fusions of ommatidia (Fig. 1A and E) (13). Expression of Dab1RFP on the Gl¹ background partially rescues the organization of the ommatidial array, and half the number of fusions were observed (Fig. 1B, F, and I). To determine if the rescue was dependent upon Dab1 tyrosine phosphorylation, we examined whether expression of Dab1RFP-5F ameliorated the eye phenotype. Expression of neither Dab1RFP-5F nor RFP alters the appearance of the Gl¹ mutant eye or significantly reduces the number of fusions observed (Fig. 1C, D, G, H, and I). The expression level of the phenylalanine-substituted Dab1RFP-5F molecule was similar to that of unsubstituted Dab1 (data available upon request). Taken together, these data suggest that the phosphorylation sites are required for the observed effect on external eye morphology.

FIG. 1.

Expression of tyrosine-phosphorylated Dab1RFP partially rescues the Gl¹ external eye phenotype. (A and E) Scanning electron micrograph of an Elav-Gal4/+; +/+; Gl¹/+ eye shows the characteristic small rough eye. At higher magnification the ommatidia appear disorganized, with variable sizes and frequent fusions. The bristles are also often duplicated. Bar, 100 μm. Panels E to H are magnified ×5 compared to panel A. (B and F) Gl¹ flies overexpressing phosphorylated Dab1RFP (UAS-Dab1RFP/Elav-Gal4; +/+; Gl¹/+) have fuller and smoother eyes, with a net improvement in the arrangement of ommatidial facets, fewer fusions, and better organized bristles. (C, D, G, and H) Flies expressing a phosphorylation-defective Dab1RFP-5F mutant (Elav-Gal4/+; UAS-Dab1RFP-5F/+; Gl¹/+) (C and G) or RFP (UAS-RFP/Elav-Gal4; +/+; Gl¹/+) (D and H) on the Gl¹ background show no significant improvement of the external eye morphology. (I) The number of fusions in a 156- by 118-μm area of the eye was counted. Five or six animals were examined for each mutant. The data were analyzed by unpaired t tests (*, P < 0.0001; ♦, P = 0.1525; •, P = 0.6807). The error bars are standard errors of the means.

In the experiments outlined above, we used a relatively weak Elav-Gal4 driver line to direct expression of the UAS-Dab1RFP transgene in the photoreceptor cells in the eye and other postmitotic neurons (56). Dab1RFP expression under control of this driver does not lead to eye roughening, nor does it alter the organization of the photoreceptor cells within the individual ommatidia (data not shown). Interestingly, using the stronger GMR driver, overexpression of Dab1RFP but not Dab1RFP-5F produces a rough eye phenotype (12, 40). Neither Dab1RFP nor Dab1RFP-5F expression under the control of the GMR driver rescued the Gl¹mutant phenotype (data not shown). This could imply that expression of Dab1 above a certain threshold hampers rescue. Alternatively, the expression profile of the Elav driver may be critical for rescue. The Elav-Gal4 driver induces expression soon after neurons become postmitotic (56). This precedes expression from the GMR-Gal4 driver, which is restricted to the eye but includes support cells in addition to neurons (12). No alteration in the differentiation of the photoreceptor cells was observed in flies expressing either Dab1RFP or Dab1RFP-5F under either the Elav-Gal4 or GMR-Gal4 driver by thin-layer electron microscopy (data not shown), suggesting that the observed effects were due to changes in cell properties rather than cell identity.

Dab1 expression improves nuclear position in the Glued background.

In wild-type larvae, photoreceptor cell clusters are positioned in a highly organized manner with their nuclei anchored in the upper third of the retina (Fig. 2A). As previously described, in Gl¹ mutants the majority of photoreceptor cell bodies and nuclei fail to position correctly; instead, numerous nuclei migrate into the optic stalk, with only a thin process remaining in the retina (compare Fig. 2B with A) (52). In Dab1RFP- or Dab1RFP-5F-expressing larvae, Dab1 was observed throughout the cytoplasm of photoreceptor cells (Fig. 2D and data not shown). Moreover, we observed that Dab1RFP expression improves the nuclear position of photoreceptor cells within the imaginal disk (compare Fig. 2B with D and Fig. 3C with D). In particular, we observed fewer nuclei in the lower third of the retina and more nuclei in the upper two-thirds compared to Gl¹ mutants in the absence of Dab1RFP expression (Fig. 3G). A large number of nuclei did remain in the optic stalk with Dab1RFP expression. This may reflect confounding problems in the Gl¹ mutant photoreceptors that are not rescued by altering the balance of the microtubule motor proteins. This rescue is comparable to rescue observed previously by loss-of-function mutations in the kinesin heavy chain (khc) gene, in both the external eye morphology and the position of nuclei within the retina and optic stalk (52). In contrast to Dab1RFP, expression of Dab1RFP-5F or RFP failed to rescue the photoreceptor position (Fig. 3E, F, and G). This suggests that Dab1 regulates motor proteins that control nuclear placement, in a phosphorylation-dependent manner.

FIG. 2.

Photoreceptor cell differentiation and axonal projection in third-instar imaginal disks. (A) A wild-type third-instar eye disk stained with a photoreceptor-specific antibody (anti-Chaoptin, green), a neuronal nuclear marker (anti-Elav, blue) and phalloidin (red) demonstrates the stereotypical arrangement of photoreceptor clusters in the developing retina. The photoreceptor cell bodies are anchored in the upper two-thirds of the retina, between the actin-rich apical and basal surfaces. The optic stalk is free of nuclei. (B and C) In the Elav/+; +/+; Gl¹/+ (B) or UAS-RFP/Elav; +/+; Gl¹/+ (C) eye disks, the photoreceptors span the retina but many of the cell bodies and nuclei have migrated into the optic stalk. (D) However, in the UAS-Dab1RFP/Elav; +/+; Gl¹/+ imaginal disks, fewer nuclei were observed in the lower one-third of the elongated retinal cells. Dab1RFP expression was observed throughout the cytoplasm of UAS-Dab1RFP/Elav;+/+;Gl¹/+ photoreceptor cells (labeled with anti-Dab1 in red, phalloidin in green, and anti-Elav in blue). Bar, 10 μm.

FIG. 3.

Expression of phosphorylated Dab1RFP rescues the nuclear misplacement defect in the Gl¹ mutant third-instar eye disk. (A) Wild-type photoreceptor nuclei (labeled with anti-Elav, green) are positioned in clusters under the actin-rich apical surface of the developing retina (red). Note that all nuclei are found in the top two-thirds of the retina and none in the optic stalk (bottom). Bar, 10 μm. (B) Schematic representation of a wild-type eye disk showing the arrangement of photoreceptor cells. MF, morphogenetic furrow; OS, optic stalk. (C) In the Gl¹ animal the nuclei are scattered throughout the retina, and many of them migrate in the optic stalk, under the actin-rich basal surface of the retina. (D) Expression of phosphorylated Dab1RFP in the photoreceptor cells under the Elav-Gal4 driver (see Fig. 2) results in fewer nuclei trapped in the lower one-third of the retina and more found in their normal position closer to the apical surface. (E and F) Neither the phosphorylation-defective Dab1RFP-5F mutant (E) nor RFP alone (F) rescued the Gl¹ nuclear phenotype. (G) The positions of individual nuclei were measured in four to six animals for each mutant (average of 300 to 500 nuclei per mutant). For each nucleus, the values were presented as a ratio of the total distance between the basal and apical surfaces over the distance of the individual nucleus to the apex. The values were grouped in two bins, with the basal bin (values of 1.5 or less) representing nuclei found in the lower one-third of the retina. The data were analyzed by unpaired t tests (*, P = 0.0353; ⧫, P = 0.3673; •, P = 0.6201). The error bars are standard errors of the means. WT, wild type.

In mouse, Lis1, a dynein-interacting protein, binds tyrosine-phosphorylated Dab1 (3). Drosophila Lis (DLis) has been demonstrated to regulate the position of photoreceptor nuclei (45). We have investigated possible effects of Dab1RFP expression on the external eye morphology of dominant DLis mutants; however, no rescue was observed (data not shown). If Dab1RFP is operating through DLis to alter the position of photoreceptor nuclei, perhaps reducing the dose of DLis lessens the influence of Dab1. Alternatively, the profound effects of DLis mutation on multiple cell types may make it intractable to improvement by Dab1RFP expression.

APPL deficiency enhances Dab1 overexpression phenotypes.

Dab1 also interacts with proteins implicated in kinesin function. Dab1 binds a sequence motif in the C-terminal domain of amyloid precursor protein, YENPTY, which matches the Dab1 PTB domain binding consensus (23, 50). Interestingly, the YENPTY motif is identical between mouse APP and Drosophila APPL proteins. The physical interaction between Dab and APP family members is likely conserved through evolution, since Drosophila Dab interacts with mammalian APP (35). It has recently been shown that both APPL and human APP promote neurite outgrowth in response to injury in the Drosophila central nervous system, suggesting conserved roles for this family of proteins (32). In Drosophila, APPL acts as a kinesin cargo receptor, and interference with its function leads to cargo jam phenotypes (15, 49). Mammalian APP has been suggested to have a similar function; however, this has recently been disputed (15, 28, 31). To examine whether APP family members influence Dab1 function, we generated flies that overexpress Dab1RFP in the visual system on an Appl-deficient genetic background. Broad, high-level expression of Dab1RFP in the developing Drosophila visual system using the strong GMR-Gal4 driver results in a rough eye phenotype (Fig. 4B and F) (40). Appl is expressed in the developing photoreceptor cells, but there is no morphological defect in the visual system of Appl-null flies (33, 34). We observed, however, that expression of Dab1RFP in Appl-deficient flies results in a worsening of the external eye morphology (Fig. 4D and H), characterized by a loss of organization of the ommatidial array and an increase in ommatidial fusions. Therefore in the absence of APPL, flies appear to be less able to compensate for the effects of high-level Dab1RFP expression.

FIG. 4.

Loss of APPL enhances the rough eye phenotype in Dab1RFP-overexpressing flies. (A and E) The scanning electron micrograph of an adult UAS-RFP/GMR-Gal4 control fly shows smooth external eye morphology with a linear array of ommatidial lens facets and interommatidial bristles. Bars, 100 μm (A) and 20 μm (E).(B and F) The UAS-Dab1RFP/GMR-Gal4 eye is rough, with disruptions in the linear arrangement of ommatidial lens facets and occasional fusions. (C and G) On the Appl-deficient background, UAS-RFP/GMR-Gal4 adults do not demonstrate any eye roughening. (D and H) However, expression of UAS-Dab1-RFP/GMR-Gal4 on the Appl-deficient background leads to worsening of the rough eye compared to that in panel B. The eye is characterized by a circular arrangement of the ommatidial lens facets compared to the relatively linear pattern observed in the UAS-Dab1-RFP/GMR-Gal4 flies, as well as more frequent ommatidial fusions.

In order to assess whether this effect on eye roughening was accompanied by aberrations in nuclear or cell placement, we examined cross sections of adult heads with the fluorescent nuclear stain DAPI. Characteristic patterns were observed in the control flies (Fig. 5A). Expression of Dab1RFP resulted in aberrations in the placement of nuclei in the visual system (Fig. 5B). To determine whether this required the Dab1 tyrosine phosphorylation sites, we examined the effects of overexpression of Dab1RFP-5F. Eyes from these transgenic animals showed no external eye phenotype, as previously reported (40), and the nuclear organization was relatively unaffected (Fig. 5C). Therefore, the tyrosine phosphorylation sites on Dab1 are required for the nuclear or cellular misplacement anomaly in eyes of transgenic flies with wild-type levels of Appl. However, expression of a Dab1 mutant with a defective PTB domain, Dab1RFP-158V, resulted in a nuclear placement phenotype (Fig. 5D). This mutation has previously been shown to reduce the affinity of the interaction between Dab1 and the YENPTY sequence found in the cytoplasmic domain of APPL (23).

FIG. 5.

Phosphorylated Dab1RFP alters nuclear or cellular positioning in the adult retina. (A) DAPI staining of control UAS-RFP/GMR-Gal4 eyes shows nuclei forming an arc pattern containing apical and basal nuclei with a nucleus-free region in the middle. Bar, 100 μm. (B) Adult UAS-Dab1RFP/GMR-Gal4 retinas had alterations in the nuclear position of the apical nuclei. (C) Minor disturbances of the outer layer of nuclei were observed in UAS-Dab1RFP-5F/GMR-Gal4 flies. (D) The arrangement of nuclei in UAS-Dab1RFP-158V/GMR-Gal4 flies was similar in severity to the Dab1RFP phenotype. (E) Appl deficiency causes only subtle, localized disturbances of the nuclear rows in Drosophila expressing RFP. (F) However, Appl deficiency in conjunction with the UAS-Dab1RFP/GMR-Gal4 transgenes caused a dramatic increase in the number of misplaced nuclei, with a loss of the central nucleus-free area. (G) The nuclear placement in Dab1RFP-5F-expressing retinas was also more erratic on the Appl mutant background. (H) In contrast, the phenotype of Dab1RFP-158V/GMR-Gal4 was relatively unaltered by Appl deficiency.

To determine whether the enhanced external eye phenotype that resulted from expressing Dab1RFP on an Appl-deficient background was accompanied by nuclear or cellular misplacement, we examined horizontal cross sections of the eyes (Fig. 5). Appl-deficient flies showed only small, localized anomalies in nuclear position in the retina (Fig. 5E). Expression of Dab1RFP in the absence of APPL, however, resulted in pronounced anomalies, with nuclei scattered throughout the retina (Fig. 5F). Surprisingly, expression of Dab1RFP-5F on the Appl-deficient background also resulted in a more severe phenotype (Fig. 5G), suggesting a reduced dependency on tyrosine phosphorylation. The phenotype caused by Dab1RFP-158V expression, however, was not dramatically altered by loss of APPL (Fig. 5H). This suggests that the influence of APPL is direct and that it acts to alter Dab1 function through a PTB domain-dependent process. Neither the total expression level nor the tyrosine phosphorylation states were altered by loss of Appl (data available upon request). This suggests that the observed phenotypic enhancements are not due to altered protein stability or tyrosine phosphorylation. Together these results point to a Dab1 activity that regulates cell or nuclear position, which is buffered by the presence of APPL protein. Interestingly in the absence of APPL, Dab1 tyrosine phosphorylation appears to be less critical.

In the above-described experiment, the nuclei were displaced towards the optic stalk by high-level Dab1 expression from the GMR driver, which becomes especially apparent in the Appl mutant background. In contrast, with low-level expression of Dab1 using the Elav driver, the effect was to rescue the nuclei towards the apical pole in the Glued mutant background. Differential effects caused by protein expression level are common in the study of components of the molecular motors. For instance, overexpression of dynamitin, a subunit of dynactin, disrupts dynactin function, while stoichiometric levels are required for function (11). APP overexpression interferes with kinesin function, resulting in a vesicle jam phenotype, which is partially rescued by reduced doses of dynactin (15). Overexpression of various components of the motors apparently sequesters critical subunits away from the complexes, hindering their function. APPL may act to sequester Dab1 and reduce interference with microtubule motor proteins in transgenic flies expressing high levels of Dab1RFP.

Overexpression of APP attenuates Reelin-induced Dab1 phosphorylation in mammalian neurons.

One obvious question raised by this work is whether APP family members affect Dab1 function in mammalian systems. In mammals, there are three family members (APP, APLP1, and APLP2) that have recently been demonstrated to influence neuronal positioning. The disruption of all three genes leads to lissencephaly type II, which is distinct from the cortical dysplasia caused by mutations in the components of the Reelin pathway (17). Neither Dab1 nor Reelin levels appear to be significantly altered in these mice. This suggests that APP family proteins are not essential for Reelin activity; however, it does not rule out possible roles for APP in the down-regulation or fine-tuning of Reelin signaling.

To address whether expression levels of APP family members influence Reelin signaling, we compared the response of cortical neurons from embryonic mice that overexpress human APP (25) to the response of normal neurons. Equal numbers of wild-type and hAPP-overexpressing neurons, harvested in parallel from the same litter, were stimulated with control- or Reelin-conditioned medium. Dab1 tyrosine phosphorylation was assayed at various times after stimulation. We observed that Dab1 tyrosine phosphorylation was induced by Reelin stimulation in both wild-type and hAPP-expressing neurons. However, the hAPP-overexpressing neurons responded less robustly to Reelin, and Dab1 tyrosine phosphorylation reached levels that were approximately 30 percent of those observed in wild-type neurons (Fig. 6A and B). These observations suggest that increased levels of APP expression may quench Reelin signaling.

FIG. 6.

APP transgenic neurons respond less robustly to Reelin stimulation. (A) Primary neuronal cultures from embryonic day 15.5 wild-type or hAPP transgenic littermates were stimulated with control (C) or Reelin-enriched (R) medium for 2 to 120 min as indicated. Reelin-induced Dab1 tyrosine phosphorylation (pY) was apparent after 2 min of treatment, with a peak at 30 min, followed by signal decrease over the next 1.5 h in wild-type neurons. In contrast, stimulation of hAPP-expressing neurons failed to induce equivalent levels of Dab1 tyrosine phosphorylation. Similar time courses and total phosphorylation levels were observed in five paired experiments. (B) Quantification of Dab1 tyrosine phosphorylation normalized to actin from the experiment shown in panel A demonstrated that hAPP expression reduces Dab1 tyrosine phosphorylation to a value of less than half of that observed from wild-type neurons.

The APP transgenic line used in this experiment harbors the so-called Swedish mutation that leads to Alzheimer's disease in humans and causes memory deficits in this mouse line after 8 months of age (24). This mutation increases the generation of the Aβ fragment of APP. During the Reelin stimulation assays, medium is exchanged prior to addition of Reelin, so it is unlikely that there is significant Aβ or other secreted forms of APP in the stimulation experiment shown in Fig. 6A. To determine if Aβ generated in the hAPP cultures might influence signaling, we added exogenous Aβ to Reelin stimulation assay mixtures and measured Dab1 tyrosine phosphorylation. Inclusion of Aβ with Reelin-conditioned medium did not significantly affect the Reelin-induced Dab1 tyrosine phosphorylation (data not shown). Therefore, it seems unlikely that Aβ itself is interfering with Reelin signaling in our assay and more likely that the full-length APP or C-terminal fragments of APP are responsible for the decreased response to Reelin. This suggests that the level of APP family members determines the strength of Reelin signaling in some contexts. To date, APP transgenic animals have not been reported to have phenotypes that are consistent with compromises in the components of the Reelin-signaling pathway. Therefore, in normal animals Reelin signaling may be sufficiently robust to overcome effects of APP overexpression. We are currently examining the effects of APP overexpression on a mouse line with hypomorphic mutations at the Dab1 locus to determine the physiological consequences of APP expression on Dab1 activity.

Recently it has been demonstrated that overexpression of human APP, APLP1, and APLP2 causes phenotypes in Drosophila that are augmented by overexpression of and suppressed by reductions in Drosophila Dab (35). These phenotypes resemble Notch gain-of-function phenotypes in the mechano-sensory organ, and roles for Dab in Notch signaling have been proposed previously (14). Here we found an antagonistic relationship between APP family members and Dab1 function. This relationship could explain observations that despite continued expression of Reelin and the receptors in adult animals, Dab1 tyrosine phosphorylation is dramatically reduced after birth, at times when APP expression levels are increasing (20, 29, 38). In mouse, loss of Dab1 is associated with augmented Tau phosphorylation in some genetic backgrounds. Qualitative trait locus mapping identified a chromosomal region in the in vicinity of APP that correlates with increased Tau phosphorylation (7). It will be interesting to determine if APP and Dab1 interact genetically in mouse and how this influences Reelin signaling and Tau phosphorylation.

Partial rescue of the dominant Glued mutant phenotype by a tyrosine-phosphorylated Dab1 molecule was similar to the rescue by dominant mutations in khc (52). The tyrosine phosphorylation-defective mutant Dab1-5F did not functionally rescue. The most straightforward explanations of this are that Dab1 either acts to enhance Dynein motor function or acts to suppress kinesin function (Fig. 7), in a phosphorylation-dependent manner. It has previously been demonstrated that tyrosine-phosphorylated Dab1 binds to Lis1, an evolutionarily conserved regulatory component of dynein-dynactin motor complexes. It is possible that through the formation of a complex with Lis1, Dab1 is capable of activating the nuclear positioning function of dynein motor proteins.

FIG. 7.

A model for the role of Dab1 in nuclear placement. Enhancement of dynein-dynactin-related activity and suppression of Kinesin function are possible mechanisms by which tyrosine-phosphorylated Dab1 could act to partially rescue the Glued mutant phenotype.

A subset of neurons migrate through a mechanism of nuclear-somal translocation, while others have been demonstrated to move by locomotion. Nuclear movements are obviously regulated during nuclear-somal translocation, but they may also be critical to other modes of migration (51). It has recently been demonstrated that Lis1 and Dcx, two genes involved in human lissencephaly, play roles in the coupling of the nucleus to the centrosome (46). Failure to maintain this coupling appears to compromise the ability of neurons to migrate. In mouse mutants lacking various components of the Reelin signaling pathway, a subset of neurons, such as the Purkinje cells in the cerebellum, fail to migrate far enough to reach their final destinations. Reelin is most likely to regulate microtubule motor proteins through Dab1 in this subset of neurons. The regulation of nuclear positioning is a complex problem involving proteins conserved through evolution and likely influenced by a number of signaling events, including signaling that regulates Dab1 function.