Interaction of reelin signaling and Lis1 in brain development - Nature

1 downloads 0 Views 2MB Size Report
Oct 26, 2003 - reelin signaling pathway and Lis1. To explore genetic interactions between the Reln pathway and. Pafah1b1, we produced compound mutant ...
© 2003 Nature Publishing Group http://www.nature.com/naturegenetics

LETTERS

Interaction of reelin signaling and Lis1 in brain development Amir H Assadi1–3,12, Guangcheng Zhang1,2,12, Uwe Beffert4, Robert S McNeil1,2, Amy L Renfro1–3, Sanyong Niu1,2, Carlo C Quattrocchi1,2,5, Barbara A Antalffy6, Michael Sheldon2,7, Dawna D Armstrong6, Anthony Wynshaw-Boris8, Joachim Herz4, Gabriella D’Arcangelo1,2,9,10 & Gary D Clark1,2,9,11 Loss-of-function mutations in RELN (encoding reelin) or PAFAH1B1 (encoding LIS1) cause lissencephaly, a human neuronal migration disorder1. In the mouse, homozygous mutations in Reln result in the reeler phenotype, characterized by ataxia and disrupted cortical layers2. Pafah1b1+/– mice have hippocampal layering defects, whereas homozygous mutants are embryonic lethal3. Reln encodes an extracellular protein that regulates layer formation by interacting with VLDLR and ApoER2 (Lrp8) receptors4–6, thereby phosphorylating the Dab1 signaling molecule7–10. Lis1 associates with microtubules and modulates neuronal migration11. We investigated interactions between the reelin signaling pathway and Lis1 in brain development. Compound mutant mice with disruptions in the Reln pathway and heterozygous Pafah1b1 mutations had a higher incidence of hydrocephalus and enhanced cortical and hippocampal layering defects. Dab1 and Lis1 bound in a reelin-induced phosphorylation-dependent manner. These data indicate genetic and biochemical interaction between the reelin signaling pathway and Lis1. To explore genetic interactions between the Reln pathway and Pafah1b1, we produced compound mutant mice (Fig. 1) and observed a higher incidence of progressive hydrocephalus in combined mutants compared with Pafah1b1+/– mice. We successfully crossed Reln, Vldlr, Lrp8 or Dab1 knockout mice with Pafha1b1+/– mice and noted significantly higher incidences of hydrocephalus in these compound mutants. Notably, compound heterozygous Vldlr+/–Pafah1b1+/– and Dab1+/– Pafah1b1+/– mice also had a significantly higher incidence of hydrocephalus, whereas no progressive hydrocephalus was observed in mice that were null or heterozygous with respect to Vldlr or Dab1 alone. These observations were consistent with previous descriptions of localization of Dab1, ApoER2, VLDLR and LIS1 to ependymal precursors12–14. In the present study,

we confirmed VLDLR expression in the ependymal cells of the third ventricle and aqueduct. The progressive hydrocephalus phenotype in Pafah1b1+/– mice (ref. 3) or compound mutants was the result of an obstruction of the ventricular drainage system caused by defects in the ependymal lining of the third ventricle and the Aqueduct of Sylvius. Hydrocephalus was not observed in Reln mutants. These results indicate that mutations in the Reln pathway enhance the Pafah1b1 progressive hydrocephalus phenotype. Histochemical analysis of brains of Dab1 Pafah1b1 double mutant mice provided further evidence for genetic interactions of the Reln and Pafah1b1 signaling pathways in hippocampal and cortical layer formation (Fig. 2). We examined comparable hippocampal and cortical sections in these mice by choosing paramedian sagittal sections lateral to and caudal to the corpus callosum, as this structure was preserved in all mutants. As previously noted, Dab1 heterozygotes had no discernable hippocampal, cortical or cerebellar abnormalities7. Pafah1b1+/– mutants had mild hippocampal disorganization but only subtle cortical abnormalities3. The compound heterozygous Dab1+/– Pafah1b1+/– mutants, however, had more disruption of hippocampal lamination. The absence of hippocampal laminar defects in Dab1+/– mice and the greater laminar disruptions in double Dab1+/– Pafah1b1+/– heterozygotes compared with single Pafah1b1+/– heterozygotes point to an epistatic relationship of these genes. Hippocampi of Dab1–/– Pafah1b1+/– mice resembled those of Reln–/– Pafah1b1+/– double mutants (data not shown), with heterotopic clusters of horizontally oriented cells. Analysis of mutant cortical phenotypes further supported an epistatic relationship between Dab1 and Pafah1b1 (Fig. 2). We used C-Neu immunoreactivity to label the large pyramidal neurons of layer 5 (ref. 15) and calbindin immunoreactivity to label interneurons that normally populate layers 2 and 3 (ref. 4) in the adult cortex of siblings (except for Pafah1b1+/– mice, which were not siblings). C-Neu immunohistochemistry patterns in Dab1+/– and Pafah1b1+/– cortices

1Cain

Foundation Laboratories and 2Department of Pediatrics, Baylor College of Medicine, Houston, Texas 77030, USA. 3University of Texas School of Public Health, Houston, Texas, USA. 4Department of Molecular Genetics, University of Texas Southwestern Medical Center, Dallas, Texas, USA. 5Program in Neuroscience, Università degli Studi di Brescia, Italy. 6Department of Pathology and 7Division of Hematology-Oncology and Texas Children’s Cancer Center, Baylor College of Medicine, Houston, Texas, USA. 8Department of Pediatrics and Medicine, University of California San Diego, San Diego, California, USA. 9Division of Neuroscience, 10Program in Developmental Biology and 11Department of Neurology, Baylor College of Medicine, Houston, Texas, USA. 12These authors contributed equally to this work. Correspondence should be addressed to G.D.C. ([email protected]). Published online 26 October 2003; doi:10.1038/ng1257

270

VOLUME 35 | NUMBER 3 | NOVEMBER 2003 NATURE GENETICS

© 2003 Nature Publishing Group http://www.nature.com/naturegenetics

LETTERS resembled that observed in wild-type cortex. In Dab1+/– Pafah1b1+/– double heterozygotes, however, we observed a marked dispersion of layer 5 neurons and of calbindin-positive interneurons. The relative positional distributions of calbindin-positive cells in cortices from wild-type (data not shown), Dab1+/– and Pafah1b1+/– mice were similar. The distribution of calbindin-positive cells was shifted in cortices from Dab1+/– Pafah1b1+/– mice, however; we observed a considerable number of these cells in the lower cortex, which we interpreted as a partial inversion of cortex4. The distribution of both C-Neu-positive and calbindin-positive cells in cortices of Dab1–/– mice was severely disrupted, and Dab1–/– Pafah1b1+/– mice had similar disruptions. To determine whether the cortical phenotypes of the compound Dab1 Pafah1b1 heterozygotes were significantly different from those of the individual heterozygotes, we analyzed the cumulative distribution of calbindin-positive cells using the Kolmogorov-Smirnov two-sample test (Fig. 2). To validate this analysis, we compared cortices from two wild-type and two Pafah1b1+/– mice and observed no significant differences in probability distributions (P ≥ 0.05; data not shown). A similar analysis identified significant differences between Pafah1b1+/– and compound heterozygous mice and between Dab1+/– and compound heterozygous mice (Fig. 2; P < 0.001 that the distributions are the same). The distribution of calbindin-positive cells did not differ significantly between Dab1–/– and Dab1–/– Pafah1b1+/– mice (Fig. 2), suggesting that the Dab1-null phenotype predominates in the Dab1–/– Pafah1b1+/– combination. This result and similar results in Reln–/–

a

b

c

d

j

k

e

Pafah1b1+/– mice (data not shown) suggest that the disruption of Pafah1b1 did not enhance the Reln- or Dab1-null cortical phenotypes. We interpret the results of the cortical analyses to indicate a genetic interaction in cortical development between Pafah1b1 and Dab1. The enhancement of phenotypes associated with Pafah1b1+/– by Reln signaling disruptions and specifically by heterozygosity with respect to Dab1 prompted us to investigate potential biochemical interactions between Dab1 and Lis1 (Fig. 3). About 15% of hemagglutinin-tagged Dab1 (Dab1-HA) coimmunoprecipitated with green fluorescent protein–tagged Lis1 (GFP-Lis1) when these proteins were coexpressed in COS-7 cells. Reelin phosphorylates Dab1 on tyrosine residues 198 and 220 (refs. 9,10,16,17). To determine whether these residues are also important for Dab1 interaction with Lis1, we replaced them individually or in combination with phenylalanine. Mutant Dab1(Y198F) and Dab1(Y220F) retained interactions with Lis1, but the Dab1(Y198F/ Y220F) double mutant protein lost interaction with Lis1. We tested whether Pafah1b1 mutations that are known to cause lissencephaly in humans18 might also affect the interaction of Lis1 with Dab1. One mutant with a severe human phenotype, Lis1H149R, showed no interaction with Dab1. In contrast, a mutant associated with a more subtle phenotype, Lis1(S169P), did interact with Dab1. Mutant Lis1 previously used to disrupt the β-subunit G-protein-like structure of Lis1 (Lis1(S152W/V400W)18) did not impair the interaction of Dab1 and Lis1. Thus, disruption of Lis1 binding to Dab1 correlated with the severity of human disease.

f

g

h

i

m

l

NATURE GENETICS VOLUME 35 | NUMBER 3 | NOVEMBER 2003

Figure 1 Progressive hydrocephalus in combined mutants caused by periventricular and aqueductal defects. (a) Typical wild-type mouse and (b) Pafah1b1+/– mutant at P16 with a dome-shaped skull typical of the progressive hydrocephalus phenotype. (c–e) Hematoxylin and eosin staining of (c) a coronal section from a P16 Pafah1b1+/– mouse head with hydrocephalus and a coronal section of the brainstem at the level of the Aqueduct of Sylvius (arrow, A) from (d) a P0 wild-type mouse and (e) a Pafah1b1+/– mice (arrowheads indicate defects in the ependymal lining of the Aqueduct of Sylvius). (f–i) Coronal sections from (f) a P0 wild-type mouse, with the third ventricle (III) indicated by arrows (boxed area expanded in h), and (g) a P0 Pafah1b1+/– mouse. Arrowheads in g indicate a stenotic area devoid of ependyma, and the box (expanded in i) shows an area of poststenotic dilatation. (j,k) VLDLR immunoreactivity at the level of the third ventricle in (j) a wild-type mouse, showing localization of this receptor to ependyma, and (k) a Vldlr-null mouse. (l) VLDLR immunoreactivity at the level of the Aqueduct of Sylvius in a wild-type mouse. Scale bars = 100 µm (d,e,h,i,l) or 200 µm (f,g,j,k). (m) Development of progressive hydrocephalus in Pafah1b1+/– heterozygous mice lacking Reln, lipoprotein receptors or Dab1. Genotypes are indicated in the table below the bar chart. Asterisks indicate that the frequency of observations was statistically significant (P < 0.001) compared with the frequency of hydrocephalus in Pafah1b1+/– mice.

271

© 2003 Nature Publishing Group http://www.nature.com/naturegenetics

LETTERS

a

f

k

p

u

z

b

g

l

q

v

aa

c

h

m

r

w

bb

d

i

n

s

x

cc

e

j

o

t

y

dd

ee

ff

gg

Figure 2 Enhanced hippocampal and cortical malformations in Pafah1b1 Dab1 compound heterozygotes. Histochemical examination was done at P16 in siblings (except for Pafah1b1+/– mice) without apparent progressive hydrocephalus. (a–e) Cresyl violet–stained sagittal sections of P16 hippocampus at 4× magnification. DG, dentate gyrus; CA1–CA3, cornu ammonis 1–3. (f–j) Calbindin immunohistochemistry marks the dentate gyrus and mossy fiber projections from the dentate gyrus into CA3 regions of P16 hippocampus. Arrows in c and h indicate the disrupted dentate gyrus in Dab1+/– Pafah1b1+/– mice, and arrowheads indicate the more disturbed CA3 and mossy fiber projections in Dab1–/– Pafah1b1+/– double mutants. Scale bars = 250 µm (a,f) and apply to all hippocampal images. (k–y) Cresyl violet, C-Neu and Calbindin immunohistochemical stains of cortices of indicated compound mutants. C-Neu-positive cells (layer 5) were observed in a band, indicated by the bracket in p–t. (k,p,u) Dab1 heterozygotes had no obvious abnormal phenotypes, whereas the cortex of Pafah1b1 heterozygotes had a columnar appearance (l,q,v). (m,r,w) Cortical phenotypes of Dab1 and Pafah1b1 double heterozygotes were more severe than those of Pafah1b1 heterozygotes; the C-Neu staining was noticeably more dispersed and calbindin staining showed labeling in deeper layers. Dab1 homozygotes had a reeler phenotype in cortex (n,s,x). (o,t,y) Cortices of Dab1–/– Pafah1b1+/– double mutants resembled those of Reln Pafah1b1 double mutants (data not shown) with clustering of cells. The distribution of calbindin-positive neurons in Dab1–/– Pafah1b1+/– mice was inverted and was not significantly different from that in Dab1 homozygotes. Scale bars = 250 µm (k,p,u) and apply to all panels in the respective columns. (z–dd) Histograms plot the relative distance of calbindin-positive cells from the ventricle to the pial surface in 10% bins. (ee–gg) The distributions of calbindin-positive cells were converted to cumulative probability plots. The cumulative probability of observing cells less than or equal to the relative position on the abscissa were plotted for (ee) Pafah1b1+/– versus Pafah1b1+/– Dab1+/–, (ff) Dab1+/– versus Pafah1b1+/– Dab1+/– and (gg) Pafah1b1+/– Dab1–/– versus Dab1–/–.

272

VOLUME 35 | NUMBER 3 | NOVEMBER 2003 NATURE GENETICS

LETTERS

© 2003 Nature Publishing Group http://www.nature.com/naturegenetics

a

b

Figure 3 Biochemical interaction of Dab1 with Lis1. (a) Lis1 interacts specifically with Dab1 when Dab1 has been phosphorylated on Tyr198 and Tyr220, as induced by reelin signaling. Dab1-HA or mutant forms of Dab1-HA carrying phenylalanine substitutions at residue 198 (Dab1(Y198F)-HA), residue 220 (Dab1(Y220F)-HA) or both (Dab1(Y198F/Y220F)-HA) were coexpressed in COS-7 cells with GFP-Lis1. Immunoprecipitates of Dab1-HA or of mutant forms of Dab1-HA were immunoblotted with antibody to GFP and immunoprecipitates of GFP-Lis1 were immunoblotted with antibody to HA. The whole-cell lysates of each cotransfection were immunoblotted with antibody to phosphotyrosine (phospho-Y). (b) A mutation in Pafah1b1 that causes human lissencephaly results in loss of Dab1 binding. GFP-Lis1 or mutant forms of GFP-Lis1 with substitutions at residue 149 (GFP-Lis1(H149R), which has a severe human phenotype), residue 152 (GFP-Lis1(S152W)), residue 169 (GFP-Lis1(S169P), which has a subtle human phenotype) and residue 400 (GFP-Lis1(V400W)) were expressed in COS-7 cells. Dab1-HA immunoprecipitates were immunoblotted with antibody to GFP to detect coimmunoprecipitated GFP-Lis1. The whole-cell lysates of each transfection were immunoblotted with antibodies to either GFP or HA. IB, immunoblot; IP, immunoprecipitate; WCL, whole-cell lysate.

To further examine the role of phosphorylation in the interaction between Dab1 and Lis1, we produced wild-type Dab1-HA and mutant Dab1(Y198F/Y220F)-HA in a cell-free expression system with FLAG-Lis1 either with or without recombinant Src kinase (Fig. 4). Immunoprecipitation with antibody to FLAG showed minimal interaction of Dab1 and Lis1 in the absence of Src kinase, but their binding was strongly increased in the presence of Src.

Binding was markedly decreased for the double mutant Dab1(Y198F/Y220F)-HA but not for the single mutant proteins. We reasoned that Dab1 and Lis1 might not interact in Reln–/– brain. We readily observed coimmunoprecipitation of native Lis1 and Dab1 in brain lysates from wild-type and Reln+/– mice, where ∼11% of Dab1 was involved in this complex, but no interaction in brain lysates from Reln–/– mice. As previously reported9, Dab1 was

a

c

b

d

Figure 4 Reelin-induced tyrosine phosphorylation of Dab1 is required for Lis1-Dab1 binding. (a) In vitro translated wild-type or mutant Dab1-HA constructs were incubated with FLAG-Lis1 purified from 293T cells, phosphorylated using Src and coimmunoprecipitated with antibody to FLAG. The phosphorylated proteins were separated by SDS-PAGE, transferred to nitrocellulose membrane and immunoblotted with antibody to HA. The individual protein inputs from each transfection were immunoblotted with antibody to FLAG to detect coprecipitated tagged Lis1. (b) In vitro translated wild-type of mutant Dab1-HA constructs were phosphorylated using Src and immunoblotted with antibody to phosphotyrosine (phospho-Y). All proteins were phosphorylated by Src. (c,d) Brain lysates of E16 wild-type, Reln heterozygous and Reln homozygous mutant mice were either (c) immunoprecipitated with antibody to Dab1 and immunoblotted with antibody to Lis1 or (d) immunoprecipitated with antibody to Lis1 and immunoblotted with antibody to Dab1. IB, immunoblot; IP, immunoprecipitate; WCL, whole-cell lysate.

NATURE GENETICS VOLUME 35 | NUMBER 3 | NOVEMBER 2003

273

© 2003 Nature Publishing Group http://www.nature.com/naturegenetics

LETTERS Figure 5 Cellular colocalization of Lis1 and Dab1. (a–c) Laser confocal image of a cortical neuron from E16 mouse brain immunostained for Lis1 (green) and for Dab1 (red) with overlay. (d) An image of cortical neurons from Dab1 knockout mouse with DAPI counterstain showing lack of immunoreactivity to Dab1 under the same conditions as a–c. (e–l) Expressed fusion proteins in COS-7 cells; expression of protein is designated by green or red lettering. COS-7 cells expressing only Dab1 showed predominant membrane localization of protein, whereas cells expressing Lis1 showed typical perinuclear distribution (f). (g–i) Predominate colocalization of Lis1 and Dab1 observed in the perinuclear region. Similarly, Lis1 showed some membrane colocalization with Dab1 (j–l). Laser confocal images showed colocalization of Dab1 and Lis1 in a single plane. Scale bars = 20 µm.

a

b

c

d

e

f

g

h

i

k

l

overexpressed in Reln–/– mice, whereas Lis1 was expressed at normal levels, suggesting that the lack of interaction between these j proteins does not result from loss of expression. We examined Dab1 phosphorylation in Pafah1b1 heterozygotes compared with wild-type mice and observed no change in Dab1 phosphorylation (data not shown). To further confirm binding between Dab1 and Lis1, we carried out laser confocal analysis of hippocampal neurons doubly labeled with antibodies to Dab1 and Lis1 (Fig. 5). Examination of single optical sections indicated that the two proteins colocalized in the perinuclear cytoplasm. This localization of Lis1 was similar to that reported previously for Lis1, dynein and NudC19. In COS-7 cells transfected with single plasmids, Dab1 was expressed primarily at the plasma membrane whereas Lis1 was expressed in a perinuclear distribution. When the two genes were coexpressed, however, Dab1 also appeared in a perinuclear location, overlapping the Lis1 signal. Confocal analysis of doubly transfected cells in a single optical plane confirmed this finding. These data suggest that Dab1 was partially pulled to a Lis1-rich compartment, further indicating interactions between these proteins. Enhancement of the Pafah1b1+/– hydrocephalus phenotype by reelin signaling disruptions and the more disturbed lamination of hippocampus and cortex in Dab1+/– Pafah1b1+/– mice give evidence for epistatic relationships between Reln signaling and Pafah1b1. The

biochemical interaction of Dab1 with Lis1 is the most probable mechanism by which these genetic interactions occur. Reelin signaling converges on Dab1 and results in tyrosine phosphorylation on residues 198 and 220; this leads to the interaction between phosphorylated Dab1 and Lis1.

Figure 6 Schematic summary of reelin and Lis1 signaling. Reelin binding to lipoprotein receptors results in Dab1 phosphorylation by Src family kinases (SFK) and interaction of Dab1 with Lis1. The data presented here are consistent with either a sequential activation of Dab1 and Lis1 or a significant cross-talk between these proteins in the regulation of ependymal cell placement and in the formation of laminar structures in brain. We cannot exclude the possibility that Lis1 is a downstream effector of Dab1 and that Lis1 is in the reelin signaling pathway. We also cannot exclude the possibility that Lis1 and Dab1 are in parallel pathways that are closely coordinated through the interactions uncovered here. Phosphorylated Dab1 additionally interacts with PI3-kinase28, whereas Lis1 functionally interacts with the α subunits of the Pafah1b complex, Pafah1b2 (α2) and Pafah1b3 (α1)18,29. Lis1 also interacts with Nudel/cytoplasmic dynein complex24,26,30, NudC19 and mNude23 to regulate lamination in the central nervous system (CNS).

274

VOLUME 35 | NUMBER 3 | NOVEMBER 2003 NATURE GENETICS

© 2003 Nature Publishing Group http://www.nature.com/naturegenetics

LETTERS Because homozygous deficiency of Pafah1b1 is lethal early during embryogenesis3, it is not possible with the tools currently available to ascertain whether Lis1 is an obligate component of the reelin pathway that functions downstream of Dab1 or is a component of just one of several branches that constitute the intracellular response to reelin. But the similarity of the lissencephaly phenotype in human cortices affected by PAFAH1B1 and RELN mutations suggests that LIS1 may be required for regulating crucial steps of reelindependent neuronal positioning. Tyrosine-phosphorylated Dab1 probably acts through additional signaling partners, such as the microtubule-associated protein Tau and PI3K, to coordinate Reln signaling and microtubule-dependent transport6,20 (Fig. 6). Similarly, Lis1 may exert its additional functions through interactions with dynein21–26. We propose that reelin and Lis1 are integrated components of a larger coordinated signaling system that regulates neuronal migration, nuclear distribution and cortical layering and maintains the integrity of the ependymal lining of the ventricles in the mammalian brain. METHODS

Mice. Reeler mice were on a C57BL/6 × C3H background; Vldlr, Lrp8 and Dab1 knockout mice were on a hybrid C57BL/6 × 129S6/SvEv background; and Pafah1b1 mutants (we used Pafah1b1neo, a null allele3) were on a 129S6/SvEv background at the beginning of matings. The mice used were produced over a period of three years. We genotyped Pafah1b1 by PCR as described previously3 and genotyped Lrp8, Vldlr, Dab1 and Reln by PCR as described4,7,27. Immunohistochemistry and brain histology. We fixed cells in 4% paraformaldehyde, rinsed them in phosphate-buffered saline, blocked them and then incubated them overnight at 4 °C with rabbit antibody to Dab1 CT-38 (gift from T. Curran, St. Jude Children’s Research Hospital, Memphis, Tennessee, USA; diluted 1:500). After washing them, we incubated the cells with Alexa Fluor 594 (Molecular Probes) and DAPI for 1.5 h at ambient temperature. For Lis1 staining, we used a goat polyclonal antibody (Santa Cruz; diluted 1:1,000) and a secondary antibody conjugated to fluorescein isothiocyanate. For brain histology, we dissected the brain tissues, perfused them with 4% paraformaldehyde, embedded them in paraffin, sectioned them (5 µm) and either stained them with hematoxylin and eosin or with cresyl violet or labeled them with antibody to calbindin4 or to C-Neu as described15. Quantitative analysis of the neocortex in mutant mice. We labeled four paramedian sagittal sections caudal to the corpus callosum from siblings of each genotype with calbindin. We used Metamorph software (Universal Imaging) to threshold, count and generate a positional (y) coordinate for each immunopositive cell. We normalized these coordinates to the total cortical distance and expressed the values as percentage distance from ventricle (y = 0) to the pia (y = 100.0) surface. We expressed the data as cumulative probability plots and used the Kolmogorov-Smirnov two-sample statistic (SSPS software) to determine whether two distributions were significantly different. The Kolmogorov-Smirnov statistic measures the maximum vertical distance (D) between two cumulative probability distributions of m and n number of observations, respectively. We used P < 0.05 to determine whether two distributions were significantly different and P ≥ 0.05 to conclude that two distributions were similar. Cell culture and transfections. We cultured 293T cells or COS-7 cells (American Type Culture Collection) in the presence of Dulbecco’s modified Eagle medium containing 10% fetal bovine serum and penicillin/streptomycin at 37 °C, 5% CO2. We transfected cells with FuGENE 6 (Roche) and cultured them for 24–48 h. We prepared primary cultures of hippocampal neurons from embryonic day (E)-18 mice and plated them on coverslips coated with poly-Dlysine. We then transferred the coverslips to a dish containing a neuronal/glial monolayer and cultured them for 5 d in B27/Neurobasal A medium.

NATURE GENETICS VOLUME 35 | NUMBER 3 | NOVEMBER 2003

Immunoprecipitation and western-blot analysis. We extracted proteins from transfected cells and from brain tissues of E16 mice in lysis buffer (1× phosphate-buffered saline, 5 mM EDTA, 1% Triton X-100 and protease inhibitor cocktail tablet (Roche)). We immunoprecipitated fusion constructs from cells with antibodies to HA (Sigma) or GFP (Clontech) and Protein A/G agarose beads (Pierce). We used antibodies to Dab1 (Novus) or Lis1 (Santa Cruz) to immunoprecipitate native proteins from brain lysates. We separated immunoprecipitated proteins by SDS-PAGE, immunoblotted them and incubated them with the appropriate primary antibodies overnight at 4 °C and then with the corresponding horseradish peroxidase–conjugated secondary antibodies for 1–2 h at room temperature. Proteins were detected using ECL chemiluminescence (Amersham). Generation of the mutant Pafah1b1 and Dab1 constructs. We carried out site-directed mutagenesis with the Stratagene Quickchange Site-Directed Mutagenesis Kit using mouse Pafah1b1 cDNA subcloned into pEGFP-C3 (Clontech) or pCMV-Tag-2 (FLAG; Stratagene). We modified the mouse Dab1 cDNA by PCR to express an HA tag and subcloned it into pcDNA3.1(+) (Invitrogen). Primer sequences are available on request. Dab1 in vitro expression and binding. We carried out in vitro translation of cDNAs encoding wild-type Dab1-HA, Dab1(Y198F)-HA, Dab1(Y220F)-HA and Dab1(Y198F/Y220F)-HA using the TNT Quick Couple Transcription/ Translation System (Promega). We added antibody to HA for immunoprecipitation as described above. Different Dab1-HA constructs were phosphorylated using the Src Assay Kit (Upstate). We then eluted the proteins from the beads with 0.1 M glycine, pH 2.5 (diluted 1:40), and mixed them with purified FLAGLis1 protein beads for 1–2 h. We separated the coimmunoprecipitated proteins by SDS-PAGE, immunoblotted them and detected them as described above. Fluorescence microscopy and confocal imaging. We visualized protein expression patterns using a Nikon Eclipse TE 200 inverted microscope equipped with phase-contrast and fluorescence settings (Nikon). We acquired images with a Photometrics CoolSnap CCD digital camera (Roper Scientific) and analyzed them using Metaview Imaging software (Universal Imaging Corporation). We acquired confocal images of COS-7 cells transfected with constructs encoding DsRed-Dab1 and EGFP-Lis1 (GFP-Lis1) and of immunolabeled neurons using a Fluoview FV300 confocal laser scanning unit mounted on a BX50WI fixed-stage upright microscope and Fluoview software (all from Olympus America). We acquired GFP images by excitation with an argon laser (488 nm) and a 510–550-nm band-pass emission filter set. We acquired DsRed images by excitation with a krypton laser (568 nm line) and a 585-nm longpass emission filter set. Image noise reduction was accomplished using a Kallman accumulation average setting of 4. We acquired eight Z-series images for each cell sequentially for each wavelength at 0.5-µm increments. We processed maximum projection images of Z-series stacks using both Fluoview and Photoshop 7.0 (Adobe Systems) software. ACKNOWLEDGMENTS The authors thank C. Walsh, G. Eichele and H. Zoghbi for critical reading of the manuscript; T. Curran for the gift of many reagents; J. Hayes and W.-L. Niu for technical assistance. This study was supported by grants from the US National Institutes of Health (to G.D.C., G.D., A.W.-B., J.H.), the Alzheimer Association and the Humboldt Foundation (to J.H.). U.B. was a fellow of the Human Frontier Science Program during part of this work. COMPETING INTERESTS STATEMENT The authors declare that they have no competing financial interests. Received 18 May; accepted 29 September 2003 Published online at http://www.nature.com/naturegenetics/

1. Reiner, O. et al. Isolation of a Miller-Dieker lissencephaly gene containing G protein β-subunit-like repeats. Nature 364, 717–721 (1993). 2. D’Arcangelo, G. et al. A protein related to extracellular matrix proteins deleted in the mouse reeler. Nature 374, 719–723 (1995). 3. Hirotsune, S. et al. Graded reduction of Pafah1b1 (Lis1) activity results in neuronal migration defects and early embryonic lethality. Nat. Genet. 19, 333–339 (1998). 4. Trommsdorff, M. et al. Reeler/Disabled-like disruption of neuronal migration in knockout mice lacking the VLDL receptor and ApoE receptor 2. Cell 97, 689–701 (1999).

275

© 2003 Nature Publishing Group http://www.nature.com/naturegenetics

LETTERS 5. D’Arcangelo, G. et al. Reelin is a ligand for lipoprotein receptors. Neuron 24, 471–479 (1999). 6. Hiesberger, T. et al. Direct binding of reelin to VLDL receptor and ApoE receptor 2 induces tyrosine phosphorylation of Disabled-1 and modulates Tau phosphorylation. Neuron 24, 481–489 (1999). 7. Howell, B.W., Hawkes, R., Soriano, P. & Cooper, J. Neuronal position in the developing brain is regulated by mouse disabled-1. Nature 389, 733–737 (1997). 8. Sheldon, M. et al. Scrambler and yotari disrupt the disabled gene and produce a reeler-like phenotype in mice. Nature 389, 730–733 (1997). 9. Howell, B.W., Herrick, T.M. & Cooper, J.A. Reelin-induced phosphorylation of Disabled 1 during neuronal positioning. Genes Dev. 13, 643–648 (1999). 10. Keshvara, L., Benhayon, D., Magdaleno, S. & Curran, T. Identification of Reelininduced sites of tyrosyl phosphorylation on Disabled 1. J. Biol. Chem. 276, 16008–16014 (2001). 11. Sapir, T., Elbaum, M. & Reiner, O. Reduction of microtubule catastrophe events by LIS1, platelet-activating factor acetylhydrolase subunit. EMBO J. 16, 6977–6984 (1998). 12. Luque, J.M., Morante-Oria, J. & Fairen, A. Localization of ApoER2, VLDLR and Dab1 in radial glia: groundwork for a new model of reelin action during cortical development. Brain Res. Dev. Brain Res. 140, 195–203 (2003). 13. Tramontin, A.D., Garcia-Verdugo, J.M., Lim, D.A. & Alvarez-Buylla, A. Postnatal development of radial glia and the ventricular zone (VZ): a continuum of the neural stem cell compartment. Cereb. Cortex 13, 580–587 (2003). 14. Clark, G.D., Mizuguchi, M., Antalffy, B., Barnes, J. & Armstrong, D. Predominant localization of the LIS family of gene products to Cajal-Retzius cells and ventricular neuroepithelium in the developing human cortex. J. Neuropathol. Exp. Neurol. 56, 1044–1052 (1997). 15. Miller, M.W. & Pitts, F.A. Neurotrophin receptors in the somatosensory cortex of the mature rat: co-localization of p75, trk, isoforms and c-neu. Brain Res. 852, 355–366 (2000). 16. Howell, B.W., Herrick, T.M., Hildebrand, J.D., Zhang, Y. & Cooper, J.A. Dab1 tyrosine phosphorylation sites relay positional signals during mouse brain development. Curr. Biol. 10, 877–885 (2000). 17. Bock, H.H. & Herz, J. Reelin activates Src family tyrosine kinases in neurons. Curr.

276

Biol. 13, 18–26 (2003). 18. Sweeney, K.J., Clark, G.D., Prokscha, A., Dobyns, W.B. & Eichele, G. Lissencephaly associated mutations suggest a requirement for the PAFAH1B heterotrimeric complex in brain development. Mech. Dev. 92, 263–271 (2000). 19. Aumais, J.P. et al. NudC is highly expressed during neurogenesis and associates with Lis1 and the dynein motor at the leading edge of migrating neurons. J. Neurosci. 21, 1–7 (2001). 20. Brich, J. et al. Genetic modulation of tau phosphorylation in the mouse. J. Neurosci. 23, 187–192 (2003). 21. Kitagawa, M. et al. Direct association of LIS1, the lissencephaly gene product, with a mammalian homologue of a fungal nuclear distribution protein, rNUDE. FEBS Lett. 479, 57–62 (2000). 22. Efimov, V.P. & Morris, N.R. The LIS1-related NUDF protein of Aspergillus nidulans interacts with the coiled-coil domain of the NUDE/RO11 protein. J. Cell Biol. 150, 681–688 (2000). 23. Feng, Y. et al. LIS1 regulates CNS lamination by interacting with mNudE, a central component of the centrosome. Neuron 28, 665–679 (2000). 24. Niethammer, M. et al. Nudel is a novel Cdk5 substrate that associates with Lis1 and cytoplasmic dynein. Neuron 28, 697–711 (2000). 25. Wynshaw-Boris, A. & Gambello, M.J. LIS1 and dynein motor function in neuronal migration and development. Genes Dev. 15, 639–651 (2001). 26. Sweeney, K.J., Prokscha, A. & Eichele, G. NudE-L, a novel Lis1-interacting protein, belongs to a family of vertebrate coiled-coil proteins. Mech. Dev. 101, 21–33 (2001). 27. D’Arcangelo, G., Miao, G.G. & Curran, T. Detection of the reelin breakpoint in reeler mice. Brain Res. Mol. Brain Res. 39, 234–236 (1996). 28. Bock, H.H. et al. PI3-Kinase interacts with the adaptor protein Dab1 in response to Reelin signaling and is required for normal cortical lamination. J. Biol. Chem. 278, 38772–38779 (2003). 29. Yan, W. et al. Previously uncharacterized roles of platelet-activating factor acetylhydrolase 1b complex in mouse spermatogenesis. Proc. Natl. Acad. Sci. USA 100, 7189–7194 (2003). 30. Sasaki, S. et al. A LIS1/NUDEL/cytoplasmic dynein heavy chain complex in the developing and adult nervous system. Neuron 28, 681–696 (2000).

VOLUME 35 | NUMBER 3 | NOVEMBER 2003 NATURE GENETICS