Reelin Activates Src Family Tyrosine Kinases in Neurons

CORE Metadata, citation and similar papers at core.ac.uk

Provided by Elsevier - Publisher Connector Current Biology, Vol. 13, 18–26, January 8, 2003, 2003 Elsevier Science Ltd. All rights reserved. PII S0960-9822(02)01403-3 Reelin Activates Src Family Tyrosine Kinases in Neurons

Hans H. Bock* and Joachim Herz* phorylation of Dab1 [3, 4], which is essential for trans- Department of Molecular Genetics mission of the positional cue that is conveyed by Reelin University of Texas Southwestern Medical Center to the migrating cell. Dallas, Texas 75390 The nature of the tyrosine kinase or kinases that mediate the phosphorylation of Dab1 has not been ascer- tained. We have previously proposed a model in which Summary a coreceptor, which could either be a transmembrane tyrosine kinase or a membrane protein that can recruit Background: Reelin is a large signaling molecule that a non-receptor tyrosine kinase to the membrane, also regulates the positioning of neurons in the mammalian interacts with Reelin and is thereby brought into proxim- brain. Transmission of the Reelin signal to migrating ity to Dab1 [5]. Cadherin-related neuronal receptors embryonic neurons requires binding to the very-low- (CNRs) [6] and members of the integrin family of adhe- density lipoprotein receptor (VLDLR) and the apolipo- sion proteins [7] have been suggested as such core- protein E receptor-2 (apoER2). This induces tyrosine ceptors, but further genetic and biochemical data are phosphorylation of the adaptor protein Disabled-1 required to establish the respective roles of these recep- (Dab1), which interacts with a shared sequence motif in tors in this signaling pathway. the cytoplasmic tails of both receptors. However, the A role for Src family kinases (SFKs) in Reelin signaling kinases that mediate Dab1 tyrosine phosphorylation and has been suggested by several lines of evidence. First, the intracellular pathways that are triggered by this event the mammalian form of Dab1 was originally identified remain unknown. as a Src-interacting protein in a yeast two-hybrid screen Results: We show that Reelin activates members of the by Howell and colleagues [8]. Second, mice that are Src family of non-receptor tyrosine kinases (SFKs) and genetically deficient in the SFK Fyn show neuronal posi- that this activation is dependent on the Reelin receptors tioning defects in the hippocampal CA3 region and the apoER2 and VLDLR and the adaptor protein Dab1. Dab1 dentate gyrus [9, 10]. Third, Fyn has been reported to is tyrosine phosphorylated by SFKs, and the kinases bind to the cytoplasmic tails of the CNRs [11], and, themselves can be further activated by phosphorylated fourth, integrins, which have been shown to be present Dab1. Increased Dab1 protein expression in fyn-defi- in a complex with the Reelin receptors VLDLR and cient mice implies a response to impaired Reelin signal- apoER2 [7], have also been recognized as activators of ing that is also observed in mice lacking Reelin or its SFKs [12]. receptors. However, fyn deficiency alone does not com- Here, we have investigated the role of SFKs in Reelin pound the neuronal positioning defect of vldlr-or signaling. Our results demonstrate that Reelin activates apoer2-deficient mice, and this finding suggests func- SFKs through a mechanism that depends on the Reelin tional compensation by other SFKs. receptors VLDLR and apoER2 and the adaptor protein Conclusions: Our results show that Dab1 is a physio- Dab1. Efficient tyrosine phosphorylation of Dab1 re- logical substrate as well as an activator of SFKs in neu- quires SFK activity, and SFKs can undergo autoactiva- rons. Based on genetic evidence gained from multiple tion in the presence of phosphorylated Dab1. Upregu- strains of mutant mice with defects in Reelin signaling, lation of Dab1 expression in fyn-deficient neurons is we conclude that activation of SFKs is a normal part of consistent with a response mechanism by which the cell the cellular Reelin response. attempts to compensate for reduced kinase availability. Taken together, these results suggest that SFKs func- Introduction tion as intrinsic components of the Reelin signaling pathway. The proper positioning of neurons in the CNS depends on a signaling pathway that involves the large secreted signaling protein Reelin, at least two receptors for this Results and Discussion protein on the surface of migrating neurons, and the cytosolic adaptor protein Disabled-1 (Dab1) [1, 2]. Two Activation of Src Family Kinases by Dab1 members of the low-density lipoprotein (LDL) receptor It has been shown that Src family kinases (SFKs) can family of apolipoprotein E (ApoE) receptors, the very- phosphorylate Dab1 in vitro and can interact with tyro- low-density lipoprotein receptor (VLDLR) and the ApoE sine-phosphorylated Dab1 through their SH2 domain [4, receptor-2 (apoER2), have been shown to function as 8, 13]. We questioned whether Dab1 could activate SFKs Reelin receptors. Dab1 interacts with a conserved Asn- by established mechanisms, i.e., either by displacement Pro-X-Tyr tetra amino acid motif (NPxY, where x repre- of the inhibitory phosphorylated tyrosine residue 529 sents any amino acid) in the cytoplasmic domain of both from the SH2 domain or by an SH3 domain-mediated receptors. Binding of Reelin to the extracellular domains interaction with proline-rich sequences [14]. To test this, of VLDLR and apoER2 rapidly induces tyrosine phos- we transfected HEK-293 cells with expression plasmids for c-Src (Figure 1A, lane 1) and Dab1 (lane 3) alone or *Correspondence: [email protected]; joachim.herz@ with both plasmids (lane 2). Simultaneous expression utsouthwestern.edu of both proteins led not only to increased tyrosine phos- Reelin Activates Src Family Tyrosine Kinases 19

Figure 1. Activation of SFKs by Coexpression with Dab1 in Transfected Cells (A) HEK-293 cells were transfected with plasmids containing c-Src (lane 1), Dab1 (lane 3), or were cotransfected with c-Src and Dab1 (lane 2). Tyrosine-phosphorylated proteins (5 ␮g total protein/lane) were detected by immunoblotting with an ␣-phosphotyrosine antibody (4G10). Erk served as a loading control. (B) An analogous experiment was performed with c-Fyn instead of c-Src. Immunoblotting with an antibody directed against a conserved phosphotyrosine residue in the activation domain of SFKs (␣-pY418) was used to monitor Dab1-dependent Fyn activation. Cdk5 served as a loading control. phorylation of Dab1, but also to a dramatic overall in- Dab1 in cultured cortical neurons, we were able to moni- crease in total cellular Src kinase activity. In contrast, tor Reelin signaling by direct immunoblotting of crude expression of recombinant c-Src alone only modestly cell lysates from wild-type embryonic neurons with an increased cellular phosphotyrosine levels (lane 1). In- anti-phosphotyrosine antibody without the need for creased tyrosine kinase activity was also observed when prior immunoprecipitation (see Figure S1 in the Supple- Dab1 and another Src family kinase, c-Fyn, were coex- mentary Material available with this article online). Using pressed in HEK-293 cells (Figure 1B, lane 3). This corre- this simplified assay, Reelin-induced tyrosine phosphor- -fold increase in phosphorylation of a ylation of Dab1 was observed not only in cortical neu-4ف lated with an tyrosine residue in the activation loop. This residue is rons from mouse embryos, as had been previously dem- conserved in all SFK members (referred to as Y418, its onstrated [3], but also in neurons that had been prepared position in human Src) and stimulates SFK activity in the from newborn (P0.5) pups (Figure S1); this finding sug- phosphorylated state (reviewed in [14]). These results gests that the Reelin signaling pathway remains func- suggest that Dab1 and SFKs can mutually activate each tional beyond the birth period of neurons during embry- other if the local concentration exceeds a critical thresh- onic development. Reelin expression continues in the old. Physiologically, this could be achieved by ligand- adult brain [16], where it was recently shown to modulate induced clustering at the cell surface, for instance, by synaptic plasticity [17]. Taken together, these findings oligomerized Reelin [15]. suggest functions of this signaling pathway that go beyond the regulation of cortical lamination. Treatment of cortical neurons with Reelin induced not Reelin-Induced Activation of SFKs in Primary only tyrosine phosphorylation of Dab1 (Figure 2A, upper fold, but also the activation of SFKs by-7ف Cortical Neurons panel) by Tyrosine phosphorylation of Dab1 has been shown to increasing the level of tyrosine phosphorylation at the fold. In contrast, no-4ف represent a key step in Reelin signaling and is required activating Y418 residue by for the physiological function of Reelin during brain de- change in phosphorylation levels was observed at the velopment [1, 3, 13]. As it had previously been shown inhibitory autophosphorylation site of Src (Y529 in hu- that tyrosine-phosphorylated Dab1 interacts with the man; Y534 in mouse Src). The phosphorylation state of Src SH2 domain [8], and coexpression of Dab1 and SFKs the main activating residue (Y397) of focal adhesion in cultured cells greatly enhances overall tyrosine kinase kinase (FAK), another non-receptor tyrosine kinase that activity, we wanted to investigate a possible involve- mediates signaling by integrins and thereby regulates ment of SFKs in Reelin signaling in primary neurons. By cellular adhesion and migration [18], was determined as improving the sensitivity and specificity of the biochemi- a control and did not change upon Reelin stimulation. cal assay developed by Howell and collegues [3], which Total cellular levels of c-Src, FAK, and the serine/threo- detects Reelin-induced tyrosine phosphorylation of nine kinases Cdk5 and Akt also remained unaffected. Current Biology 20

Figure 2. Reelin-Induced Activation of SFKs in Primary Embryonic Cortical Neurons (A) Immunoblot analysis of crude cell lysates from neurons treated with control medium or Reelin. Reelin-induced Dab1 tyrosine phosphorylation was detected with the 4G10 antibody (also see Figure S1). Activating phosphorylation of SFKs in response to Reelin was revealed by using ␣-pY418. No change in the phosphorylation state at the inhibitory tyrosine residue (Y529) in Src was observed. Phosphorylation at the major activation site (Y397) of another non-receptor tyrosine kinase and Src substrate, focal adhesion kinase (FAK), as well as total levels of c-Src, FAK, and the serine/threonine kinases Cdk5 and Akt/PKB remained unchanged. (B) Fyn is tyrosine phosphorylated in response to Reelin in primary cortical neurons. Fyn was immunoprecipitated from total cell lysates of neurons treated with control medium or with Reelin. Precipitated proteins were immunoblotted with 4G10 (anti-phosphotyrosine, upper panel). Increased overall tyrosine phosphorylation of Fyn in Reelin- treated neurons was paralleled by increased phosphorylation in the activation loop (pY418). Equal precipitation and loading was verified by immunoblotting with ␣-Fyn (lower panel).

Since the Y418 epitope is highly conserved between are overexpressed in cultured cells (Figure 1). Reelin the different SFKs, this antibody did not allow us to stimulation of cultured neurons resulted in phosphoryla- distinguish which family members were specifically acti- tion of Dab1 and activation of SFKs (Figure 2). To investi- vated by the Reelin signal. Circumstantial evidence sug- gate whether Dab1 might also be a substrate for SFKs in gested a particular role for the SFK Fyn in the Reelin vivo, we used a panel of commercially available tyrosine pathway. Fyn is expressed in regions of the developing kinase inhibitors (Figure S2). The pyrazolopyrimidines brain that are affected in the Reelin-deficient reeler mice PP1 and PP2, two tyrosine kinase inhibitors with the and enriched in embryonic growth cones [19], the pre- highest relative selectivity for SFKs [21], but not the sumptive Reelin-responsive site of migrating neurons structurally related inactive compound PP3, blocked (our unpublished data), together with other SFKs [20]. Reelin-induced Dab1 tyrosine phosphorylation in a Furthermore, Fyn has been reported to interact with the dose-dependent manner. In contrast, Herbimycin A and cytoplasmic tails of putative Reelin coreceptors [6, 7]. STI 571 (Gleevec), both inhibitors of c-Abl as well as Moreover, mutant mice lacking Fyn display structural several other tyrosine kinases, were ineffective (Figure abnormalities in the hippocampus that resemble, but S2). This result argues against a major role of c-Abl in are less severe than, those seen in reeler mice [9, 10]. Dab1 tyrosine phosphorylation but does not exclude a Immunoprecipitation of Fyn and immunoblot analysis downstream function of c-Abl. with the nonselective anti-phosphotyrosine antibody Because none of the commercially available inhibitors 4G10 (Figure 2B, upper panel) and with the activation including PP1 and PP2 are completely selective for any loop-specific anti-pY418 antibody (middle panel) re- specific tyrosine kinase, these results are consistent vealed that Fyn was indeed tyrosine phosphorylated in with, but do not in themselves prove, a direct role of SFKs response to Reelin treatment (10-fold and 3-fold in- in Reelin-mediated signaling and Dab1 phosphorylation crease, respectively). Fyn immunoprecipitation from in vivo. Reelin-treated and control samples was quantitative and We thus sought to obtain genetic evidence to support identical (lower panel). a role of SFKs in Reelin-induced signaling in vivo. Mice that lack only one of several SFKs show none of the Pharmacological Inhibition of Dab1 Tyrosine gross cortical lamination defects that are characteristic Phosphorylation and Effect of Fyn Deficiency for mutants with defects in the Reelin signaling pathway on Reelin Signaling in Cortical Neurons [22]. An exception are fyn-deficient mice that show a Our results so far had suggested that Dab1 can be both subtle malformation of the dentate gyrus [9, 10]. We a substrate and an activator of SFKs when both proteins therefore focused our attention on the fyn mutants and Reelin Activates Src Family Tyrosine Kinases 21

Figure 3. Tyrosine Phosphorylation of Dab1 by Reelin Is Not Impaired in Fyn-Deficient Neurons (A) Total cellular proteins from cortical embryonic neurons containing two (lane 5), one (lanes 2 and 3), or no (lanes 1 and 4) functional fyn alleles were separated by SDS gelelectrophoresis and were analyzed by immunoblotting with the indicated antibodies. The relative amount of Fyn immunoreactivity in the samples correlates with the number of intact fyn alleles. Total cellular amounts of VLDLR, apoER2, FAK, and Cdk5 remain unchanged irrespective of the genotype, whereas Dab1 protein levels are significantly increased in fyn-deficient neurons. (B) Cortical neurons from fyn-deficient (lanes 1 and 2) and wild-type (lanes 3 and 4) embryos were incubated with control medium (lanes 1 and 3) or Reelin (lanes 2 and 4). Endogenously produced Reelin in the control samples causes the low background phosphorylation of Dab1. Reelin-induced Dab1 tyrosine phosphorylation occurs normally in the absence of Fyn (lane 2), but phosphorylation of the activation domain of neuronal SFKs (␣-pY418) is greatly reduced. Expression of endogenous Src is slightly reduced in Fyn-deficient neurons (lanes 1 and 2), while Yes remains unchanged. Cdk5 served as a loading control. investigated whether the loss of Fyn had any measurable of the typical neuroanatomical malformations that occur effect on the known components of the Reelin signaling in mice in which the Reelin signaling pathway is com- pathway. pletely disrupted [1, 23]. However, Reelin-induced phos- Figure 3A shows an immunoblot analysis of neocorti- phorylation at the activating tyrosine residue of SFKs cal brain extracts from wild-type, fynϩ/Ϫ, and fynϪ/Ϫ em- was reduced by almost 5-fold in Fyn-deficient neurons fold in homo- (compare lanes 2 and 4). Yet the relative, but not the-6ف bryos. Dab1 expression was increased fold absolute, increase in Y418 phosphorylation upon Reelin-2ف zygous fyn-deficient mice (lanes 1 and 4) and fold increase) was at least-7ف ,in heterozygotes (lanes 2 and 3) compared to wild-type stimulation (lanes 1 and 2 animals (lane 5). Intriguingly, there was also a small, as robust as that seen in wild-type neurons (lanes 3 and fold increase). These findings are also consistent-3ف ,fold), increase in the smaller 4-2ف but reproducible (by unglycoslylated form of one of the Reelin receptors, the with a dominant role of Fyn in Reelin signaling, because VLDLR, although total protein levels were unchanged. the phosphorylation site-specific antibody (␣-pY 418) Expression of the second Reelin receptor, apoER2, was does not distinguish between different SFK members. unaffected, and total levels of FAK and Cdk5, which is Thus, the overall reduced levels of cellular Y418-phos- also involved in the regulation of neuronal migration, phorylated SFKs upon Reelin stimulation appear to re- were also unaffected. Increase of Dab1 expression also flect the contribution of Fyn to this signaling pathway. occurs in mice lacking Reelin or its receptors [1, 5] and Compensation for the absence of Fyn apparently occurs likely reflects an attempt of the cell to adjust the “gain” through an increase in Dab1 levels, but does not involve of the expected Reelin signal input. upregulation of the related SFKs Yes or Src. In fact, in %50ف We next isolated primary cortical neurons from wild- expression of the latter was even reduced by type (Figure 3B, lanes 3 and 4) and fyn-deficient (lanes fyn knockout neurons (Figure 3B, lanes 1 and 2). 1 and 2) mice and tested the ability of Reelin to induce tyrosine phosphorylation of Dab1 as well as phosphory- Reelin-Mediated Activation of SFKs lation of other SFKs in the activation loop when Fyn was Requires Dab1 absent. As in the brain extracts, Dab1 expression was The increased expression of Dab1 in the brains of fyn increased, but Reelin-induced tyrosine phosphorylation knockout mice and in fyn-deficient neurons suggests a of Dab1 occurred normally in fyn-deficient cortical neu- functional role of Fyn in Reelin signaling and further rons (lane 2); this finding is consistent with the absence supports a physiological role for SFKs in this signaling Current Biology 22

its receptors VLDLR and apoER2 [24, 25]. To confirm that Reelin binding to its receptors is required for the Dab1-dependent activation of SFKs, we exposed wild- type neurons to Reelin in the absence (Figure 5A, lanes 1–3) or presence (lane 4) of GST-RAP, a fusion protein between glutathione-S-transferase (GST) and the receptor-associated protein (RAP). The latter is a universal inhibitor of ligand binding to LDL receptor family members [26] including the VLDLR and apoER2. We had previously shown that GST-RAP antagonizes Reelin binding to both receptors and greatly reduces Reelin- dependent Dab1 phosphorylation [24]. Consistent with our previous findings, GST-RAP reduced Dab1 phosphorylation (Figure 5A). The low resid- ual phosphorylation of Dab1 by Reelin in the presence of GST-RAP is explained by incomplete inhibition due fold higher binding affinity of Reelin to the-20ف to the receptors [24]. Nevertheless, the Reelin-dependent increase of SFK phosphorylation (compare lanes 1 and 4, no change) was completely abrogated. GST alone served as a negative control. Dab1 phosphorylation and SFK activation were also dependent upon the presence of functional Reelin receptors on the neurons. Cells lacking both vldlr and apoer2 were unable to induce Dab1 or SFK tyrosine phosphorylation in response to Reelin (Figure 5B, lane 2), while cells lacking vldlr but expressing functional Figure 4. Activation of SFKs by Reelin Is Blocked in dab1-Deficient apoER2 activated Dab1 and SFKs almost normally (lane -fold increase). Total Src or Fyn levels were unal-3ف ,Cortical Neurons 4 (A) Total levels of Fyn and Src are unaffected by dab1 genotype in tered in neurons from double-deficient mice, whereas E15.5 embryonic brains containing two (lanes 1 and 6), one (lanes Dab1 protein expression was increased by approxi- 2 and 3), or no (lanes 4 and 5) functional dab1 alleles. Cdk5 served mately 10-fold (lanes 1 and 2). as a loading control. Taken together, these data show that SFKs function (B) Reelin-induced phosphorylation of neuronal SFK activation domains (␣-pY418) is normal in dab1ϩ/ϩ (lanes 1 and 2), but absent in downstream of the Reelin receptor-Dab1 signaling com- dab1Ϫ/Ϫ, neurons (lanes 3 and 4). plex. However, pharmacological inhibition of SFKs also interferes with Dab1 tyrosine phosphorylation, sug- pathway. To obtain additional genetic evidence for a gesting that SFKs also participate in Dab1 tyrosine functional relationship between components of the Ree- phosphorylation. These findings are consistent with a lin signaling pathway and SFKs, we investigated the model in which Reelin-induced aggregation of the re- effect of dab1 inactivation on Fyn expression and SFK ceptors and Dab1 at the plasma membrane generates activation. Neither Fyn nor Src protein levels were al- a “supercritical density” that results in mutual activation tered in the brains of dab1 knockout mice (Figure 4A, of Dab1 and its binding partner and effector, i.e., SFKs. lanes 4 and 5). Dab1 protein was reduced by approxi- It is possible, although not absolutely required, that a mately 50% in dab1ϩ/Ϫ brains (lanes 2 and 3) compared coreceptor (e.g., CNRs or integrins) is recruited into the to wild-type (lanes 1 and 6). complex by Reelin and serves as a primer to facilitate the As expected, the Reelin-induced, tyrosine-phosphor- initial reaction (Figure 6). This model would be consistent -kDa, which corresponds to with the results shown in Figure 1 in which overexpres 80ف ylated protein band at Dab1 (also see Figure S1), was absent from lysates of sion of Dab1 and an SFK, but not of either protein alone, dab1Ϫ/Ϫ neurons (Figure 4B, lane 4). SFK activation by enhanced Dab1 tyrosine kinase phosphorylation as well Reelin was monitored by using the ␣-pY418 antibody as total cellular tyrosine kinase activity. and was normal in wild-type neurons (lanes 1 and 2, fold increase), but it was completely abolished in-7ف dab1Ϫ/Ϫ neurons (lane 4, no change). However, the basal Combined Genetic Deficiency of fyn levels of Y418 phosphorylation were identical in the and Reelin Receptors dab1ϩ/ϩ and dab1Ϫ/Ϫ neurons, indicating that dab1 defi- Extensive redundancy between Src kinase family mem- ciency abrogates Reelin responsiveness, but has no bers impedes the dissection of their relative functional other discernible impact on the steady state of SFK importance in various physiological processes. Triple activation in embryonic neurons. mutant mice that completely lack the major neuronally expressed SFKs Src, Yes, and Fyn [27] are embryonic Activation of SFKs by Reelin Requires the Reelin lethal around E9.5, too early to be useful to study the Receptors VLDLR and apoER2 effect of SFKs on brain development. In order to test Tyrosine phosphorylation and thus activation of Dab1 whether it might be possible to obtain genetic evidence in response to Reelin requires the binding of Reelin to for a potential specific role of Fyn in Reelin signaling by Reelin Activates Src Family Tyrosine Kinases 23

Figure 5. Reelin-Mediated Activation of SFKs Requires Binding to ApoE Receptors (A) Cortical neurons were incubated in the absence (lane 1) or presence of Reelin (lanes 2–4) and in the absence (lanes 1–3) or presence (lane 4) of receptor-associated protein fused to glutathione-S-transferase (GST-RAP), a universal inhibitor of ligand binding to LDL receptor family members. GST protein was used as a control (lane 3). Reelin-induced tyrosine phosphorylation of Dab1 (upper panel) and SFKs (␣-pY418) is greatly reduced by the presence of RAP (lane 4). (B) Reelin-mediated SFK activation is impaired in neurons from vldlr and apoer2 knockout mice. Cortical neurons were isolated from apoer2/ vldlr double knockout (lanes 1 and 2) or vldlr-deficient mice (lanes 3 and 4). Treatment with Reelin (lanes 2 and 4) was compared to control medium (lanes 1 and 3) and induced Dab1 and SFK phosphorylation only in vldlrϪ/Ϫ;apoer2ϩ/ϩ (lane 4), but not in neurons lacking both Reelin receptors (lane 2). Total Dab1 protein levels were greatly increased in double-deficient neurons (lanes 1 and 2), whereas total levels of Src and Fyn were not affected by the genotype of the neurons. Cdk5 served as a loading control. Neurons were obtained from individual littermate embryos. epistasis, we crossed vldlrϪ/Ϫ or apoer2Ϫ/Ϫ mice with abnormalities beyond those already present in the re- fynϪ/Ϫ animals. Mice lacking only one of these receptors spective single-knockout mice (Figure S3). This finding display characteristic hypomorphic and spatially re- underscores the high degree of functional redundancy stricted phenotypes [5]. A combined deficiency in one among SFKs and the remarkable degree to which neu- of the receptors and a tyrosine kinase involved in Dab1 rons can compensate for the loss of some of the molecu- phosphorylation might conceivably compound the ob- lar components that relay and interpret the Reelin signal served phenotype overall or in specific regions of the within the cell. brain. However, mice double mutant for either fyn and The absence of a compounded phenotype in the dou- apoer2 or for fyn and vldlr were viable and did not exhibit ble mutants may reflect an ability of the cell to adjust any gross neurological phenotype or neuroanatomical the responsiveness of the cellular processes that are

Figure 6. Hypothetical Model for SFK Activa- tion by Reelin Oligomerized Reelin induces clustering of the structurally closely related VLDLR and apoER2. This results in a high local density of Dab1 at the cytoplasmic leaflet. Dab1 may activate SFKs directly, or coreceptors (e.g., CNRs or integrins) may serve as primers to initiate the reaction. Tyrosine-phosphorylated Dab1 can recruit additional SFKs through SH2 domain/phosphotyrosine inter- actions. This would result in rapid and local- ized amplification of Reelin-induced tyrosine kinase activity. Blue ovals indicate alterna- tively spliced cysteine-rich repeats in the Reelin binding domain of apoER2. Current Biology 24

targeted by the Reelin-activated SFKs. Thus, the relative Animals Ϫ Ϫ Ϫ Ϫ change of SFK activity, rather than the absolute amount, The vldlr / ;apoer2 / mutant mice [5] were maintained on a mixed ϫ Ϫ/Ϫ may be critical to elicit the intended physiological effect. C57BL/6J Sv129Ev strain background. The fyn mice [31] were purchased from the Jackson Laboratory, and the dab1Ϫ/Ϫ mutant This could include remodeling of the cytoskeleton dur- mice were a kind gift from Jon Cooper and Brian Howell and have ing the period of neuronal migration [28]. been described elsewhere [32]. All animals were maintained in ac- Experimental constraints allow us only to reasonably cordance with National Institutes of Health and institutional animal estimate the relative quantitative change of SFK activa- care guidelines. Wild-type mice were maintained on the mixed tion in whole-cell extracts. Thus, we can draw no conclu- C57BL/6J ϫ Sv129Ev strain background. For neuronal cultures, sions about the absolute change of SFK activity at spe- male and female mice were mated, and, if a vaginal plug was observed the next morning, the developmental stage was designated cific subcellular locations in response to Reelin. The E0.5. finding that the levels of phosphorylated Dab1 are not Ϫ Ϫ affected by the reduced SFK activation in fyn / neurons Production of Recombinant Proteins (Figure 3B) can be explained by the increased availability Recombinant Reelin was produced as described in the Supplemen- of unphosphorylated Dab1 substrate. Together, these tary Material, and receptor-associated protein (RAP) was prepared mechanisms could account for this remarkable degree as described previously [26]. of functional compensation. Cell Culture and Transfection It is likely that SFK activation by Reelin reflects only HEK-293 cells (CRL-1573, ATCC) were grown in DMEM supple- one branch of a complex cytoplasmic signaling pathway mented with 10% fetal calf serum and were transfected by calcium that may not manifest itself primarily in the regulation phosphate precipitation (MBS kit) (Stratagene) with 1 ␮g pcDNA3.1 of neuronal positioning but, for instance, in the modula- zeo(ϩ) mDab555 [33] and 1 ␮g pCMV5-cSrc (murine neuronal) or tion of synaptic plasticity. Our findings that Reelin en- with pRc/CMV-cFyn (human, T: obtained from Gottfried Baier, Inns- bruck) per 60-mm dish. Transfections with the respective empty hances synaptic plasticity, and that mice genetically vectors served as controls. After 36–48 hr, cells were harvested by deficient in Reelin receptors display LTP defects [17] aspirating the medium, rinsing with phosphate-buffered saline, and that closely resemble those seen in fyn-deficient mice adding 500 ␮l lysis buffer (20 mM Tris-buffered saline [pH 7.5], 2 [9], are consistent with such a hypothesis. mM EDTA, 1% Triton X-100 supplemented with 1 tablet protease inhibitor cocktail [Roche Molecular Biochemicals]/10 ml, and phosphatase inhibitor cocktails I and II [Sigma] at the recommended Conclusions dilutions) per 60-mm dish. Lysates were cleared by centrifugation We have shown that tyrosine-phosphorylated Dab1 can at 20,000 ϫ g, adjusted for protein content, mixed with 1 volume activate SFKs and that Reelin induces SFK activation 2ϫ gel-loading buffer (2% sodium dodecyl sulfate [SDS], 10% glyc- in embryonic cortical neurons. This requires the Reelin erol, 0.25 M Tris-HCl [pH 6.8], 0.02% bromophenol blue, 5% receptors apoER2 and VLDLR as well as the cytoplasmic ␤-mercaptoethanol), boiled for 5 min, and stored at Ϫ20ЊC or used adaptor protein Dab1. Results with pharmacological in- directly for immunoblot analysis. hibitors of non-receptor tyrosine kinases are consistent Neuronal Cultures with a functionally redundant role of SFKs in Reelin- Primary cortical neurons were prepared essentially as described induced Dab1 tyrosine phosphorylation. Increased previously [3] (also see the Supplementary Material). Dab1 protein expression in neurons from fynϪ/Ϫ embryos suggests a prominent role of Fyn in Reelin signaling. Reelin Stimulation of Cortical Embryonic Neurons Reelin-induced activation of Fyn and related SFKs in and Treatment with Inhibitors Reelin-containing or control concentrated media were diluted corre- neurons further implies a functional role for this family sponding to a final estimated Reelin concentration of approximately of tyrosine kinases in brain development or neuronal 5 nM and were incubated at 37ЊC for 25 min if not indicated other- function in the mature brain. Taken together with the wise. Treatment was terminated by aspirating the medium, rinsing complementary findings by Arnaud et al. [29], the pres- the cells with PBS containing 50 mM NaF, 25 mM ␤-glycerophos- ent findings have thus revealed another part of the path- phate, and 2 mM sodium orthovanadate (Sigma), and adding ice- way by which Reelin shapes the developing nervous cold lysis buffer (PBS [pH 7.4], 2 mM EGTA, 5 mM EDTA, 1% [v/v] Triton X-100, 0.5% [w/v] SDS, 0.25% [w/v] sodium deoxycholate system and regulates neuronal cell function. and protease and phosphatase inhibitor cocktails as described above; 16 ␮l/cm2 surface area). Lysates were adjusted for protein Experimental Procedures content, mixed with twice-concentrated gel-loading buffer, and used for immunoblot analysis. Stocks of the phosphatase inhibitors Reagents were prepared in dimethyl sulphoxide (DMSO; Sigma) (10 mM; herbi- The tyrosine kinase inhibitors radicicol, herbimycin A, PP1, and PP2 mycin A was dissolved at 1 mM). Neurons were pretreated for 60 were purchased from Biomol; the inactive PP2 analog PP3 was min (herbimycin A: 18 hr) at the indicated final inhibitor concentra- purchased from Sigma. tions before adding the Mock- or Reelin-containing media.

Antibodies Immunoprecipitation The rabbit polyclonal antisera raised against the 13 carboxyl-termi- For immunoprecipitation of Dab1 or Fyn protein, neurons plated on nal amino acids of murine Dab1, VLDLR, and ApoER2 have been 100 mm-dishes were harvested in PBS (pH 7.4) (1% Triton X-100, described elsewhere [5, 24]. The monoclonal anti-phosphotyrosine 0.1% SDS and 0.1% deoxycholate, 2 mM EDTA) containing protease antibody 4G10, the Src monoclonal antibody GD11, and a Fyn rabbit and phosphatase inhibitors (1 ml per dish). Lysates were cleared polyclonal antibody were purchased from Upstate Biotechnology. by centrifugation and were incubated with 10 ␮l/ml rabbit nonim- Phosphorylation site-specific antibodies against p-Src (Y418), p-Src mune or anti-Dab1 serum, or 4 ␮g/ml anti-Fyn monoclonal antibody (Y529), and p-FAK (Y397) were obtained from Biosource. Antibodies on ice. Immune complexes were precipitated by adding 50 ␮l protein against Erk and AKT were purchased from Cell Signaling Technol- A (Sigma)- or protein G (Amersham)-sepharose slurry (correspond- ogy; a mouse monoclonal antibody against Fyn and the Cdk5 antibody ing to 25 ␮l packed beads, washed three times with lysis buffer) were from Santa Cruz Biotechnology. The mouse monoclonal anti- per milliliter of lysate. The IP supernatant was collected, and the Reelin antibody (G10) was a generous gift from Andre Goffinet [30]. beads were washed once with 500 ␮l lysis buffer, twice with PBS, Reelin Activates Src Family Tyrosine Kinases 25

and were resuspended and boiled in 30 ␮l twice-concentrated gel learning, and hippocampal development in fyn mutant mice. loading buffer. Aliquots of the lysate, the IP supernatants, and the Science 258, 1903–1910. extracted beads were analyzed by immunoblotting (see the Supple- 10. Yagi, T., Aizawa, S., Tokunaga, T., Shigetani, Y., Takeda, N., mentary Material). and Ikawa, Y. (1993). A role for Fyn tyrosine kinase in the suckling behaviour of neonatal mice. Nature 366, 742–745. Protein Determination and Quantitation 11. Kohmura, N., Senzaki, K., Hamada, S., Kai, N., Yasuda, R., Wata- Protein concentrations of cellular lysates were determined with a nabe, M., Ishii, H., Yasuda, M., Mishina, M., and Yagi, T. (1998). detergent-compatible protein assay based on the Lowry method Diversity revealed by a novel family of cadherins expressed in (Bio-Rad). Relative intensities of protein bands on immunoblots were neurons at a synaptic complex. Neuron 20, 1137–1151. determined by using NIH image 1.62 software. 12. Schlaepfer, D.D., and Hunter, T. (1998). Integrin signalling and tyrosine phosphorylation: just the FAKs? Trends Cell Biol. 8, Histology 151–157. Three- to five-week-old animals were perfused transcardially with 13. Howell, B.W., Herrick, T.M., Hildebrand, J.D., Zhang, Y., and warm phosphate-buffered saline (PBS), followed by 4% paraformal- Cooper, J.A. (2000). Dab1 tyrosine phosphorylation sites relay dehyde. Sagittal paraffin sections in the region of the cerebellar positional signals during mouse brain development. Curr. Biol. peduncle (7 ␮m) were stained with hematoxylin and eosin (H&E). 10, 877–885. 14. Brown, M.T., and Cooper, J.A. (1996). Regulation, substrates Supplementary Material and functions of src. Biochim. Biophys. Acta 1287, 121–149. Supplementary Material including data relevant to assay proce- 15. Utsunomiya-Tate, N., Kubo, K., Tate, S., Kainosho, M., Kata- dures, pharmacological inhibition of Reelin-activated tyrosine ki- yama, E., Nakajima, K., and Mikoshiba, K. (2000). Reelin mole- nase, and histology of mutant mouse brains is available at http:// cules assemble together to form a large protein complex, which images.cellpress.com/supmat/supmatin.htm. is inhibited by the function-blocking CR-50 antibody. Proc. Natl. Acad. Sci. USA 97, 9729–9734. Acknowledgments 16. Pesold, C., Impagnatiello, F., Pisu, M.G., Uzunov, D.P., Costa, E., Guidotti, A., and Caruncho, H.J. (1998). Reelin is preferentially We are indebted to Wen-Ling Niu, Jenny Hayes, Lisa Beatty, and expressed in neurons synthesizing gamma-aminobutyric acid LaMetria Blair for invaluable technical support, to Jim Richardson, in cortex and hippocampus of adult rats. Proc. Natl. Acad. Sci. Jeff Stark and John Shelton for help with the histological analysis, USA 95, 3221–3226. to Tom Curran and Eckardt Fo¨ rster for Reelin expression vectors 17. Weeber, E.J., Beffert, U., Jones, C., Christian, J.M., Fo¨ rster, E., and for sharing the Reelin-expressing 293 cell line, Gottfried Baier Sweatt, J.D., and Herz, J. (2002). Reelin and ApoE receptors for the Fyn expression plasmid, and to Axel Ullrich and Petra May cooperate to enhance hippocampal synaptic plasticity and for critical suggestions. This work was supported by grants from the learning. J. Biol. Chem. 277, 39944–39952. National Institutes of Health (HL20948, NS43408, and HL63762), the 18. Ilic, D., Damsky, C.H., and Yamamoto, T. (1997). Focal adhesion Alzheimer Association, and the Perot Family Foundation. J.H. is kinase: at the crossroads of signal transduction. J. Cell Sci. 110, the recipient of a Wolfgang-Paul Award by the Humboldt Foundation 401–407. and a Visiting Professor at the Center for Molecular Biology at the 19. Bixby, J.L., and Jhabvala, P. (1993). Tyrosine phosphorylation University of Heidelberg. H.B. is supported by a fellowship from the in early embryonic growth cones. J. Neurosci. 13, 3421–3432. Deutsche Forschungsgemeinschaft. 20. Helmke, S., and Pfenninger, K.H. (1995). Growth cone enrich- ment and cytoskeletal association of non-receptor tyrosine ki- Received: August 14, 2002 nases. Cell Motil. Cytoskeleton 30, 194–207. Revised: September 26, 2002 21. Hanke, J.H., Gardner, J.P., Dow, R.L., Changelian, P.S., Bris- Accepted: October 18, 2002 sette, W.H., Weringer, E.J., Pollok, B.A., and Connelly, P.A. Published: January 8, 2003 (1996). Discovery of a novel, potent, and Src family-selective tyrosine kinase inhibitor. Study of Lck- and FynT-dependent T References cell activation. J. Biol. Chem. 271, 695–701. 22. Stein, P.L., Vogel, H., and Soriano, P. (1994). Combined deficien- 1. Rice, D.S., and Curran, T. (2001). Role of the reelin signaling cies of Src, Fyn, and Yes tyrosine kinases in mutant mice. Genes pathway in central nervous system development. Annu. Rev. Dev. 8, 1999–2007. Neurosci. 24, 1005–1039. 23. Herz, J., and Bock, H.H. (2002). Lipoprotein receptors in the 2. Herz, J. (2001). The LDL receptor gene family: (un)expected nervous system. Annu. Rev. Biochem. 71, 405–434. signal transducers in the brain. Neuron 29, 571–581. 24. Hiesberger, T., Trommsdorff, M., Howell, B.W., Goffinet, A., 3. Howell, B.W., Herrick, T.M., and Cooper, J.A. (1999). Reelin- Mumby, M.C., Cooper, J.A., and Herz, J. (1999). Direct binding induced tryosine phosphorylation of disabled 1 during neuronal of Reelin to VLDL receptor and ApoE receptor 2 induces tyrosine positioning. Genes Dev. 13, 643–648. phosphorylation of disabled-1 and modulates tau phosphoryla- 4. Keshvara, L., Benhayon, D., Magdaleno, S., and Curran, T. tion. Neuron 24, 481–489. (2001). Identification of reelin-induced sites of tyrosyl phosphor- 25. D’Arcangelo, G., Homayouni, R., Keshvara, L., Rice, D., Sheldon, ylation on disabled 1. J. Biol. Chem. 276, 16008–16014. M., and Curran, T. (1999). Reelin is a ligand for lipoprotein recep- 5. Trommsdorff, M., Gotthardt, M., Hiesberger, T., Shelton, J., tors. Neuron 24, 471–479. Stockinger, W., Nimpf, J., Hammer, R.E., Richardson, J.A., and 26. Herz, J., Goldstein, J.L., Strickland, D.K., Ho, Y.K., and Brown, Herz, J. (1999). Reeler/Disabled-like disruption of neuronal mi- M.S. (1991). 39-kDa protein modulates binding of ligands to low gration in knockout mice lacking the VLDL receptor and ApoE density lipoprotein receptor-related protein/alpha 2-macro- receptor 2. Cell 97, 689–701. globulin receptor. J. Biol. Chem. 266, 21232–21238. 6. Senzaki, K., Ogawa, M., and Yagi, T. (1999). Proteins of the CNR 27. Klinghoffer, R.A., Sachsenmaier, C., Cooper, J.A., and Soriano, family are multiple receptors for Reelin. Cell 99, 635–647. P. (1999). Src family kinases are required for integrin but not 7. Dulabon, L., Olson, E.C., Taglienti, M.G., Eisenhuth, S., McGrath, PDGFR signal transduction. EMBO J. 18, 2459–2471. B., Walsh, C.A., Kreidberg, J.A., and Anton, E.S. (2000). Reelin 28. Thomas, S.M., Soriano, P., and Imamoto, A. (1995). Specific and binds ␣3␤1 integrin and inhibits neuronal migration. Neuron 27, redundant roles of Src and Fyn in organizing the cytoskeleton. 33–44. Nature 376, 267–271. 8. Howell, B.W., Gertler, F.B., and Cooper, J.A. (1997). Mouse dis- 29. Arnaud, L., Ballif, B.A., Fo¨ rster, E., and Cooper, J.A. (2003). Fyn abled (mDab1): a Src binding protein implicated in neuronal tyrosine kinase is a critical regulator of Disabled-1 during brain development. EMBO J. 16, 121–132. development. Curr. Biol. 13, 9–17. 9. Grant, S.G., O’Dell, T.J., Karl, K.A., Stein, P.L., Soriano, P., and 30. de Bergeyck, V., Naerhuyzen, B., Goffinet, A.M., and Lambert Kandel, E.R. (1992). Impaired long-term potentiation, spatial de Rouvroit, C. (1998). A panel of monoclonal antibodies against Current Biology 26

reelin, the extracellular matrix protein defective in reeler mutant mice. J. Neurosci. Methods 82, 17–24. 31. Stein, P.L., Lee, H.M., Rich, S., and Soriano, P. (1992). pp59fyn mutant mice display differential signaling in thymocytes and peripheral T cells. Cell 70, 741–750. 32. Howell, B.W., Hawkes, R., Soriano, P., and Cooper, J.A. (1997). Neuronal position in the developing brain is regulated by mouse disabled-1. Nature 389, 733–737. 33. Trommsdorff, M., Borg, J.P., Margolis, B., and Herz, J. (1998). Interaction of cytosolic adaptor proteins with neuronal apolipoprotein E receptors and the amyloid precursor protein. J. Biol. Chem. 273, 33556–33560.

Note Added in Proof

The data referred to throughout as “unpublished data” are now in press: Beffert, U., Morfini, G., Bock, H.H., Reyna, H., Brady, S.T., and Herz, J. (2002). Reelin-mediated signaling locally regulates protein kinase B/Akt and glycogen synthase kinase 3␤. J. Biol. Chem. 277, 49958–49964.