During neuronal development, netrin and its receptors UNC5 and DCC (deleted in colorectal cancer) guide axonal growth cones in navigating to their targets. Netrin also plays important roles in the regulation of cell migration, tissue morphogenesis and tumor growth. Here, we show that netrin induces UNC5 tyrosine phosphorylation and that this effect of netrin is dependent on its co-receptor DCC. UNC5 tyrosine phosphorylation is known to be important for netrin to induce cell migration and axonal repulsion. Src tyrosine kinase activity is required for netrin to stimulate UNC5 tyrosine phosphorylation in neurons and transfected cells. The SH2 domain of Src kinase directly interacts with the cytosolic domain of UNC5 in a tyrosine-phosphorylation-dependent manner. Furthermore, the tyrosine kinase focal adhesion kinase (FAK) is also involved in netrin-induced UNC5 tyrosine phosphorylation. Both Src and FAK can phosphorylate UNC5. Our data suggest a model in which netrin stimulates UNC5 tyrosine phosphorylation and signaling in a manner dependent on the co-receptor DCC, through the recruitment of Src and FAK kinases.
Netrin is a laminin-like secreted protein originally identified as an axonal guidance molecule (Hedgecock et al., 1990; Ishii et al., 1992; Kennedy et al., 1994; Serafini et al., 1994). Netrins have also been implicated in cell migration (Hebrok and Reichardt, 2004; Killeen et al., 2002; Su et al., 2000; Wadsworth et al., 1996), tissue morphogenesis (Hinck, 2004; Liu et al., 2004b; Lu et al., 2004; Salminen et al., 2000; Srinivasan et al., 2003; Wadsworth et al., 1996), and tumor growth (Arakawa, 2004; Mazelin et al., 2004; Mehlen and Fearon, 2004).
Both DCC (deleted in colorectal cancer) and UNC5 are demonstrated netrin receptors (Hedgecock et al., 1990; Hong et al., 1999; Livesey and Hunt, 1997; Stein and Tessier-Lavigne, 2001; Tessier-Lavigne and Goodman, 1996). Netrin stimulates axon growth and causes growth cone attraction in DCC-expressing neurons (Bashaw and Goodman, 1999; Chan et al., 1996; Hong et al., 1999; Jarjour et al., 2003; Keino-Masu et al., 1996; Merz and Culotti, 1999; Stein and Tessier-Lavigne, 2001). By contrast, UNC5-expressing neurons are repulsed by netrin gradients (Bashaw and Goodman, 1999; Hamelin et al., 1993; Hedgecock et al., 1990; Hong et al., 1999; Jarjour et al., 2003; Leonardo et al., 1997; Leung-Hagesteijn et al., 1992). It is generally thought that UNC5 receptor alone transduces short range signals whereas DCC transduces long range signals important for neuron migration, neurite growth and axon repulsion (Keleman and Dickson, 2001; Kennedy, 2000). Therefore, DCC may enhance the signaling function of UNC5 by forming a complex with UNC5.
DCC and UNC5 have also been reported to play important roles in non-neuronal systems. For instance, the protein level of DCC is dramatically decreased in some cancer cell lines such as colorectal cancer cells (Fearon et al., 1990; Gao et al., 1993; Itoh et al., 1993; Kennedy et al., 1994; Kikuchi-Yanoshita et al., 1992; Uchino et al., 1992). Also, in some cell types both DCC and UNC5 are found to induce apoptosis in the absence of netrin, whereas netrin acts as a survival factor to inhibit cell death (Arakawa, 2004; Calalb et al., 1995; Forcet et al., 2001; Llambi et al., 2001; Mazelin et al., 2004; Mehlen and Fearon, 2004; Mehlen et al., 1998). Netrin and its receptors (DCC and UNC5) are also reported to play a role in controlling morphogenesis of the epithelial and vascular system and in lung development (Calalb et al., 1995; Forcet et al., 2001; Hebrok and Reichardt, 2004; Hinck, 2004; Liu et al., 2004b; Llambi et al., 2001; Lu et al., 2004; Mehlen et al., 1998; Nikolopoulos and Giancotti, 2005; Park et al., 2004).
Although the importance of netrin and its receptors (DCC/UNC5) in organ development, including neuronal and non-neuronal network systems, has been well characterized, the mechanisms by which signals are transduced downstream of the receptors are only partially understood. DCC has been proposed to signal as a tyrosine kinase-coupled receptor (Killeen et al., 2002; Tong et al., 2001). DCC forms a physical and functional complex with focal adhesion kinase (FAK) and Src family members (Li et al., 2004; Liu et al., 2004a; Meriane et al., 2004; Ren et al., 2004). Netrin stimulation activates both FAK and Src kinases and induces phosphorylation of tyrosine residue Y1420 in DCC. Mutation of the tyrosine phosphorylation site in DCC diminishes the ability of DCC to mediate the netrin-induced growth stimulating and attractive effects on spinal neurons in vitro. These results have established an essential function of tyrosine phosphorylation and tyrosine kinases in DCC mediated netrin signaling (Li et al., 2004; Liu et al., 2004a; Meriane et al., 2004; Ren et al., 2004).
Much less is known about the mechanism of UNC5 signaling. Tong et al. have reported that RCM, an UNC5-related netrin receptor, is tyrosine phosphorylated in transfected mammalian cells in the presence of DCC and netrin (Tong et al., 2001). Co-expression of Src also induces RCM tyrosine phosphorylation. However, it is not clear whether Src plays a role in UNC5 phosphorylation and how netrin modulates UNC5 phosphorylation under physiological conditions. The kinase(s) responsible for netrin-induced UNC5 phosphorylation has not been identified. However, tyrosine residue Y482 in UNC5 has been mapped as a phosphorylation site important for UNC5 function in C. elegans (Killeen et al., 2002).
FAK and Src interact and positively regulate each other in response to many extracellular stimuli (Cobb et al., 1994; Schlaepfer et al., 1999; Tatosyan and Mizenina, 2000; Thomas and Brugge, 1997). For example, FAK and Src are known to be activated by extracellular matrix stimulation through integrin receptors (Cary and Guan, 1999; Guan, 1997). FAK can bind to Src and possibly enhance Src activity, whereas Src can phosphorylate and increase FAK kinase activity. Extensive genetic and cell biological studies have implicated both FAK and Src in neuronal growth, axon guidance and cell migration through regulating the signal transduction pathways downstream of guidance receptors (Brouns et al., 2001; Carter et al., 2002; Kalia et al., 2004; Knoll and Drescher, 2004; Li et al., 2004; Liu et al., 2004a; Palmer et al., 2002; Ren et al., 2004).
In this report, we show that netrin induces UNC5 tyrosine phosphorylation and that DCC greatly enhances UNC5 tyrosine phosphorylation in response to netrin stimulation. We found that FAK and Src family kinases are required for netrin to stimulate this effect. Our data suggest a model in which FAK and Src play critical roles in UNC5 and DCC signaling in response to netrin.
Netrin induces UNC5 tyrosine phosphorylation
UNC5 tyrosine phosphorylation has been observed in transfected HEK293 cells. Furthermore, tyrosine phosphorylation of endogenous UNC5 has been reported in C. elegans (Killeen et al., 2002; Tong et al., 2001). To determine whether netrin stimulates tyrosine phosphorylation of endogenous UNC5 in neuronal cells, we tested spinal cord neurons isolated from mouse E15 embryos. The dissociated neurons were cultured in vitro and stimulated with netrin. UNC5 was immunoprecipitated and tyrosine phosphorylation of UNC5 was detected by anti-phosphotyrosine immunoblot. We observed that netrin induced a rapid tyrosine phosphorylation of UNC5 in spinal cord neurons (Fig. 1A). To test whether Src family kinases play a role in tyrosine phosphorylation of UNC5, neurons were pretreated with PP2, a Src family kinase inhibitor, before the addition of netrin. We found that the netrin-stimulated UNC5 tyrosine phosphorylation was inhibited by PP2. These results suggest that Src family tyrosine kinases play an essential role in netrin-induced UNC5 tyrosine phosphorylation.
We further tested whether netrin can stimulate UNC5 tyrosine phosphorylation in transfected HEK293 cells. Our data showed that netrin had slightly increased UNC5 tyrosine phosphorylation when UNC5 was expressed alone in HEK293 cells (Fig. 1B, lanes 1-4). The lack of netrin stimulation could be due to the lack of DCC, which associates with UNC5 to form the functional repulsive netrin receptor complex. Co-expression of DCC enhanced UNC5 tyrosine phosphorylation. Importantly, netrin significantly increased UNC5 tyrosine phosphorylation when both UNC5 and DCC were co-expressed (Fig. 1B, lanes 5-8). These observations are consistent with a model in which DCC enhances the netrin-induced growth cone repulsion and the complex formation of UNC5 and DCC plays an important role in netrin signaling in vivo.
To examine the effect of UNC5 on cell morphology, Cos-7 cells were transfected with UNC5, DCC and their C-terminus truncation or their combination as shown in Fig. 1C. After AP-netrin treatment, we observed that UNC5 caused Cos-7 cells to shrink and elongate, and UNC5 and DCC together caused Cos-7 cells to shrink and form long filopodia-like spike shapes. The morphological change in Cos-7 requires the intracellular domain of UNC5 because the intracellular domain deletion mutant of UNC5C (ΔC) no longer promotes this cell shrinking and spike formation (Fig. 1C). We used the Trypan Blue staining method to monitor cell viability to exclude the possibility that these shrinking cells were dying. Our observations are consistent with reports that some receptors, such as plexin A1 and B1, induce Cos-7 cell shrinkage (Takahashi et al., 1999).
The P3 domain in DCC is essential for UNC5 tyrosine phosphorylation
The DCC intracellular domain consists of three domains, P1, P2 and P3, which are evolutionally conserved. To test which domain in DCC is important for UNC5 tyrosine phosphorylation, we co-expressed DCC truncation mutants ΔP1 (residues 1147-1171), ΔP2 (residues 1335-1356) and ΔP3 (residues 1412-1447) with UNC5 and stimulated with netrin. We observed that deletion of the P3 domain dramatically decreased UNC5 tyrosine phosphorylation whereas deletion of the P1 or P2 domain had little effect (Fig. 2A). These data indicate that the P3 domain in DCC is required for DCC to enhance UNC5 tyrosine phosphorylation.
We have demonstrated that the C-terminal region of the DCC P3 domain is essential for interaction with FAK (Li et al., 2004). Consistently with our previous observation, the P3 domain deletion mutant in DCC abolished interaction with FAK (Fig. 2B). Therefore, the above results imply that FAK plays a role in netrin-induced and DCC-mediated UNC5 tyrosine phosphorylation.
In order to test the DCC-dependent UNC5 tyrosine phosphorylation under conditions where UNC5 was not overexpressed, dorsal spinal cord neurons were examined. We found that netrin treatment induced a significant increase of UNC5 tyrosine phosphorylation (Fig. 1A, Fig. 2C). To confirm the involvement of DCC in netrin-induced UNC5 phosphorylation, a function-blocking DCC antibody directed against the extracellular domain was added to the culture for 1 hour before netrin treatment. The results indicate that DCC antibody pretreatment blocked netrin-induced UNC5 tyrosine phosphorylation (Fig. 2C). Our data demonstrate that DCC is required for UNC5 tyrosine phosphorylation in response to netrin.
Focal adhesion kinase is required for UNC5 tyrosine phosphorylation
To further test the function of FAK in UNC5 receptor activation, we co-expressed Myc-FAK with HA-UNC5 and DCC in HEK293 cells. We observed that FAK modestly increased UNC5 tyrosine phosphorylation (Fig. 2D, lanes 1 and 7). Co-expression of DCC and FAK caused a dramatic increase of UNC5 tyrosine phosphorylation that was further stimulated by netrin (Fig. 2D). These data support a role for FAK in UNC5 phosphorylation and provide evidence for the positive effect of DCC on UNC5 phosphorylation. To provide further evidence of the importance of FAK in UNC5 phosphorylation, we performed experiments by co-expressing FRNK, a naturally existing dominant negative form of FAK encoding residues 693-1052, with UNC5 and DCC. We found that FRNK blocked UNC5 tyrosine phosphorylation in response to netrin (Fig. 2E). These results are consistent with a positive role of FAK in UNC5 activation in vivo. We further confirmed that both DCC and UNC5 were phosphorylated in response to netrin treatment and this phosphorylation could be blocked by either PP2 pretreatment or co-expression of FRNK (Fig. 2F).
FAK-knockout mouse embryonic fibroblasts (MEFs) (Schlaepfer et al., 1999) were examined to further support a role for FAK in UNC5 phosphorylation. Netrin failed to stimulate UNC5 tyrosine phosphorylation in FAK-knockout MEF cells, whereas UNC5 phosphorylation was observed in response to netrin stimulation in wild-type MEF cells (Fig. 3A). Importantly, re-expression of FAK in FAK-knockout MEF cells restored tyrosine phosphorylation of UNC5 and netrin stimulation (Fig. 3B). Taken together, our data demonstrate that FAK is an important component required for UNC5 receptor phosphorylation in response to netrin.
Importance of Src family kinases in UNC5 phosphorylation
Src family kinases are known to be involved in DCC receptor activation. We have also found that Src inhibitor PP2 blocked tyrosine phosphorylation of endogenous UNC5 in neurons in response to netrin (Fig. 1A). Consistently, it has been reported that Src kinase phosphorylates UNC5. We examined the involvement of Src in UNC5 phosphorylation in transfected HEK293 cells. Treatment with PP2 inhibitor effectively blocked UNC5 tyrosine phosphorylation induced by netrin (Fig. 4A), which supports a critical role of Src kinase in tyrosine phosphorylation of UNC5. To examine whether Src kinase causes UNC5 tyrosine phosphorylation, we conducted experiments in which UNC5 is coexpressed with Src. Src expression caused a dramatic increase of UNC5 tyrosine phosphorylation whereas PP2 inhibited Src-induced UNC5 tyrosine phosphorylation (Fig. 4B). The kinase-dead mutant Src-K/M did not induce UNC5 tyrosine phosphorylation. When high levels of Src were expressed, UNC5 tyrosine phosphorylation was high and could not be further stimulated by netrin (data not shown). However, low levels of Src expression induced a modest increase of UNC5 phosphorylation that could be further stimulated by netrin (data not shown).
If Src family kinases are required for UNC5 tyrosine phosphorylation induced by netrin, then netrin should not stimulate UNC5 tyrosine phosphorylation in Src-family-kinase-knockout cells. To confirm our prediction, we investigated the UNC5 tyrosine phosphorylation in Src, Fyn and Yes triple-knockout MEF cells (Li et al., 2004; Lowell and Soriano, 1996) in response to netrin treatment. We observed that netrin induced UNC5 tyrosine phosphorylation in wild-type MEF cells but not in the triple-knockout MEF cells (Fig. 5A). To further confirm that the failure of netrin in stimulating UNC5 tyrosine phosphorylation is due to the lack of Src family kinases, we reexpressed low levels of Src kinase in the triple-knockout MEF cells. We observed that re-expression of Src restored netrin-induced UNC5 tyrosine phosphorylation (Fig. 5B).
The interaction between UNC5 and Src depends on the SH2 domain and kinase activity of Src
UNC5 tyrosine phosphorylation has been implicated to be important in netrin signaling in C. elegans (Killeen et al., 2002; Tong et al., 2001). However, the mechanisms by which Src and FAK kinases phosphorylate UNC5 are still not clear. Therefore, we sought to answer this question by exploring the possibility that Src and FAK directly interact with UNC5. Under our co-immunoprecipitation conditions, we observed that Src co-immunoprecipitated with UNC5 (Fig. 6A) whereas FAK did not (data not shown). The kinase-dead mutant Src (K/M) did not interact with UNC5, suggesting that Src kinase activity and UNC5 tyrosine phosphorylation are important in mediating the UNC5-Src interaction (Fig. 6A). This result also implies that the SH2 domain in Src is responsible for interaction with tyrosine phosphorylated UNC5. Indeed, we found that mutation of the SH2 domain but not the SH3 domain of Src decreased its association with UNC5 significantly, which suggests that the SH2 domain in Src contributes to the interaction with UNC5 (Fig. 6B).
We then wanted to examine whether the interaction between Src and UNC5 occurs under physiological conditions. To do this we used lysates of spinal cord neurons isolated from mouse embryos for co-immunoprecipitation. We found that UNC5 was co-immunoprecipitated with Src but not the negative control (Fig. 6C), and consistently, Src was coimmunoprecipitated by UNC5 antibody (Fig. 6D). These results provided compelling data that Src interacts with UNC5 in vivo.
To map which region in UNC5 is responsible for the interaction between UNC5 and Src, we performed in vitro pull-down experiments by using MBP-fused UNC5 purified from bacteria expressing internal deletion mutants of UNC5, including M1 (aa 441-460), M2 (aa 551-594), M3 (aa 632-670), M4 (aa 675-720), M5 (aa 751-800) and M6 (aa 842-931). We found that the M2 mutant significantly decreased interaction with Src whereas other deletion mutants, including M3, M4, M5 and M6, increased interaction with Src (Fig. 6E). These data indicate that tyrosine residues deleted in the M2 mutant play a role in the interaction with Src.
FAK and Src directly phosphorylate UNC5 in vitro
To determine whether FAK or Src can directly phosphorylate UNC5, we carried out in vitro phosphorylation experiments using purified UNC5 recombinant protein as a substrate. Both Src and FAK were immunoprecipitated from transfected HEK293 cells under highly stringent immunoprecipitation conditions. We found that both FAK and Src could phosphorylate UNC5, although Src phosphorylated UNC5 more efficiently than FAK (Fig. 7A). Kinase-inactive FAK-KM did not phosphorylate UNC5. To exclude the possibility that the presence of any co-precipitated Src in the FAK immunoprecipitates might be responsible for UNC5 phosphorylation, we included the Src inhibitor PP2. PP2 did not inhibit FAK-induced UNC5 phosphorylation in vitro. As predicted, PP2 potently inhibited Src autophosphorylation and its ability to phosphorylate UNC5 (Fig. 7A).
To determine tyrosine phosphorylation sites in UNC5 by Src, we performed in vitro phosphorylation experiments. Src was immunoprecipitated from transfected HEK293 cells under highly stringent conditions whereas UNC5 was purified from bacterial expression. The in vitro kinase results showed that phosphorylation of the UNC5 mutant M1 (aa 441-460) and M3 (aa 632-670) decreased significantly whereas phosphorylation of the other mutants was not significantly changed (Fig. 7B). Tyrosine residues Y449 and Y454 are deleted in the M1 mutant, and tyrosine residues Y649 and Y667 are deleted in the M3 mutant. Therefore, these four sites, or at least one in each deletion mutant, are major sites of in vitro tyrosine phosphorylation of UNC5 by Src kinase. To confirm the functional importance of the tyrosine residues Y649 and Y667 in UNC5 in UNC5-induced cell shrinkage, we transfected wild-type, M3 and intracellular domain deletion mutants into Cos-7 cells. We observed that the M3 mutant is much less effective than the wild-type UNC5 in inducing cell shrinkage in Cos-7 cells (Fig. 7C). These results are consistent with the observation that the M3 mutant shows a decreased level of tyrosine phosphorylation. Together the above biochemical experiments indicate that both Src and FAK contribute to UNC5 tyrosine phosphorylation and reveal the importance of tyrosine phosphorylation for UNC5 function.
UNC5 is a transmembrane protein identified as a functional receptor for netrin, which is essential for repulsive events guiding the axon growth cone away from netrin sources in the developing neuronal system (Keleman and Dickson, 2001; Leonardo et al., 1997; Leung-Hagesteijn et al., 1992; Merz et al., 2001). Genetic studies in C. elegans have shown that UNC5 tyrosine phosphorylation plays an important biological role in neuronal axon guidance and gonad migration, two processes known to be regulated by the Unc-6/netrin, Unc-40/DCC and UNC5 receptor complexes (Hedgecock et al., 1990; Hong et al., 1999; Killeen et al., 2002; Su et al., 2000). Furthermore, tyrosine phosphorylation of mammalian UNC5 has been detected in transfected cells (Killeen et al., 2002; Tong et al., 2001). The importance of tyrosine phosphorylation in netrin signaling is further supported by a recent discovery that mutation in the clr-1 tyrosine phosphatase in C. elegans enhances the biological function of Unc-6/netrin (Chang et al., 2004). However, whether UNC5 tyrosine phosphorylation is regulated by netrin and which role DCC plays in these signaling events remains unknown. Moreover, the kinase(s) responsible for netrin to induce UNC5 tyrosine phosphorylation has not been identified.
In this report, we present data indicating that FAK and Src family kinases play an important role in netrin-induced UNC5 tyrosine phosphorylation. Netrin stimulation induces endogenous UNC5 tyrosine phosphorylation in cultured neurons. Inhibition of Src family kinases blocks UNC5 tyrosine phosphorylation in response to netrin. Furthermore, in transfected cells, we show that DCC is required for netrin to stimulate UNC5 tyrosine phosphorylation. These observations are consistent with the current model in which the DCC-UNC5 receptor complex is important for neurons to display repulsive responses in the presence of a netrin gradient. Interestingly, UNC5 can repel the axon growth cone in the presence or absence of DCC but, with DCC, UNC5 can function over a longer range and more effectively (Keleman and Dickson, 2001). These observations are consistent with our observation that DCC enhances the ability of netrin to induce UNC5 tyrosine phosphorylation, therefore, increasing netrin signaling through UNC5. It is worth noting that netrin induces a weak increase of UNC5 tyrosine phosphorylation in the absence of DCC (Fig. 2B). This observation is consistent with the notion that UNC5 alone induces a weak response to netrin, whereas the presence of DCC enhances the repulsive effect of UNC5.
We observed that both Src and FAK can directly phosphorylate UNC5 in vitro and that co-expression of either Src or FAK with UNC5 increases UNC5 phosphorylation. These in vitro studies suggest that either FAK or Src is sufficient for UNC5 tyrosine phosphorylation. However, experiments relying on endogenous Src or FAK suggest that both Src and FAK are required for netrin to stimulate UNC5 phosphorylation efficiently. For example, PP2 inhibits UNC5 phosphorylation in neurons stimulated by netrin. The FAK-interaction-defective DCC-ΔP3 cannot enhance UNC5 phosphorylation. The importance of Src and FAK in netrin-induced UNC5 tyrosine phosphorylation is further supported by the fact that netrin is unable to stimulate UNC5 phosphorylation in either FAK-/- or Src/Fyn/Yes-/-/- MEF cells.
To explain the above observations, we propose that activation of the DCC-UNC5 receptor complex by netrin results in the recruitment and activation of Src and FAK to the receptor complex under physiological conditions (Fig. 8). Upon netrin treatment, DCC recruits FAK to the UNC5/DCC complex, possibly through a mechanism in which DCC directly associates with FAK. Although UNC5 can directly associate with Src, both DCC and FAK contribute to Src recruitment (Fig. 8). FAK and Src are known to positively regulate each other. Activation of these two kinases leads to tyrosine phosphorylation of the netrin receptors DCC and UNC, and possibly downstream signaling molecules. Phosphorylation of DCC and UNC5 could create binding sites for recruitment of downstream signaling molecules with SH2 domains or PTB domains. Therefore, the netrin receptors DCC and UNC5 function as tyrosine-kinase-coupled receptors by activating Src and FAK in the receptor complex.
We have also attempted to map the phosphorylation sites in UNC5. Point mutation experiments indicated that UNC5 is phosphorylated at multiple sites (data not shown). Therefore, we used truncation mutants to map the major phosphorylation sites in UNC5. Our results indicated that Y449, Y454, Y649 and Y667 are possible sites of tyrosine phosphorylation in UNC5 by Src (Fig. 7B). Genetic studies in C. elegans have provided evidence that UNC5 tyrosine phosphorylation is functionally important. Mutation of the major tyrosine phosphorylation site Y482, corresponding to the Y449 in mammalian UNC5, severely compromises the function of UNC5 in neuronal axon guidance and gonad migration in C. elegans (Killeen et al., 2002). The importance of Src in UNC5 signaling is also supported by our observation that Src-1 plays an important role in UNC5 signaling in the regulation of C. elegans gonad migration and commissural neuron guidance (Lee et al., 2005). We propose that tyrosine phosphorylation of UNC5 may create binding sites for downstream signaling molecules that play an important role in the generation of a repulsive signal, whereas DCC alone produces an attractive signal in response to netrin. Further experiments are needed to elucidate the molecular mechanism by which UNC5 converts the DCC-mediated attraction to repulsion.
Materials and Methods
Plasmid constructs and reagents
The neuronal isoform of Src (gi: 625219) was constructed by PCR from a human brain cDNA library, and subcloned into pcDNA3 (Invitrogen). FAK and related plasmids were described previously (Cooper et al., 2003). Src SH3 (127/128WW/RR), SH2 (178R/A) and kinase-dead (298K/M) mutants, as well as DCC ΔP1 (1147-1170), DCC ΔP2 (1335-1356), DCC ΔP3 (1412-1447), Y1420F and other expression constructs, were constructed by mutagenesis (Stratagene). HA-UNC5C was constructed by amplification of UNC5C (gi:3789764) from a human brain cDNA library followed by subcloning into pcDNA3 tagged with HA at C-terminus. MBP-UNC5 tagged at the C-terminus with HA and human netrin tagged at the C-terminus with alkaline phosphatase (AP) were constructed by PCR and subcloned into the vectors pMAL-p2 and pcDNA3, respectively. UNC5 deletion mutants including M1 (aa 441-460), M2 (aa 551-594), M3 (aa 632-670), M4 (aa 675-720), M5 (aa 751-800) and M6 (aa 842-931) were made by PCR using a site-directed mutagenesis kit (Strategene). HEK293 cells stably expressing chicken Myc-netrin were a generous gift from M. Tessier-Lavigne (Stanford University, CA). Human His-tagged netrin was constructed by PCR.
We used the following antibodies in our experiments: anti-DCC (PharMingen and Oncogene), pan-anti-Src, anti-FAK (Santa Cruz Biotechnology), anti-Src (Oncogene, Santa Cruz Biotechnology and Upstate Biotechnology), anti-phospho-Src (Bioscience and Cell Signaling), anti-hemagglutinin (HA) (Covance), anti-phospho-Tyr (PY20 from Transduction Laboratories and 4G10 from Upstate Biotechnology) and anti-UNC5 polyclonal antibody (a generous gift from Tony Pawson, Samuel Lunenfeld Research Institute).
Immunoprecipitation and in vitro binding
HEK293 cells were transfected using the Lipofectamine (Invitrogen) method. For immunoprecipitation (IP), cells were lysed 48 hours after transfection in mild lysis buffer (MLB: 20 mM Tris-Cl pH 7.4, 100 mM NaCl, 1% NP-40, 0.1 mM phenylmethylsulfonyl fluoride, 5 μg/ml aprotinin and 5 μg/ml leupeptin) and subjected to IP with 1 μg of antibody. Two hours later, protein A/G-agarose beads (Santa Cruz Biotechnology) were added for 1.5 hours and then washed four times prior to addition of 1× SDS sample buffer. For phosphotyrosine immunoblot experiments, PLC buffer (50 mM HEPES, pH 7.5, 150 mM NaCl, 10% glycerol, 1% Triton X-100, 1.5 mM MgCl2, 1 mM EGTA, 10 mM NaPPi, 100 mM NaF, 1 mM vanadate and protease inhibitors) was used in cell lysis. To inhibit Src family kinase activity, the inhibitor PP2 (Calbiochem) was added to serum-starved cells at a concentration of 0.2-5 μM in 1% DMSO for 1 hour. Tyrosine phosphorylation was determined by immunoblotting using PY20 or 4G10 antibody.
Mouse embryo brains at E15 were dissected and brain tissue was lysed in MLB buffer. Lysate was immunoprecipitated by Src or UNC5 antibodies and the association of UNC5 or Src was then determined by immunoblotting with UNC5 or Src-specific antibodies.
In vitro kinase assay
The kinase assay was carried out essentially as described previously. HA-FAK and Src were obtained by immunoprecipitation from expression of transiently transfected HEK293 cells. MBP-UNC5 purified from Escherichia coli was used as the substrate. The phosphorylation levels of UNC5 by FAK or Src were determined by phosphoimager. The nitrocellulose membrane was stained by Coomassie Blue to determine MBP-UNC5 levels.
Dorsal spinal cords were dissected from E15 mouse embryos. The isolated spinal cords were triturated to generate dissociated neuronal cells in HBSS buffer. The dissociated cells were cultured in L15 medium supplemented with 0.5% B27, 1% N2 and 30 mM glucose in plates, which were coated with laminin. For primary cortical neuronal culture, the cortex was dissected from E13-E14 mouse embryonic brains. The dissociated neurons were cultured in six-well plates coated with poly-L-lysine and laminin. The cells were cultured overnight and serum starved for about 6 hours before stimulation with netrin. Use of laboratory animals was approved by the University of Michigan Committee on the Use and Care of Animals (Protocol #8719).
We thank Jun-Lin Guan (Cornell University) for the FAK and S/F/Y-knockout MEF cells; Tony Pawson (Samuel Lunenfeld Research Institute) for the anti-UNC5 antibody and M. Tessier-Lavigne (Stanford University) for the chicken Myc-netrin stable cell line. We also thank Robert Kruger for carefully reading the manuscript. This work is supported by grants from the National Institutes of Health.