1. Developmental Biology
Download icon

N-cadherin-regulated FGFR ubiquitination and degradation control mammalian neocortical projection neuron migration

  1. Elif Kon
  2. Elisa Calvo-Jiménez
  3. Alexia Cossard
  4. Youn Na
  5. Jonathan A Cooper
  6. Yves Jossin  Is a corresponding author
  1. Institute of Neuroscience, Université Catholique de Louvain, Belgium
  2. Fred Hutchinson Cancer Research Center, United States
Research Article
  • Cited 0
  • Views 699
  • Annotations
Cite this article as: eLife 2019;8:e47673 doi: 10.7554/eLife.47673

Abstract

The functions of FGF receptors (FGFRs) in early development of the cerebral cortex are well established. Their functions in the migration of neocortical projection neurons, however, are unclear. We have found that FGFRs regulate multipolar neuron orientation and the morphological change into bipolar cells necessary to enter the cortical plate. Mechanistically, our results suggest that FGFRs are activated by N-Cadherin. N-Cadherin cell-autonomously binds FGFRs and inhibits FGFR K27- and K29-linked polyubiquitination and lysosomal degradation. Accordingly, FGFRs accumulate and stimulate prolonged Erk1/2 phosphorylation. Neurons inhibited for Erk1/2 are stalled in the multipolar zone. Moreover, Reelin, a secreted protein regulating neuronal positioning, prevents FGFR degradation through N-Cadherin, causing Erk1/2 phosphorylation. These findings reveal novel functions for FGFRs in cortical projection neuron migration, suggest a physiological role for FGFR and N-Cadherin interaction in vivo and identify Reelin as an extracellular upstream regulator and Erk1/2 as downstream effectors of FGFRs during neuron migration.

https://doi.org/10.7554/eLife.47673.001

Introduction

The mammalian neocortex has a complicated structure, with multiple layers of different types of neurons linked together in microcircuits. Cortical neurons are generated by progenitor divisions at or near the ventricular zone (VZ), then migrate to defined positions in the developing cortical plate (CP) where they differentiate and connect among themselves and to other brain regions. In general, projection neurons arise from the local VZ, below the CP, while interneurons migrate tangentially from the ganglionic eminences (Parnavelas, 2000; Rash and Grove, 2006). The projection neurons are layered in birth order, such that neurons born early in development are below those born later (Takahashi et al., 1999). Correct neuron layering is important: major defects in projection neuron migration may underlie lissencephalies and heterotopias coupled with mental retardation while milder defects are associated with mental disabilities including dyslexia, schizophrenia, epilepsy, and autism spectrum disorders (Bielas et al., 2004; Francis et al., 2006; Romero et al., 2018). A great deal is known about the distinct neuron migration paths and the signals that regulate them but a significant number of patients with developmental disorders lack a diagnosis. It is thus crucial to identify further causes of disease and understand them mechanistically.

Migrating projection neurons have been visualized in fetal mouse brains in which individual neurons have been labeled in utero (Tabata and Nakajima, 2001; Nadarajah and Parnavelas, 2002; Noctor et al., 2004). Migration occurs in distinct phases. First, neurons move radially from the VZ into the lower intermediate zone (IZ). They then become multipolar and despite an apparent irregular movement, move towards the CP (Nadarajah et al., 2003; Tabata and Nakajima, 2003; Noctor et al., 2004; Jossin and Cooper, 2011). When multipolar neurons reach the upper IZ, they become bipolar and traverse the CP by glial-guided locomotion. The lower IZ therefore constitutes a multipolar migration zone (MMZ) and the upper IZ and CP a radial migration zone (RMZ). Movement out of the MMZ is regulated by many signals, which, if disrupted, can lead to altered layering (Cooper, 2014; Kon et al., 2017). One signal that regulates exit from the MMZ is Reelin, an extracellular matrix protein that activates the small GTPases Rap1A and RalA to upregulate surface levels of neural cadherin (NCad, CDH2) (Jossin, 2011; Jossin and Cooper, 2011). Inhibiting Reelin, Rap1 or NCad interferes with the orientation of multipolar neurons, increases sideways and downwards movements, and increases time spent in the MMZ. Defects in multipolar migration may contribute to the altered layering in Reelin mutant mice and humans (Lambert de Rouvroit and Goffinet, 1998; Hong et al., 2000).

Fibroblast growth factors (FGFs) and their receptors (FGFRs) are important during the development of many tissues and for wound healing, tissue repair and metabolism after birth (Beenken and Mohammadi, 2009; Ornitz and Itoh, 2015). There is wide functional redundancy between family members, with specificity conferred by cell-type-specific expression of FGFs and FGFRs and alternative splicing of FGFRs. At the cellular level, FGFRs regulate cell proliferation, migration, differentiation, survival and cell shape. At the molecular level, FGF and heparan sulfate proteoglycan bind to FGFRs and induce FGFR dimerization and activation. Activated FGFRs autophosphorylate on multiple cytoplasmic tyrosine residues, followed by recruitment and phosphorylation of a variety of downstream signaling proteins (Ornitz and Itoh, 2015). Following activation, FGFRs are down-regulated by ubiquitination, endocytosis and lysosomal degradation (Katzmann et al., 2002; Haugsten et al., 2005). Defects in FGFR activation or down-regulation can lead to anomalous signaling and are associated with developmental defects, metabolic disorders and cancer (Wesche et al., 2011; Ornitz and Itoh, 2015).

In addition to FGFs and heparan sulfate, cell surface proteins including neural cadherin (NCad or CDH2), epidermal cadherin (ECad or CDH1), L1 cell adhesion molecule (L1CAM) and neural CAM (NCAM) can also bind to and activate FGFRs (Williams et al., 1994; Williams et al., 2001; Suyama et al., 2002; Brown et al., 2016). Both NCad and FGFRs are highly expressed during the epithelial-mesenchymal transition of cancer cells and their interaction may be important for metastasis. Indeed, tumor cells artificially over-expressing NCad require FGFR activity for metastasis. Nevertheless, the role of NCad-dependent FGFR activation in metastasis pathology remains unclear (Hulit et al., 2007; Qian et al., 2014).

In the developing cerebral cortex FGFR1-3 are expressed in the VZ and MMZ (Iwata and Hevner, 2009; Hébert, 2011) (Allen Institute for Brain Science. Allen Developing Mouse Brain Atlas. Available from: http://developingmouse.brain-map.org/) and FGFRs have been associated with neurodevelopmental diseases including schizophrenia (O'Donovan et al., 2009; Terwisscha van Scheltinga et al., 2013), epilepsy (Coci et al., 2017; Okazaki et al., 2017), autism spectrum disorders (Wentz et al., 2014; Coci et al., 2017) and lissencephaly (Tan and Mankad, 2018), suggesting possible roles in neuron migration. However, analysis of cortical neuron migration in FGFR mutant mice has been inconclusive for two reasons. First, functional redundancy may suppress the phenotypes of loss of function mutants (Beenken and Mohammadi, 2009; Hébert, 2011; Ornitz and Itoh, 2015). Second, FGFRs are needed for developmental steps that occur before migration, such as regional patterning of the cortex, neurogenesis and radial glia differentiation (Shin et al., 2004; Rash and Grove, 2006; Mason, 2007; Kang et al., 2009; Paek et al., 2009). Therefore, the roles of FGFRs in neocortical neuron migration are unclear.

In this paper we report that FGFR1-3 have overlapping functions during the multipolar migration in vivo. FGFRs are required downstream of Rap1 for multipolar cells to orient towards the CP, adopt bipolar morphology, and migrate out of the MMZ. We found that Rap1-dependent NCad upregulation stabilizes FGFRs by inhibiting K27- and K29-linked polyubiquitination and lysosomal degradation and that NCad-FGFR cis interaction (on the same cell) is involved. Consequently, FGFRs accumulate and are activated, resulting in prolonged activation of Erk1/2 when neurons are stimulated in vitro with Reelin. In vivo inhibition of K27-linked polyubiquitination or overexpression of FGFRs rescues the migration of neurons with inhibited Rap1. Inhibition of Erk1/2 activity in the developing cerebral cortex induces a similar phenotype as FGFR or Rap1 inhibition. These data reveal a novel function of FGFRs in cortical projection neuron migration and the control of its activity by ubiquitination and NCad interaction in vivo. To our knowledge, this is the first physiological role for FGFR-NCad interaction during tissue development. Furthermore, we identified FGFRs as mediating Reelin activation of Erk1/2 to control migration during the multipolar phase. These findings provide insights into FGFR mutation-related inherited brain diseases.

Results

FGFRs are required for multipolar neurons to orient correctly and become bipolar in vivo

To avoid potential functional redundancy, we tested the importance of FGFRs in neuron migration by inhibiting all family members. Cytoplasmic domain deletion mutants of FGFR1-3 are dominant negative (DN) because they form non-functional heterodimers with all FGFR family members (Ueno et al., 1992). To avoid effects on neurogenesis, DN mutants were expressed from the NeuroD promoter, which is activated after cells leave the VZ (Jossin and Cooper, 2011). Apical neural stem cells located at the VZ were electroporated in utero (Tabata and Nakajima, 2001) at embryonic day E14.5 with DN FGFR1-3 along with GFP and the positions of daughter cells were monitored 3 days later at E17.5. While most control neurons expressing GFP alone had entered the RMZ, neurons over-expressing DN mutant but not full-length FGFR1-3 were arrested in the MMZ (Figure 1a). These results suggest that the FGFR1-3 cytoplasmic domains are important for multipolar migration. The knock-down of FGFR1 or FGFR2 using specific shRNAs also induced an arrest of cells at the MMZ, with a more pronounced phenotype when the two receptors are downregulated together (Figure 1b, Figure 1—figure supplement 1). The knock-down of FGFR3 resulted in a small, statistically non-significant effect on cell positioning (Figure 1b, Figure 1—figure supplement 1). These results suggest that FGFRs work redundantly with a prominent role for FGFR1 and FGFR2.

Figure 1 with 1 supplement see all
FGFRs regulate projection neuron migration in vivo.

(a,b) shRNA against FGFR1, 2 or 3 and dominant-negative (DN) but not wildtype FGFR1, 2 or 3 induce an accumulation of neurons at the MMZ. (a) pNeuroD-FGFR(DN) and pNeuroD-FGFR plasmids, expressed in neurons, and (b) shRNA against FGFR1, 2 or 3 or negative control shRNA (shCtrl) were co-electroporated in utero with pCAG-GFP, expressed in progenitors and neurons, at embryonic day (E) 14.5. (a,b) Three days later, cryosections were prepared and labeled for DAPI (blue) and GFP (green). The cerebral wall was subdivided into radial morphology zone (RMZ), multipolar morphology zone (MMZ) and VZ. Graphs show the percentage of cells in the RMZ. n = 5 Control, 7 FGFR1(DN), 4 FGFR2(DN), 4 FGFR3(DN), 3 FGFR1, 3 FGFR2, 4 FGFR3, 6 shCtrl, 4 shFGFR1, 4 shFGFR2, 4 shFGFR3, 5 shFGFR1+2, 3 shFGFR1+3, 3 shFGFR2+3. P values: FGFR1(DN): 9.6E-6, FGFR2(DN): 4.2E-6, FGFR3(DN): 4.0E-5, FGFR1: 0.245, FGFR2: 0.170, FGFR3: 0.353, shFGFR1: 3.0E-4, shFGFR2: 2.9E-3, shFGFR3: 0.169, shFGFR1+2: 6.4E-5, shFGFR1+3: 1.0E-4, shFGFR2+3: 3.2E-3. Error bars, s.e.m. ***p<0.001, **p<0.01, *p<0.05, NS, not significant Scale bar 100 µm.

https://doi.org/10.7554/eLife.47673.002

To test whether inhibition of FGFRs alters cell proliferation, fate or apoptosis, we examined marker expression 2 days after electroporation. At this stage there is no significant difference between FGFR-inhibited and control cells in their position in the cortex, with most GFP+ cells located within the IZ. FGFR1(DN) had no effect on the proportion of GFP+Ki67+ proliferative cells, GFP+Sox2+ apical neural stem cells or GFP+Tbr2+ basal progenitors (Figure 2a). FGFR-inhibited neurons were correctly specified, as shown by the normal expression of Satb2, a marker for upper layer neurons born at the time of the electroporation. Immunostaining for activated caspase-3 showed no increase in cell death (Figure 2a).

Inhibiting FGFRs in post-mitotic neurons has no effect on proliferation and differentiation but regulates multipolar neuron orientation and morphology.

In utero electroporation was performed at embryonic day E14.5 and analyzed 2 days later (a–f) or 3 days later (g–i). (a) Inhibition of FGFRs did not affect cell division (Ki67), apical (Sox2) or basal (Tbr2) progenitor cells, neuronal commitment (Satb2), or survival (cleaved Caspase-3). Expression of CherryFP (red) alone (control) or with FGFR1(DN) as indicated. After immunostaining for the indicated markers (green), the results were quantified by counting the number of labeled electroporated cells in a constant area of each section and averaged across sections from at least three different embryos for each antibody. Values are normalized to control (100%). (mean ± s.e.m.). NS, not significant. Scale bars, 50 μm for Ki67, Sox2 and Tbr2, 25 µm for Satb2, 100 μm for cleaved Caspase 3. (b, c, d) Inhibition of FGFR did not affect the number of neurites or the length to width morphology of multipolar cells. (b) High magnification of GFP+ multipolar neurons within the MMZ following overexpression of GFP or FGFR(DN). (c) Proportion of GFP+ cells with the indicated number of neurites within the MMZ. (d) Ratio of length/width of the GFP+ cells within the MMZ as an indicator of cell shape. P value: 0.196. (mean ± s.e.m.). NS, not significant. Arrows indicate the neurites, arrowheads indicate the axons. Scale bar 10 µm e) FGFR-inhibited neurons are disoriented. Golgi staining (green) of MMZ neurons (purple). The figure shows examples of multipolar neurons with their Golgi facing the CP (white arrows) or facing other directions (white arrowheads). The percentage of cells with Golgi facing the cortical plate was calculated (mean ± s.e.m.). *p<0.05, P value: 0.013. Scale bar 10 µm. (f) FGFR inhibition affects the multipolar to radial transition. Computer-based reconstruction of GFP+ neurons morphologies at the multipolar to radial transition zone (MRT) and the lower RMZ. The graph shows the percentage of bipolar radially oriented neurons. Scale bar 30 µm. Error bars, s.e.m. ***p<0.001, P value: 6.5E-6. (g, h, i) Inhibition of FGFR did not affect the length of the leading process and the length-to-width morphology of radially migrating cells. (g) High magnification of GFP+ bipolar neurons within the RMZ following overexpression of GFP or FGFR(DN). (h) Length of the leading process of GFP+ bipolar cells within the RMZ. P value: 0.180. (i) Ratio of length/width of the GFP+ cells within the RMZ as an indicator of cell shape. P value: 0.155 Arrows indicate the leading process, Scale bar 10 µm. (mean ± s.e.m.). NS, not significant.

https://doi.org/10.7554/eLife.47673.005

To gain insight into the mechanism underlying the migration defect, we analyzed the morphology of migrating neurons. Analysis of the morphology revealed no difference in the number of neurites or in the cell body length-to-width ratio of FGFR-inhibited multipolar neurons compared to control multipolar neurons (Figure 2b–d). However, while most control multipolar cells had their Golgi apparatus oriented towards the CP, fewer FGFR-inhibited neurons had their Golgi facing the CP, suggesting a failure to orient correctly (Figure 2e). In addition, while most control electroporated neurons at the multipolar to radial transition zone had transformed into bipolar cells, FGFR-inhibited neurons were still mostly multipolar (Figure 2f). Yet, the few FGFR-inhibited bipolar neurons migrating in the RMZ exhibited no difference in the length of the leading process and the cell body length-to-width ratio compared to control cells and possess an axon at the rear (Figure 2g–i). These results suggest that FGFRs are required for multipolar neurons to orient correctly, become bipolar, exit the MMZ, and enter the RMZ. For simplicity we will call this phenotype a defect in multipolar migration.

Rap1 and NCad regulate FGFRs protein levels to control multipolar migrating neurons in vivo

Since the phenotype induced by dominant-negative FGFRs resembles that induced by inhibiting Reelin receptors, NCad or Rap1 (Jossin and Cooper, 2011), there may be a common mechanism. Therefore, we tested for epistasis by over-expressing FGFRs when Rap1 is inhibited by the Rap1 GTPase-activating protein (Rap1GAP). The migration defect induced by Rap1GAP was partly suppressed by over-expression of wild-type FGFR1, 2 or 3 (Figure 3a). This suggests that signals from the Reelin-Rap1-NCad pathway may require FGFRs to stimulate multipolar migration, perhaps in parallel with or downstream of NCad.

Figure 3 with 1 supplement see all
Rap1 and NCad regulate FGFR levels and function in multipolar migrating neurons.

(a) FGFR1, 2 and 3 partially rescue the neuronal migration phenotype induced by Rap1 inhibition. E14.5 embryos were electroporated in utero with pCAG-GFP, pNeuroD vector or pNeuroD-Rap1GAP (RG), and pNeuroD-FGFR1, 2 or 3 as shown. Cryosections were prepared 3 days later and labeled for DAPI (blue) and GFP (green). The cerebral wall was subdivided into radial morphology zone (RMZ), multipolar morphology zone (MMZ) and VZ. Graphs show the percentage of cells in the RMZ (mean ± s.e.m.). ***p<0.001; *p<0.05, P values: Rap1GAP (RG): 9.8E-8, RG+FGFR1: 7.0E-4, RG+FGFR2: 3.0E-4, RG+FGFR3: 0.020 (n = 4 Control, 4 Rap1GAP (RG), 7 RG+FGFR1, 7 RG+FGFR2, 4 RG+FGFR3). (b) Protein abundance of FGFR1-GFP is regulated by Rap1 in vivo. E14.5 embryos were electroporated in utero with a mixture of pCAG-CherryFP, pNeuroD-FGFR1-GFP and either vector or pNeuroD-Rap1GAP. Two days later, mCherry and FGFR1-GFP were detected by epifluorescence. The graphs show mean and standard deviation of image intensity measured across lines drawn through the center of the cell body for eight neurons in each case. (c) Embryonic cortical neurons were electroporated to overexpress pCAG-NCad-HA or with a control plasmid, cultured for 2 days then analyzed for the protein level of NCad-HA and endogenous FGFR1 by Western blot.

https://doi.org/10.7554/eLife.47673.007

NCad can bind to, stabilize, and activate FGFR1 in cell culture (Suyama et al., 2002; Sanchez-Heras et al., 2006), providing a potential mechanism for FGFR activation in multipolar neurons. Therefore, we tested whether Rap1 regulates FGFR protein abundance in vivo. Since we could not reliably detect endogenous FGFR by immunofluorescence, we co-electroporated FGFR1-GFP and CherryFP with Rap1GAP or control plasmid in utero. The level of FGFR1-GFP in CherryFP+ neurons was reduced when Rap1 was inhibited (Figure 3b). In addition, over-expressing NCad-HA in cultured primary neurons increased the level of endogenous FGFR1 (Figure 3c), consistent with NCad mediating Rap1-dependent FGFR1 stabilization in vivo. As expected, expressing NCad but not ECad increased protein levels of all three FGFRs in cultured cells (Figure 3—figure supplement 1). Control experiments showed that FGFR inhibition did not change the protein abundance of NCad and did not perturb NCad homophilic interaction properties (Figure 3—figure supplement 1b). NCad was still able to accumulate at cell-cell junctions in the presence of FGFR1(DN) (Figure 3—figure supplement 1c).

These results extend previous reports that NCad can increase FGFRs protein levels in cell culture to an in vivo developmental system.

NCad homophilic adhesion is dispensable for the multipolar migration and for increasing FGFR protein levels

If NCad regulates FGFRs during multipolar migration, NCad-mediated cell-cell adhesion may be dispensable. To test this possibility, we generated a mutant NCad that is incapable of forming homophilic cell-cell adhesion. W161 of NCad (numbered from the initiator methionine, corresponding to W2 in the mature protein) is required for NCad-NCad binding between cells (Tamura et al., 1998; Pertz et al., 1999). As expected, NCadW161A did not bind NCad expressed on different cells (Figure 4—figure supplement 1a lane 2) but still bound NCad expressed on the same cell (Figure 4—figure supplement 1b lane 2). Remarkably, NCadW161A rescued the movement of Rap1-inhibited neurons in vivo (Figure 4), and increased FGFR1 protein level in vitro (Figure 4—figure supplement 1c). NCadW161A was expressed at the same level as NCad (Figure 4—figure supplement 2). These results suggest that NCad function in multipolar migration is independent of NCad-NCad trans interactions but may require NCad binding to FGFR.

Figure 4 with 2 supplements see all
NCad homophilic binding mutant NCadW161A but not ECad rescues multipolar migration of Rap1-inhibited neurons.

E14.5 embryos were electroporated in utero with pCAG-GFP, pNeuroD-Rap1GAP (RG), and pNeuroD vector, NCad, NCadW161A or ECad. Cryosections were prepared 3 days later and labeled for DAPI (blue) or GFP (green). The graph shows the percentage of cells in the RMZ (mean ± s.e.m.). P values: RG+NCad: 9.8E-6, RG+ NCadW161A: 9.3E-6, RG+ECad: 0.213 (n = 5 Rap1GAP (RG), 5 RG+NCad, 6 RG+NCadW161A, 5 RG+ECad). Error bars, s.e.m. ***p<0.001, NS, not significant. Scale bar 100 µm.

https://doi.org/10.7554/eLife.47673.010

NCad EC4 is required for NCad-FGFR Cis interaction and multipolar migration in vivo

To test whether NCad-FGFR binding is necessary to increase FGFR protein levels and rescue migration, we generated an NCad mutant that does not bind FGFRs. To do this, we deleted NCad extracellular domain 4 (EC4), previously reported to mediate NCad-FGFR binding (Williams et al., 2001). NCadΔEC4 no longer bound to FGFR1 in transfected cells (Figure 5a), although it retained homophilic binding to co-transfected NCad (Figure 5b). Also, NCadΔEC4 failed to increase the protein abundance of co-transfected FGFR1 or to activate FGFRs, as observed by an increase in FGFR auto-phosphorylation on tyrosines 653/654 and phosphorylation of the well-known downstream signaling kinases Erk1/2 (Figure 5c). Finally, an FGFR inhibitor (Nakanishi et al., 2014) prevented NCad-induced FGFR auto-phosphorylation and phosphorylation of Erk1/2 (Figure 5c). We also found that NCad binds FGFRs in cis, on the same cell, but not trans, between cells (Figure 5—figure supplement 1). Together, these results show that NCad cis interaction with FGFRs induces FGFR accumulation and FGFR-dependent Erk1/2 phosphorylation in cell culture. Importantly, NCadΔEC4 did not rescue the migration of Rap1-inhibited neurons (Figure 5d). The requirement for EC4 to bind and activate FGFRs and to rescue migration supports the idea that cell autonomous NCad-FGFR binding and activation are required to stimulate multipolar migration in vivo.

Figure 5 with 1 supplement see all
NCad-FGFR cis interaction through NCad EC4 is required for multipolar migration in vivo.

(a) NCad EC4 is required for FGFR1 binding in vitro. Cells were transfected with pCAG-FGFR1-Myc and pCAG-NCad-HA, NCadΔEC4-HA or vector. To equalize FGFR1-Myc expression, half the amount of FGFR1-Myc was transfected with wildtype NCad. One day later, cells were lysed and immunoprecipitated with anti-HA. Lysates and co-immunoprecipitated proteins were analyzed by Western blot. (b) EC4 is dispensable for NCad homophilic binding. Cells were transfected with pCAG-NCad-FLAG and pCAG-NCad-HA, NCadΔEC4-HA or vector. One day later, cells were lysed and immunoprecipitated with anti-HA. Lysates and co-immunoprecipitated proteins were analyzed by Western blot. (c) NCad increases FGFR protein level dependent on EC4, and increases FGFR and Erk1/2 phosphorylation dependent on EC4 and FGFR kinase activity. HEK293T cells were transfected with equal amounts of pCAG-FGFR1-Myc DNA and either pCAG-NCad-HA, pCAG-NCadΔEC4-HA or vector. 24 hr after transfection, the specific FGFR inhibitor Debio1347 was used at 5 µM for 2 hr. Lysates were analyzed by Western blot using the indicated antibodies. Experiments a–c) were repeated independently three times with similar results. (d) NCad EC4 is required for the multipolar migration. E14.5 embryos were electroporated in utero with pCAG-GFP and pNeuroD-Rap1GAP (RG), pNeuroD-NCadΔEC4-HA or vector. Cryosections were prepared three days later and labeled for DAPI (blue) and GFP (green). The graph shows the percentage of cells in the RMZ. n = 4 control, 4 Rap1GAP (RG), 6 RG+ NCadΔEC4. P value: 0.116. Scale bar 100 µm. Error bars, s.e.m., NS, not significant.

https://doi.org/10.7554/eLife.47673.014

NCad but not ECad domains 1–2 increase FGFR protein levels and promote multipolar migration in vivo

Cell culture studies differ as to whether FGFRs bind to both NCad and ECad or only NCad (Williams et al., 2001; Brown et al., 2016). Since we found that NCad but not ECad rescues multipolar migration in vivo (Figure 4a), despite being expressed at similar levels (Figure 4—figure supplement 2), we re-examined interactions between FGFRs, NCad and ECad in cell culture. In our hands, NCad, NCadW161A and ECad all bound FGFRs in transfected cells (Figure 6—figure supplement 1) but only NCad and NCadW161A increased the protein levels of co-transfected FGFRs (Figure 3—figure supplement 1a and Figure 4—figure supplement 1c). Thus, NCad but not ECad rescue of multipolar migration correlates with the ability to increase FGFR protein abundance while receptors interaction is necessary but not sufficient.

Our results suggest that a unique feature of NCad, not shared with ECad, is required to increase FGFR protein levels and stimulate migration. We identified the critical region of NCad by the use of NCad/ECad chimeras. Classic cadherins are composed of an extracellular domain (ECD) with five extracellular cadherin (EC) repeats and a highly conserved intracellular domain (ICD) that interacts with signaling proteins. We switched the entire ECD and ICD of NCad and ECad, creating ENCad and NECad (Figure 6a). NECad but not ENCad was able to increase FGFR protein abundance (Figure 6a). Moreover, NECad but not ENCad rescued the positional defect observed when Rap1 is inhibited in vivo (Figure 6b). This suggests that the specificity of NCad to protect FGFR from degradation and to function in multipolar migration in vivo lies in the ECD. To map the NCad-specific function more closely, we prepared two other chimeric proteins: ENNCad has the EC1, EC2 and the first half of EC3 domains of ECad and the remainder of NCad, while NENCad has the second half of EC3, EC4 and EC5 domains of ECad and the remainder of NCad. NENCad but not ENNCad was able to increase FGFR protein level and rescue the positional defect observed when Rap1 is inhibited in vivo (Figure 6). These results show that EC4-5 of either NCad or ECad can protect FGFRs from degradation and stimulate migration provided that EC1-2 and part of EC3 are derived from NCad. Overall, these data demonstrate that cadherin EC4 interaction with FGFR is necessary but not sufficient to stabilize and activate FGFRs and regulate multipolar migration in vivo. Additional unique features carried by NCad EC1-2 are also needed.

Figure 6 with 1 supplement see all
NCad EC1-2 are required to increase FGFR protein levels and stimulate multipolar migration in vivo.

(a) NCad EC1-2 are necessary to increase FGFR protein abundance. Cells were transfected to express the indicated proteins. 2 days later, protein levels were observed by Western blot. Similar results were obtained from three independent experiments. The figure includes a schematic representing the chimeric proteins used. (b) NCad EC1-2 promote neuronal migration in vivo. In utero electroporation at embryonic day E14.5 and analysis 3 days later. pNeuroD plasmids coding for the indicated proteins and pCAG-GFP were co-electroporated. The graph shows the percentage of cells in the RMZ (mean ± s.e.m.). P values: RG+NCad: 9.8E-6, RG+ECad: 0.215, RG+NECad: 5.4E-4, RG+ENCad: 0.080, RG+NENCad: 4.0E-3, RG+ENNCad: 0.206. n = 5 Rap1GAP (RG), 5 RG+NCad, 5 RG+ECad, 8 RG+NECad, 5 RG+ENCad, 5 RG+NENCad, 5 RG+ENNCad. Error bars, s.e.m. ***p<0.001, **p<0.01, NS, not significant.

https://doi.org/10.7554/eLife.47673.017

FGFR K27/K29-linked polyubiquitination and lysosomal degradation control multipolar migration in vivo

The Rap1/NCad-dependent increase in FGFR protein in vivo (Figure 3b,c) and in vitro (Figure 3—figure supplement 1a) suggests that NCad may inhibit FGFR degradation, as observed in some cancer cells (Suyama et al., 2002). Degradation of many cell surface receptors involves ubiquitination of their cytoplasmic domains and targeting to the lysosome (Piper et al., 2014). When FGFR1-Myc was co-expressed with HA-ubiquitin, a ladder of FGFR1-ubiquitin conjugates could be immunoprecipitated with antibodies against Myc (Figure 7a lane 3) or HA (data not shown). FGFR ubiquitination was inhibited by co-expressed NCad but not NCadΔEC4 (Figure 7a) or ECad (data not shown), consistent with NCad binding inhibiting FGFR ubiquitination. Ubiquitin ladders can result from the addition of single ubiquitin moieties at many sites (multi-monoubiquitination) or addition of ubiquitin chains (polyubiquitination). Seven lysine (K) residues on the ubiquitin molecule may be used for polyubiquitination, resulting in diverse outcomes for the target protein (Ikeda and Kerppola, 2008; Fushman and Wilkinson, 2011; Sadowski et al., 2012). To test whether over-expressed FGFR1-Myc is multi-monoubiquitinated or polyubiquitinated, we inhibited polyubiquitination by co-over-expressing a ubiquitin mutant (UbiK0), in which all 7 K residues substituted to arginine (R) (Lim et al., 2005). When co-expressed with FGFR1, HA-UbiK0 but not HA-UbiWT increased FGFR1 protein level and decreased FGFR1 ubiquitination (Figure 7b) suggesting polyubiquitination. To identify the specific polyubiquitin linkage, we used ubiquitin mutants that contain single K to R substitutions. We found that over-expressing UbiK27R or UbiK29R but not other mutants increased FGFR protein level to the same level as that induced by the presence of NCad-HA (Figure 7c). Co-expression of UbiK27R or UbiK29R with NCad did not have any cumulative effect, suggesting that FGFR is degraded following attachment of K27- and K29-linked polyubiquitin and that NCad inhibits this process. To test whether K27 or K29 linkages are sufficient or whether both are needed, we used ubiquitin mutants where all lysines except 27 or 29 are mutated to arginine (UbiK27 and UbiK29). Co-expressing together UbiK27 and UbiK29, which allows only K27 and K29 linkages, did not inhibit FGFR degradation (compare to UbiWT, Figure 7d). However, expressing either UbiK27 or UbiK29 inhibited FGFR degradation, suggesting that both K27 and K29 ubiquitin linkages are required for FGFR degradation. Importantly, preventing FGFR degradation in vivo by the overexpressing UbiK27R in utero partially rescued the multipolar migration of Rap1-inhibited neurons (Figure 7e).

Figure 7 with 1 supplement see all
FGFRs K27/K29-linked polyubiquitination and lysosomal degradation controls multipolar neuronal migration in vivo.

(a) NCad but not NCadΔEC4 inhibits FGFR1 ubiquitination. Cells were transfected with pCAG-FGFR1-Myc and pCAG-NCad, NCadΔEC4-HA or vector. One day later, cells were lysed and proteins immunoprecipitated with anti-Myc. Lysates and immunoprecipitates were analyzed with Western blotting using antibodies to Myc and ubiquitin. To equalize FGFR1-Myc levels, half the amount of DNA was used for FGFR1-Myc when expressed with NCad-HA. (b) HA-UbiKO but not HA-UbiWT increased FGFR1 protein level and decreased FGFR1 ubiquitination. Cells were transfected with pCAG-FGFR1-Myc and HA-UbiKO or HA-UbiWT. One day later, cells were lysed and proteins immunoprecipitated with anti-Myc. Lysates and immunoprecipitates were analyzed with Western blotting using antibodies to Myc and HA. (c) Inhibition of K27- and K29-linked polyubiquitination increases FGFR1 protein level. Ubiquitin-GFP mutants in which one lysine is mutated into arginine were used to identify lysine residues required for polyubiquitin chain formation. Co-translational cleavage detaches the GFP and frees the terminal glycine of ubiquitin for subsequent conjugation (Boname et al., 2010). The cleaved GFP was used to quantify ubiquitin mutant expression. The graph shows the relative FGFR1-Myc band intensity when expressed in the presence or absence of NCad-HA and in the presence of an ubiquitin mutant as indicated (mean ± s.e.m.). P values: UbiK6R: 0.026, UbiK11R: 0.034, UbiK27R: 0.490, UbiK29R: 0.466, UbiK33R: 0.036, UbiK48R: 0.032, UbiK63R: 0.024, UbiWT: 4.8E-4. n = 4 UbiK6R, 3 UbiK11R, 3 UbiK27R, 3 UbiK29R, 4 UbiK33R, 4 UbiK48R, 3 UbiK63R, 8 UbiWT. (d) FGFR1 levels remain normal only when both K27- and K29-linked polyubiquitination are permitted. HA-Ubiquitin mutants in which all but one lysine is mutated into arginine were used to allow only one type of polyubiquitin chain formation (UbiK27 and UbiK29). (e) Inhibition of K27-linked polyubiquitin chain formation in vivo rescues the migration defect of Rap1GAP-expressing cells. In utero electroporation at embryonic day E14.5 and analysis 3 days later. Plasmids coding for the indicated proteins and GFP were co-electroporated. The graph shows the percentage of cells in the RMZ (mean ± s.e.m.). P values: RG+UbiK27R: 1.7E-3, RG+UbiWT: 0.471. n = 4 Rap1GAP (RG), 8 RG+UbiK27R, 6 RG+UbiWT. (f) Endogenous FGFR1 is degraded by the lysosome in vivo. Primary embryonic cortical neurons were cultured in the presence of 250 nM proteasome inhibitor epoxomycin or 300 µM lysosome inhibitor leupeptin for 4 hr and analyzed by Western blot. Similar results were obtained in three independent experiments. Scale bar 100 µm. *p<0.05, **p<0.01 ***p<0.001, NS not significant.

https://doi.org/10.7554/eLife.47673.020

The requirement for both K27 and K29 polyubiquitin linkages suggested that FGFR may be targeted for lysosomal degradation (Chastagner et al., 2006; Ikeda and Kerppola, 2008; Zotti et al., 2011). Indeed, adding lysosome inhibitor leupeptin but not proteasome inhibitor epoxomycin to primary cortical embryonic neurons increased the protein abundance of endogenous FGFR1 (Figure 7f). Leupeptin but not epoxomycin also increased levels of transfected FGFR1 to the same extent as over-expressed NCad, suggesting that NCad protects FGFRs from lysosomal degradation (Figure 7—figure supplement 1). Lysosome inhibition did not have a cumulative effect on FGFR1 in cells over-expressing NCad, confirming that NCad protects FGFRs from degradation by the lysosome (Figure 7—figure supplement 1). We verified that the proteasome inhibitor was active and induced the accumulation of ubiquitinated proteins (Figure 7—figure supplement 1). Overall, these results suggest that NCad regulates multipolar migration in vivo by inhibiting FGFR K27- and K29-linked polyubiquitination and degradation through the lysosome, thereby raising FGFR protein levels. This raises the question of whether FGFR levels are increased when NCad is upregulated by Reelin.

Reelin induces NCad-dependent FGFR and Erk1/2 activation in cortical neurons

To test whether Reelin activation of the Rap1-NCad pathway increases FGFR protein levels and signaling, we stimulated primary cortical embryonic neurons with partly-purified Reelin or Mock conditioned media and assayed FGFR1 levels and Erk1/2 phosphorylation. As a positive control for FGFR1 signaling, neurons were treated for 20 min with FGF2, which activated Erk1/2 phosphorylation (Figure 8a). As expected, treating for 20 min with Reelin induced phosphorylation of Dab1, a known rapid effect of Reelin (Howell et al., 1999), but did not activate Erk1/2 or increase FGFR1 levels (Figure 8a). However, Reelin did increase FGFR1 protein abundance after 15 hr, consistent with inhibition of FGFR degradation and slow accumulation of FGFR1 over time (Figure 8b lane 4). Remarkably, 15 hr Reelin treatment also activated Erk1/2, to a similar extent as 20 min exposure to EGF or FGF2 (Figure 8b lanes 2 and 3). FGFR1 protein level and Erk1/2 phosphorylation were also increased by 15 hr treatment with R3-6 (Figure 8c), a fragment of Reelin that is produced in vivo and is necessary and sufficient for Reelin regulation of neuron migration (Jossin et al., 2004; Jossin et al., 2007).

Figure 8 with 1 supplement see all
Reelin induces Rap1- and NCad-dependent FGFR and Erk1/2 activity in cortical neurons.

E16.5 mouse cortical neurons were cultured for 3 days then stimulated with FGF2 or Reelin for different times. All experiments were repeated three times with similar results. (a) Short-term Reelin stimulation does not increase FGFR1 protein level or Erk1/2 phosphorylation. Neurons were stimulated for 20 min with 75 ng/ml FGF2, Mock- or Reelin-conditioned media. (b) Long-term Reelin stimulation increases FGFR1 protein level and FGFR1 and Erk1/2 phosphorylation dependent on FGFR1 kinase activity. Neurons were stimulated for 15 hr with Mock- or Reelin-conditioned media or for 20 min with FGF2 or EGF. FGFR inhibitor Debio1347 was used at a concentration of 5 µM for a total of 17 hr before cell lysis. (c) Reelin fragment R3-6 induces FGFR1 accumulation and Erk1/2 activation. Neurons were stimulated for 15 hr with Mock, R3-6 or Reelin-conditioned medium. (d) Long-term Reelin stimulation of FGFR1 protein level and Erk phosphorylation requires Rap1 and NCad. Neurons were electroporated with pCAG-Rap1GAP, pCAG-NCadDN, or vector, incubated for 2 days, then stimulated with Mock- or Reelin-conditioned media for 15 hr and analyzed by Western blotting.

https://doi.org/10.7554/eLife.47673.023

Reelin-induced Erk1/2 phosphorylation but not FGFR1 accumulation was completely abrogated by FGFR kinase inhibitor Debio1347. Debio1347 was specific because it inhibited Erk1/2 activation by 20 min FGF2 but did not inhibit Erk1/2 activation by EGF (Figure 8b) or Reelin-induced Dab1 phosphorylation (Figure 8—figure supplement 1). These results suggest that Reelin-induced Erk1/2 phosphorylation is dependent on FGFR activity and correlates with its effect on FGFR protein levels.

To test whether FGFR1 protein upregulation and Erk1/2 activation require Rap1 or NCad, cortical neurons were transfected with Rap1GAP or NCadDN, both of which inhibit NCad accumulation at the plasma membrane and inhibit multipolar migration (Jossin and Cooper, 2011). Inhibiting Rap1 or NCad prevented Reelin-induced FGFR1 protein increase and Erk1/2 activation (Figure 8d). Taken together, these results show that Reelin or R3-6 can activate FGFR1 signaling and Erk1/2 through the Rap1-NCad pathway.

Erk1 and Erk2 regulate neuron migration in vivo

Since Reelin, Rap1, NCad and FGFRs are all important to activate Erk1/2 and stimulate multipolar migration, we tested whether Erk1/2 are needed for multipolar migration in vivo. We overexpressed dominant-negative mutants of Erk1 or Erk2, each of which inhibits both family members (Watts et al., 1998; Li et al., 1999; Zampieri et al., 2007). Over-expression of either Erk1DN or Erk2DN, but not wildtype Erk1 or Erk2, induced partial migration arrest of multipolar neurons (Figure 9a). Overall our results suggest that multipolar migration of projection neurons requires a Reelin-Rap1-NCad-FGFR-Erk1/2 pathway.

Erk1 and Erk2 regulate multipolar migration in vivo.

(a) Erk1/2 inhibition impairs multipolar migration. E14.5 embryos were electroporated in utero with pCAG-GFP and dominant-negative (DN) or wildtype pNeuroD-Erk1/2. Cryosections were prepared three days later. The graph shows the percentage of cells in the RMZ (mean ± s.e.m.). P values: Erk1(DN): 1.4E-3, Erk2(DN): 1.2E-3, Erk1: 0.497, Erk2: 0.173. n = 4 Control, 6 Erk1(DN), 5 Erk2(DN), 6 Erk1, 4 Erk2. Scale bar 100 µm; Error bars, s.e.m, **p<0.01, NS, not significant. (b) Working model. (1) Reelin, its central fragment R3-6, and possibly other signals activate Rap1 in multipolar neurons. Rap1 upregulates NCad on the cell surface. (2) The NCad fourth cadherin extracellular domain (EC4, orange color) binds FGFRs. This binding, together with a specific function of NCad EC1 and EC2 (green color) prevents FGFR polyubiquitination by mixed K27- and K29-linked polyubiquitin chains and lysosomal degradation. (3) Decreased FGFR ubiquitination causes FGFR accumulation and persistent activation of FGFR signaling pathways, including Erk1/2. Erk1/2 and maybe other effectors are required for the multipolar migration in vivo. The mechanisms by which NCad EC1-2 regulate FGFR stability and by which Erk1/2 regulate migration remain unknown. See text for discussion.

https://doi.org/10.7554/eLife.47673.025

Discussion

Although much is known about the role of FGFRs in apical neural stem cells during neurogenesis and regional patterning of the cortex, the functions and regulation of FGFRs in neuronal migration have not been elucidated. Our findings implicate FGFRs in a signaling pathway that regulates the orientation of projection neuron multipolar migration during the development of the cerebral cortex (see our working model Figure 9b). When FGFRs are inhibited, multipolar neurons accumulate in the multipolar migration zone. This includes a defect in the orientation of multipolar migration with fewer cells facing the CP, followed by a delay in the multipolar to radial morphology transition and a consequent accumulation of cells at the MMZ. Previous studies indicated that the orientation of multipolar migration and the subsequent bipolar morphology transition is triggered by Reelin, or the R3-6 fragment, diffusing from the outer part of the cortex (Jossin et al., 2004; Jossin et al., 2007; Uchida et al., 2009; Jossin and Cooper, 2011). Reelin activates Rap1, and this in turn upregulates NCad on the neuron surface (Jossin and Cooper, 2011). However, the mechanism by which NCad stimulates the multipolar migration was unknown. One possibility was that NCad on migrating neurons engages in homophilic binding interactions with NCad on surrounding cells (Kawauchi et al., 2010; Matsunaga et al., 2017). However, our new evidence leads us to propose a model in which NCad-NCad trans interactions are not required. Instead, NCad binds in cis to FGFRs on the same cell. This cell-autonomous binding inhibits FGFR K27/K29 polyubiquitination and lysosomal degradation, resulting in a large increase in FGFR abundance and prolonged activation of FGFR and Erk1/2 that is required for the multipolar migration. FGFR-dependent Erk1/2 stimulation may therefore be a key signal for orienting multipolar neurons towards the CP.

FGFRs have been implicated in other developmental cell movements, including migration of mesodermal and tracheal cells in Drosophila (Gisselbrecht et al., 1996; Lebreton and Casanova, 2016), neuroblasts in the mouse olfactory bulb (Zhou et al., 2015) and keratinocytes during repair of epidermal injury (Meyer et al., 2012). A possible role for FGFRs in cortical projection neuron migration has not been reported previously, perhaps because FGFRs are needed for early telencephalon patterning and neurogenesis, and because of functional redundancy (Kang et al., 2009; Paek et al., 2009). Cortical layering was defective in a mouse model expressing dominant-negative FGFR1 during neurogenesis but this could be secondary to the defective radial glia processes (Shin et al., 2004). By inhibiting FGFR signaling in postmitotic neurons, we avoided effects on neurogenesis or radial glia processes. Under these conditions, neurons were delayed in the multipolar migration zone, with randomized orientation, suggesting that FGFRs provide directional information. Conventionally, FGFRs are activated by FGFs and heparan sulfate proteoglycans (Ornitz and Itoh, 2015), so directional information could be provided by an FGF gradient, serving as a chemorepellent or attractant, as in the Drosophila trachea (Lebreton and Casanova, 2016). However, a large number of FGFs are expressed in the developing cortex (Ford-Perriss et al., 2001), making it challenging to identify which, if any, may be involved. It is also possible that NCad activates FGFRs in the absence of FGFs. Reelin could stimulate FGFR-dependent Erk1/2 activity in cultured neurons in the absence of added FGF, suggesting that external FGF may not be needed in vivo. The co-clustering of FGFRs with NCad may be sufficient to activate the receptor independently of FGF, with FGFRs acting as ‘catalytic subunits’ downstream of NCad. A similar mechanism has been suggested for activation by the co-receptor Klotho (Lee et al., 2018).

Our cell culture experiments suggest that NCad stabilizes cell-surface FGFR by inhibiting K27/K29-linked polyubiquitination and lysosomal degradation. Importantly, inhibiting K27-linked polyubiquitination by expression of a mutant ubiquitin rescued the migration of Rap1-inhibited neurons, suggesting that this mechanism also occurs in vivo. The involvement of K27 and K29 is unusual. Polyubiquitin chains assemble through any of seven lysine residues on the ubiquitin molecule, resulting in diverse outcomes for the target protein (Fushman and Wilkinson, 2011). While K11 and K48 polyubiquitination is linked to proteasomal degradation and K63 polyubiquitination to lysosomal degradation, the roles of K6-, K27-, K29-, and of K33-linked ubiquitin chains are less clear (Sadowski et al., 2012). K27 and K29 ubiquitin linkages have been implicated in protein interactions and protein degradation (Chastagner et al., 2006; Ikeda and Kerppola, 2008; Zotti et al., 2011; Fei et al., 2013; Zhou et al., 2013; Birsa et al., 2014; Liu et al., 2014). We found here that both K27 and K29 ubiquitin linkages are necessary for FGFR lysosomal degradation. This may involve mixed linear or mixed branched ubiquitin chains or single K27 and K29 chains attached to different lysine residues on the cytoplasmic tail of FGFRs (Yau and Rape, 2016). The E3 ligase and mechanism of lysosomal targeting remains unclear.

ECad also binds FGFRs but, unlike NCad, does not inhibit FGFR ubiquitination, protect FGFRs from degradation, or regulate multipolar migration in vivo. Using mutated receptors and chimeric proteins, we found that FGFRs bind both ECad and NCad and that binding is necessary but not sufficient to stabilize FGFRs and stimulate multipolar migration. Binding requires the fourth NCad EC domain, but FGFR stabilization and neuronal migration rescue also require the first two NCad domains. Their function is unclear. They may bind to a third partner in the complex that helps NCad inhibit FGFR degradation.

Our results indicated that Reelin does not activate ERK signaling directly but does so by stabilizing FGFR. ERK phosphorylation was delayed, presumably due to the time necessary to accumulate sufficient FGFR. A prolonged Erk1/2 activation might be required to induce a signal as is the case for EGF-induced neuronal differentiation of PC12 cells (Traverse et al., 1994). For instance, Erk1/2 signaling has been linked to transcription of matrix metalloproteinases genes (Westermarck and Kahari, 1999) and some metalloproteinases are essential for the organization of the cerebral cortex (Jorissen et al., 2010). Interestingly, Erk1 and Erk2 double mutant mice exhibit a defect in cortical lamination (Imamura et al., 2010). A cell-autonomous function of Erk1/2 signaling could not be determined because of a failure in maintenance of the radial glia scaffolding. However, our results indicate that Erk1/2 activity is required in migrating neurons during the multipolar phase.

While we were writing this manuscript, FGFR2 was reported to regulate neuronal migration and spine density (Szczurkowska et al., 2018). Defective mice showed impaired core behaviors related to autism spectrum disorders. Interestingly, the Reelin pathway also regulates spine density and has been linked to autism (Niu et al., 2008; Lammert and Howell, 2016). Reelin was also reported to regulate the migration of dopaminergic neurons into the substantia nigra, with induction of bipolar morphology (Vaswani et al., 2019). Additional research will be needed to elucidate the mechanism of FGFRs in Reelin-induced migration and spine density and the link with autism spectrum disorders.

Materials and methods

Key resources table
Reagent type (species) or resourceDesignationSource or referenceIdentifiersAdditional information
Strain, strain background (Escherichia coli)One shot TOP10Thermo Fisher ScientificCat #: C404010Chemically Competent Cells
Strain, strain background (Mus musculus)CD1Charles River Laboratories022
Cell line (Homo sapiens)HEK293T cellsATCCCRL-3216
Cell line (Mus musculus)Embryonic primary mouse cortical neuronThis paperN/APrimary culture at E16,5.
AntibodyAnti-HA.11 clone 16B12 (Mouse monoclonal)EurogeneticCat# MMS-101R-500, RRID: AB_10063630WB (1:8000)
IF (1:100)
IP (1:400)
AntibodyAnti-Myc (Rabbit polyclonal)Cell Signaling TechnologyCat# 2272, RRID: AB_10692100WB (1:5000)
AntibodyAnti-Myc-tag clone 9B11 (Mouse monoclonal)Cell Signaling TechnologyCat# 2276, RRID: AB_331783IP (1:200)
AntibodyAnti mono- and polyubiquitinated conjugated, clone FK2 (Mouse monoclonal)Enzo Life ScienceCat# BML-PW8810-0500, RRID: AB_2051891WB (1:1000)
AntibodyAnti-B-actin (Mouse monoclonal)Thermo Fisher ScientificCat# MA5-15739, RRID: AB_10979409WB (1:5000)
AntibodyAnti- p44/42 MAPK (Erk1/2) (Rabbit polyclonal)Cell Signaling TechnologyCat# 9102, RRID: AB_330744WB (1:5000)
AntibodyAnti-Phospho-p44/42 MAPK (Erk1/2) (Thr202/Tyr204) clone D13.14.4E (Rabbit monoclonal)Cell Signaling TechnologyCat# 4370, RRID: AB_2315112WB (1:5000)
AntibodyAnti-FGFR1(D8E4) XP (Rabbit monoclonal)Cell Signaling TechnologyCat# 9740, RRID: AB_11178519WB (1:500)
AntibodyAnti- Dab-1 (E1) (Mouse)Jossin et al., 2004N/AWB (1:1000)
AntibodyPhospho-Tyrosine (P-Tyr-100) #9411(Mouse Monoclonal)Cell Signaling TechnologyCat# 9411, RRID: AB_331228WB (1:1000)
AntibodyAnti-Reelin(G10) Mousede Bergeyck et al., 1997N/AWB (1:1000)
AntibodyAnti-DYKDDDDK (FLAG) Tag (FG4R)
(Mouse monoclonal)
Thermo Fisher ScientificCat# MA5-15255, RRID: AB_2537646WB (1:5000)
AntibodyAnti-GFP (Rabbit polyclonal)Thermo Fisher
Scientific
Cat# A-11122, RRID: AB_221569WB (1:5000)
AntibodyAnti-Ki67 (Mouse monoclonal)BD BiosciencesCat# 556003, RRID: AB_396287IF (1:100)
AntibodyAnti- Sox2 (L1D6A2)
(Mouse monoclonal)
Cell Signaling TechnologyCat# 4900, RRID: AB_10560516IF (1:100)
AntibodyAnti-Tbr2 (Rabbit polyclonal)AbcamCat# ab23345, RRID: AB_778267IF (1:100)
AntibodyAnti-Satb2 (mouseAbcamCat# ab51502
RRID: AB_882455
IF (1:100)
AntibodyAnti-cleaved caspase3 (Rabbit polyclonal)Cell Signaling TechnologyCat# 9661, RRID: AB_2341188IF (1:100)
AntibodyGM130 (Mouse monoclonal)BD BiosciencesCat# 610823, RRID: AB_398142IF (1:100)
AntibodyAnti-mouse IgG, HRP-linked Antibody (horse)Cell Signaling TechnologyCat# 7076, RRID: AB_330924WB (1:5000)
AntibodyAnti-rabbit IgG, HRP-linked Antibody
(goat polyclonal)
Cell Signaling TechnologyCat# 7074, RRID: AB_2099233WB (1:5000)
AntibodyAnti-Mouse IgG (H+L) Cross-Adsorbed Secondary Antibody, Alexa Fluor 488 (Goat Polyclonal)invitrogenCat# A-11001, RRID: AB_2534069IF (1:100)
AntibodyAnti-Rabbit IgG (H+L) Antibody, Alexa Fluor 488 Conjugated (Goat polyclonal)invitrogenCat# A-11008, RRID: AB_143165IF (1:100)
AntibodyAnti-Mouse IgG (H+L) Antibody, Alexa Fluor 568 Conjugated
(goat polyclonal)
invitrogenCat# A-11004, RRID: AB_2534072IF (1:100)
AntibodyAnti-Rabbit IgG (H+L) Cross-Adsorbed Secondary Antibody, Alexa Fluor 568
(Goat polyclonal)
invitrogenCat# A-11011, RRID: AB_143157IF (1:100)
AntibodyAnti-Mouse IgG (H+L) Cross-Adsorbed Secondary Antibody, Alexa Fluor 647 (Goat polyclonal)invitrogenCat# A-21235, RRID: AB_2535804IF (1:100)
AntibodyAnti-Rabbit IgG (H+L) Cross-Adsorbed Secondary Antibody, Alexa Fluor 647 (Goat polyclonal)invitrogenCat# A-21244, RRID: AB_2535812IF (1:100)
Recombinant DNA reagentNeuroD: Rap1GAPJossin and Cooper, 2011N/A
Recombinant DNA reagentNeuroD: NCad
CAG: NCad
Jossin and Cooper, 2011N/A
Recombinant DNA reagentpBS mFgfr1 (CT#92)AddgeneRRID: Addgene_14005mFgfr1 subcloned into pCAG and pNeuroD vectors.
Recombinant DNA reagentCAG:FGFR2
NeuroD:FGFR2
This paperN/A
Recombinant DNA reagentCAG:FGFR3
NeuroD-FGFR3
This paperN/A
Recombinant DNA reagentCAG:FGFR1-DN
NeuroD:FGFR1-DN
This paperN/ADeletion of the ICD, replace by GFP
Recombinant DNA reagentCAG:FGFR2-DN
NeuroD:FGFR2-DN
This paperN/ADeletion of the ICD, replace by GFP
Recombinant DNA reagentCAG:FGFR3-DN
NeuroD:FGFR3-DN
This paperN/ADeletion of the ICD, replace by GFP
Recombinant DNA reagentpRK5:HA-ubiquitin-KOAddgeneRRID: Addgene_17603
Recombinant DNA reagentpRK5:HA-ubiquitin-K27AddgeneRRID: Addgene_22902
Recombinant DNA reagentpRK5:HA-ubiquitin-K29AddgeneRRID: Addgene_22903
Recombinant DNA reagentpHRSIN:6HIS-UbiWT-GFPSchaller et al., 2014N/A
Recombinant DNA reagentpHRSIN:6HIS-UbiK6R-GFPSchaller et al., 2014N/A
Recombinant DNA reagentpHRSIN:6HIS-UbiK11R-GFPSchaller et al., 2014N/A
Recombinant DNA reagentpHRSIN:6HIS-UbiK27R-GFPSchaller et al., 2014N/A
Recombinant DNA reagentpHRSIN:6HIS-UbiK29R-GFPSchaller et al., 2014N/A
Recombinant DNA reagentpHRSIN:6HIS-UbiK33R-GFPSchaller et al., 2014N/A
Recombinant DNA reagentpHRSIN:6HIS-UbiK48R-GFPSchaller et al., 2014N/A
Recombinant DNA reagentpHRSIN:6HIS-UbiK63R-GFPSchaller et al., 2014N/A
Recombinant DNA reagentpFLAG-CMV-hErk1AddgeneRRID: Addgene_49328hErk1 subcloned into pCAG and pNeuroD vectors
Recombinant DNA reagentpFLAG-CMV-hErk1K71RAddgeneRRID: Addgene_49329hErk1K71R subcloned into pCAG and pNeuroD vectors
Recombinant DNA reagentp3XFLAG-CMV7-Erk2AddgeneRRID: Addgene_39223Erk2 subcloned into pCAG and pNeuroD vectors
Recombinant DNA reagentp3XFLAG-CMV7-Erk2_KRAddgeneRRID: Addgene_39224Erk2_KR subcloned into pCAG and pNeuroD vectors
Recombinant DNA reagentNeuroD:NCadΔEC4
CAG: NCadΔEC4
This paperN/ADeletion of residues515–604
Recombinant DNA reagentCAG:NCadDN
NeuroD:NCadDN
Jossin and Cooper, 2011N/ADeletion of residues 99–708
Recombinant DNA reagentCAG: NcadW161A
NeuroD: NcadW161A
This paperN/ACodon 161 TGG replaced by GCG
Recombinant DNA reagentCAG:Ecad
NeuroD:ECad
This paperN/A
Recombinant DNA reagentCAG:ENCad
NeuroD: ENCad
This paperN/A
Recombinant DNA reagentCAG:NECad
NeuroD: NECad
This paperN/A
Recombinant DNA reagentCAG: NENCad
NeuroD: NENCad
This paperN/A
Recombinant DNA reagentCAG: ENNcad
NeuroD: ENNcad
This paperN/A
Recombinant DNA reagentpLKO.1.shFGFR1SigmaCat #: TCRN0000378435Target sequence:
5’-CTGGCTGGAGTCTCCGAATAT-3’
Recombinant DNA reagentpLKO.1.shFGFR2SigmaCat #: TRCN0000023715Target sequence:
5’-GCCAGGGATATCAACAACATA-3’
Recombinant DNA reagentpLKO.1.shFGFR3SigmaCat #: TRCN0000363373Target sequence: 5’CCACTTCAGTGTGCGTGTAAC-3’
Recombinant DNA reagentpCMV:Reelin and pCMV:R3-6Jossin et al., 2004N/A
Peptide, recombinant proteinhFGF-2PeproTechCat #: 100-18B
Peptide, recombinant proteinmEGF-2PeproTechCat #: 315–09
Chemical compound, drugProtease inhibitor cocktailRocheCat# 05056489001
Chemical compound, drugPhosphatase inhibitor cocktailRocheCat #: A32957
Chemical compound, drugB27invitrogenCat #: 17504–044
Chemical compound, drugEpoxomycinSigmaCat #: E3652
Chemical compound, drugLeupeptinCarl RothCat #: CN33
Chemical compound, drugDebio1347SelleckchemCat #: S7665
Chemical compound, drugPenicillin-streptomycinGibcoCat #: 11548876
Commercial assay or kitPlasmid DNA Purification Mini Prep KitIntron BiotechnologyCat# 17098
Commercial assay or kitQuick Gel extraction KitThermo Fisher ScientificCat# K2100-12
Commercial assay or kitHiPure Plasmid Maxiprep KitThermo Fisher ScientificCat# K2100-07
Software, algorithmImage JNIHN/A
OtherDAPI stainingSigmaCat #: D9542
OtherPolyJet In Vitro DNA Transfection ReagentSignagenCat #: SL100688
OtherDynabeads protein AinvitrogenCat #: 10001D
OtherDynabeads protein GİnvitrogenCat #: 1003D
OtherSuper signal West Pico PLUS chemuluminescent substrateThermo ScientificCat #: 34578
OtherO.C.T.VWRCat # 361603E
OtherDMEM-F-12GibcoCat #: 21331–020
OtherDMEM, high glucoseGibcoCat #: 41965–039
OtherCL-X Posure filmThermo fisherCat #: 34091

Mice

CD1 mice were bred in standard conditions and animal procedures were carried out in accordance with European guidelines and approved by the animal ethics committee of the Université catholique de Louvain.

In utero electroporation

Request a detailed protocol

In utero microinjection and electroporation was performed at E14.5 essentially as described (Tabata and Nakajima, 2001), using timed pregnant CD-1 mice. Timed-pregnant mice were anesthetized and each uterus was exposed under sterile conditions. Plasmid solutions containing 1 µg/µl of DNA were injected into the lateral ventricles of the embryos using a heat-pulled capillary. Needles for injection were pulled from Wiretrol II glass capillaries (Drummond Scientific) and calibrated for 1 μl injections. DNA solutions were mixed in 10 mM Tris, pH 8.0, with 0.01% Fast Green. Forceps-type electrodes (Nepagene) with 5 mm pads were used for electroporation (five 50 ms pulses of 45 V using the ECM830 electroporation system, Harvard Apparatus). Embryos were placed back into the abdominal cavity, and mice were sutured.

Histology and immunofluorescence

Request a detailed protocol

Embryos were collected at E16.5 or E17.5. Brains were dissected and successful electroporations were chosen by epifluorescence microscopy. Positive brains were fixed in a 3.7% paraformaldehyde (PFA) in phosphate-buffered saline (PBS) solution and cryoprotected in a 30% sucrose/PBS solution overnight at 4°C. Brains were frozen in optimal cutting temperature compound (OCT), and sectioned with a cryostat at 14-μm-thickness. Sections were placed on slides, permeabilized for 30 min in 0.4% Triton X-100/PBS then blocked for 30 min with 5% normal goat serum (NGS) in 0.4% Triton X100/PBS. Primary antibodies were diluted in 0.4% Triton X100/PBS incubated on slides overnight at 4°C. Sections were washed 3 times for 5 min in 0.4% Triton X100/PBS. Secondary antibodies were diluted in 0.4% Triton X100/PBS and incubated for 1 hr at room temperature. Nuclei were stained with 4,6-diamidino-2-phenylindole (DAPI). Slides were washed three times as before and coverslipped with Fluoroshield with 1, 4-Diazabicyclo [2.2.2] octane (Sigma) as an anti-fade reagent. Images were acquired with an Olympus FV1000 confocal microscope.

Isolation, culture and nucleofection of primary Cortical Neuron

Request a detailed protocol

Neurons were dissected from E16.5 mouse embryo telencephalons. Cells were plated in DMEM-F12 medium (Fisher) with 2% B27 supplement (Fisher) and Penicillin-Streptomycin (Fisher) on 12 well-plate coated with poly-D-lysine (Sigma) and E-C-L (entactin-collagen IV-laminin) Cell Attachment Matrix (Upstate Biotechnology) at a density of 2 × 106 cells per dish. Cultures were maintained at 37°C in a 5% CO2 incubator. After 2 days in culture, neurons were stimulated with partly-purified Reelin or Mock-conditioned media (see below), EGF (Pepro Tech, 100 ng/mL), FGF-2 (Pepro Tech, 75 ng/mL). Debio1347 (5 µM, Selleckchem) was used to inhibit FGFR1, 2 and 3 (Nakanishi et al., 2014; Nakanishi et al., 2015). Cells were lysed with ice-cold NP-40 buffer (150 mM NaCl, 50 mM Tris-HCl (pH 8.0), 1% NP-40, 5 mM EDTA) supplemented with protease and phosphatase inhibitor cocktail (Roche). For plasmid DNA transfection, cells were Amaxa nucleoporated with 5 µg of plasmid DNA in 0.2 mm cuvette using A-033 program and 100 µl of electroporation buffer containing 120 mM Na2HPO4/NaH2PO4 pH 7.2, 5 mM NaCl, 5 mM KCl, 20 mM MgCl2, and 0.5 mM reduced glutathione.

Production of recombinant reelin and R3-6

Request a detailed protocol

HEK293T cells cultured in Dulbecco modified Eagle medium (Fisher) with 10% fetal bovine serum (Fisher) were transfected with the Reelin or R3-6 cDNA constructs (Jossin et al., 2004), using Polyjet (Tebu-Bio). After 24 hr, the medium was replaced with a serum-free medium, which was collected 2 days later and stored at 4°C in the presence of a protease inhibitor cocktail (Complete, Roche). Prior to use, the supernatants were concentrated using Amicon Ultra columns with 100,000-molecular weight cutoff filters (Millipore) to reach the approximate concentration of 400 pM, which was estimated as described previously (Jossin et al., 2004), and dialyzed against culture medium by drop dialysis (Millipore VSWP02500). Mock solutions were prepared from control transfected HEK293T cells and used to control for potential co-purifying proteins.

Immunoprecipitation and western blot

Request a detailed protocol

Transfected 293T cells were lysed with ice-cold NP-40 buffer supplemented with protease and phosphatase inhibitor cocktail (Roche). Lysates were clarified by centrifugation at 14,000Xg for 10 min at 4°C. Antibodies were added to the lysates for 2 hr at 4°C. Dynabeads protein A or protein G magnetic beads (Invitrogen) were washed three times in PBS then blocked in 1% BSA/PBS for 2 hr at 4°C. Beads were washed twice with PBS and once in NP-40 buffer then added into cell lysate mixture and incubated overnight at 4°C. Beads were washed three times with NP-40 lysis buffer. Proteins were eluted by boiling for 5 min in polyacrylamide gel electrophoresis loading buffer and analyzed by sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis.

Proteins were separated by SDS-gel electrophoresis then transferred to nitrocellulose membrane (Amersham Biosciences) by electroblotting. Membranes were blocked in 5% skimmed milk and 0.05% Tween 20 in PBS for 1 hr and incubated overnight at 4°C with antibodies. After three washing steps in PBS with 0.05% Tween 20, membranes were incubated with horseradish peroxidase-conjugated secondary antibodies (DAKO) in blocking solution for 1 hr at room temperature and washed three times. Membranes were treated with the SuperSignal West Pico chemiluminescent substrate (Pierce) and exposed to Hyperfilm ECL (Amersham Biosciences).

Antibodies

The following antibodies were used for immunofluorescence, immunoprecipitation or biochemistry: mouse anti-HA.11 clone 16B12 monoclonal antibody (Eurogentec), rabbit anti-myc (Cell Signaling), anti mono- and poly-ubiquitinated antibody clone FK2 (Enzo), mouse anti-β-Actin(Thermo Pierce), rabbit anti-Erk1/2 (Cell Signaling), rabbit anti p44/42 Erk (Thr 202/Tyr 204) monoclonal antibody (Cell Signaling), rabbit anti-FGFR1(D8E4) monoclonal antibody (Cell Signaling), rabbit anti-FGFR(phosphor-Tyr653/654) (Cell Signaling), mouse anti-Dab1 (E1) (Jossin et al., 2004), mouse anti-phospho-tyrosine antibody (Cell Signaling), mouse anti-Reelin (G10) (de Bergeyck et al., 1997), mouse anti-Flag (Thermo Pierce), rabbit anti-GFP (Invitrogen), mouse anti-Ki67 (Beckton Dickinson), mouse anti-Sox2 (Cell Signaling), Rabbit anti Tbr2 (Abcam), mouse anti-Satb2 (Abcam), rabbit anti-cleaved caspase 3 (Cell Signaling), mouse GM130 (Beckton Dickinson).

Goat secondary antibodies labeled with Alexa 488, 568, and 647 (Invitrogen) for immunofluorescence. Goat anti-mouse or anti-rabbit horseradish peroxidase-conjugated secondary antibodies (Cell Signaling) for biochemistry.

Vector constructions

Request a detailed protocol

Rap1GAP and NCad sequences inserted into the pNeuroD vector were described previously (Jossin and Cooper, 2011). Plasmid containing the coding sequences for FGFR1 (Addgene plasmid # 14005) was used as template to insert the sequences into pCAG or pNeuroD vectors. FGFR2 and FGFR3 were amplified from E16.5 embryonic mouse cortex. Dominant negative forms of FGFR1, 2 and 3 contain the transmembrane and extracellular domains while the intracellular domains where replace by GFP. pRK5-HA-Ubiquitin-KO was a gift from Ted Dawson (Addgene plasmid # 17603). pRK5-HA-Ubiquitin-K27 and pRK5-HA-Ubiquitin-K29 were a gift from Sandra Weller (Addgene plasmid # 22902 and # 22903). Wild-type and all single lysine to arginine ubiquitin mutant vectors pHRSIN-6HIS-UbiWT-GFP, UbiK6R-GFP, UbiK11R-GFP, UbiK27R-GFP, UbiK29R-GFP, UbiK33R-GFP, UbiK48R-GFP, UbiK63R-GFP were provided by R Rezsohazy with the permission of M Malim (Schaller et al., 2014). pFLAG-CMV-hErk1, pFLAG-CMV-hErk1K71R, p3xFlag-CMV7-Erk2, p3xFlag-CMV7-Erk2_KR were a gift from Melanie Cobb (Addgene plasmid # 49328, # 49329, # 39223, # 39224) and subcloned into pCAG and pNeuroD vectors. NCadΔEC4 was made by PCR using junction primers at the KpnI site already present in the sequence resulting in the deletion of residues 515–604. NCadW161A was made by site-directed mutagenesis with oligo 5’-GCT CTA CAA AGG CAG AAG CGA GAC GCG GTC ATC CCG CCA ATC AAC-3’ and its reverse complement, changing codon 161 from TCG to GCG and introducing a silent mutation to destroy an EarI restriction site that was used for screening. NCadDN contains the transmembrane and intracellular domains (deletion of residues 99–708) and was described before (Jossin and Cooper, 2011). ENCad and NECad were made by using PCR using junction primers that inserted SpeI sites five residues N-terminal to the transmembrane domain (between ECad codons 704 and 705, and between NCad codons 717 and 720). NENCad and ENNCad were made using overlap extension PCR with recombination junctions between residues 420–421 of NCad and 415–416 of ECad. All cadherin constructs were of murine origin, terminated with an HA tag and cloned into pCAG and pNeuroD vectors. Codon numbers are given from the initiator methionine. Several FGFR targeting shRNAs (pLKO.1 plasmids either from Sigma or generously provided by Slobodan Beronja and shRNA expression vectors kindly provided by Laura Cancedda and Giovanni Piccoli; Szczurkowska et al., 2018) were tested and the most efficient shRNAs were selected for in vivo experiments. The most effective shRNAs are shFGFR1: clone ID TRCN0000378435 (target sequence 5’-CTGGCTGGAGTCTCCGAATAT-3’), shFGFR2: clone ID TRCN0000023715 (target sequence 5’-GCCAGGGATATCAACAACATA-3’) and shFGFR3: clone ID TRCN0000363373 (target sequence 5’-CCACTTCAGTGTGCGTGTAAC-3’).

Cell line culture

Request a detailed protocol

HEK293T cells (ATCC) maintained in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% FBS, 100 IU/mL penicillin, and 100 μg/mL streptomycin were transfected using Polyjet (Tebu-bio). Cells were cultured at 37°C under 5% CO2, and are mycoplasma-free. The following inhibitors were used: Epoxomycin, (Sigma), Leupeptin (Roth), Debio1347 (Selleckchem).

Statistical analysis

Request a detailed protocol

Statistical analysis made use of Student’s t-test across N samples, where N is the number of embryos or experiments as defined in the figure legends.

References

  1. 1
  2. 2
  3. 3
  4. 4
  5. 5
  6. 6
  7. 7
  8. 8
  9. 9
  10. 10
  11. 11
  12. 12
  13. 13
  14. 14
  15. 15
  16. 16
  17. 17
  18. 18
  19. 19
  20. 20
  21. 21
  22. 22
  23. 23
  24. 24
  25. 25
  26. 26
  27. 27
  28. 28
  29. 29
  30. 30
  31. 31
  32. 32
  33. 33
  34. 34
  35. 35
  36. 36
  37. 37
  38. 38
  39. 39
  40. 40
  41. 41
  42. 42
  43. 43
  44. 44
  45. 45
  46. 46
  47. 47
  48. 48
  49. 49
  50. 50
    The fibroblast growth factor signaling pathway
    1. DM Ornitz
    2. N Itoh
    (2015)
    Wiley Interdisciplinary Reviews: Developmental Biology 4:215–266.
    https://doi.org/10.1002/wdev.176
  51. 51
  52. 52
  53. 53
  54. 54
  55. 55
  56. 56
  57. 57
  58. 58
  59. 59
  60. 60
  61. 61
  62. 62
  63. 63
  64. 64
  65. 65
  66. 66
  67. 67
  68. 68
  69. 69
  70. 70
  71. 71
  72. 72
    A truncated form of fibroblast growth factor receptor 1 inhibits signal transduction by multiple types of fibroblast growth factor receptor
    1. H Ueno
    2. M Gunn
    3. K Dell
    4. A Tseng
    5. L Williams
    (1992)
    The Journal of Biological Chemistry 267:1470–1476.
  73. 73
  74. 74
  75. 75
  76. 76
  77. 77
  78. 78
  79. 79
  80. 80
  81. 81
  82. 82
  83. 83
  84. 84

Decision letter

  1. Marianne E Bronner
    Senior and Reviewing Editor; California Institute of Technology, United States
  2. Simon Hippenmeyer
    Reviewer; Austria Institute of Science and Technology, Austria

In the interests of transparency, eLife includes the editorial decision letter and accompanying author responses. A lightly edited version of the letter sent to the authors after peer review is shown, indicating the most substantive concerns; minor comments are not usually included.

[Editors’ note: a previous version of this study was rejected after peer review, but the authors submitted for reconsideration. The first decision letter after peer review is shown below.]

Thank you for submitting your work entitled "FGFR ubiquitination and degradation controls neuronal migration in vivo" for consideration by eLife. Your article has been reviewed by two peer reviewers, and the evaluation has been overseen by Ivan Dikic as the Reviewing and Senior Editor. The following individual involved in review of your submission have agreed to reveal their identity: Mary E Hatten (Reviewer #2).

Our decision has been reached after consultation between the reviewers. Based on these discussions and the individual reviews below, we regret to inform you that your work will not be considered further for publication in eLife.

Summary:

The manuscript studies the role of FGF receptors in the migration of cortical projection neurons. The authors show that interaction of FGFR with N-Cadherin, CDH2, seems to protect FGFR from ubiquitination and subsequent degradation by lysosome. This leads to sustained FGFR levels and increased Erk1/2 downstream signaling which the authors propose regulates neuronal migration. Reelin seems important for FGFR-mediated Erk1/2 signaling, presumably by stabilizing and/or recruiting N-Cadherin to the plasma membrane.

While a number of interesting observations are presented, none directly assays FGFR function in cortical neuron migration. Rather, the data appear to show that disruption of FGFR and/or FGFR signaling stalls neuronal development at the multipolar stage, thereby preventing further differentiation and/or migration to form the neuronal layers. The experimental design and performance were nice, but the experiments are carried out mostly in HEK293 cells, not on developing cortical neurons. In addition, from the currently presented data it is difficult to see a coherent model for how the FGFRs actually function in locomotion (or a polarity defect at the multipolar stage). In total, the authors failed to demonstrate a direct function for FGFRs in migration in neurons.

The essential revisions to improve this work include a number of issues indicated in the reviews. In particular we feel that providing live-imaging data to resolve whether and to which extent FGFR signaling is required for migration versus polarization/differentiation is needed. In addition the authors need to repeat the biochemistry/pathways experiments in primary cortical cells and also draw a more coherent model of their functional relevance in vivo. Taken together, substantial amount of additional data is required in order to reach the high standard of eLife. If the authors wish to submit the work again to eLife following their revisions fulfilling the raised points at a later time the manuscript will be considered as a new submission.

Reviewer #1:

Kon and colleagues address important questions in cortical development and convincingly demonstrate an essential function for FGFR signaling in radial projection neuron migration. By using in utero electroporation (for gene knockdown) and structure function analysis, the authors elucidate mechanistic details. They show that interaction of FGFR with N-Cadherin, CDH2, seems to protect FGFR from ubiquitination and subsequent degradation by lysosome. This leads to sustained FGFR levels and increased Erk1/2 downstream signaling which regulates neuronal migration. Interestingly, Reelin is an important upstream regulatory component and seems important for FGFR-mediated Erk1/2 signaling, presumably by stabilizing and/or recruiting N-Cadherin to the plasma membrane.

Overall the manuscript by Kon et al. investigates timely aspects in cortical development. The manuscript is well written and the findings illustrated nicely. Some of the claims could be further substantiated by addressing the following points:

1) The authors propose that FGFR signaling is important for polarization and/or orientation of neurons in the MMZ. However the actual data supporting this claim directly is a bit slim. In Figure 1B the authors stain and assay for the location of Golgi but it is very difficult to appreciate the defect in orientation in relation to the cortical plate. Perhaps more important would be to document and quantify the disoriented direction of multipolar migration by life-imaging. This piece of data will be critical since it will provide mechanistic insight at the cellular level and further contribute to the conceptual advance of the study.

2) Figure 5B could be improved by showing the real microscopic images that relate to the quantification chart.

3) The Discussion ends a bit abruptly and a conclusive paragraph is needed. The authors should also discuss their findings in a somewhat broader context. They should elaborate how their findings extend and complement the current model of FGFR function in neuronal migration in vivo, which was recently published by the Cancedda lab.

Reviewer #2:

The manuscript by Cooper and colleagues "FGFR ubiquitination and degradation controls neuronal migration in vivo" presents a series of experiments on cortical development the stated purpose of which is to investigate the role of FGF receptors in the migration of cortical projection neurons. While a number of interesting observations are presented, none directly assays FGFR function in cortical neuron migration and the various pathways investigated do not inform a coherent model of how FGFRs would directly function in neuronal locomotion. Rather, the data appear to show that disruption of FGFR function stalls neuronal development at the multipolar stage, thereby preventing further differentiation and/or migration to form the neuronal layers. Also, while the pathways that impact FGFR dependent ERK1/2 phosphorylation and degradation are interesting vis a vis downstream signaling pathways, and the experiments are nicely done, the experiments are carried out in HEK293 cells, not on developing cortical neurons. The paper fails to demonstrate a direct function for FGFRs in migration.

Specific issues include the following:

1) Expression of a dominant negative form of the receptor is not as clean as a conditional genetic experiment in cortical projection neurons. As is true with many electroporation approaches, the identity of the neuronal populations that are affected are not specified.

2) There is no direct evidence that the FGFR family members act redundantly during the transition from the multipolar to bipolar stage and migration into the cortical plate. Can the dominant negative FGFR dimerize with receptors other than FGFR family members or N-Cad? Such interactions could contribute to the phenotypes observed.

3) Does NCad-dEC4 bind to receptors other than FGFRs?

4) Assays to identify binding partners, e.g. IP/mass spec experiments using lysates from primary cortical neurons expressing the dominant negative FGFR and NCad-dEC4 proteins, would be more informative.

5) While the HEK293T cells are a great starting point for the biochemistry, it is critical to perform the protein interaction and expression assays using purified cortical neurons or at least cultured primary cortical neurons. The fact that these changes, or lack of changes in some cases, occur in HEK293T cells does not guarantee that this is the case in post-mitotic cortical neurons.

6) There appears to be a disconnect between the Rap1/CDH2/FGFR and Reelin/FGFR/ERK1 and 2 signaling pathways in this manuscript that are bridged by the following sentence and the authors' previous paper "The Rap1-dependent upregulation of NCad is triggered by Reelin, an extracellular ligand present in the MMZ (Jossin and Cooper, 2011)". This needs to be more clearly stated.

7) Why is there no control image for Figure 1B to show the Golgi orientation of control cells?

8) Figure 2A and 3D: Why is FGFR expression not visible in all lanes on the WB when the protein is overexpressed? If it is a matter of exposure time, it would be preferable to use longer exposures in order to see bands in all lanes that indicate FGFR expression. It would also be helpful to see loading controls like in Figure 3D.

9) Figure 3A and 3B: In contrast to Figure 2C, there is no increase in FGFR1-Myc when co-expressed with NCad-HA. If the DNA concentration is adjusted as noted in the legend, it would still be expected that the NCad-HA and NCad-W161A-HA lanes show similar levels of FGFR1-Myc, which they do not. FGFR1 appears even lower when co-expressed with NCad-HA. On the other hand, NCad-dEC4, which is proposed to not interact with FGFR1, seems to increase FGFR1 expression. I would suggest to show blots with consistent data, or remove those claims in the text.

10) The authors state that NCad-dEC4 does not interact with FGFR1 and is unable to rescue migration defects in Rap1GAP cells, which indicates that NCad interacts with FGFR1 through its EC4 domain. However, the data in Figure 5 shows that NCad increases FGFR1 expression and promotes neuronal migration through its EC1-2 domains. The same figure also shows that the EC4 domain-containing chimeric E/NCad does not increase FGFR or promote migration. This is confusing. If NCad does interact with FGFR1 via EC4 and promotes migration via EC1-2, it would suggest that homophilic binding of NCad or heterophilic binding to other proteins, in addition to heterophilic NCad:FGFR binding, plays a role in migration. Nevertheless, the authors present data indicating that homophilic NCad interactions do not have a role migration. These data need clarification.

11) Were the ERK1/2 constructs expressed in cortical neurons selectively expressed in post-mitotic cells?

12) The Discussion ends rather abruptly.

[Editors’ note: what now follows is the decision letter after the authors submitted for further consideration.]

Thank you for sending your article entitled "FGFR ubiquitination and degradation control neuronal migration in vivo" for peer review at eLife. Your article is being evaluated by two peer reviewers, and the evaluation is being overseen by a Reviewing Editor and Marianne Bronner as the Senior Editor.

The reviewers are somewhat split regarding the revision of this manuscript. In particular, reviewer 2 raises important issues regarding the specificity of the FGFR loss-of-function experiments and well as conclusions drawn from the imaging. In particular, it would be important to include a more specific conditional knock-down and better imaging including live imaging.

Given the list of essential revisions, including new experiments, the editors and reviewers invite you to respond within the next two weeks with an action plan and timetable for the completion of the additional work. We plan to share your responses with the reviewers and then issue a binding recommendation. The full reviews are listed below.

Reviewer #1:

Kon and colleagues address important questions in cortical development and convincingly demonstrate an essential function for FGFR signaling in radial projection neuron migration. More specifically the authors show that FGFR-mediated signaling regulates multipolar migration and the switch to bipolar morphology. By using in utero electroporation (for gene knockdown) and structure function analysis, the authors elucidated mechanistic details. They show that interaction of FGFR with N-Cadherin, CDH2, seems to protect FGFR from ubiquitination and subsequent degradation by lysosome. This leads to sustained FGFR levels and increased Erk1/2 downstream signaling. Interestingly, Reelin is an important upstream regulatory component and seems important for FGFR-mediated Erk1/2 signaling, presumably by stabilizing and/or recruiting N-Cadherin to the plasma membrane. The study by Kon et al. investigates timely aspects in cortical development.

This version of the manuscript improved, is very well written and the findings illustrated nicely. The authors added new and important data which greatly support and strengthen the main conclusions of the manuscript. The authors made an effort to clarify the writing in general and now explicitly state that FGFR signaling controls multipolar migration and subsequent switch to bipolar morphology. Given the convincing data and additional experimental evidence the current data set supports the main conclusions of the manuscript very well, even without direct live-imaging data. Overall this study is likely to be of great interest to the broader neuroscience community.

Reviewer #2:

Kon et al. revised their manuscript ʺFGFR ubiquitination and degradation controls neuronal migration in vivoʺ to reflect their observation that cortical neurons appear to stall at the multipolar stage when FGFR signaling is perturbed by transgenic expression of DN FGFR. They have done a nice job of streamlining the biochemical and in utero electroporation epistasis experiments, and performing more of the biochemical experiments in cultured cortical neurons as requested, to work out the signaling mechanism involved. While the manuscript overall has improved, there are several issues that I feel do not make it suitable for publication in eLife at this time and may be more suitable for a more specialized journal.

First, I still worry that expression of a DN FGFR, as opposed to conditional genetic loss of function experiments, may have some off-target effects. In addition, the finding by Szczurkowska et al., 2018, that FGFR2 regulates neuronal migration and spine density suggests that the FGFRs may not act redundantly.

Second, I do not agree that the terminology "multipolar migration" is justified without more detailed analysis of the cellular mechanism involved. Better high magnification/resolution images of cell morphology at this stage, combined with live imaging, would be required to determine why and how these cells are stalling.

https://doi.org/10.7554/eLife.47673.029

Author response

[Editors’ note: the author responses to the first round of peer review follow.]

Summary:

The manuscript studies the role of FGF receptors in the migration of cortical projection neurons. The authors show that interaction of FGFR with N-Cadherin, CDH2, seems to protect FGFR from ubiquitination and subsequent degradation by lysosome. This leads to sustained FGFR levels and increased Erk1/2 downstream signaling which the authors propose regulates neuronal migration. Reelin seems important for FGFR-mediated Erk1/2 signaling, presumably by stabilizing and/or recruiting N-Cadherin to the plasma membrane.

While a number of interesting observations are presented, none directly assays FGFR function in cortical neuron migration. Rather, the data appear to show that disruption of FGFR and/or FGFR signaling stalls neuronal development at the multipolar stage, thereby preventing further differentiation and/or migration to form the neuronal layers. The experimental design and performance were nice, but the experiments are carried out mostly in HEK293 cells, not on developing cortical neurons. In addition, from the currently presented data it is difficult to see a coherent model for how the FGFRs actually function in locomotion (or a polarity defect at the multipolar stage). In total, the authors failed to demonstrate a direct function for FGFRs in migration in neurons.

The essential revisions to improve this work include a number of issues indicated in the reviews. In particular we feel that providing live-imaging data to resolve whether and to which extent FGFR signaling is required for migration versus polarization/differentiation is needed. In addition the authors need to repeat the biochemistry/pathways experiments in primary cortical cells and also draw a more coherent model of their functional relevance in vivo. Taken together, substantial amount of additional data is required in order to reach the high standard of eLife. If the authors wish to submit the work again to eLife following their revisions fulfilling the raised points at a later time the manuscript will be considered as a new submission.

We thank the editor and reviewers for the summary. In response we have performed additional experiments and rewritten the text to improve logical flow. In summary:

1) We agree that disrupting FGFR signaling stalls or delays neurons at the multipolar stage. While our new data show that the phenotype is not due to a defect in differentiation, the evidence points to a failure before multipolar neurons become bipolar. This is the same phenotype as we described before when we inhibited Reelin receptors, Rap1 or N‐cadherin (Ncad) (Jossin and Cooper, 2011). In those experiments, we tracked neurons in slice cultures and found that migration speed was normal but the migration paths were more erratic, with more neurons moving down or sideways compared with controls. While we now lack the facilities to image neurons in slice culture, we show that blocking FGFR signaling has a similar effect as blocking Rap1 or NCad on neuron morphology and orientation near the top of the multipolar zone. Moreover, FGFR expression rescues multipolar neurons with inactive Rap1. Since Rap1 and NCad‐dependent FGFR activation by Reelin stimulates Erk, and since inhibiting Erk also delays neurons at the multipolar stage, we suspect that FGFRs may be needed for Erk‐dependent signals activated by Reelin in multipolar neurons. These signals might include expression of genes required for polarization, orientation or migration, but it would likely require single cell RNA‐Seq to identify a distinct differentiation state. To avoid misleading the reader as to the precise nature of the migration defect, we have now replaced “defect in neuronal migration” with “defect in multipolar migration”.

2) The original paper presented in vivo results, from in utero electroporation, showing:

a) Dominant‐negative FGFRs disturb the migration of neurons with multipolar morphology and impair their orientation towards the cortical plate.

b) Wildtype FGFR expression rescues migration of neurons with inhibited Rap1.

c) Ncad‐mediated cell‐cell adhesion is not needed for neuron migration in vivo.

d) NCad that cannot bind FGFRs does not rescue neurons with inhibited Rap1.

e) Over‐expressing ubiquitin K27R inhibits FGFR downregulation and rescues Rap1inhibited neurons.

f) While both ECad and NCad bind FGFRs, ECad does not prevent FGFR degradation and does not rescue multipolar migration in vivo unless it carries the NCad EC1‐2 region.

g) Dominant‐negative but not wildtype Erk1 or Erk2 inhibits multipolar neuron migration in vivo.

These in vivo experiments were complemented by experiments with cultured embryonic cortical neurons showing that long‐term stimulation with Reelin increases endogenous FGFR1 expression and activates Erk, dependent on FGFR kinase activity, and that endogenous FGFRs are degraded by the lysosome pathway. Transfected HEK293 cells were used for biochemical experiments to validate the NCad mutants used, to confirm that NCad binds to and inhibits ubiquitination and degradation of FGFRs, and to identify the ubiquitin linkage in polyubiquitinated FGFR. In restructuring the paper, we have moved many of the HEK293 experiments to supplementary figures, to focus attention on the in vivo and neuron culture experiments. We also provide the following new data:

a) Neuron proliferation, apoptosis, and differentiation state, as determined with

Sox2, Tbr2 and Satb2, are unaffected when FGFRs are inhibited in vivo (new Figure 2A).

b) Inhibition of FGFRs in vivo affects the morphological switch from multipolar to bipolar (new Figure 2C).

c) Transfected FGFR levels decrease when Rap1 is inhibited in utero (new Figure 3B).

d) Endogenous FGFR levels increase when NCad is expressed in primary neurons (new Figure 3C).

e) Increased FGFR protein levels and activation of Erk in Reelin‐stimulated neurons requires Rap1 and NCad (new Figure 8D).

Other figures have been updated with improved data. We hope that the reviewers will understand that repeating all the biochemical experiments in neurons would be very time consuming and will excuse the remaining HEK293 experiments.

Reviewer #1:

[…] Overall the manuscript by Kon et al. investigates timely aspects in cortical development. The manuscript is well written and the findings illustrated nicely. Some of the claims could be further substantiated by addressing the following points:

1) The authors propose that FGFR signaling is important for polarization and/or orientation of neurons in the MMZ. However the actual data supporting this claim directly is a bit slim. In Figure 1B the authors stain and assay for the location of Golgi but it is very difficult to appreciate the defect in orientation in relation to the cortical plate. Perhaps more important would be to document and quantify the disoriented direction of multipolar migration by life-imaging. This piece of data will be critical since it will provide mechanistic insight at the cellular level and further contribute to the conceptual advance of the study.

We agree that live imaging would be a great addition to this study. Unfortunately, we no longer have the facilities for live imaging. Nevertheless, we improved the quality of the images provided in the new Figure 2B so the reader could better appreciate the phenotype. We went further in the description of the phenotype by showing a delay in the morphological transition from multipolar to bipolar cell (new Figure 2C). While the phenotype is very similar to that caused by inhibiting Rap1 or NCad (Jossin and Cooper, 2011),, we have bolstered the evidence that FGFR function is linked to the Reelin/Rap1/NCad pathway by adding new data: Reelin increases FGFR signaling in primary cortical neuron cultures, in an NCad‐dependent manner; Rap1 inhibition in vivo decreases FGFR protein level in multipolar neurons; NCad over‐expression in neurons increases FGFR protein level; FGFR inhibition does not alter cell fate, proliferation or survival in vivo.

2) Figure 5B could be improved by showing the real microscopic images that relate to the quantification chart.

We thank the reviewer for their comments. The requested pictures have been added (new Figure 6B).

3) The Discussion ends a bit abruptly and a conclusive paragraph is needed. The authors should also discuss their findings in a somewhat broader context. They should elaborate how their findings extend and complement the current model of FGFR function in neuronal migration in vivo, which was recently published by the Cancedda lab.

The interesting paper from Szczurkowska et al. shows that inhibiting FGFR expression impairs migration of late‐born neurons and decreases spine density after neurons mature. The first result is expected in light of our observation of a defect in multipolar migration. The second result also suggests parallels with Reelin signaling, because Reelin also increases spine density. Szczurkowska et al. have the added twist that another cell adhesion molecule, NEGR1 also inhibits FGFR degradation. This suggests that at least two pathways – Reelin/Rap1/NCad and NEGR1 – stabilize FGFRs and increase signaling. We have added a concluding paragraph commenting on this new paper.

Reviewer #2:

The manuscript by Cooper and colleagues "FGFR ubiquitination and degradation controls neuronal migration in vivo" presents a series of experiments on cortical development the stated purpose of which is to investigate the role of FGF receptors in the migration of cortical projection neurons. While a number of interesting observations are presented, none directly assays FGFR function in cortical neuron migration and the various pathways investigated do not inform a coherent model of how FGFRs would directly function in neuronal locomotion. Rather, the data appear to show that disruption of FGFR function stalls neuronal development at the multipolar stage, thereby preventing further differentiation and/or migration to form the neuronal layers.

We never stated that FGFRs were involved in neuronal locomotion but may have implied it by our use of the generic term “neuronal migration”. We use now the more precise term “multipolar migration”. Cortical projection neurons go through two types of migration: multipolar migration and bipolar migration (or locomotion). Our results show that FGFRs are important for the multipolar migration that occurs before locomotion. The results provided here suggest that the Reelin/Rap1/Ncad pathway regulates FGFRs to control multipolar migration and our previous studies on Rap1‐inhibited neurons revealed no alteration in locomotion. We are sorry that it was confusing and we have made changes in the text to clarify this issue. Please see general point 1 above.

We have now added immunostaining results to show that neuronal differentiation, proliferation and survival are not affected in vivo (Figure 2A). However, we cannot exclude that the neurons are in a subtly different stage of differentiation that may be revealed by detailed gene expression analysis.

Also, while the pathways that impact FGFR dependent ERK1/2 phosphorylation and degradation are interesting vis a vis downstream signaling pathways, and the experiments are nicely done, the experiments are carried out in HEK293 cells, not on developing cortical neurons. The paper fails to demonstrate a direct function for FGFRs in migration.

In the previous version of the paper, we made use of cultured embryonic cortical neurons to show that long‐term stimulation with Reelin increases endogenous FGFR1 expression and activates Erk, dependent on FGFR kinase activity, and that endogenous FGFRs undergo lysosomal, not proteasomal, degradation. In the re‐submission, we have added new data showing that NCad over‐expression in cortical neurons increases endogenous FGFR1 protein level (new Figure 3C), that activation of Erk in Reelin simulated cortical neurons requires Rap1 and NCad (new Figure 8D), and that FGFR levels decrease when Rap1 is inhibited in utero (Figure 3B). HEK293 cells, which share features with neurons (Shaw et al. 2002, PMID 11967234), were used to validate the NCad mutants used, to confirm that NCad binds to and inhibits ubiquitination and degradation of FGFRs, and to identify the K27/29 ubiquitin linkage in polyubiquitinated FGFR. Such an extensive series of transfection experiments would be challenging in primary neurons. We feel that the preponderance of evidence supports the model depicted in Figure 9B, but there are gaps in our knowledge that we discuss frankly in the text.

Specific issues include the following:

1) Expression of a dominant negative form of the receptor is not as clean as a conditional genetic experiment in cortical projection neurons.

It is true that the use of dominant‐negative forms has disadvantages but conditional knockout with post‐mitotic neuronal drivers such as Dcx‐Cre may not be effective if there is perdurance of FGFR mRNA or protein. In addition, FGFR1‐3 have overlapping activators and downstream effectors. By using the NeuroD promoter to inhibit the whole family of FGFRs we can rapidly inhibit FGFR signaling in migrating neurons without disrupting neurogenesis. There were no effects of electroporating the full-length FGFRs under similar conditions, suggesting that FGFR cytoplasmic domain functions are required for normal migration.

As is true with many electroporation approaches, the identity of the neuronal populations that are affected are not specified.

We agree with this limitation. The neuronal populations affected are those derived from apical progenitors in the neocortical VZ at the time of electroporation and their progeny. This includes postmitotic neurons and basal progenitors. However, the number of Tbr2 positive basal progenitors in the electroporated area was not affected by DN FGFRs (new Figure 2A). In addition, there was no change in the number of Satb2 positive neurons destined for the upper CP, or in Sox2 positive apical neural stem cells. Cell proliferation and cell survival were also unaffected (new Figure 2A).

2) There is no direct evidence that the FGFR family members act redundantly during the transition from the multipolar to bipolar stage and migration into the cortical plate.

FGFR1‐3 are all expressed in the developing neocortex. Single mutants have subtle defects and the triple mutant has no telencephalon, precluding investigation of neuron migration (Paek et al., 2009; Iwata and Hevner, 2009). However, our use of dominant negatives, which inhibit the family by forming heterodimers (Ueno et al., 1992), suggests that one or more family members is important for migration. Moreover, each family member has the ability to rescue Rap1‐inhibited neurons. They thus act interchangeably under these conditions. The Wikipedia definition of genetic redundancy is “situations where a given biochemical function is redundantly encoded by two or more genes”. However, since we have not been able to formally prove that each family member functions in multipolar migration when expressed at endogenous level in vivo, we now say the evidence “suggests” that the FGFR family members act redundantly.

Can the dominant negative FGFR dimerize with receptors other than FGFR family members or N-Cad? Such interactions could contribute to the phenotypes observed.

The dominant‐negative mutants that were used have an intact extracellular domain but truncated cytoplasmic domain. They have the potential to interfere with endogenous FGFRs by competing for extracellular partners, including FGFs, NCad, NCAM, L1CAM and NEGR1, or by inhibiting endogenous FGFR kinase activation by forming mixed heterodimers. However, ability of both NCad and FGFR to rescue migration of Rap1 inhibited neurons, coupled with the lack of rescue by NCadΔEC4 and the finding that dominant‐negative NCad inhibits Reelin‐induced FGFR activation in neurons makes us favor NCad as the relevant partner.

3) Does NCad-dEC4 bind to receptors other than FGFRs?

We are not aware of any studies. However, we found that NCadΔEC4 still formed homophilic dimers with NCad, so it is not grossly misfolded.

4) Assays to identify binding partners, e.g. IP/mass spec experiments using lysates from primary cortical neurons expressing the dominant negative FGFR and NCad-dEC4 proteins, would be more informative.

In a different set of experiments, we did undertake large scale mass spec screens for neuronal proteins that bind NCadW161A (the mutant that does not form trans homodimers) but not ECad. We used both BioID (with extracellular BirA) and affinity purification mass spec. Neither screen detected additional cell surface receptors.

5) While the HEK293T cells are a great starting point for the biochemistry, it is critical to perform the protein interaction and expression assays using purified cortical neurons or at least cultured primary cortical neurons. The fact that these changes, or lack of changes in some cases, occur in HEK293T cells does not guarantee that this is the case in post-mitotic cortical neurons.

The interaction, ubiquitination and protein degradation experiments were indeed performed in HEK293T cells because this kind of approach is very difficult in cultured neurons. HEK293T cells share features with neurons and are easy to transfect (Shaw et al., 2002). Nevertheless, we have added new data to the paper showing that endogenous FGFR1 protein level increases when NCad is overexpressed in cultured cortical neurons (new Figure 3C), that inhibiting Rap1 in neurons in utero decreases FGFR protein level (new Figure 3B) and that the effect of Reelin on FGFRs in cortical neurons depends on NCad and Rap1 (new Figure 8D).

6) There appears to be a disconnect between the Rap1/CDH2/FGFR and Reelin/FGFR/ERK1 and 2 signaling pathways in this manuscript that are bridged by the following sentence and the authors' previous paper "The Rap1-dependent upregulation of NCad is triggered by Reelin, an extracellular ligand present in the MMZ (Jossin and Cooper, 2011)". This needs to be more clearly stated.

We agree that the link was unclear. We have performed a new experiment which demonstrates that NCad is important for the effect of Reelin on FGFR protein level and Erk phosphorylation (Figure 8D).

7) Why is there no control image for Figure 1B to show the Golgi orientation of control cells?

A control image has been added.

8) Figure 2A and 3D: Why is FGFR expression not visible in all lanes on the WB when the protein is overexpressed? If it is a matter of exposure time, it would be preferable to use longer exposures in order to see bands in all lanes that indicate FGFR expression. It would also be helpful to see loading controls like in Figure 3D.

The requested long exposure for FGFR has been added along with the loading control in new Figure 3—figure supplement 1. Long exposure for FGFR has been added to the new Figure 8A and B.

9) Figure 3A and 3B: In contrast to Figure 2C, there is no increase in FGFR1-Myc when co-expressed with NCad-HA. If the DNA concentration is adjusted as noted in the legend, it would still be expected that the NCad-HA and NCad-W161A-HA lanes show similar levels of FGFR1-Myc, which they do not. FGFR1 appears even lower when co-expressed with NCad-HA. On the other hand, NCad-dEC4, which is proposed to not interact with FGFR1, seems to increase FGFR1 expression. I would suggest to show blots with consistent data, or remove those claims in the text.

We varied the amount of FGFR DNA to compensate for FGFR stabilization by different cadherins and make sure we had enough protein to test for co‐immunoprecipitation. However, as the reviewer noted, we over‐compensated in the experiments shown. We have repeated the experiments and now show data where FGFR is equally expressed (Figure 5A, Figure 6—figure supplement 1).

10) The authors state that NCad-dEC4 does not interact with FGFR1 and is unable to rescue migration defects in Rap1GAP cells, which indicates that NCad interacts with FGFR1 through its EC4 domain. However, the data in Figure 5 shows that NCad increases FGFR1 expression and promotes neuronal migration through its EC1-2 domains. The same figure also shows that the EC4 domain-containing chimeric E/NCad does not increase FGFR or promote migration. This is confusing. If NCad does interact with FGFR1 via EC4 and promotes migration via EC1-2, it would suggest that homophilic binding of NCad or heterophilic binding to other proteins, in addition to heterophilic NCad:FGFR binding, plays a role in migration. Nevertheless, the authors present data indicating that homophilic NCad interactions do not have a role migration. These data need clarification.

This is exactly what we are claiming in this manuscript. Binding of NCad EC4 to FGFR is necessary but not sufficient to inhibit FGFR degradation. Indeed, ECad also interacts with FGFR through its EC4 but is unable to prevent FGFR degradation or promote neuronal migration in vivo. An NCad molecule containing EC4‐5 from ECad still protects FGFR from degradation and promotes neuron migration in vivo. The reciprocal switch, putting EC4 of NCad into ECad, does not give to ECad the ability to prevent FGFR degradation or to promote neuronal migration in vivo. The protection from degradation and promotion of migration map to EC1‐2 even though it needs the NCad‐FGFR interaction through EC4. This EC1‐2 function is not NCad‐NCad homophilic binding. We suggest that another protein interacting with NCad but not ECad EC1‐2 might be involved. Discovering this third partner is however beyond the scope of this manuscript.

11) Were the ERK1/2 constructs expressed in cortical neurons selectively expressed in post-mitotic cells?

Yes, we used the NeuroD promoter, as stated in the figure legend (Figure 9A).

12) The Discussion ends rather abruptly.

We thank the reviewer for the comment, and have added a concluding paragraph.

[Editors' note: the author responses to the re-review follow.]

Kon et al. revised their manuscript ʺFGFR ubiquitination and degradation controls neuronal migration in vivoʺ to reflect their observation that cortical neurons appear to stall at the multipolar stage when FGFR signaling is perturbed by transgenic expression of DN FGFR. They have done a nice job of streamlining the biochemical and in utero electroporation epistasis experiments, and performing more of the biochemical experiments in cultured cortical neurons as requested, to work out the signaling mechanism involved. While the manuscript overall has improved, there are several issues that I feel do not make it suitable for publication in eLife at this time and may be more suitable for a more specialized journal.

First, I still worry that expression of a DN FGFR, as opposed to conditional genetic loss of function experiments, may have some off-target effects. In addition, the finding by Szczurkowska et al., 2018, that FGFR2 regulates neuronal migration and spine density suggests that the FGFRs may not act redundantly.

We addressed the issue of specificity of FGFRs downregulation and overlapping functions of FGFRs in multipolar migration. We specifically knock‐down FGFR1, 2 and 3. We tested at least 6 shRNA for each receptor and selected the most efficient. The efficiency of the shRNAs was tested in cell culture and in vivo as explained in Figure 1—figure supplement 1.

We found that the in vivo knock‐down of FGFR1 and 2 induces an arrest of cells at the MMZ, with a more pronounced phenotype when the two receptors are downregulated together (Figure 1B). The knock‐down of FGFR3 resulted in a small, statistically non‐significant effect on cell positioning (Figure 1B). These results suggest that FGFRs work redundantly with a prominent role for FGFR1 and FGFR2.

Reviewer #2 was also concerned that FGFR(DN) could affect NCad. We found that in cell culture, FGFR(DN) does not change NCad protein expression level, does not reduce NCad homophilic interaction and does not prevent NCad accumulation at cell‐cell junctions (Figure 3—figure supplement 1B).

Second, I do not agree that the terminology "multipolar migration" is justified without more detailed analysis of the cellular mechanism involved. Better high magnification/resolution images of cell morphology at this stage, combined with live imaging, would be required to determine why and how these cells are stalling.

To gain insight into the mechanism underlying the migration defect, we analyzed the morphology of migrating neurons. Analysis of the morphology revealed no difference in the number of neurites or in the cell body length‐to‐width ratio of FGFR‐inhibited multipolar neurons compared to control multipolar neurons

(Figure 2B-D). In addition, the few FGFR‐inhibited bipolar neurons migrating in the RMZ exhibited no difference in the length of the leading process and the cell body length‐to‐width ratio compared to control cells and possess an axon at the rear (Figure 2G=I).

https://doi.org/10.7554/eLife.47673.030

Article and author information

Author details

  1. Elif Kon

    Laboratory of Mammalian Development & Cell Biology, Institute of Neuroscience, Université Catholique de Louvain, Brussels, Belgium
    Contribution
    Data curation, Formal analysis, Validation, Investigation, Visualization, Performed most of the biochemistry and histology experiments, Assisted with preparing the manuscript
    Competing interests
    No competing interests declared
  2. Elisa Calvo-Jiménez

    Laboratory of Mammalian Development & Cell Biology, Institute of Neuroscience, Université Catholique de Louvain, Brussels, Belgium
    Contribution
    Data curation, Formal analysis, Validation, Investigation, Visualization, Assisted with the biochemistry and histology experiments, Performed the primary cortical neuron nucleofection experiments, Assisted with preparing the manuscript
    Competing interests
    No competing interests declared
  3. Alexia Cossard

    Laboratory of Mammalian Development & Cell Biology, Institute of Neuroscience, Université Catholique de Louvain, Brussels, Belgium
    Contribution
    Investigation, Visualization, Assisted with the biochemistry and histology experiments, Assisted with preparing the manuscript
    Competing interests
    No competing interests declared
  4. Youn Na

    Division of Basic Sciences, Fred Hutchinson Cancer Research Center, Seattle, United States
    Contribution
    Investigation, Assisted with the biochemistry and histology experiments, Assisted with preparing the manuscript
    Competing interests
    No competing interests declared
  5. Jonathan A Cooper

    Division of Basic Sciences, Fred Hutchinson Cancer Research Center, Seattle, United States
    Contribution
    Writing—review and editing, Reviewed and edited the final version of the manuscript, Assisted with preparing the manuscript
    Competing interests
    Senior editor, eLife
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-8626-7827
  6. Yves Jossin

    Laboratory of Mammalian Development & Cell Biology, Institute of Neuroscience, Université Catholique de Louvain, Brussels, Belgium
    Contribution
    Conceptualization, Resources, Data curation, Formal analysis, Supervision, Funding acquisition, Validation, Investigation, Visualization, Methodology, Writing—original draft, Project administration, Writing—review and editing, Conceived and supervised the project, Designed all experiments, Wrote the original draft of the manuscript, Reviewed and edited the final version of the manuscript, Performed the in vivo experiments
    For correspondence
    yves.jossin@uclouvain.be
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0001-8466-7432

Funding

Fonds De La Recherche Scientifique - FNRS (J.0129.15)

  • Yves Jossin

Fonds De La Recherche Scientifique - FNRS (J.0179.16)

  • Yves Jossin

Fonds De La Recherche Scientifique - FNRS (T.0243.18)

  • Yves Jossin

Fonds pour la Formation à la Recherche dans l’Industrie et dans l’Agriculture

  • Elif Kon
  • Elisa Calvo-Jimenez
  • Alexia Cossard

National Institutes of Health (R01-NS080194)

  • Jonathan A Cooper

National Institutes of Health (GM109463)

  • Jonathan A Cooper

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Acknowledgements

We thank Sasha Strait for excellent technical assistance. YJ is a Fonds National de la Recherche Scientifique (FNRS) investigator. EK, ECJ and AC are supported by Fonds pour la Formation à la Recherche dans l’Industrie et dans l’Agriculture (FRIA) fellowships. This work was supported by grants J.0129.15, J.0179.16 and T.0243.18 from the FNRS and R01-NS080194 and GM109463 from the National Institutes of Health.

Ethics

Animal experimentation: CD1 mice were bred in standard conditions and animal procedures were carried out in accordance with European guidelines and approved by the animal ethics committee of the Université Catholique de Louvain under the protocol number: 2017/UCL/MD/009.

Senior and Reviewing Editor

  1. Marianne E Bronner, California Institute of Technology, United States

Reviewer

  1. Simon Hippenmeyer, Austria Institute of Science and Technology, Austria

Publication history

  1. Received: April 13, 2019
  2. Accepted: October 1, 2019
  3. Accepted Manuscript published: October 2, 2019 (version 1)
  4. Version of Record published: October 10, 2019 (version 2)

Copyright

© 2019, Kon et al.

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.

Metrics

  • 699
    Page views
  • 143
    Downloads
  • 0
    Citations

Article citation count generated by polling the highest count across the following sources: Crossref, PubMed Central, Scopus.

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Download citations (links to download the citations from this article in formats compatible with various reference manager tools)

Open citations (links to open the citations from this article in various online reference manager services)