Calponin 2 Acts As an Effector of Noncanonical Wnt-Mediated Cell Polarization during Neural Crest Cell Migration

Cell Reports

Report Calponin 2 Acts As an Effector of Noncanonical Wnt-Mediated Cell Polarization during Neural Crest Cell Migration Ba¨rbel Ulmer,1,3 Cathrin Hagenlocher,1,3 Silke Schmalholz,1,3 Sabrina Kurz,1,3 Axel Schweickert,1 Ayelet Kohl,2 Lee Roth,2,4 Dalit Sela-Donenfeld,2,* and Martin Blum1,* 1Institute

of Zoology, University of Hohenheim, Garbenstrasse 30, 70593 Stuttgart, Germany School of Veterinary Medicine, Robert H. Smith Faculty of Agriculture, Food and Environment, The Hebrew University of Jerusalem, Rehovot 76100, Israel 3These authors contributed equally to this work 4Present address: The Weizmann Institute, Rehovot 76100, Israel *Correspondence: [email protected] (D.S.-D.), [email protected] (M.B.) http://dx.doi.org/10.1016/j.celrep.2013.02.015 2Koret

SUMMARY

Neural crest cells (NCCs) migrate throughout the embryo to differentiate into cell types of all germ layers. Initial directed NCC emigration relies on planar cell polarity (PCP), which through the activity of the small GTPases RhoA and Rac governs the actin-driven formation of polarized cell protrusions. We found that the actin binding protein calponin 2 (Cnn2) was expressed in protrusions at the leading edge of migratory NCCs in chicks and frogs. Cnn2 knockdown resulted in NCC migration defects in frogs and chicks and randomized outgrowth of cell protrusions in NCC explants. Morphant cells showed central stress fibers at the expense of the peripheral actin network. Cnn2 acted downstream of Wnt/PCP, as migration defects induced by dominant-negative Wnt11 or inhibition of RhoA function were rescued by Cnn2 knockdown. These results suggest that Cnn2 modulates actin dynamics during NCC migration as an effector of noncanonical Wnt/PCP signaling. INTRODUCTION Neural crest cells (NCCs) are found exclusively in vertebrates (Bronner and LeDouarin, 2012). Following induction during early neurulation at the boundary of the neural plate and epidermis, cells undergo epithelial-mesenchymal transition and delaminate from the forming neural tube. NCCs migrate throughout the embryo in a directed manner to differentiate into a variety of cell types ranging from neurons to pigment cells, and to contribute to diverse tissues and organs (le Douarin and Kalcheim, 1999; Krispin et al., 2010). Recently, a Wnt/planar cell polarity (PCP)-based mechanism for directed NCC locomotion prior to guidance-mediated migration was described in frog and zebrafish (De Calisto et al., 2005;

Matthews et al., 2008; Carmona-Fontaine et al., 2008b). During the initial steps of NCC migration, cells retract their protrusions and change direction following confrontation due to contact inhibition. Noncanonical Wnt signaling results in the local activation of the small GTPase RhoA at sites of cell contacts (Carmona-Fontaine et al., 2008a). Through its target Rho kinase (ROCK), RhoA antagonizes another small GTPase, Rac (Carmona-Fontaine et al., 2008b; Shoval and Kalcheim, 2012). Cell polarization thus depends on opposing intracellular gradients of RhoA and Rac, which are set up through Wnt/ PCP. In consequence, lamellipodia and filopodia form where cells have no contacts, i.e., at the edges of premigratory NCC territories, which direct cells away from each other and thus away from the closing neural tube. Recent analyses have shown that filopodia and lamellipodia are formed in a ROCK- and actin-dependent manner in delaminating NCCs (Berndt et al., 2008; Clay and Halloran, 2010). However, the link between the PCP effectors RhoA and Rac and the actin cytoskeleton has remained elusive. The calponin (Cnn) proteins represent an evolutionarily highly conserved family of actin-binding proteins (Wu and Jin, 2008). Three Cnn-like (clik) repeats mediate actin binding (Gimona and Winder, 1998; Gimona et al., 2003). In vertebrates, three isoforms have been described (Rozenblum and Gimona, 2008). Several studies have demonstrated the ability of Cnn to modulate the dynamics of the actin cytoskeleton (Winder and Walsh, 1996). Whereas Cnn1 and Cnn3 localize to and stabilize stress fibers in cultured fibroblasts, Cnn2 has been found in cell protrusions, i.e., bound to cortical actin networks (Danninger and Gimona, 2000). Currently, few in vivo data regarding Cnn’s function are available. Smooth muscle contraction was affected in the vas deferens and bladder of Cnn1 knockout mice (Matthew et al., 2000; Fujishige et al., 2002). Deletion of Cnn2 resulted in a reduction of macrophage spreading and an increase of phagocytotic activity (Huang et al., 2008). The available data are consistent with the proposed role of calponin in smooth muscle contraction and cell motility (Wu and Jin, 2008). Here we report on a function of Cnn2 in directed NCC migration in chick and frog embryos. Cell Reports 3, 615–621, March 28, 2013 ª2013 The Authors 615

RESULTS AND DISCUSSION Cnn2 Is Expressed in Migratory NCCs in Frog and Chick Cnn2 gene and protein structures are depicted in Figures S1A, S1B, and S1D. Cnn2 messenger RNA (mRNA) expression was analyzed by whole-mount in situ hybridization (WMISH). Antisense probes revealed mRNA localization in migrating NCCs (Figures 1A–1C and 1G–1I), which was confirmed by comparison with the NCC markers Xsnail, Xslug, and Xtwist in Xenopus (Figures 1D–1F) and costaining with HNK1 and Krox20 in chicks (Figures 1H, 1H0 , and 1I). In both species, Cnn2 expression was found at early stages of emigration (Figures 1A, 1G, and 1G0 ). When cells progressed ventrally, a marked attenuation of expression was observed in Xenopus, with a continued presence in subpopulations of NCC streams (Figure 1C). Variable expressions were observed at other sites (Figures 1A–1C). No localized expression was seen in mice, in agreement with a previous study (Magdaleno et al., 2006).

Figure 1. Cnn2 Is Expressed in Migratory NCCs in Frogs and Chicks (A–F) Frog. Cnn2 mRNA localization is shown in delaminating (A) and early migrating NCCs (B and C). Comparison with Xsnail (D), Xslug (E), and Xtwist (F) demonstrates that Cnn2 mRNA was found during early stages of emigration. Transcription was attenuated as cells progressed ventrally (cf. the strength of signals in proximal and distal NCCs in C and F). Note that Cnn2 remained to be expressed in a subpopulation of mandibular, hyoid, and branchial NCCs (open arrowheads). Additional sites of Cnn2 mRNA localization were found in the notochord (A0 –C0 ), heart (asterisk in C), somites, pronephros, and pronephric duct (arrowheads in C). (A0 –C0 ) Histological transverse vibratome sections (levels indicated in A–C). Frontal views in (A), (B), (D), and (E); lateral views (anterior to the left) in (C) and (F). ba, branchial; hy, hyoid; ma, mandibular; no, notochord. (G–I) Chick. Ccnn2 expression in migratory cranial and cervical NCCs (arrowheads) at HH10 (G), HH11 (H), and HH12 (I). Embryos are shown in dorsal view. Costaining with HNK1 antibody (H, red) reveals a partial overlap, demonstrating expression during early stages of emigration, as the HNK1 epitope only arises at a distance from the neural tube. (G0 and H0 ) Cranial and cervical transverse vibratome sections at levels indicated in (G) and (H). Krox20 (I, blue) costaining, which marks r3 and r5, demonstrates expression in migratory streams at r2, r4, and r6, with continued and strong expression in the hyoid stream (r4). hb, hindbrain; mb, midbrain; ov, optic vesicle; pm, paraxial mesoderm; r, rhombomere. See also Figure S1.

616 Cell Reports 3, 615–621, March 28, 2013 ª2013 The Authors

Cnn2 Is Required for NCC Migration in Frog and Chick To assess the role of Cnn2 in NCC migration, we performed morpholino oligonucleotide (MO)-mediated gene knockdowns. MOs were applied unilaterally, and contralateral sides served as internal controls (Figures 2A00 –2C00 ). Apoptosis and proliferation were not affected in frogs or chicks (Figures S2A–S2M). In Xenopus, a splice-site MO was applied (Figure S1A) to circumvent possible early phenotypes due to maternal expression (not shown). Xtwist staining was used to mark migratory NCCs. Injection of a control MO (coMO) did not affect migration (Figure 2A), whereas Xcnn2MO inhibited migration on the injected side (Figure 2B). The shift from a symmetric expression domain in coMO-injected tailbud embryos to dorsally retarded patches (Figures 2A0 and 2B0 ) was rescued by coinjection of synthetic Xcnn2 mRNA (Figures 2C and 2C0 ). At the concentrations used here, injection of Xcnn2 mRNA alone did not alter NCC migration (Figure 2D). Higher doses (>2.4 ng/embryo) caused embryonic lethality due to head and neural tube closure defects (not shown). Qualitative (percentage of migration defects) and quantitative (migration distance) evaluations of the results demonstrated highly significant effects of Xcnn2 knockdown on NCC migration (p < 0.001; Figures 2D and 2E). NCC specification was unaffected, as control (n = 167) and morphant embryos (n = 123) displayed equal Xsnail mRNA expression at stage 16 (Figures S2N–S2Q). A mere delay in NCC migration may not have lasting effects at later stages of development. When cranial cartilage formation was assessed at late-tadpole stages, a highly significant size reduction was observed in morphants (Figures S2R and S2S). Cartilage phenotypes were rescued by the parallel administration of Xcnn2 mRNA (Figure S2T) in a highly significant manner (p < 0.001; Figure S2U). In chicks, qualitatively identical results were obtained. A fluorescent MO that targeted the translational start site was used. Migrating NCCs were identified by HNK1 immunofluorescence (IF). Electroporation of coMO did not affect NCC migration (Figures 2F, 2G, 2H, and 2L). This was seen in both cranial (Figure 2F) and cervical (Figure 2G) NCCs and confirmed in histological sections (Figure 2H). In contrast, Ccnn2MO significantly

Figure 2. Cnn2 Is Required for NCC Migration in Frogs and Chicks (A–E) Frog. MO and mRNA were injected unilaterally into left animal blastomeres at the four- to eight-cell stage. Specimens were analyzed for expression of Xtwist by WMISH at the stages indicated. (A–C) NCC migration (A) was compromised upon Cnn2 knockdown (B) and rescued by coinjection of Cnn2 mRNA (C). Histological sections (A0 –C0 ) at the levels indicated (dashed lines) demonstrate the equal (A0 and C0 ) and unequal (B0 ) dorsoventral extension of Xtwistpositive NCCs. (D) Qualitative evaluation of results. Embryos were scored as WT or defective. (E) Quantitative evaluation of results. Distances of migration were measured on the injected and uninjected sides as indicated in (B). A box plot of injected/uninjected sides is shown. Note that knockdown and rescue were significant in both cases. (F–L) Chick. Fluorescent MOs (green) were unilaterally electroporated (asterisks) into the neural tube at the four- to eight-somite stage and analyzed for migratory NCCs by HNK1 staining (red) 16 hr thereafter. Migration of cranial (F) and cervical NCCs (G) was unaffected when coMO was injected, but was inhibited on the injected sides upon Ccnn2MO application (I and J). (H and K) Histological sections at cervical levels of specimens treated with coMO (H) or Ccnn2MO (K). (L) Quantitative evaluation of results. HNK1-positive cells were counted on the injected and uninjected sides of coMO- and Ccnn2MO-injected embryos (n = 6 each). Note that Cnn2 knockdown resulted in a reduction of HNK1-positives cells to 40%. Error bars represent SD. NT, neural tube; PM, paraxial mesoderm. See also Figure S2.

reduced HNK1 staining at sites of MO application (Figures 2I– 2L). Because the HNK1 epitope only becomes evident in NCCs once they emigrate from the neural tube (Del Barrio and Nieto, 2004), this result is similar to the migration defects seen in frog Xcnn2 morphants (cf. Figures 2A–2E). Taken together, these experiments unequivocally demonstrated that Cnn2 was required for NCC migration in chicks and frogs. Cnn2 Polarizes to the Leading Edge of Emigrating NCCs Expression in migrating NCCs suggested a role of Cnn2 in modulating the actin cytoskeleton. Cnn2 localization at the distal end of stress fibers and in lamellipodial protrusions of spreading NIH 3T3 cells was previously shown (Danninger and Gimona, 2000). Because Xenopus-specific Cnn2 antibodies are not available, and heterologous antibodies did not work in frogs (not shown), we cloned an N-terminal myc-fusion construct to analyze the subcellular localization of Cnn2 by IF. Synthetic mRNA was injected into animal blastomeres of Xenopus embryos at the eight-cell stage to target emigrating NCCs. We investigated NCC explants (Figure 3A) to monitor Cnn2-myc at different stages of emigration (Newgreen and Minichiello, 1995) and confirmed the NCC identity of the explants by Xtwist expression (not shown). Coinjection of lineage tracer demonstrated that explants were targeted

throughout (Figure 3B). Remarkably, Cnn2-myc was undetectable in the explant and restricted to cells at the boundary of the epithelioid cell layer (Figure 3C). Polarized Cnn2-myc signals were found at the leading edge of cells that were about to separate from the epithelioid layer to become fully mesenchymal (Figure 3C0 ). No signal was detected at the back end of such cells (Figure 3C0 ). Faint staining represented Cnn2myc localization at leading edges from cells that were out of focus (dashed triangles in Figure 3C0 ). Mesenchymal migrating cells showed strong and ubiquitous expression in the cytoplasm (Figure 3D). Occasionally, attenuation of signal was observed at contact sites with neighboring cells (Figures 3D0 and 3D00 ), which may result from contact-inhibition-mediated activation of RhoA (Carmona-Fontaine et al., 2008b). A reciprocal expression pattern was reported for the small GTPase RhoA in chicks (Groysman et al., 2008). We therefore analyzed RhoA in Xenopus explant cultures as well, and detected a qualitatively similar pattern. RhoA was strongly expressed at the cell membrane of premigratory NCCs (Figures 3E and 3E0 ). Upon delamination, RhoA disappeared from the leading edge, whereas the protein remained to be seen at the back end of cells that were about to delaminate (asterisks in Figure 3E). In mesenchymal migratory NCCs, RhoA was attenuated (Figure 3F). Cell Reports 3, 615–621, March 28, 2013 ª2013 The Authors 617

Figure 3. Cnn2 Polarizes to the Leading Edge of Emigrating NCCs (A–F) Polarized localization of Xcnn2-myc and RhoA in Xenopus NCC explant cultures. (A) Schematic diagram depicting the progression of cell emigration from NCC explants. A multilayered explant is outlined in dark gray, and the epithelioid sheath of monolayered premigratory NCCs and mesenchymal migrating NCCs is indicated in light gray. (B) Low-magnification overview of an explant culture that was injected with lineage tracer to demonstrate that injections targeted cells throughout. e, explant, m, migration area. (C and C0 ) Xcnn2-myc (red) localization to leading edge of delaminating NCC and throughout a mesenchymal cell (asterisk). Nuclei in blue (DAPI), actin in green (phalloidin staining). (D, D0 , and D00 ) Strong Xcnn2-myc staining in individual mesenchymal cells. Note that Xcnn2myc was reduced at contact sites (open arrowhead in D0 /D00 ). In (D00 ) Xcnn2-myc (red channel) is shown in white to clearly reveal the attenuation at the contact site. (E and F) Polarized expression of RhoA in the epithelioid part of explant culture (E) was lost in migrating NCCs (F). In (E0 ) RhoA (red channel) is shown in white to clearly reveal the loss of polarization at the leading edge of delaminating cells (asterisks). (G and H) Xcnn2MO rescued migration defects induced by dnWnt11. Embryos were injected unilaterally into animal blastomeres at the four- to eight-cell stage and analyzed for NCC migration by Xtwist WMISH at stage 26. (G) Migration defects. Embryos were scored as WT or defective. (H) Quantitative evaluation of results. Distances of migration were measured on the injected and uninjected sides as indicated in Figure 2B. Box plot of injected/uninjected sides. (I) Xcnn2MO suppressed Y27632-enhanced NCC migration in explants. Migration areas (m) were calculated and compared with explant size (e), as indicated in (B). Xcnn2 knockdown did not significantly alter NCC migration compared with WT explants (cf. Figure S3). ROCK inhibitor Y27632 induced excess migration, which was suppressed by parallel knockdown of Xcnn2. Scale bars in (C)–(F) represent 50 mm (C and D) and 10 mm (E, F, and C0 –E0 ).

Cnn2 Acts Downstream of Noncanonical Wnt Signaling The mutually exclusive protein localization pattern of Cnn2-myc and RhoA suggested that Cnn2 was a target of noncanonical Wnt signaling. A possible interaction of Xcnn2 and Wnt was assessed by individual and combined loss-of-function experiments in Xenopus. Wnt11 was chosen as the noncanonical ligand because of its known role in NCC migration (De Calisto et al., 2005). When a dominant-negative Wnt11 (dnWnt11) construct was injected, NCC migration was retarded, as shown by Xtwist expression (Figures 3G, 3H, and S3A–S3C; De Calisto 618 Cell Reports 3, 615–621, March 28, 2013 ª2013 The Authors

et al., 2005). Remarkably, a parallel knockdown of Xcnn2 significantly rescued this migration defect (Figures 3G, 3H, and S3D; p < 0.001). The combined loss-of-function also revealed a significant improvement of migration patterns compared with Xcnn2 knockdown (p < 0.001; Figures 3G and 3H). This experiment suggested that Xcnn2 acted downstream of Wnt11 signaling, as loss of Wnt11 was reverted by loss of Xcnn2. Next we investigated a possible role of ROCK in Wnt/Cnnmediated NCC migration. ROCK has been reported to phosphorylate chick Cnn1 at four sites in clik repeat 1 (Kaneko

et al., 2000; Abouzaglou et al., 2004). Phosphorylation of one of these sites (Ser175 in clik repeat 1; cf. Figures S1D and S1E) has been shown to result in a conformational change and, consequently, loss of actin binding (Abouzaglou et al., 2004). Based on the conservation of these sites in Cnn2, it has been argued that ROCK should phosphorylate Cnn2 as well (Figure S1E; Wu and Jin, 2008). The reciprocal RhoA protein localization (Groysman et al., 2008) together with the ROCK-induced Cnn inactivation (Kaneko et al., 2000) indicated that RhoA, through its target kinases, might result in Cnn2 inactivation and/or degradation. We studied the functional interaction of ROCK and Cnn2 during NCC migration in NCC explant cultures in Xenopus. In order to assess NCC migration in a semiquantitative manner, we measured the area covered by NCC cells following a 16– 20 hr culture period (m), the area of the original explant (e), and the migration ratio (m/e; cf. Figure 3B). Xcnn2 knockdown revealed a tendency toward a reduction of the migration ratio (p = 0.203) and thus much weaker effects compared with morphant embryos (cf. Figure 2B). Explants, however, differ from the in vivo situation because spatial and directional restrictions related to tissue architecture (e.g., arrangement of the neural tube, epidermis, and somites) as well as guidance cues are eliminated. Delamination and emigration in NCC explant cultures occurred in all directions in an unrestricted manner, which likely resulted in a much reduced impact of Cnn2 knockdown. In contrast, the ROCK inhibitor Y27632 enhanced migration significantly (Figures 3I and S3E; p < 0.001). This was in line with lossand gain-of-function experiments in the chick, which clearly implicated RhoA and ROCK as negative regulators of NCC delamination (Groysman et al., 2008; Shoval and Kalcheim, 2012). Remarkably, parallel loss of Xcnn2 through injection of Xcnn2MO reduced the Y27632-enhanced migration significantly (p = 0.03; Figures 3I and S3E). Thus, knockdown of Xcnn2 rescued the loss-of-function of both Wnt11 and ROCK, demonstrating that Cnn2 acted downstream of noncanonical Wnt signaling to promote NCC migration. The mutually exclusive nature of RhoA and Cnn2 expression may also explain the absence of Cnn2-myc from cells at the center of the explant and those surrounded by other cells within the epithelioid monolayer, despite the above-mentioned targeting of the entire explant (Figures 3B, 3C, and 3E). Interestingly, Cnn2-myc was found within the epithelioid cell layer following ROCK inhibition (Figure S3G). Therefore, RhoA activity through ROCK-mediated Cnn2 phosphorylation may result in Cnn2 degradation. In line with this notion, a Cnn2-myc fusion construct in which clik domains with potential ROCK phosphorylation sites were deleted (cf. Figure S1C) was detected throughout explants as well (Figure S3H). In summary, the expression patterns of Cnn2 mRNA and protein in migratory NCCs in frogs and chicks are compatible with a function in NCC migration and a negative regulation by RhoA. Xcnn2 Governs the Formation of the Polarized Cortical Actin Network in Migratory NCCs Next, we wondered whether Cnn2 is involved in the dynamic organization of the actin cytoskeleton in migratory NCCs. RhoA has been shown to promote the formation of central stress fibers

at the expense of the cortical actin network (Katoh et al., 2011). In chick NCC explants, inhibition of ROCK resulted in a reverse transformation, i.e., an extended cortical actin network was formed at the expense of central stress fibers (Groysman et al., 2008; Shoval and Kalcheim, 2012). When the actin cytoskeleton was analyzed in Xenopus NCCs of wild-type (WT) and Xcnn2 morphant explants, an increase in central stress fibers and a reduction of the cortical network was found (Figures 4A and 4B). Cell protrusions were analyzed in WT and morphant explants (Figures 4A0 and 4B0 ). Directionality was evaluated by determining the angles at which the protrusions were pointing. Vectors were grouped in segments of 45 each. In WT explants, the protrusions were polarized such that the majority pointed away from neighboring cells (Figure 4A0 ), i.e., in the direction of migration, as proposed by Carmona-Fontaine et al. (2008b). This directionality was severely affected in morphant specimens (Figure 4B0 ). Compilation of the vectors demonstrated a randomization of protrusion directionality in the morphant explants (Figure 4C). When the distal and proximal vectors were compared in individual cells, the distal decrease (Figure 4C, dark gray) and proximal increase (Figure 4C, light gray) were found to be highly significant (p < 0.01 and p < 0.001, respectively; Figure 4C). Because no differences were found in the number of protrusions (not shown), these data implicate Cnn2 as an actin-binding protein that acts specifically on the directionality of cell migration and not on cell migration in general. Thus, the much stronger effect on NCC migration observed in vivo may represent a secondary effect due to the failure of cells to emigrate in a directional manner. The following model emerges from our data (Figure 4D): In the epithelial cell layer of explants or in the dorsal neural tube, contact inhibition of locomotion is governed by noncanonical Wnt signaling. Active RhoA is localized to the plasma membrane and activates ROCK. ROCK in turn phosphorylates Cnn2, which results in inactivation and degradation of Cnn2. In mesenchymal cells migrating individually, noncanonical Wnt signaling is off and thus Cnn2 is not inactivated, which results in stabilization of the cortical actin network. Cells that are about to leave the epithelioid plane are polarized (i.e., with respect to noncanonical Wnt signaling), as are mesenchymal cells upon establishment of occasional contacts during migration. This polarization leads to a polar distribution of active Cnn2, which is a prerequisite for the formation of directed protrusions in explants and for directed migration of NCCs in vivo. The demonstrated interaction of Cnn2 and Wnt11/ROCK strongly suggests that Cnn2 directs NCC migration downstream of noncanonical Wnt signaling. In conclusion, Cnn2 establishes the missing link between PCP and the dynamics of the actin cytoskeleton during NCC migration. EXPERIMENTAL PROCEDURES NCC Explants NCCs were prepared from stage 16/17 neurula embryos and transferred to Danilchik’s media as previously described (Alfandari et al., 2003). Explants were transferred 20–60 min later to fibronectin-coated coverslips in six-well plates. ROCK inhibitor Y27632 (Selleck Novel Life Science Reagents) was added at 15 pM and cultures were incubated for a total of 16–20 hr. The dimensions of the explants and area of migrating neural crest were outlined and measured in ImageJ (Rasband, 2012).

Cell Reports 3, 615–621, March 28, 2013 ª2013 The Authors 619

Quantification of NCC Protrusions Single mesenchymal cells were aligned along their proximal-distal axis, distal side up. The center of the nucleus and the end of protrusions were marked in ImageJ. The angle of the vectors pointing from the nucleus to the end of the protrusions were calculated in R and merged. Wind rose diagrams depict the frequency distributions of vector angles. For box plots and statistical comparison, the percentages of proximal (light gray) or distal (dark gray) protrusions of each individual cell, representing the vectors projecting in an angle of 135 toward or away from the neighboring cell, respectively, were calculated. For further details regarding the materials and methods used in this work, see Extended Experimental Procedures. SUPPLEMENTAL INFORMATION Supplemental Information includes three figures and Extended Experimental Procedures and can be found with this article online at http://dx.doi.org/10. 1016/j.celrep.2013.02.015. LICENSING INFORMATION This is an open-access article distributed under the terms of the Creative Commons Attribution-NonCommercial-No Derivative Works License, which permits non-commercial use, distribution, and reproduction in any medium, provided the original author and source are credited. ACKNOWLEDGMENTS Thomas Thumberger developed the method for analysis of cell protrusions. Tim Ott helped with the preparation of figures. Nora Loges cloned the rescue construct and helped with the mouse ISH. Kerstin Feistel helped with the interpretation of results. B.U. was supported by a Minerva short-term research grant and was the recipient of a PhD fellowship from the Landesgraduiertenfo¨rderung Baden-Wu¨rttemberg. L.R. was supported by a fellowship from the Landtag Baden-Wu¨rttemberg for Israeli students. Collaborative work in the Blum and Sela-Donenfeld labs was funded by a grant from the Hohenheim-Rehovot partnership program. Received: July 24, 2012 Revised: January 17, 2013 Accepted: February 11, 2013 Published: March 14, 2013

Figure 4. Xcnn2 Governs the Formation of the Polarized Cortical Actin Network in Migratory NCCs

REFERENCES

(A–C) Xenopus NCC explants from WT (A) and Xcnn2 morphant (B) embryos were cultured and the actin cytoskeleton was visualized by phalloidin staining. Nuclei were counterstained with DAPI. Panels show cells derived from the epithelioid-mesenchymal transition zone of explants. Magnifications of single WT (A0 ) and morphant (B0 ) cells are shown. Broken arrows in (A0 ) and (B0 ) indicate cell protrusions. (C) Wind rose diagrams depict the distribution of protrusion angles into eight segments spanning 45 each. Note that in WT explants, cell protrusions (n = 541 from 28 cells derived from six embryos) predominantly point in the distal direction, whereas morphant vectors (n = 521 protrusions, 22 cells, 5 embryos) are randomized. Bottom: Quantification of vector analysis. Morphant cells display significantly fewer distal protrusions (dark gray) and more proximal protrusions (light gray), caused by a loss of cortical actin and a gain of stress fibers. (D) Model of RhoA and Cnn2 acting downstream of PCP to polarize the actin cytoskeleton and direct NCC migration, and a schematic depiction of different migration states in NCC explant. Left: Epithelioid premigratory NCCs are characterized by actin stress fibers. They express RhoA at the plasma membrane (red), leading to Cnn2 phosphorylation and degradation. Middle: Cells at the margin of the epithelioid cell layer, which are about to leave the lattice, are polarized and characterized by central stress fibers and a peripheral cortical actin network. RhoA and Cnn2 (green) are expressed in a mutually

Abouzaglou, J., Be´nistant, C., Gimona, M., Roustan, C., Kassab, R., and Fattoum, A. (2004). Tyrosine phosphorylation of calponins. Inhibition of the interaction with F-actin. Eur. J. Biochem. 271, 2615–2623.

620 Cell Reports 3, 615–621, March 28, 2013 ª2013 The Authors

Alfandari, D., Cousin, H., Gaultier, A., Hoffstrom, B.G., and DeSimone, D.W. (2003). Integrin alpha5beta1 supports the migration of Xenopus cranial neural crest on fibronectin. Dev. Biol. 260, 449–464. Berndt, J.D., Clay, M.R., Langenberg, T., and Halloran, M.C. (2008). Rhokinase and myosin II affect dynamic neural crest cell behaviors during epithelial to mesenchymal transition in vivo. Dev. Biol. 324, 236–244. Bronner, M.E., and LeDouarin, N.M. (2012). Development and evolution of the neural crest: an overview. Dev. Biol. 366, 2–9. Carmona-Fontaine, C., Matthews, H., and Mayor, R. (2008a). Directional cell migration in vivo: Wnt at the crest. Cell Adhes. Migr. 2, 240–242.

exclusive manner such that Cnn2 is found where the cortical actin is localized and RhoA is found in the part of the cells that harbors stress fibers. Right: Migratory mesenchymal NCCs lack stress fibers and possess peripheral actin with colocalized Cnn2 throughout. RhoA is present in migratory mesenchymal cells but absent from the cortical actin network.

Carmona-Fontaine, C., Matthews, H.K., Kuriyama, S., Moreno, M., Dunn, G.A., Parsons, M., Stern, C.D., and Mayor, R. (2008b). Contact inhibition of locomotion in vivo controls neural crest directional migration. Nature 456, 957–961.

Katoh, K., Kano, Y., and Noda, Y. (2011). Rho-associated kinase-dependent contraction of stress fibres and the organization of focal adhesions. J. R. Soc. Interface 8, 305–311.

Clay, M.R., and Halloran, M.C. (2010). Control of neural crest cell behavior and migration: Insights from live imaging. Cell Adhes. Migr. 4, 586–594.

Krispin, S., Nitzan, E., and Kalcheim, C. (2010). The dorsal neural tube: a dynamic setting for cell fate decisions. Dev. Neurobiol. 70, 796–812.

Danninger, C., and Gimona, M. (2000). Live dynamics of GFP-calponin: isoform-specific modulation of the actin cytoskeleton and autoregulation by C-terminal sequences. J. Cell Sci. 113, 3725–3736.

le Douarin, N., and Kalcheim, C. (1999). The Neural Crest (Cambridge: Cambridge University Press).

De Calisto, J., Araya, C., Marchant, L., Riaz, C.F., and Mayor, R. (2005). Essential role of non-canonical Wnt signalling in neural crest migration. Development 132, 2587–2597. Del Barrio, M.G., and Nieto, M.A. (2004). Relative expression of Slug, RhoB, and HNK-1 in the cranial neural crest of the early chicken embryo. Dev. Dyn. 229, 136–139. Fujishige, A., Takahashi, K., and Tsuchiya, T. (2002). Altered mechanical properties in smooth muscle of mice with a mutated calponin locus. Zoolog. Sci. 19, 167–174. Gimona, M., and Winder, S.J. (1998). Single calponin homology domains are not actin-binding domains. Curr. Biol. 8, R674–R675. Gimona, M., Kaverina, I., Resch, G.P., Vignal, E., and Burgstaller, G. (2003). Calponin repeats regulate actin filament stability and formation of podosomes in smooth muscle cells. Mol. Biol. Cell 14, 2482–2491. Groysman, M., Shoval, I., and Kalcheim, C. (2008). A negative modulatory role for rho and rho-associated kinase signaling in delamination of neural crest cells. Neural Dev. 3, 27. Huang, Q.-Q., Hossain, M.M., Wu, K., Parai, K., Pope, R.M., and Jin, J.-P. (2008). Role of H2-calponin in regulating macrophage motility and phagocytosis. J. Biol. Chem. 283, 25887–25899. Kaneko, T., Amano, M., Maeda, A., Goto, H., Takahashi, K., Ito, M., and Kaibuchi, K. (2000). Identification of calponin as a novel substrate of Rhokinase. Biochem. Biophys. Res. Commun. 273, 110–116.

Magdaleno, S., Jensen, P., Brumwell, C.L., Seal, A., Lehman, K., Asbury, A., Cheung, T., Cornelius, T., Batten, D.M., Eden, C., et al. (2006). BGEM: an in situ hybridization database of gene expression in the embryonic and adult mouse nervous system. PLoS Biol. 4, e86. Matthew, J.D., Khromov, A.S., McDuffie, M.J., Somlyo, A.V., Somlyo, A.P., Taniguchi, S., and Takahashi, K. (2000). Contractile properties and proteins of smooth muscles of a calponin knockout mouse. J. Physiol. 529, 811–824. Matthews, H.K., Marchant, L., Carmona-Fontaine, C., Kuriyama, S., Larraı´n, J., Holt, M.R., Parsons, M., and Mayor, R. (2008). Directional migration of neural crest cells in vivo is regulated by Syndecan-4/Rac1 and non-canonical Wnt signaling/RhoA. Development 135, 1771–1780. Newgreen, D.F., and Minichiello, J. (1995). Control of epitheliomesenchymal transformation. I. Events in the onset of neural crest cell migration are separable and inducible by protein kinase inhibitors. Dev. Biol. 170, 91–101. Rasband, W.S. (2012). ImageJ. U.S. National Institutes of Health, Bethesda, MD, USA. http://imagej.nih.gov/ij/. Rozenblum, G.T., and Gimona, M. (2008). Calponins: adaptable modular regulators of the actin cytoskeleton. Int. J. Biochem. Cell Biol. 40, 1990–1995. Shoval, I., and Kalcheim, C. (2012). Antagonistic activities of Rho and Rac GTPases underlie the transition from neural crest delamination to migration. Dev. Dyn. 241, 1155–1168. Winder, S.J., and Walsh, M.P. (1996). Calponin. Curr. Top. Cell. Regul. 34, 33–61. Wu, K.-C., and Jin, J.P. (2008). Calponin in non-muscle cells. Cell Biochem. Biophys. 52, 139–148.

Cell Reports 3, 615–621, March 28, 2013 ª2013 The Authors 621