Cdc42 regulates the cellular localization of Cdc42ep1 in controlling neural crest cell migration

Cdc42 regulates the cellular localization of Cdc42ep1 in controlling neural crest cell migration Abstract The member of Rho family of small GTPases Cdc42 plays important and conserved roles in cell polarity and motility. The Cdc42ep family proteins have been identified to bind to Cdc42, yet how they interact with Cdc42 to regulate cell migration remains to be elucidated. In this study, we focus on Cdc42ep1, which is expressed predominantly in the highly migratory neural crest cells in frog embryos. Through morpholino-mediated knockdown, we show that Cdc42ep1 is required for the migration of cranial neural crest cells. Loss of Cdc42ep1 leads to rounder cell shapes and the formation of membrane blebs, consistent with the observed disruption in actin organization and focal adhesion alignment. As a result, Cdc42ep1 is critical for neural crest cells to apply traction forces at the correct place to migrate efficiently. We further show that Cdc42ep1 is localized to two areas in neural crest cells: in membrane protrusions together with Cdc42 and in perinuclear patches where Cdc42 is absent. Cdc42 directly interacts with Cdc42ep1 (through the CRIB domain) and changes in Cdc42 level shift the distribution of Cdc42ep1 between these two subcellular locations, controlling the formation of membrane protrusions and directionality of migration as a consequence. These results suggest that Cdc42ep1 elaborates Cdc42 activity in neural crest cells to promote their efficient migration. actin cytoskeleton, neural crest, Rho GTPases, Cdc42 effector protein 1, cell migration Introduction The Rho family of small GTPases, RhoA, Rac1, and Cdc42, are important for cellular behaviors, including cell adhesion, cell shape changes, migration, and cytokinesis (Nobes and Hall, 1995). During embryonic development, Rho GTPases play critical roles in gastrulation, neural tube closure, and neural crest migration. The neural crest is a highly motile cell population specific to vertebrates. They migrate long distances in vertebrate embryos to give rise to craniofacial bones and cartilages as well as many other derivatives (Le Douarin, 1980; Dupin et al., 2006). Similar to other directional cell migration, neural crest cell migration requires cell polarization with protrusions extending at the cell front and stress fibers retracting the cell rear. At the molecular level, this polarity is mediated through the activation of small Rho GTPases. Activation of Rac1 and Cdc42 leads to the formation of lamellipodia at the leading edge, whereas activation of RhoA at the rear results in retraction of the cell (Matthews et al., 2008). At cell–cell contact, activation of members of the Wnt planar cell polarity pathway activates RhoA, and local RhoA activity antagonizes Rac1 and, therefore, inhibits the formation of cell protrusions (Carmona-Fontaine et al., 2008). To achieve the multiplexed activities of Rho GTPases, numerous downstream effector proteins exist and mediate a subset of their activities. A common motif found in the majority of Cdc42/Rac1 effector proteins is the conserved 16-aa Cdc42/Rac interactive-binding (CRIB) domain, which serves as a binding site for Cdc42 and/or Rac (Burbelo et al., 1995). Many proteins containing this motif have been identified and proved to be effector proteins for Cdc42/Rac1, such as p21 activated kinases (PAKs), mixed lineage kinases (MLKs), and Cdc42 effector proteins (Cdc42eps or CEPs). CEPs are a family of poorly understood Cdc42 effector proteins, also named binder of the Rho GTPases (Borgs) (Joberty et al., 1999). CEPs only exist in vertebrates, possibly mediating Cdc42 in vertebrate specific structures. There are only a handful of studies on CEPs during embryonic development. In mouse blastocysts, CEP1 (Borg5) enhances trophectoderm differentiation by promoting the sorting of trophectodermal cells to the outer layer (Vong et al., 2010). CEP1 also promotes microvascular angiogenesis by regulating directional migration of endothelial cells and is required in retinal and cardiac development in mouse (Liu et al., 2014). CEP2 (Borg1) plays a role in Xenopus gastrulation, by promoting the involution of mesoderm and the cell–cell adhesion of non-neural ectoderm (Nelson and Nelson, 2004). In cell culture, CEPs have been reported to involve in cell protrusions and actin cytoskeleton organization. CEP1 can induce membrane ruffling in Cos-7 cells, whereas induce long actin-based protrusions in NIH 3T3 fibroblasts (Burbelo et al., 1999). In primary keratinocytes, CEP2 and CEP5 (Borg3) expression reduced F-actin localization to adherence junctions with an increase in thin stress fibers (Hirsch et al., 2001). Despite physically binding to Cdc42, little is known of how they interact with Cdc42 in regulating cell motility. In this study, we characterize the activity of Cdc42ep1 (CEP1) in neural crest migration in Xenopus laevis. We show that in contrast to broadly expressed Cdc42, CEP1 is predominantly expressed in cranial neural crest (CNC) cells in Xenopus embryos. Consistent to its expression pattern, loss of CEP1 inhibited CNC cell migration. CEP1 knockdown not only affected cell morphology, but also interrupted the organization of functional structures for cell migration, such as the lamellipodia protrusions and focal adhesions, and as a result, affected cell traction on extracellular matrix. We further examined its interaction with Cdc42. CEP1 co-localizes with Cdc42 in membrane protrusions, but also localizes to perinuclear patches where Cdc42 is not present. Through binding to CRIB domain of CEP1, Cdc42 influences the cellular distribution of CEP1 between the two locations and affects cell behaviors. Our findings suggest that CEP1 plays both Cdc42 dependent and independent roles in regulating neural crest cell migration, and Cdc42 level controls the balance between the two activities. Results CEP1 is predominantly expressed in neural crest cells during Xenopus embryogenesis Cdc42 is critical for cellular behaviors and is involved in multiple cell and tissue movements. Consistent with its importance in multiple developmental processes, Cdc42 is broadly expressed during embryogenesis (Choi and Han, 2002; Lucas et al., 2002). To determine how CEP1 may mediate the activities of Cdc42, we first examined the expression of CEP1 during early development of Xenopus embryos. In situ hybridization analysis was performed and the results were summarized in Figure 1. During gastrulation, CEP1 is weakly expressed in the animal pole in presumptive ectoderm (data not shown). At the end of gastrulation, CEP1 is expressed in mesodermal cells around the blastopore, and in lateral borders of the neural plate, where neural crest cells are specified. CEP1 expression persists in neural crest cells as they separate from neural epithelium, segregate into three lobes, and commence migration at neurula stages (arrows and arrowhead). Neural crest cells continue to express CEP1 as they migrate into the branchial arches in tailbud stages (open arrowheads). From mid-tailbud stages, CEP1 transcript is also detected in anterior somites. The abundance of CEP1 transcripts in premigratory and migrating neural crest cells suggests that CEP1 maybe involved in regulating neural crest development. Figure 1 View largeDownload slide CEP1 is expressed in neural crest cells during frog embryogenesis. At gastrula stage (st.12.5 and st.14), CEP1 is expressed at the lateral borders of the neural plate, where neural crest cells are specified (arrows). Its expression is restricted to CNC at neurula stages (st.18) and continues in CNC cells as they commence migration (st.20; arrowhead) and migrate into branchial arches (open arrowheads). At tailbud stages (st.24 and st.26), CEP1 expression is also detected in somites. All embryos are oriented with anterior to the left. Dorsal views are shown in the upper panels and lateral views are shown in the lower panels. Figure 1 View largeDownload slide CEP1 is expressed in neural crest cells during frog embryogenesis. At gastrula stage (st.12.5 and st.14), CEP1 is expressed at the lateral borders of the neural plate, where neural crest cells are specified (arrows). Its expression is restricted to CNC at neurula stages (st.18) and continues in CNC cells as they commence migration (st.20; arrowhead) and migrate into branchial arches (open arrowheads). At tailbud stages (st.24 and st.26), CEP1 expression is also detected in somites. All embryos are oriented with anterior to the left. Dorsal views are shown in the upper panels and lateral views are shown in the lower panels. CEP1 is required for CNC migration, but not for neural crest specification Since CEP1 is expressed in both premigratory and migrating neural crest cells, CEP1 may potentially regulate both neural crest specification and migration. To determine which process CEP1 is involved, we next performed loss-of-function experiments using a translation blocking morpholino oligomer (MO) against the Xenopus laevis CEP1 gene. To validate the efficiency of CEP1-MO, CEP1(UTR)-EGFP containing the MO recognition site was overexpressed into 2-cell stage embryos with or without CEP1-MO. In situ hybridization analysis against CEP1 RNA was performed and showed that all embryos received ectopic CEP1 RNA. While CEP1(UTR)-EGFP expressing embryos showed strong EGFP signal, those co-injected with CEP1-MO showed much reduced or no EGFP signal, suggesting that the translation of the protein was blocked (Figure 2C). CEP1-EGFP without the 5′UTR can not be blocked by CEP1-MO, and is used in rescue experiments. Figure 2 View largeDownload slide CEP1 is required for CNC cell migration. (A) Embryos were injected with control-MO, CEP1-MO (10 ng), or CEP1-MO with CEP1 RNA (0.1 ng) into one of the dorsal animal cells at 8-cell stage, together with a lineage tracer (nβGal, red staining, marked by *). In situ hybridization analysis was performed at neurula stage (st.14) and tailbud stage (st.23). Arrows point to reduced neural crest migration indicated by Sox10/Twist expression. Embryos are oriented with anterior to the left. Dorsal views at neurula stage and lateral views of both uninjected and injected sides of the same embryo at tailbud stage are shown. (B) Percentage of embryos with defective neural crest migration. While CEP1-MO lead to defective migration in >90% of the embryos, adding CEP1 RNA rescued the rate of defective embryos to 45%. Fisher’s exact test was performed between control and CEP1-MO conditions and P < 0.05. χ2-test was performed between CEP1-MO and rescue conditions and P < 0.05. (C) CEP1-MO efficiently blocked the translation of CEP1. CEP1-EGFP containing MO recognition sequence in the 5′UTR was overexpressed in 2-cell stage embryos. While CEP1 RNA was expressed in all embryos as reflected by in situ hybridization, addition of CEP1-MO blocked the protein translation as reflected by the fluorescence signal of EGFP tag. CEP1-EGFP without 5′UTR can not be blocked by the morpholino. Figure 2 View largeDownload slide CEP1 is required for CNC cell migration. (A) Embryos were injected with control-MO, CEP1-MO (10 ng), or CEP1-MO with CEP1 RNA (0.1 ng) into one of the dorsal animal cells at 8-cell stage, together with a lineage tracer (nβGal, red staining, marked by *). In situ hybridization analysis was performed at neurula stage (st.14) and tailbud stage (st.23). Arrows point to reduced neural crest migration indicated by Sox10/Twist expression. Embryos are oriented with anterior to the left. Dorsal views at neurula stage and lateral views of both uninjected and injected sides of the same embryo at tailbud stage are shown. (B) Percentage of embryos with defective neural crest migration. While CEP1-MO lead to defective migration in >90% of the embryos, adding CEP1 RNA rescued the rate of defective embryos to 45%. Fisher’s exact test was performed between control and CEP1-MO conditions and P < 0.05. χ2-test was performed between CEP1-MO and rescue conditions and P < 0.05. (C) CEP1-MO efficiently blocked the translation of CEP1. CEP1-EGFP containing MO recognition sequence in the 5′UTR was overexpressed in 2-cell stage embryos. While CEP1 RNA was expressed in all embryos as reflected by in situ hybridization, addition of CEP1-MO blocked the protein translation as reflected by the fluorescence signal of EGFP tag. CEP1-EGFP without 5′UTR can not be blocked by the morpholino. To examine the loss of function effect of CEP1, CEP1-MO together with a lineage tracer (nβGal) was injected into one of the dorsal animal blastomeres of 8-cell stage embryos, leaving the uninjected contralateral side as an internal control. The embryos were collected at either late gastrula stage (stage 14) when neural crest cells are specified, or early tailbud stage (stage 23) when CNC cells have migrated into branchial arches, and neural crest cells in these embryos were marked by in situ hybridization against neural crest-specific marker genes Sox10 and Twist. At stage 14, Sox10 expression was observed in a bilaterally symmetric pattern in both control and CEP1-MO injected embryos (Figure 2A), suggesting that CEP1 is not required for neural crest specification. At stage 23, Sox10 and Twist expressing CNC cells migrated into branchial arches on both sides of the control embryo, with a clear segregation of the three migratory lobes (the third lobe will split to two at late tailbud stages). In contrast, neural crest gene expression on CEP1-MO expressing side was impaired (90.5%, n = 63), indicating a defect in neural crest migration. CEP1-MO receiving neural crest cells migrated shorter distances comparing to cells on the contralateral side, and failed to segregate into distinct lobes. When CEP1 RNA was co-expressed with CEP1-MO, the expression pattern of Sox10 and Twist was rescued (55% rescued to normal, n = 40), confirming that the migration defects were specific to the loss of CEP1. These results suggest that CEP1 is required for neural crest migration without affecting neural crest specification. To determine whether CEP1 is required cell autonomously by neural crest cells for their migration, a grafting experiment was performed. CNC explants from EGFP-labeled donor embryos were transplanted isotopically and isochronically into wild-type host embryos where the endogenous neural crest tissue was removed (Figure 3). When grafted control CNC cells migrated efficiently to the branchial arches and segregated into three to four migratory streams, CEP1-MO expressing grafts migrated much shorter distances ventrally and failed to form distinct migratory streams. These cells seemed to lose their interaction during this collective migration and appeared rounder in shape. When CEP1 was co-expressed, the migration of CNC grafts was significantly rescued. Figure 3 View largeDownload slide CEP1 is required cell-autonomously for neural crest migration. (A) GFP-labeled CNC explants were dissected and transplanted into unlabeled host embryos and their migration in the host embryos was imaged at late tailbud stages. Fluorescence and DIC merged images and fluorescence images alone are shown side by side. Defects in CEP1-MO expressing cells to migrate ventrally and to segregate into distinct streams (marked by numbers) were rescued by co-expressing CEP1 RNA. (B) The relative distance of lateral migration to the entire D-V length was calculated, and the number of distinct migratory stream was counted for each CNC graft and summarized in the bar graph. Both the migration distance and the number of segregated streams are significantly reduced by CEP1 knockdown, while adding CEP1 RNA significantly rescued both defects (student’s t-test, P < 0.01). Figure 3 View largeDownload slide CEP1 is required cell-autonomously for neural crest migration. (A) GFP-labeled CNC explants were dissected and transplanted into unlabeled host embryos and their migration in the host embryos was imaged at late tailbud stages. Fluorescence and DIC merged images and fluorescence images alone are shown side by side. Defects in CEP1-MO expressing cells to migrate ventrally and to segregate into distinct streams (marked by numbers) were rescued by co-expressing CEP1 RNA. (B) The relative distance of lateral migration to the entire D-V length was calculated, and the number of distinct migratory stream was counted for each CNC graft and summarized in the bar graph. Both the migration distance and the number of segregated streams are significantly reduced by CEP1 knockdown, while adding CEP1 RNA significantly rescued both defects (student’s t-test, P < 0.01). To quantitate these results, the relative distance of neural crest migration and the number of migratory streams were calculated in each grafted embryo and summarized in Figure 3B. While control neural crest grafts on average migrated 81% of the D-V axis and segregated into 3.2 lobes, CEP1-MO expressing grafts only migrated 46.5% along the D-V axis and segregated into 1.3 lobes. Coinjection of CEP1 RNA efficiently rescued the migration defects such that neural crest cells on average traveled 74.1% of the D-V length and segregated into 2.5 lobes, again confirming that the migration defects observed were specific to the loss of CEP1. CEP1 is required for cranial cartilage formation Since craniofacial cartilage and bones are important derivatives of CNC cells, we next asked whether the migration defects caused by CEP1-MO had later consequences on cartilage formation. Alcian blue staining was performed with late stage tadpoles (~stage 45) to examine the formation of cartilage elements. The results showed that neural crest derived mandibular, hyoid, and branchial arch cartilages were malformed on CEP1-MO expressing side (Figure 4). They were much smaller in size and sometimes completely missing. Thus, consistent with the notion that neural crest cells need to migrate to proper destinations in order to differentiate into correct structures, defects in CNC migration result in a general failure in facial cartilage formation. Figure 4 View largeDownload slide CEP1-MO inhibits cranial cartilage formation. Cranial cartilages from control- or CEP1-MO-injected tadpoles were stained with Alcian blue. Ventral views are shown with injected sides marked by *. Cartilage on CEP1-MO-injected side was malformed or even completely missing. M, Meckel’s cartilage formed by mandibular stream; CH, ceratohyal cartilage formed by hyoid stream; CB, ceratobranchial cartilage formed by third and fourth branchial arch streams. Figure 4 View largeDownload slide CEP1-MO inhibits cranial cartilage formation. Cranial cartilages from control- or CEP1-MO-injected tadpoles were stained with Alcian blue. Ventral views are shown with injected sides marked by *. Cartilage on CEP1-MO-injected side was malformed or even completely missing. M, Meckel’s cartilage formed by mandibular stream; CH, ceratohyal cartilage formed by hyoid stream; CB, ceratobranchial cartilage formed by third and fourth branchial arch streams. CEP1 regulates the dynamic rearrangements of actin filaments in neural crest cells To determine how CEP1 regulates neural crest cell migration, a CNC explant assay, which allows for closer examination of neural crest cell behaviors, was performed. CNC explants were dissected from control or CEP1-MO injected embryos at stages 13–14, plated on fibronectin-coated coverslips, and the cell shape changes were delineated by co-expressed membrane-tethered EGFP. Control neural crest cells displayed polarized shapes and formed multiple membrane protrusions. The protrusions extended or retracted dynamically as the cell changed its direction of migration. In contrast, CEP1-MO expressing cells were much rounder in shape, and formed fewer protrusions (Supplementary Figures S1 and S2). Since actin machinery is the driving force for cell shape changes and cell locomotion, we next co-expressed EGFP-Utrophin (EGFP fusion with the actin-binding domain of Utrophin) to label actin filaments in neural crest cells. As shown in Figure 5A, actin filaments were assembled and disassembled rapidly in control neural crest cells. Actin filaments were enriched in areas where protrusions were extending, and diminished where protrusions were retracting (follow arrows or arrowheads in different time frames). When fluorescence intensities of EGFP-labeled actin filaments along the long axis of a cell were plotted at different time points, we observed dynamic changes in actin filaments throughout the cell as well as active migration of the cell (Figure 5E). Time-lapse movies show that actin filaments translocate rapidly in different areas of the cells and their fluorescence intensity changes in a wave-like manner (Supplementary Movie S1). In sharp contrast, when CEP1-MO was expressed, actin filaments formed ring-like structures around the cell cortex and displayed restrained movement (Figure 5B and Supplementary Movie S2). The intensity of actin filaments shifted between the rings in a ripple-like manner (white open arrowheads). Rather than forming lamellipodia or filopodia protrusions, CEP1-MO expressing cells formed membrane blebs (Figure 5B and C; double peaks in Figure 5E), indicating a weaker association between the actin cortex and plasma membrane. Actin intensity plots from another 15 control and 15 CEP1-MO expressing cells were analyzed and the results confirmed that loss of CEP1 dramatically restricted the localization of actin filaments (Figure 5D). Rather than distributing along the entire axis of the cell, CEP1-MO expressing cells show significantly lower actin intensity at the cell center. The defects in actin organization and lamellipodia formation in these cells were also consistent with a decrease in cell translocation (purple-shaded cell in Figure 5A and B). When CEP1 was added back to the cells, the cell shapes and actin dynamics were significantly rescued (Figure 5C and Supplementary Movie S3). Figure 5 View largeDownload slide CEP1-MO disrupts the organization of actin filaments. CNC explants receiving EGFP-Utrophin with or without CEP1-MO (5 ng) were dissected and plated on FN-coated cover slides and the dynamics of actin filaments during cell migration were recorded by time-lapse microscopy. (A) Image frames of control neural crest explants over 5.5 min are shown. White or yellow arrows and arrowheads follow the same protrusion over time. (B) Image frames of CEP1-MO expressing cells. White open arrowheads point to the same actin bundle over time and yellow open arrowheads point to the membrane blebs. One cell in both control and CEP1-MO movie frames is shaded in purple to show cell shape changes and translocation over time. A’–A” and B’–B” are higher magnification views of corresponding boxed area. Images were taken at 40×. Scale bar, 20 μm. (C) The number of each type of protrusions formed was counted and averaged in the bar graph. While control neural crest cells formed 1.5 lamellipodia and 3.8 filopodia on average, blocking CEP1 reduced the formation of both protrusions to 0.3 and 0.8 per cell, respectively, and increased the formation of membrane blebs to 1.3 per cell. Co-expressing CEP1 rescued the formation of cell protrusions and restored the numbers to 1.0, 2.5, and 0.3, respectively. Student’s t-test was performed and both MO inhibition and rescue were significant (P < 0.01). (D) Intensity plots for EGFP-labeled actin filaments were generated for 15 control and 15 CEP1-MO expressing cells. Solid lines reflect the mean signal intensity at locations relative to the length of the cell, and dashed lines above and below mark the standard deviation at each location. There is no significant difference in actin intensity at the cell periphery, but significant difference at cell center. (E) Actin fluorescence intensities across one control and one CEP1-MO cell were plotted over 4 min. In control graph, dashed lines mark the positions of cell protrusions at the beginning and the end of the experiment, reflecting leftward migration of the cell. In CEP1-MO graph, blebs reflected by the double peaks are marked by arrows. The cell is not migrating, but shrinking towards the end. Figure 5 View largeDownload slide CEP1-MO disrupts the organization of actin filaments. CNC explants receiving EGFP-Utrophin with or without CEP1-MO (5 ng) were dissected and plated on FN-coated cover slides and the dynamics of actin filaments during cell migration were recorded by time-lapse microscopy. (A) Image frames of control neural crest explants over 5.5 min are shown. White or yellow arrows and arrowheads follow the same protrusion over time. (B) Image frames of CEP1-MO expressing cells. White open arrowheads point to the same actin bundle over time and yellow open arrowheads point to the membrane blebs. One cell in both control and CEP1-MO movie frames is shaded in purple to show cell shape changes and translocation over time. A’–A” and B’–B” are higher magnification views of corresponding boxed area. Images were taken at 40×. Scale bar, 20 μm. (C) The number of each type of protrusions formed was counted and averaged in the bar graph. While control neural crest cells formed 1.5 lamellipodia and 3.8 filopodia on average, blocking CEP1 reduced the formation of both protrusions to 0.3 and 0.8 per cell, respectively, and increased the formation of membrane blebs to 1.3 per cell. Co-expressing CEP1 rescued the formation of cell protrusions and restored the numbers to 1.0, 2.5, and 0.3, respectively. Student’s t-test was performed and both MO inhibition and rescue were significant (P < 0.01). (D) Intensity plots for EGFP-labeled actin filaments were generated for 15 control and 15 CEP1-MO expressing cells. Solid lines reflect the mean signal intensity at locations relative to the length of the cell, and dashed lines above and below mark the standard deviation at each location. There is no significant difference in actin intensity at the cell periphery, but significant difference at cell center. (E) Actin fluorescence intensities across one control and one CEP1-MO cell were plotted over 4 min. In control graph, dashed lines mark the positions of cell protrusions at the beginning and the end of the experiment, reflecting leftward migration of the cell. In CEP1-MO graph, blebs reflected by the double peaks are marked by arrows. The cell is not migrating, but shrinking towards the end. CEP1 is required for proper force transmission between cell and extracellular matrix When CEP1 was knocked down in neural crest cells, actin organization, lamellipodia formation, and cell translocation were compromised. To understand how CEP1 mediates such effects, we next looked at the mechanical basis for cell migration. During cell migration, the actin filaments polymerize at the front of the cell to push the cell forward, while relative movements between myosin II and actin filaments contract and pull the bulk of the cell towards the front. Both movements rely on the dynamic assembly and disassembly of focal adhesions (Nobes and Hall, 1995). Focal adhesions are large protein complexes that anchor a cell (and actin filaments) to the extracellular matrix. At the cell's leading edge, focal complexes stabilize the forming membrane protrusions; in the center, focal adhesions impede the retrograde movement of actin; and at cell rear, focal adhesions disassemble and proteins in the complex are recycled to the front. Here, we examined the formation of focal adhesions using antibodies against phosphorylated tyrosine (Zamir et al., 1999). In control cells, long focal adhesions were assembled in well-aligned arrays in the leading edge of cells (arrows and insets in Supplementary Figure S3). There were not many focal adhesions located in the center or trailing edge of the cells, consistent with their disassembly in these areas. Frog neural crest cells imbricate on each other in a scale-like manner, and some focal adhesions observed in the center of cell clusters were actually in lamellipodia underneath the neighboring cell. In contrast, focal adhesions formed in CEP1-MO cells were generally shorter in length. They sometimes appeared in the center of an isolated cell, or in disorganized arrays in the cell periphery (arrows and insets in Supplementary Figure S3). Since focal adhesions align with cell contractility, differences in focal adhesion arrangement and localization may correspond to a disoriented force transmission, leading to decreased cell translocation. To directly elucidate the propulsive forces mediated through focal adhesions, we performed traction force microscopy. Control or CEP1-MO expressing CNC explants were plated onto fibronectin-coated hydrogels. The traction force the cells applied on the hydrogel was reflected by the strain energy absorbed by the hydrogel, which can be calculated by measuring the displacements of fluorescent beads embedded in the surface layer of the hydrogel (Yeung et al., 2005; Kovari et al., 2016). As shown in Figure 6 and Supplementary Movies S4 and S5, we observed three types of traction force distribution around the cells, corresponding to three different types of cell behaviors. First were well-spread cells with large and dynamic protrusions, and there was high strain energy absorbed by the gel underneath these protrusions (Figure 6A). In addition, the orientation of the strain was always towards the cell body (or opposite to the direction the protrusion extended). The second were rounder cells that formed membrane blebs. They usually did not exert much stress underneath the blebs, which is predictable since the blebs do not adhere to the matrix. Instead, there was high strain energy underneath the center of the cell, possibly corresponding to the focal adhesions there (Figure 6B). However, the orientation of the strain was random, reflecting a defect in directionality. The third were cells displaying transitioning behaviors, and the spatial distribution of traction forces was also shifting between the cell peripheries and the cell center. Over 200 control and CEP1-MO expressing cells were categorized into these three types of cell behaviors and the results are shown in Figure 6C. While 69% of control cells displayed spreading behavior, only 22% of CEP1-MO expressing cells spread on hydrogel. Instead, 63% of CEP1-MO cells displayed membrane blebs. In addition to the differences in their spatial distribution, the magnitude of the strain energy was also reduced under blebbing cells. The average strain energy produced by each cell over the course of the movie was calculated (Figure 6D). While the average strain energies were similar under cells with the same phenotype, there was a significant increase in the strain energy under spreading cells. Figure 6 View largeDownload slide CEP-MO affected force transmission during neural crest cell migration. (A and B) Traction force microscopy was performed and the strain energy received by extracellular matrix was calculated and plotted on the image. In control cells, high strain energy was observed underneath lamellopodia protrusions (arrows). In CEP1-MO cells, high strain energy was observed underneath the cell center rather than underneath membrane blebs (arrows). (C) Over 200 control and CEP1-MO cells were categorized into three behavioral types, i.e. spreading, intermediate, and blebbing. χ2-test confirmed that CEP1-MO decreased the percentage of spreading cells and increased the percentage of blebbing cells significantly (P < 0.05). (D) Average strain energy underneath cells was calculated over the course of the movie and compared between different behavioral types using Fisher’s least significant difference test. While the hydrogel beneath control or CEP1-MO cells of the same behavior absorbed similar amount of strain energy, there were significant differences in strain energy beneath cells with different behaviors. Figure 6 View largeDownload slide CEP-MO affected force transmission during neural crest cell migration. (A and B) Traction force microscopy was performed and the strain energy received by extracellular matrix was calculated and plotted on the image. In control cells, high strain energy was observed underneath lamellopodia protrusions (arrows). In CEP1-MO cells, high strain energy was observed underneath the cell center rather than underneath membrane blebs (arrows). (C) Over 200 control and CEP1-MO cells were categorized into three behavioral types, i.e. spreading, intermediate, and blebbing. χ2-test confirmed that CEP1-MO decreased the percentage of spreading cells and increased the percentage of blebbing cells significantly (P < 0.05). (D) Average strain energy underneath cells was calculated over the course of the movie and compared between different behavioral types using Fisher’s least significant difference test. While the hydrogel beneath control or CEP1-MO cells of the same behavior absorbed similar amount of strain energy, there were significant differences in strain energy beneath cells with different behaviors. CEP1 is localized to two distinct locations during neural crest cell migration To better understand how CEP1 mediates its activities in actin organization and cell migration, we next examined the subcellular localization of CEP1 in migrating neural crest cells. EGFP-CEP1 and RFP-Utrophin were co-expressed in neural crest cells and their relative distribution during neural crest cell migration was recorded by time-lapse microscopy. As shown in Figure 7, CEP1 is distributed rather broadly in neural crest cells. One fraction of CEP1 is localized to the periphery of the cell. It was detected in plasma membrane, at the base of filopodia protrusions, and in a punctate manner along actin filaments in lamellipodia protrusions (see enlarged insets A’–A”, B’–B”, and C’). In addition, another fraction of CEP1 is localized to the center of the cell, close to the nucleus. They were organized in a fragmented or filamentous manner in the cytoplasm and concentrated to a dense patch near the nucleus (arrows). In mice endothelial cells, CEP1 (Borg5) has been described to associate with septins in the perinuclear region and promote the persistent directional migration of endothelial cells (Liu et al., 2014). In neural crest cells, CEP1 may play a similar role in maintaining the direction of cell migration at the perinuclear location. Figure 7 View largeDownload slide CEP1 localization in migrating neural crest cells. EGFP-CEP1 (0.05 ng) and RFP-Utrophin were co-expressed in neural crest cells. (A and B) CEP1 located to membrane protrusions and cell membrane, as well as in the perinuclear region in filamentous or condensed patch (arrows). A’–A” and B’–B” are enlarged insets of the boxed areas in A and B, showing co-localization of CEP1 with actin along lamellipodia protrusion, cell membrane, and at the base of filopodia protrusion. (C) For better visualization of CEP1 localization, one isolated neural crest cell was shown with GFP channel alone. CEP1 is localized to cell membrane (arrowheads) and in a punctate manner in protrusions (C’). Images were taken at 63×. Scale bar, 20 μm. Figure 7 View largeDownload slide CEP1 localization in migrating neural crest cells. EGFP-CEP1 (0.05 ng) and RFP-Utrophin were co-expressed in neural crest cells. (A and B) CEP1 located to membrane protrusions and cell membrane, as well as in the perinuclear region in filamentous or condensed patch (arrows). A’–A” and B’–B” are enlarged insets of the boxed areas in A and B, showing co-localization of CEP1 with actin along lamellipodia protrusion, cell membrane, and at the base of filopodia protrusion. (C) For better visualization of CEP1 localization, one isolated neural crest cell was shown with GFP channel alone. CEP1 is localized to cell membrane (arrowheads) and in a punctate manner in protrusions (C’). Images were taken at 63×. Scale bar, 20 μm. CEP1 interacts with Cdc42 and regulates the cellular localization of each other during neural crest cell migration To understand how the activities of CEP1 at both locations are regulated, we next examined its interaction with Cdc42. Biochemical analysis in cell culture indicates that CEP1 can physically bind to Cdc42 with high affinity through its CRIB domain (Burbelo et al., 1999; Joberty et al., 1999). However, where and how CEP1 interacts with Cdc42 during cell migration remains unknown. Since cell migration is a highly dynamic process, we expect that the interaction between CEP1 and Cdc42 to be transient and dynamic as well. To determine where they interact during cell migration, we expressed fluorescence fusion proteins for Cdc42 and CEP1 in neural crest cells and directly observed their localization during cell migration. As shown in Figure 8A and Supplementary Movie S6, CEP1 and Cdc42 overlapped in membrane protrusions (arrows), but not in perinuclear region (arrowheads). Such localization pattern suggests that CEP1 may play both Cdc42 dependent and Cdc42 independent activities in neural crest cells. Figure 8 View largeDownload slide CEP1 and Cdc42 interact with each other during neural crest migration. (A) Co-localization of CEP1 and Cdc42 (0.05 ng each). Cdc42 co-localizes with CEP1 in membrane protrusions (arrows and inset), but not in perinuclear region (arrowheads). Images were taken at 60×. Scale bar, 20 μm. (B) The cellular localization of Cdc42 or CEP1 at increased (0.2 ng of RNA) or decreased (5 ng of MO) level of CEP1 or Cdc42. Arrowheads in EGFP-Cdc42 panels point to high level of Cdc42 in protrusions and cell membrane. In Cherry-CEP1 panels, arrows point to CEP1 in perinuclear region, while closed and open arrowheads point to cell protrusions and plasma membrane, respectively. Enlarged insets show one cell (in dashed box) under each condition. Images were taken at 63×. Scale bar, 20 μm. (C) Numbers of membrane protrusions in control, CEP1 overexpression, and CEP1-MO cells were counted and summarized in the bar graph. Both increase and decrease of membrane protrusions are significant (student’s t-test, P < 0.01). (D) The cellular distribution of CEP1 under control, Cdc42 overexpression, and Cdc42 knockdown conditions were compared. The intensities of Cherry-CEP1 fluorescence signals in 16–17 cells were analyzed along a line across each cell (see examples in B), and the data were normalized against the length of the cell. Solid lines are average signal intensity under each condition and dashed lines are average ± standard deviation. Gray double-headed arrows point to regions of significant differences. (E) CEP1 lacking the CRIB domain can not interact with Cdc42. EGFP-CEP1(ΔCRIB) is localized to the nucleus, and changes in Cdc42 level do not alter its localization. In contrast, EGFP-CEP1 is mainly localized to membrane protrusions when Cdc42 is co-expressed. Scale bar, 20 μm. Figure 8 View largeDownload slide CEP1 and Cdc42 interact with each other during neural crest migration. (A) Co-localization of CEP1 and Cdc42 (0.05 ng each). Cdc42 co-localizes with CEP1 in membrane protrusions (arrows and inset), but not in perinuclear region (arrowheads). Images were taken at 60×. Scale bar, 20 μm. (B) The cellular localization of Cdc42 or CEP1 at increased (0.2 ng of RNA) or decreased (5 ng of MO) level of CEP1 or Cdc42. Arrowheads in EGFP-Cdc42 panels point to high level of Cdc42 in protrusions and cell membrane. In Cherry-CEP1 panels, arrows point to CEP1 in perinuclear region, while closed and open arrowheads point to cell protrusions and plasma membrane, respectively. Enlarged insets show one cell (in dashed box) under each condition. Images were taken at 63×. Scale bar, 20 μm. (C) Numbers of membrane protrusions in control, CEP1 overexpression, and CEP1-MO cells were counted and summarized in the bar graph. Both increase and decrease of membrane protrusions are significant (student’s t-test, P < 0.01). (D) The cellular distribution of CEP1 under control, Cdc42 overexpression, and Cdc42 knockdown conditions were compared. The intensities of Cherry-CEP1 fluorescence signals in 16–17 cells were analyzed along a line across each cell (see examples in B), and the data were normalized against the length of the cell. Solid lines are average signal intensity under each condition and dashed lines are average ± standard deviation. Gray double-headed arrows point to regions of significant differences. (E) CEP1 lacking the CRIB domain can not interact with Cdc42. EGFP-CEP1(ΔCRIB) is localized to the nucleus, and changes in Cdc42 level do not alter its localization. In contrast, EGFP-CEP1 is mainly localized to membrane protrusions when Cdc42 is co-expressed. Scale bar, 20 μm. To determine the functional interactions between CEP1 and Cdc42, the level of one protein was manipulated and its impact on the intracellular distribution of the other protein was examined (Figure 8B and Supplementary Movies S7–S12). When CEP1 was downregulated by morpholino, neural crest cells adopted a rounder morphology without forming obvious membrane protrusions. When CEP1 was overexpressed, cells looked relatively normal, despite that they made protrusions at different orientations. Figure 8C summarizes the numbers of membrane protrusions formed by cells expressing different levels of CEP1 and both the increase and the decrease of CEP1 level significantly changed the number of protrusions. The changes in Cdc42 localization were not very obvious. In control cells, Cdc42 was localized in cell membrane and membrane protrusions (arrowheads). When CEP1 overexpression led to more protrusions in each cell, Cdc42 was also enriched in different protrusions. In contrast, when CEP1 was knocked down, Cdc42 was distributed more evenly throughout the cell membrane. Since Cdc42 needs to be in its GTP-bound active state to mediate downstream effects, we next examined whether the distribution of activated Cdc42 was influenced by the level of CEP1. To this end, we expressed a reporter GFP-wGBD (Cdc42-binding domain of N-WASP) that only binds to active Cdc42 in neural crest cells (Benink and Bement, 2005). In control and CEP1-MO cells, we observed no change in the overall level of active Cdc42, consistent with the notion that CEP1 does not have kinase activity to directly activate Cdc42. In control cells, GFP-wGBD was enriched in the leading edge, and was constantly moving between different areas of the membrane protrusions (Supplementary Movie S13). In CEP1-MO expressing cells, despite changes in cell morphology, GFP-wGBD was expressed around the cell periphery, especially at areas membrane blebs were forming (Supplementary Movie S14). This demonstrates that CEP1 influences the localization of active Cdc42 to regulate cell polarity and the formation of membrane protrusions. Conversely, when Cdc42 was overexpressed, a large portion of CEP1 was recruited to Cdc42 in cell membrane and in membrane protrusions (open arrowheads and arrowheads in Figure 8B), regardless of the orientation of the membrane. When Cdc42 was downregulated by morpholino, CEP1 was largely relocated near the nucleus (arrows), leaving little CEP1 at cell membrane. Signal intensities of Cherry-CEP1 across 16–17 cells under each condition were analyzed and plotted in Figure 8D. While CEP1 alone was localized to cell periphery as well as cell center (blue curve), Cdc42 concentrated them to cell periphery (green curve). When the Cdc42 level was reduced, CEP1 redistributed to the cell center (red curve). Gray arrows indicate that there are significant differences in CEP1 levels at cell peripheries between control and Cdc42-MO expressing cells, and at the cell centers between control and Cdc42 overexpressing cells. Given that Cdc42 is only expressed near the cell membrane, these results suggest that through physical binding, Cdc42 controls the balance of CEP1 between the two subcellular locations. To confirm that Cdc42 directly regulates CEP1 localization, we deleted the CRIB domain from CEP1, and fused the mutant construct with EGFP (EGFP-CEP1(ΔCRIB)). When expressed into neural crest cells, CEP1(ΔCRIB) is localized to the nucleus, regardless of the level of Cdc42 in the cells (Figure 8E). This is in sharp contrast to wild-type CEP1, which is largely localized to membrane protrusions when Cdc42 is overexpressed. These results indicate that CRIB domain is critical in mediating CEP1−Cdc42 interaction, and Cdc42 recruits CEP1 to membrane protrusions through direct binding. Discussion Cdc42eps co-evolved with neural crest, somite, cardiac microvasculature, and other vertebrate specific tissue or structures, thus possibly play important roles in the development of these tissue and structures. Here, we show that Cdc42ep1 (CEP1) is highly expressed in neural crest cells and plays important roles in their migration. Our study further demonstrates that CEP1 is localized to two subcellular areas, and Cdc42 directly controls the balance between these two CEP1 subpopulations. We showed that CEP1 is required for neural crest cell migration in vivo and in vitro. When CEP1 was knocked down, neural crest cells did not form lamellipodia protrusions, but instead made membrane blebs. This could result from several mechanisms. First, CEP1 may be required for the formation of lamellipodia or filopodia protrusions. It has been reported in fibroblasts that overexpression of CEP2 can induce ectopic membrane protrusions and this requires the presence of active Cdc42 as well as intact CRIB domain (Hirsch et al., 2001), demonstrating that CEP2 directly binds to and cooperates with Cdc42 in promoting protrusion formation. It is possible that CEP1 plays a similar role in interacting with Cdc42 and promoting membrane protrusions. This is supported by the result that changes in CEP1 level significantly changed the number of membrane protrusions made, especially filopodia protrusions (Figures 5C and 8C). Second, CEP1 may negatively regulate bleb formation. Membrane blebs, which grow as a result of intracellular pressure, are generated by actomyosin contractions (Charras and Paluch, 2008). They can be induced by increasing actomyosin contractility through activated RhoA/ROCK, or by increasing cortical tension by thickening actin cortex (Han et al., 2009; Bergert et al., 2012). In our experiments, when CEP1 was reduced, higher levels of pMLC and thick actin filament bundles were observed along cell periphery (Figure 5B and Supplementary Figure S3). Whether cortical tension was increased or not remains to be elucidated. In addition to increased cortical tension, blebs also associate with lower cellular adhesion. We have observed changes in the organization of focal adhesions and the distribution and magnitude of traction forces in CEP1-MO induced blebbing cells (Figure 6 and Supplementary Figure S3). Since cell adhesion can influence actin polymerization while actomyosin contraction generates tension, which can control the dynamics of focal adhesions, defect in one process may be amplified and reinforced through such feedback regulations. In our experiments, forming blebs instead of lamellipodia compromised the migration of neural crest cells. In contrast, zebrafish germ cells use membrane blebs in their migration and several cancer cells have been observed to made bleb-like protrusions when migrate in 3D matrix (Kardash et al., 2010; Lorentzen et al., 2011). It has been suggested that the dynamic shifts between actin protrusivity and actomyosin contractility, and the resulting transition between lamellipodia and blebs may be advantageous for cells to quickly adapt to local environment (Bergert et al., 2012). Since neural crest cells made different types of protrusions at different CEP1 levels, CEP1 may be involved in the mechanisms controlling the switch between the two types of cell protrusions. We observed that some CEP1 located near the nucleus. This is very similar to the observation of Liu et al. (2014) in mouse cardiac endothelial cells that CEP1 appeared as a prominent filamentous patch above the nucleus. They showed that CEP1 promotes the assembly and alignment of septin filaments and actin filaments and together they regulate actomyosin organization and persistent cell migration. Parallel actomyosin bundles have been reported previously to align with the direction of cell migration and are required for maintaining the directionality of migration (Lo et al., 2004). The actomyosin in the center of the cell may play a role in coupling the propulsive anterior and the resistive posterior to ensure that the entire cell undergoes coordinated migration (Guo and Wang, 2012). Septin has recently been considered the fourth component of the cytoskeleton and can bind to both actin filament and myosin II to promote the assembly of actomyosin (Joo et al., 2007; Mostowy and Cossart, 2012). CEPs can also bind to septin directly and stimulate the assembly of septin filaments (Joberty et al., 2001; Calvo et al., 2015; Farrugia and Calvo, 2016). It is likely that CEP1 in neural crest cells also interact with septin filaments to control the stability and the orientation of actomyosin, thus regulating the direction of cell migration. CEPs were identified through yeast two hybrid screen for binders of Rho GTPases, and they were confirmed to physically bind to Cdc42, but not to Rac1 or RhoA (Joberty et al., 1999). However, it has been unclear how Cdc42 and CEPs interact functionally. In fibroblast cells, expression of dominant negative Cdc42 abolished pseudopodia induced by CEPs, while in keratinocytes, expression of constitutive active Cdc42 or CEP2/CEP5 all stimulate stress fiber formation (Hirsch et al., 2001). However, it is also reported in fibroblast cells that CEP5 may compete with other Cdc42 effectors and results in inhibiting Cdc42 activity in cell spreading (Joberty et al., 1999). On the other hand, in cancer-associated fibroblasts and MDCK epithelial cells, Cdc42 inhibits the interaction between CEP3/5 and septin (Joberty et al., 2001; Calvo et al., 2015). Here, we showed that CEP1 has two subcellular localizations and Cdc42, rather than inhibiting CEP1 activity, physically attracts CEP1 away from perinuclear region to cell protrusions. Our results suggest that the Cdc42 level is critical for a proper balance of CEP1 at these locations to control cell morphology and behavior. Similarly, a recent report showed that Cdc42 activity regulates the correct positioning of CEP3 in cancer-associated fibroblasts (Farrugia and Calvo, 2017). Therefore, Cdc42 may play a common role in regulating different CEPs. In contrast to wild-type CEP1 that localized to membrane protrusions and perinuclear region, CEP1(ΔCRIB) is mainly localized to the nucleus in neural crest cells. Two questions remain unresolved. First is whether CEP1 has endogenous nucleus distribution. Many effector proteins for Rho GTPases (e.g. Arp2/3, WAVE1, cofilin, and formin) can act inside the nucleus and regulate the dynamics of nucleus actin and gene transcription (Huet et al., 2012). It is possible that when there is no active Cdc42 to interact with CEP1 in the cytoplasm, CEP1 can translocate into nucleus to interact with nucleus actin. Second is whether CEP1 can regulate actin organization independent of Cdc42. Since CEP1(ΔCRIB) mainly locates to the nucleus, overexpression of CEP1(ΔCRIB) failed to regulate actin organization, reflected by no hyperpigmentation as seen in wild-type CEP1 overexpression. However, Calvo et al. (2015) reported that in addition to low affinity to Cdc42, CRIB-defective CEP3 also has lower binding affinity to actin and septins. Similarly, CEP1(ΔCRIB) may also display impaired interaction with actin, septin, or other proteins, thus hindered their activity in actin regulation. Materials and methods Embryo manipulations, MOs, and RNA preparation Xenopus laevis embryos were obtained and microinjected with morpholino and RNA as previously described (Nie et al., 2009). CEP1-MO (5′-GGTTCATTGTTCCTTCTTTTTCTGA-3′) hybridizes to −18 to 7 position relative to the translational start site of Xenopus CEP1 (GenBank Accession No. NM_1114777), Cdc42-MO (5′-CTACACATTTAATTGTCTGCATGGC-3′) hybridizes to −3 to 22 position relative to the translational start site of Xenopus Cdc42 (GenBank Accession No. NM_1085899), and standard control MO (Gene tools, Philomath, OR) were used in the study. Xenopus CEP1 and Cdc42 were subcloned into pCS2 + 8NEGFP or pCS2 + 8NmCherry vectors (addgene) to generate fusion constructs. GFP-wGBD was a gift from William Bement (Addgene plasmid #26734). To generate Xenopus CEP1 with the 5′UTR for MO recognition (CEP1(UTR)) and mutation that lacks the CRIB domain (CEP1(ΔCRIB)), site-directed mutagenesis was performed using CEP1-EGFP as a template. All experimental procedures were performed according to USDA Animal Welfare Act Regulations and have been approved by Institutional Animal Care and Use Committee, in compliance of Public Health Service Policy. Red-Gal staining, in situ hybridization, immunohistochemistry, and cartilage staining Red-Gal staining, in situ hybridization, immunohistochemistry, and cartilage staining were performed as previously describe (Nie et al., 2009). Detailed procedures are provided in Supplementary Materials and methods. Primary antibodies used in the study include anti-phosphotyrosine antibody (4G10; EMD Millipore), diluted at 1:200; anti-Arp2 (Thermo PA5-19760), diluted at 1:1000; and anti-pMLC (Abcam ab2480), diluted at 1:400. FITC-conjugated secondary antibody was used at 1:1000. CNC explant culture, grafting, and microscopy CNC explants receiving different MOs or RNAs encoding fluorescent proteins were dissected from stages 13–14 embryos as previously described (Borchers et al., 2000; Alfandari et al., 2003; DeSimone et al., 2005) and cultured and imaged as previously described (Nie et al., 2009). See Supplementary Materials and methods for details. Traction force microscopy The preparation of polyacrylamide hydrogel substrates and the analysis process used to implement traction force microscopy have been described in detail previously (Yeung et al., 2005; Kovari et al., 2016) and are summarized in Supplementary Materials and methods. The Matlab code for TFM is available at https://github.com/dkovari/TFMatlab. The mean strain energy was calculated for ~30 control and CEP1-MO cells over time. The mean strain energy from samples with different phenotypes were compared with multiple comparison test using Fisher’s least significant difference (Williams and Abdi, 2010), after confirming that different samples were unequal with one-way analysis of variance (ANOVA) test. Supplementary material Supplementary material is available at Journal of Molecular Cell Biology online. Acknowledgements We would like to thank Dr Marianne Bronner (California Institute of Technology, Pasadena, USA) for her comments and discussion. We would also like to acknowledge the support from the National Institutes of Health, National Science Foundation, and Georgia Institute of Technology. Funding This work is supported by the National Institutes of Health (R00DE022796 to S.N.) and National Science Foundation (DMR-0955811 to J.E.C. and PHY-0848797 to J.E.C. and D.T.K.). Conflict of interest: none declared. References Alfandari , D. , Cousin , H. , Gaultier , A. , et al. . ( 2003 ). Integrin α5β1 supports the migration of Xenopus cranial neural crest on fibronectin . Dev. Biol. 260 , 449 – 464 . Google Scholar CrossRef Search ADS PubMed Benink , H.A. , and Bement , W.M. ( 2005 ). Concentric zones of active RhoA and Cdc42 around single cell wounds . J. Cell Biol. 168 , 429 – 439 . Google Scholar CrossRef Search ADS PubMed Bergert , M. , Chandradoss , S.D. , Desai , R.A. , et al. . ( 2012 ). Cell mechanics control rapid transitions between blebs and lamellipodia during migration . Proc. Natl Acad. Sci. USA 109 , 14434 – 14439 . Google Scholar CrossRef Search ADS Borchers , A. , Epperlein , H.H. , and Wedlich , D. ( 2000 ). An assay system to study migratory behavior of cranial neural crest cells in Xenopus . Dev. Genes Evol. 210 , 217 – 222 . Google Scholar CrossRef Search ADS PubMed Burbelo , P.D. , Drechsel , D. , and Hall , A. ( 1995 ). A conserved binding motif defines numerous candidate target proteins for both Cdc42 and Rac GTPases . J. Biol. Chem. 270 , 29071 – 29074 . Google Scholar CrossRef Search ADS PubMed Burbelo , P.D. , Snow , D.M. , Bahou , W. , et al. . ( 1999 ). MSE55, a Cdc42 effector protein, induces long cellular extensions in fibroblasts . Proc. Natl Acad. Sci. USA 96 , 9083 – 9088 . Google Scholar CrossRef Search ADS Calvo , F. , Ranftl , R. , Hooper , S. , et al. . ( 2015 ). Cdc42EP3/BORG2 and septin network enables mechano-transduction and the emergence of cancer-associated fibroblasts . Cell Rep. 13 , 2699 – 2714 . Google Scholar CrossRef Search ADS PubMed Carmona-Fontaine , C. , Matthews , H.K. , Kuriyama , S. , et al. . ( 2008 ). Contact inhibition of locomotion in vivo controls neural crest directional migration . Nature 456 , 957 – 961 . Google Scholar CrossRef Search ADS PubMed Charras , G. , and Paluch , E. ( 2008 ). Blebs lead the way: how to migrate without lamellipodia . Nat. Rev. Mol. Cell Biol. 9 , 730 – 736 . Google Scholar CrossRef Search ADS PubMed Choi , S.C. , and Han , J.K. ( 2002 ). Xenopus Cdc42 regulates convergent extension movements during gastrulation through Wnt/Ca2+ signaling pathway . Dev. Biol. 244 , 342 – 357 . Google Scholar CrossRef Search ADS PubMed DeSimone , D.W. , Davidson , L. , Marsden , M. , et al. . ( 2005 ). The Xenopus embryo as a model system for studies of cell migration . Methods Mol. Biol. 294 , 235 – 245 . Google Scholar PubMed Dupin , E. , Creuzet , S. , and Le Douarin , N.M. ( 2006 ). The contribution of the neural crest to the vertebrate body . Adv. Exp. Med. Biol. 589 , 96 – 119 . Google Scholar CrossRef Search ADS PubMed Farrugia , A.J. , and Calvo , F. ( 2016 ). The Borg family of Cdc42 effector proteins Cdc42EP1–5 . Biochem. Soc. Trans. 44 , 1709 – 1716 . Google Scholar CrossRef Search ADS PubMed Farrugia , A.J. , and Calvo , F. ( 2017 ). Cdc42 regulates Cdc42EP3 function in cancer-associated fibroblasts . Small GTPases 8 , 49 – 57 . Google Scholar CrossRef Search ADS PubMed Guo , W.H. , and Wang , Y.L. ( 2012 ). A three-component mechanism for fibroblast migration with a contractile cell body that couples a myosin II-independent propulsive anterior to a myosin II-dependent resistive tail . Mol. Biol. Cell 23 , 1657 – 1663 . Google Scholar CrossRef Search ADS PubMed Han , Y. , Eppinger , E. , Schuster , I.G. , et al. . ( 2009 ). Formin-like 1 (FMNL1) is regulated by N-terminal myristoylation and induces polarized membrane blebbing . J. Biol. Chem. 284 , 33409 – 33417 . Google Scholar CrossRef Search ADS PubMed Hirsch , D.S. , Pirone , D.M. , and Burbelo , P.D. ( 2001 ). A new family of Cdc42 effector proteins, CEPs, function in fibroblast and epithelial cell shape changes . J. Biol. Chem. 276 , 875 – 883 . Google Scholar CrossRef Search ADS PubMed Huet , G. , Skarp , K.P. , and Vartiainen , M.K. ( 2012 ). Nuclear actin levels as an important transcriptional switch . Transcription 3 , 226 – 230 . Google Scholar CrossRef Search ADS PubMed Joberty , G. , Perlungher , R.R. , and Macara , I.G. ( 1999 ). The Borgs, a new family of Cdc42 and TC10 GTPase-interacting proteins . Mol. Cell. Biol. 19 , 6585 – 6597 . Google Scholar CrossRef Search ADS PubMed Joberty , G. , Perlungher , R.R. , Sheffield , P.J. , et al. . ( 2001 ). Borg proteins control septin organization and are negatively regulated by Cdc42 . Nat. Cell Biol. 3 , 861 – 866 . Google Scholar CrossRef Search ADS PubMed Joo , E. , Surka , M.C. , and Trimble , W.S. ( 2007 ). Mammalian SEPT2 is required for scaffolding nonmuscle myosin II and its kinases . Dev. Cell 13 , 677 – 690 . Google Scholar CrossRef Search ADS PubMed Kardash , E. , Reichman-Fried , M. , Maitre , J.L. , et al. . ( 2010 ). A role for Rho GTPases and cell-cell adhesion in single-cell motility in vivo . Nat. Cell Biol. 12 , 47 – 53 . Google Scholar CrossRef Search ADS PubMed Kovari , D.T. , Wei , W.B. , Chang , P. , et al. . ( 2016 ). Frustrated phagocytic spreading of J774A-1 macrophages ends in Myosin II-dependent contraction . Biophys. J. 111 , 2698 – 2710 . Google Scholar CrossRef Search ADS PubMed Le Douarin , N. ( 1980 ). Migration and differentiation of neural crest cells . Curr. Top. Dev. Biol. 16 , 31 – 85 . Google Scholar CrossRef Search ADS PubMed Liu , Z. , Vong , Q.P. , Liu , C. , et al. . ( 2014 ). Borg5 is required for angiogenesis by regulating persistent directional migration of the cardiac microvascular endothelial cells . Mol. Biol. Cell 25 , 841 – 851 . Google Scholar CrossRef Search ADS PubMed Lo , C.M. , Buxton , D.B. , Chua , G.C. , et al. . ( 2004 ). Nonmuscle myosin IIb is involved in the guidance of fibroblast migration . Mol. Biol. Cell 15 , 982 – 989 . Google Scholar CrossRef Search ADS PubMed Lorentzen , A. , Bamber , J. , Sadok , A. , et al. . ( 2011 ). An ezrin-rich, rigid uropod-like structure directs movement of amoeboid blebbing cells . J. Cell Sci. 124 , 1256 – 1267 . Google Scholar CrossRef Search ADS PubMed Lucas , J.M. , Nikolic , I. , and Hens , M.D. ( 2002 ). cDNA cloning, sequence comparison, and developmental expression of Xenopus rac1 . Mech. Dev. 115 , 113 – 116 . Google Scholar CrossRef Search ADS PubMed Matthews , H.K. , Marchant , L. , Carmona-Fontaine , C. , et al. . ( 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 . Google Scholar CrossRef Search ADS PubMed Mostowy , S. , and Cossart , P. ( 2012 ). Septins: the fourth component of the cytoskeleton . Nat. Rev. Mol. Cell Biol. 13 , 183 – 194 . Google Scholar PubMed Nelson , K.K. , and Nelson , R.W. ( 2004 ). Cdc42 Effector Protein 2 (XCEP2) is required for normal gastrulation and contributes to cellular adhesion in Xenopus laevis . BMC Dev. Biol. 4 , 13 . Google Scholar CrossRef Search ADS PubMed Nie , S. , and Bronner , M.E. ( 2015 ). Dual developmental role of transcriptional regulator Ets1 in Xenopus cardiac neural crest vs. heart mesoderm . Cardiovasc. Res. 106 , 67 – 75 . Google Scholar CrossRef Search ADS PubMed Nobes , C.D. , and Hall , A. ( 1995 ). Rho, rac, and cdc42 GTPases regulate the assembly of multimolecular focal complexes associated with actin stress fibers, lamellipodia, and filopodia . Cell 81 , 53 – 62 . Google Scholar CrossRef Search ADS PubMed Vong , Q.P. , Liu , Z. , Yoo , J.G. , et al. . ( 2010 ). A role for borg5 during trophectoderm differentiation . Stem Cells 28 , 1030 – 1038 . Google Scholar CrossRef Search ADS PubMed Williams , L.J. , and Abdi , H. ( 2010 ). Fisher’s least significant difference (LSD) test. In: Salkind , N.J. (ed.). Encyclopedia of Research Design . Thousand Oaks, CA : SAGE Publications Inc. , 1 – 6 . Yeung , T. , Georges , P.C. , Flanagan , L.A. , et al. . ( 2005 ). Effects of substrate stiffness on cell morphology, cytoskeletal structure, and adhesion . Cell Motil. Cytoskel. 60 , 24 – 34 . Google Scholar CrossRef Search ADS Zamir , E. , Katz , B.Z. , Aota , S. , et al. . ( 1999 ). Molecular diversity of cell-matrix adhesions . J. Cell Sci. 112(Pt 11) , 1655 – 1669 . Google Scholar PubMed © The Author (2017). Published by Oxford University Press on behalf of Journal of Molecular Cell Biology, IBCB, SIBS, CAS. All rights reserved. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Journal of Molecular Cell Biology Oxford University Press

Cdc42 regulates the cellular localization of Cdc42ep1 in controlling neural crest cell migration

Loading next page...
 
/lp/ou_press/cdc42-regulates-the-cellular-localization-of-cdc42ep1-in-controlling-LjBVx51jO4
Publisher
Oxford University Press
Copyright
© The Author (2017). Published by Oxford University Press on behalf of Journal of Molecular Cell Biology, IBCB, SIBS, CAS. All rights reserved.
ISSN
1674-2788
eISSN
1759-4685
D.O.I.
10.1093/jmcb/mjx044
Publisher site
See Article on Publisher Site

Abstract

Abstract The member of Rho family of small GTPases Cdc42 plays important and conserved roles in cell polarity and motility. The Cdc42ep family proteins have been identified to bind to Cdc42, yet how they interact with Cdc42 to regulate cell migration remains to be elucidated. In this study, we focus on Cdc42ep1, which is expressed predominantly in the highly migratory neural crest cells in frog embryos. Through morpholino-mediated knockdown, we show that Cdc42ep1 is required for the migration of cranial neural crest cells. Loss of Cdc42ep1 leads to rounder cell shapes and the formation of membrane blebs, consistent with the observed disruption in actin organization and focal adhesion alignment. As a result, Cdc42ep1 is critical for neural crest cells to apply traction forces at the correct place to migrate efficiently. We further show that Cdc42ep1 is localized to two areas in neural crest cells: in membrane protrusions together with Cdc42 and in perinuclear patches where Cdc42 is absent. Cdc42 directly interacts with Cdc42ep1 (through the CRIB domain) and changes in Cdc42 level shift the distribution of Cdc42ep1 between these two subcellular locations, controlling the formation of membrane protrusions and directionality of migration as a consequence. These results suggest that Cdc42ep1 elaborates Cdc42 activity in neural crest cells to promote their efficient migration. actin cytoskeleton, neural crest, Rho GTPases, Cdc42 effector protein 1, cell migration Introduction The Rho family of small GTPases, RhoA, Rac1, and Cdc42, are important for cellular behaviors, including cell adhesion, cell shape changes, migration, and cytokinesis (Nobes and Hall, 1995). During embryonic development, Rho GTPases play critical roles in gastrulation, neural tube closure, and neural crest migration. The neural crest is a highly motile cell population specific to vertebrates. They migrate long distances in vertebrate embryos to give rise to craniofacial bones and cartilages as well as many other derivatives (Le Douarin, 1980; Dupin et al., 2006). Similar to other directional cell migration, neural crest cell migration requires cell polarization with protrusions extending at the cell front and stress fibers retracting the cell rear. At the molecular level, this polarity is mediated through the activation of small Rho GTPases. Activation of Rac1 and Cdc42 leads to the formation of lamellipodia at the leading edge, whereas activation of RhoA at the rear results in retraction of the cell (Matthews et al., 2008). At cell–cell contact, activation of members of the Wnt planar cell polarity pathway activates RhoA, and local RhoA activity antagonizes Rac1 and, therefore, inhibits the formation of cell protrusions (Carmona-Fontaine et al., 2008). To achieve the multiplexed activities of Rho GTPases, numerous downstream effector proteins exist and mediate a subset of their activities. A common motif found in the majority of Cdc42/Rac1 effector proteins is the conserved 16-aa Cdc42/Rac interactive-binding (CRIB) domain, which serves as a binding site for Cdc42 and/or Rac (Burbelo et al., 1995). Many proteins containing this motif have been identified and proved to be effector proteins for Cdc42/Rac1, such as p21 activated kinases (PAKs), mixed lineage kinases (MLKs), and Cdc42 effector proteins (Cdc42eps or CEPs). CEPs are a family of poorly understood Cdc42 effector proteins, also named binder of the Rho GTPases (Borgs) (Joberty et al., 1999). CEPs only exist in vertebrates, possibly mediating Cdc42 in vertebrate specific structures. There are only a handful of studies on CEPs during embryonic development. In mouse blastocysts, CEP1 (Borg5) enhances trophectoderm differentiation by promoting the sorting of trophectodermal cells to the outer layer (Vong et al., 2010). CEP1 also promotes microvascular angiogenesis by regulating directional migration of endothelial cells and is required in retinal and cardiac development in mouse (Liu et al., 2014). CEP2 (Borg1) plays a role in Xenopus gastrulation, by promoting the involution of mesoderm and the cell–cell adhesion of non-neural ectoderm (Nelson and Nelson, 2004). In cell culture, CEPs have been reported to involve in cell protrusions and actin cytoskeleton organization. CEP1 can induce membrane ruffling in Cos-7 cells, whereas induce long actin-based protrusions in NIH 3T3 fibroblasts (Burbelo et al., 1999). In primary keratinocytes, CEP2 and CEP5 (Borg3) expression reduced F-actin localization to adherence junctions with an increase in thin stress fibers (Hirsch et al., 2001). Despite physically binding to Cdc42, little is known of how they interact with Cdc42 in regulating cell motility. In this study, we characterize the activity of Cdc42ep1 (CEP1) in neural crest migration in Xenopus laevis. We show that in contrast to broadly expressed Cdc42, CEP1 is predominantly expressed in cranial neural crest (CNC) cells in Xenopus embryos. Consistent to its expression pattern, loss of CEP1 inhibited CNC cell migration. CEP1 knockdown not only affected cell morphology, but also interrupted the organization of functional structures for cell migration, such as the lamellipodia protrusions and focal adhesions, and as a result, affected cell traction on extracellular matrix. We further examined its interaction with Cdc42. CEP1 co-localizes with Cdc42 in membrane protrusions, but also localizes to perinuclear patches where Cdc42 is not present. Through binding to CRIB domain of CEP1, Cdc42 influences the cellular distribution of CEP1 between the two locations and affects cell behaviors. Our findings suggest that CEP1 plays both Cdc42 dependent and independent roles in regulating neural crest cell migration, and Cdc42 level controls the balance between the two activities. Results CEP1 is predominantly expressed in neural crest cells during Xenopus embryogenesis Cdc42 is critical for cellular behaviors and is involved in multiple cell and tissue movements. Consistent with its importance in multiple developmental processes, Cdc42 is broadly expressed during embryogenesis (Choi and Han, 2002; Lucas et al., 2002). To determine how CEP1 may mediate the activities of Cdc42, we first examined the expression of CEP1 during early development of Xenopus embryos. In situ hybridization analysis was performed and the results were summarized in Figure 1. During gastrulation, CEP1 is weakly expressed in the animal pole in presumptive ectoderm (data not shown). At the end of gastrulation, CEP1 is expressed in mesodermal cells around the blastopore, and in lateral borders of the neural plate, where neural crest cells are specified. CEP1 expression persists in neural crest cells as they separate from neural epithelium, segregate into three lobes, and commence migration at neurula stages (arrows and arrowhead). Neural crest cells continue to express CEP1 as they migrate into the branchial arches in tailbud stages (open arrowheads). From mid-tailbud stages, CEP1 transcript is also detected in anterior somites. The abundance of CEP1 transcripts in premigratory and migrating neural crest cells suggests that CEP1 maybe involved in regulating neural crest development. Figure 1 View largeDownload slide CEP1 is expressed in neural crest cells during frog embryogenesis. At gastrula stage (st.12.5 and st.14), CEP1 is expressed at the lateral borders of the neural plate, where neural crest cells are specified (arrows). Its expression is restricted to CNC at neurula stages (st.18) and continues in CNC cells as they commence migration (st.20; arrowhead) and migrate into branchial arches (open arrowheads). At tailbud stages (st.24 and st.26), CEP1 expression is also detected in somites. All embryos are oriented with anterior to the left. Dorsal views are shown in the upper panels and lateral views are shown in the lower panels. Figure 1 View largeDownload slide CEP1 is expressed in neural crest cells during frog embryogenesis. At gastrula stage (st.12.5 and st.14), CEP1 is expressed at the lateral borders of the neural plate, where neural crest cells are specified (arrows). Its expression is restricted to CNC at neurula stages (st.18) and continues in CNC cells as they commence migration (st.20; arrowhead) and migrate into branchial arches (open arrowheads). At tailbud stages (st.24 and st.26), CEP1 expression is also detected in somites. All embryos are oriented with anterior to the left. Dorsal views are shown in the upper panels and lateral views are shown in the lower panels. CEP1 is required for CNC migration, but not for neural crest specification Since CEP1 is expressed in both premigratory and migrating neural crest cells, CEP1 may potentially regulate both neural crest specification and migration. To determine which process CEP1 is involved, we next performed loss-of-function experiments using a translation blocking morpholino oligomer (MO) against the Xenopus laevis CEP1 gene. To validate the efficiency of CEP1-MO, CEP1(UTR)-EGFP containing the MO recognition site was overexpressed into 2-cell stage embryos with or without CEP1-MO. In situ hybridization analysis against CEP1 RNA was performed and showed that all embryos received ectopic CEP1 RNA. While CEP1(UTR)-EGFP expressing embryos showed strong EGFP signal, those co-injected with CEP1-MO showed much reduced or no EGFP signal, suggesting that the translation of the protein was blocked (Figure 2C). CEP1-EGFP without the 5′UTR can not be blocked by CEP1-MO, and is used in rescue experiments. Figure 2 View largeDownload slide CEP1 is required for CNC cell migration. (A) Embryos were injected with control-MO, CEP1-MO (10 ng), or CEP1-MO with CEP1 RNA (0.1 ng) into one of the dorsal animal cells at 8-cell stage, together with a lineage tracer (nβGal, red staining, marked by *). In situ hybridization analysis was performed at neurula stage (st.14) and tailbud stage (st.23). Arrows point to reduced neural crest migration indicated by Sox10/Twist expression. Embryos are oriented with anterior to the left. Dorsal views at neurula stage and lateral views of both uninjected and injected sides of the same embryo at tailbud stage are shown. (B) Percentage of embryos with defective neural crest migration. While CEP1-MO lead to defective migration in >90% of the embryos, adding CEP1 RNA rescued the rate of defective embryos to 45%. Fisher’s exact test was performed between control and CEP1-MO conditions and P < 0.05. χ2-test was performed between CEP1-MO and rescue conditions and P < 0.05. (C) CEP1-MO efficiently blocked the translation of CEP1. CEP1-EGFP containing MO recognition sequence in the 5′UTR was overexpressed in 2-cell stage embryos. While CEP1 RNA was expressed in all embryos as reflected by in situ hybridization, addition of CEP1-MO blocked the protein translation as reflected by the fluorescence signal of EGFP tag. CEP1-EGFP without 5′UTR can not be blocked by the morpholino. Figure 2 View largeDownload slide CEP1 is required for CNC cell migration. (A) Embryos were injected with control-MO, CEP1-MO (10 ng), or CEP1-MO with CEP1 RNA (0.1 ng) into one of the dorsal animal cells at 8-cell stage, together with a lineage tracer (nβGal, red staining, marked by *). In situ hybridization analysis was performed at neurula stage (st.14) and tailbud stage (st.23). Arrows point to reduced neural crest migration indicated by Sox10/Twist expression. Embryos are oriented with anterior to the left. Dorsal views at neurula stage and lateral views of both uninjected and injected sides of the same embryo at tailbud stage are shown. (B) Percentage of embryos with defective neural crest migration. While CEP1-MO lead to defective migration in >90% of the embryos, adding CEP1 RNA rescued the rate of defective embryos to 45%. Fisher’s exact test was performed between control and CEP1-MO conditions and P < 0.05. χ2-test was performed between CEP1-MO and rescue conditions and P < 0.05. (C) CEP1-MO efficiently blocked the translation of CEP1. CEP1-EGFP containing MO recognition sequence in the 5′UTR was overexpressed in 2-cell stage embryos. While CEP1 RNA was expressed in all embryos as reflected by in situ hybridization, addition of CEP1-MO blocked the protein translation as reflected by the fluorescence signal of EGFP tag. CEP1-EGFP without 5′UTR can not be blocked by the morpholino. To examine the loss of function effect of CEP1, CEP1-MO together with a lineage tracer (nβGal) was injected into one of the dorsal animal blastomeres of 8-cell stage embryos, leaving the uninjected contralateral side as an internal control. The embryos were collected at either late gastrula stage (stage 14) when neural crest cells are specified, or early tailbud stage (stage 23) when CNC cells have migrated into branchial arches, and neural crest cells in these embryos were marked by in situ hybridization against neural crest-specific marker genes Sox10 and Twist. At stage 14, Sox10 expression was observed in a bilaterally symmetric pattern in both control and CEP1-MO injected embryos (Figure 2A), suggesting that CEP1 is not required for neural crest specification. At stage 23, Sox10 and Twist expressing CNC cells migrated into branchial arches on both sides of the control embryo, with a clear segregation of the three migratory lobes (the third lobe will split to two at late tailbud stages). In contrast, neural crest gene expression on CEP1-MO expressing side was impaired (90.5%, n = 63), indicating a defect in neural crest migration. CEP1-MO receiving neural crest cells migrated shorter distances comparing to cells on the contralateral side, and failed to segregate into distinct lobes. When CEP1 RNA was co-expressed with CEP1-MO, the expression pattern of Sox10 and Twist was rescued (55% rescued to normal, n = 40), confirming that the migration defects were specific to the loss of CEP1. These results suggest that CEP1 is required for neural crest migration without affecting neural crest specification. To determine whether CEP1 is required cell autonomously by neural crest cells for their migration, a grafting experiment was performed. CNC explants from EGFP-labeled donor embryos were transplanted isotopically and isochronically into wild-type host embryos where the endogenous neural crest tissue was removed (Figure 3). When grafted control CNC cells migrated efficiently to the branchial arches and segregated into three to four migratory streams, CEP1-MO expressing grafts migrated much shorter distances ventrally and failed to form distinct migratory streams. These cells seemed to lose their interaction during this collective migration and appeared rounder in shape. When CEP1 was co-expressed, the migration of CNC grafts was significantly rescued. Figure 3 View largeDownload slide CEP1 is required cell-autonomously for neural crest migration. (A) GFP-labeled CNC explants were dissected and transplanted into unlabeled host embryos and their migration in the host embryos was imaged at late tailbud stages. Fluorescence and DIC merged images and fluorescence images alone are shown side by side. Defects in CEP1-MO expressing cells to migrate ventrally and to segregate into distinct streams (marked by numbers) were rescued by co-expressing CEP1 RNA. (B) The relative distance of lateral migration to the entire D-V length was calculated, and the number of distinct migratory stream was counted for each CNC graft and summarized in the bar graph. Both the migration distance and the number of segregated streams are significantly reduced by CEP1 knockdown, while adding CEP1 RNA significantly rescued both defects (student’s t-test, P < 0.01). Figure 3 View largeDownload slide CEP1 is required cell-autonomously for neural crest migration. (A) GFP-labeled CNC explants were dissected and transplanted into unlabeled host embryos and their migration in the host embryos was imaged at late tailbud stages. Fluorescence and DIC merged images and fluorescence images alone are shown side by side. Defects in CEP1-MO expressing cells to migrate ventrally and to segregate into distinct streams (marked by numbers) were rescued by co-expressing CEP1 RNA. (B) The relative distance of lateral migration to the entire D-V length was calculated, and the number of distinct migratory stream was counted for each CNC graft and summarized in the bar graph. Both the migration distance and the number of segregated streams are significantly reduced by CEP1 knockdown, while adding CEP1 RNA significantly rescued both defects (student’s t-test, P < 0.01). To quantitate these results, the relative distance of neural crest migration and the number of migratory streams were calculated in each grafted embryo and summarized in Figure 3B. While control neural crest grafts on average migrated 81% of the D-V axis and segregated into 3.2 lobes, CEP1-MO expressing grafts only migrated 46.5% along the D-V axis and segregated into 1.3 lobes. Coinjection of CEP1 RNA efficiently rescued the migration defects such that neural crest cells on average traveled 74.1% of the D-V length and segregated into 2.5 lobes, again confirming that the migration defects observed were specific to the loss of CEP1. CEP1 is required for cranial cartilage formation Since craniofacial cartilage and bones are important derivatives of CNC cells, we next asked whether the migration defects caused by CEP1-MO had later consequences on cartilage formation. Alcian blue staining was performed with late stage tadpoles (~stage 45) to examine the formation of cartilage elements. The results showed that neural crest derived mandibular, hyoid, and branchial arch cartilages were malformed on CEP1-MO expressing side (Figure 4). They were much smaller in size and sometimes completely missing. Thus, consistent with the notion that neural crest cells need to migrate to proper destinations in order to differentiate into correct structures, defects in CNC migration result in a general failure in facial cartilage formation. Figure 4 View largeDownload slide CEP1-MO inhibits cranial cartilage formation. Cranial cartilages from control- or CEP1-MO-injected tadpoles were stained with Alcian blue. Ventral views are shown with injected sides marked by *. Cartilage on CEP1-MO-injected side was malformed or even completely missing. M, Meckel’s cartilage formed by mandibular stream; CH, ceratohyal cartilage formed by hyoid stream; CB, ceratobranchial cartilage formed by third and fourth branchial arch streams. Figure 4 View largeDownload slide CEP1-MO inhibits cranial cartilage formation. Cranial cartilages from control- or CEP1-MO-injected tadpoles were stained with Alcian blue. Ventral views are shown with injected sides marked by *. Cartilage on CEP1-MO-injected side was malformed or even completely missing. M, Meckel’s cartilage formed by mandibular stream; CH, ceratohyal cartilage formed by hyoid stream; CB, ceratobranchial cartilage formed by third and fourth branchial arch streams. CEP1 regulates the dynamic rearrangements of actin filaments in neural crest cells To determine how CEP1 regulates neural crest cell migration, a CNC explant assay, which allows for closer examination of neural crest cell behaviors, was performed. CNC explants were dissected from control or CEP1-MO injected embryos at stages 13–14, plated on fibronectin-coated coverslips, and the cell shape changes were delineated by co-expressed membrane-tethered EGFP. Control neural crest cells displayed polarized shapes and formed multiple membrane protrusions. The protrusions extended or retracted dynamically as the cell changed its direction of migration. In contrast, CEP1-MO expressing cells were much rounder in shape, and formed fewer protrusions (Supplementary Figures S1 and S2). Since actin machinery is the driving force for cell shape changes and cell locomotion, we next co-expressed EGFP-Utrophin (EGFP fusion with the actin-binding domain of Utrophin) to label actin filaments in neural crest cells. As shown in Figure 5A, actin filaments were assembled and disassembled rapidly in control neural crest cells. Actin filaments were enriched in areas where protrusions were extending, and diminished where protrusions were retracting (follow arrows or arrowheads in different time frames). When fluorescence intensities of EGFP-labeled actin filaments along the long axis of a cell were plotted at different time points, we observed dynamic changes in actin filaments throughout the cell as well as active migration of the cell (Figure 5E). Time-lapse movies show that actin filaments translocate rapidly in different areas of the cells and their fluorescence intensity changes in a wave-like manner (Supplementary Movie S1). In sharp contrast, when CEP1-MO was expressed, actin filaments formed ring-like structures around the cell cortex and displayed restrained movement (Figure 5B and Supplementary Movie S2). The intensity of actin filaments shifted between the rings in a ripple-like manner (white open arrowheads). Rather than forming lamellipodia or filopodia protrusions, CEP1-MO expressing cells formed membrane blebs (Figure 5B and C; double peaks in Figure 5E), indicating a weaker association between the actin cortex and plasma membrane. Actin intensity plots from another 15 control and 15 CEP1-MO expressing cells were analyzed and the results confirmed that loss of CEP1 dramatically restricted the localization of actin filaments (Figure 5D). Rather than distributing along the entire axis of the cell, CEP1-MO expressing cells show significantly lower actin intensity at the cell center. The defects in actin organization and lamellipodia formation in these cells were also consistent with a decrease in cell translocation (purple-shaded cell in Figure 5A and B). When CEP1 was added back to the cells, the cell shapes and actin dynamics were significantly rescued (Figure 5C and Supplementary Movie S3). Figure 5 View largeDownload slide CEP1-MO disrupts the organization of actin filaments. CNC explants receiving EGFP-Utrophin with or without CEP1-MO (5 ng) were dissected and plated on FN-coated cover slides and the dynamics of actin filaments during cell migration were recorded by time-lapse microscopy. (A) Image frames of control neural crest explants over 5.5 min are shown. White or yellow arrows and arrowheads follow the same protrusion over time. (B) Image frames of CEP1-MO expressing cells. White open arrowheads point to the same actin bundle over time and yellow open arrowheads point to the membrane blebs. One cell in both control and CEP1-MO movie frames is shaded in purple to show cell shape changes and translocation over time. A’–A” and B’–B” are higher magnification views of corresponding boxed area. Images were taken at 40×. Scale bar, 20 μm. (C) The number of each type of protrusions formed was counted and averaged in the bar graph. While control neural crest cells formed 1.5 lamellipodia and 3.8 filopodia on average, blocking CEP1 reduced the formation of both protrusions to 0.3 and 0.8 per cell, respectively, and increased the formation of membrane blebs to 1.3 per cell. Co-expressing CEP1 rescued the formation of cell protrusions and restored the numbers to 1.0, 2.5, and 0.3, respectively. Student’s t-test was performed and both MO inhibition and rescue were significant (P < 0.01). (D) Intensity plots for EGFP-labeled actin filaments were generated for 15 control and 15 CEP1-MO expressing cells. Solid lines reflect the mean signal intensity at locations relative to the length of the cell, and dashed lines above and below mark the standard deviation at each location. There is no significant difference in actin intensity at the cell periphery, but significant difference at cell center. (E) Actin fluorescence intensities across one control and one CEP1-MO cell were plotted over 4 min. In control graph, dashed lines mark the positions of cell protrusions at the beginning and the end of the experiment, reflecting leftward migration of the cell. In CEP1-MO graph, blebs reflected by the double peaks are marked by arrows. The cell is not migrating, but shrinking towards the end. Figure 5 View largeDownload slide CEP1-MO disrupts the organization of actin filaments. CNC explants receiving EGFP-Utrophin with or without CEP1-MO (5 ng) were dissected and plated on FN-coated cover slides and the dynamics of actin filaments during cell migration were recorded by time-lapse microscopy. (A) Image frames of control neural crest explants over 5.5 min are shown. White or yellow arrows and arrowheads follow the same protrusion over time. (B) Image frames of CEP1-MO expressing cells. White open arrowheads point to the same actin bundle over time and yellow open arrowheads point to the membrane blebs. One cell in both control and CEP1-MO movie frames is shaded in purple to show cell shape changes and translocation over time. A’–A” and B’–B” are higher magnification views of corresponding boxed area. Images were taken at 40×. Scale bar, 20 μm. (C) The number of each type of protrusions formed was counted and averaged in the bar graph. While control neural crest cells formed 1.5 lamellipodia and 3.8 filopodia on average, blocking CEP1 reduced the formation of both protrusions to 0.3 and 0.8 per cell, respectively, and increased the formation of membrane blebs to 1.3 per cell. Co-expressing CEP1 rescued the formation of cell protrusions and restored the numbers to 1.0, 2.5, and 0.3, respectively. Student’s t-test was performed and both MO inhibition and rescue were significant (P < 0.01). (D) Intensity plots for EGFP-labeled actin filaments were generated for 15 control and 15 CEP1-MO expressing cells. Solid lines reflect the mean signal intensity at locations relative to the length of the cell, and dashed lines above and below mark the standard deviation at each location. There is no significant difference in actin intensity at the cell periphery, but significant difference at cell center. (E) Actin fluorescence intensities across one control and one CEP1-MO cell were plotted over 4 min. In control graph, dashed lines mark the positions of cell protrusions at the beginning and the end of the experiment, reflecting leftward migration of the cell. In CEP1-MO graph, blebs reflected by the double peaks are marked by arrows. The cell is not migrating, but shrinking towards the end. CEP1 is required for proper force transmission between cell and extracellular matrix When CEP1 was knocked down in neural crest cells, actin organization, lamellipodia formation, and cell translocation were compromised. To understand how CEP1 mediates such effects, we next looked at the mechanical basis for cell migration. During cell migration, the actin filaments polymerize at the front of the cell to push the cell forward, while relative movements between myosin II and actin filaments contract and pull the bulk of the cell towards the front. Both movements rely on the dynamic assembly and disassembly of focal adhesions (Nobes and Hall, 1995). Focal adhesions are large protein complexes that anchor a cell (and actin filaments) to the extracellular matrix. At the cell's leading edge, focal complexes stabilize the forming membrane protrusions; in the center, focal adhesions impede the retrograde movement of actin; and at cell rear, focal adhesions disassemble and proteins in the complex are recycled to the front. Here, we examined the formation of focal adhesions using antibodies against phosphorylated tyrosine (Zamir et al., 1999). In control cells, long focal adhesions were assembled in well-aligned arrays in the leading edge of cells (arrows and insets in Supplementary Figure S3). There were not many focal adhesions located in the center or trailing edge of the cells, consistent with their disassembly in these areas. Frog neural crest cells imbricate on each other in a scale-like manner, and some focal adhesions observed in the center of cell clusters were actually in lamellipodia underneath the neighboring cell. In contrast, focal adhesions formed in CEP1-MO cells were generally shorter in length. They sometimes appeared in the center of an isolated cell, or in disorganized arrays in the cell periphery (arrows and insets in Supplementary Figure S3). Since focal adhesions align with cell contractility, differences in focal adhesion arrangement and localization may correspond to a disoriented force transmission, leading to decreased cell translocation. To directly elucidate the propulsive forces mediated through focal adhesions, we performed traction force microscopy. Control or CEP1-MO expressing CNC explants were plated onto fibronectin-coated hydrogels. The traction force the cells applied on the hydrogel was reflected by the strain energy absorbed by the hydrogel, which can be calculated by measuring the displacements of fluorescent beads embedded in the surface layer of the hydrogel (Yeung et al., 2005; Kovari et al., 2016). As shown in Figure 6 and Supplementary Movies S4 and S5, we observed three types of traction force distribution around the cells, corresponding to three different types of cell behaviors. First were well-spread cells with large and dynamic protrusions, and there was high strain energy absorbed by the gel underneath these protrusions (Figure 6A). In addition, the orientation of the strain was always towards the cell body (or opposite to the direction the protrusion extended). The second were rounder cells that formed membrane blebs. They usually did not exert much stress underneath the blebs, which is predictable since the blebs do not adhere to the matrix. Instead, there was high strain energy underneath the center of the cell, possibly corresponding to the focal adhesions there (Figure 6B). However, the orientation of the strain was random, reflecting a defect in directionality. The third were cells displaying transitioning behaviors, and the spatial distribution of traction forces was also shifting between the cell peripheries and the cell center. Over 200 control and CEP1-MO expressing cells were categorized into these three types of cell behaviors and the results are shown in Figure 6C. While 69% of control cells displayed spreading behavior, only 22% of CEP1-MO expressing cells spread on hydrogel. Instead, 63% of CEP1-MO cells displayed membrane blebs. In addition to the differences in their spatial distribution, the magnitude of the strain energy was also reduced under blebbing cells. The average strain energy produced by each cell over the course of the movie was calculated (Figure 6D). While the average strain energies were similar under cells with the same phenotype, there was a significant increase in the strain energy under spreading cells. Figure 6 View largeDownload slide CEP-MO affected force transmission during neural crest cell migration. (A and B) Traction force microscopy was performed and the strain energy received by extracellular matrix was calculated and plotted on the image. In control cells, high strain energy was observed underneath lamellopodia protrusions (arrows). In CEP1-MO cells, high strain energy was observed underneath the cell center rather than underneath membrane blebs (arrows). (C) Over 200 control and CEP1-MO cells were categorized into three behavioral types, i.e. spreading, intermediate, and blebbing. χ2-test confirmed that CEP1-MO decreased the percentage of spreading cells and increased the percentage of blebbing cells significantly (P < 0.05). (D) Average strain energy underneath cells was calculated over the course of the movie and compared between different behavioral types using Fisher’s least significant difference test. While the hydrogel beneath control or CEP1-MO cells of the same behavior absorbed similar amount of strain energy, there were significant differences in strain energy beneath cells with different behaviors. Figure 6 View largeDownload slide CEP-MO affected force transmission during neural crest cell migration. (A and B) Traction force microscopy was performed and the strain energy received by extracellular matrix was calculated and plotted on the image. In control cells, high strain energy was observed underneath lamellopodia protrusions (arrows). In CEP1-MO cells, high strain energy was observed underneath the cell center rather than underneath membrane blebs (arrows). (C) Over 200 control and CEP1-MO cells were categorized into three behavioral types, i.e. spreading, intermediate, and blebbing. χ2-test confirmed that CEP1-MO decreased the percentage of spreading cells and increased the percentage of blebbing cells significantly (P < 0.05). (D) Average strain energy underneath cells was calculated over the course of the movie and compared between different behavioral types using Fisher’s least significant difference test. While the hydrogel beneath control or CEP1-MO cells of the same behavior absorbed similar amount of strain energy, there were significant differences in strain energy beneath cells with different behaviors. CEP1 is localized to two distinct locations during neural crest cell migration To better understand how CEP1 mediates its activities in actin organization and cell migration, we next examined the subcellular localization of CEP1 in migrating neural crest cells. EGFP-CEP1 and RFP-Utrophin were co-expressed in neural crest cells and their relative distribution during neural crest cell migration was recorded by time-lapse microscopy. As shown in Figure 7, CEP1 is distributed rather broadly in neural crest cells. One fraction of CEP1 is localized to the periphery of the cell. It was detected in plasma membrane, at the base of filopodia protrusions, and in a punctate manner along actin filaments in lamellipodia protrusions (see enlarged insets A’–A”, B’–B”, and C’). In addition, another fraction of CEP1 is localized to the center of the cell, close to the nucleus. They were organized in a fragmented or filamentous manner in the cytoplasm and concentrated to a dense patch near the nucleus (arrows). In mice endothelial cells, CEP1 (Borg5) has been described to associate with septins in the perinuclear region and promote the persistent directional migration of endothelial cells (Liu et al., 2014). In neural crest cells, CEP1 may play a similar role in maintaining the direction of cell migration at the perinuclear location. Figure 7 View largeDownload slide CEP1 localization in migrating neural crest cells. EGFP-CEP1 (0.05 ng) and RFP-Utrophin were co-expressed in neural crest cells. (A and B) CEP1 located to membrane protrusions and cell membrane, as well as in the perinuclear region in filamentous or condensed patch (arrows). A’–A” and B’–B” are enlarged insets of the boxed areas in A and B, showing co-localization of CEP1 with actin along lamellipodia protrusion, cell membrane, and at the base of filopodia protrusion. (C) For better visualization of CEP1 localization, one isolated neural crest cell was shown with GFP channel alone. CEP1 is localized to cell membrane (arrowheads) and in a punctate manner in protrusions (C’). Images were taken at 63×. Scale bar, 20 μm. Figure 7 View largeDownload slide CEP1 localization in migrating neural crest cells. EGFP-CEP1 (0.05 ng) and RFP-Utrophin were co-expressed in neural crest cells. (A and B) CEP1 located to membrane protrusions and cell membrane, as well as in the perinuclear region in filamentous or condensed patch (arrows). A’–A” and B’–B” are enlarged insets of the boxed areas in A and B, showing co-localization of CEP1 with actin along lamellipodia protrusion, cell membrane, and at the base of filopodia protrusion. (C) For better visualization of CEP1 localization, one isolated neural crest cell was shown with GFP channel alone. CEP1 is localized to cell membrane (arrowheads) and in a punctate manner in protrusions (C’). Images were taken at 63×. Scale bar, 20 μm. CEP1 interacts with Cdc42 and regulates the cellular localization of each other during neural crest cell migration To understand how the activities of CEP1 at both locations are regulated, we next examined its interaction with Cdc42. Biochemical analysis in cell culture indicates that CEP1 can physically bind to Cdc42 with high affinity through its CRIB domain (Burbelo et al., 1999; Joberty et al., 1999). However, where and how CEP1 interacts with Cdc42 during cell migration remains unknown. Since cell migration is a highly dynamic process, we expect that the interaction between CEP1 and Cdc42 to be transient and dynamic as well. To determine where they interact during cell migration, we expressed fluorescence fusion proteins for Cdc42 and CEP1 in neural crest cells and directly observed their localization during cell migration. As shown in Figure 8A and Supplementary Movie S6, CEP1 and Cdc42 overlapped in membrane protrusions (arrows), but not in perinuclear region (arrowheads). Such localization pattern suggests that CEP1 may play both Cdc42 dependent and Cdc42 independent activities in neural crest cells. Figure 8 View largeDownload slide CEP1 and Cdc42 interact with each other during neural crest migration. (A) Co-localization of CEP1 and Cdc42 (0.05 ng each). Cdc42 co-localizes with CEP1 in membrane protrusions (arrows and inset), but not in perinuclear region (arrowheads). Images were taken at 60×. Scale bar, 20 μm. (B) The cellular localization of Cdc42 or CEP1 at increased (0.2 ng of RNA) or decreased (5 ng of MO) level of CEP1 or Cdc42. Arrowheads in EGFP-Cdc42 panels point to high level of Cdc42 in protrusions and cell membrane. In Cherry-CEP1 panels, arrows point to CEP1 in perinuclear region, while closed and open arrowheads point to cell protrusions and plasma membrane, respectively. Enlarged insets show one cell (in dashed box) under each condition. Images were taken at 63×. Scale bar, 20 μm. (C) Numbers of membrane protrusions in control, CEP1 overexpression, and CEP1-MO cells were counted and summarized in the bar graph. Both increase and decrease of membrane protrusions are significant (student’s t-test, P < 0.01). (D) The cellular distribution of CEP1 under control, Cdc42 overexpression, and Cdc42 knockdown conditions were compared. The intensities of Cherry-CEP1 fluorescence signals in 16–17 cells were analyzed along a line across each cell (see examples in B), and the data were normalized against the length of the cell. Solid lines are average signal intensity under each condition and dashed lines are average ± standard deviation. Gray double-headed arrows point to regions of significant differences. (E) CEP1 lacking the CRIB domain can not interact with Cdc42. EGFP-CEP1(ΔCRIB) is localized to the nucleus, and changes in Cdc42 level do not alter its localization. In contrast, EGFP-CEP1 is mainly localized to membrane protrusions when Cdc42 is co-expressed. Scale bar, 20 μm. Figure 8 View largeDownload slide CEP1 and Cdc42 interact with each other during neural crest migration. (A) Co-localization of CEP1 and Cdc42 (0.05 ng each). Cdc42 co-localizes with CEP1 in membrane protrusions (arrows and inset), but not in perinuclear region (arrowheads). Images were taken at 60×. Scale bar, 20 μm. (B) The cellular localization of Cdc42 or CEP1 at increased (0.2 ng of RNA) or decreased (5 ng of MO) level of CEP1 or Cdc42. Arrowheads in EGFP-Cdc42 panels point to high level of Cdc42 in protrusions and cell membrane. In Cherry-CEP1 panels, arrows point to CEP1 in perinuclear region, while closed and open arrowheads point to cell protrusions and plasma membrane, respectively. Enlarged insets show one cell (in dashed box) under each condition. Images were taken at 63×. Scale bar, 20 μm. (C) Numbers of membrane protrusions in control, CEP1 overexpression, and CEP1-MO cells were counted and summarized in the bar graph. Both increase and decrease of membrane protrusions are significant (student’s t-test, P < 0.01). (D) The cellular distribution of CEP1 under control, Cdc42 overexpression, and Cdc42 knockdown conditions were compared. The intensities of Cherry-CEP1 fluorescence signals in 16–17 cells were analyzed along a line across each cell (see examples in B), and the data were normalized against the length of the cell. Solid lines are average signal intensity under each condition and dashed lines are average ± standard deviation. Gray double-headed arrows point to regions of significant differences. (E) CEP1 lacking the CRIB domain can not interact with Cdc42. EGFP-CEP1(ΔCRIB) is localized to the nucleus, and changes in Cdc42 level do not alter its localization. In contrast, EGFP-CEP1 is mainly localized to membrane protrusions when Cdc42 is co-expressed. Scale bar, 20 μm. To determine the functional interactions between CEP1 and Cdc42, the level of one protein was manipulated and its impact on the intracellular distribution of the other protein was examined (Figure 8B and Supplementary Movies S7–S12). When CEP1 was downregulated by morpholino, neural crest cells adopted a rounder morphology without forming obvious membrane protrusions. When CEP1 was overexpressed, cells looked relatively normal, despite that they made protrusions at different orientations. Figure 8C summarizes the numbers of membrane protrusions formed by cells expressing different levels of CEP1 and both the increase and the decrease of CEP1 level significantly changed the number of protrusions. The changes in Cdc42 localization were not very obvious. In control cells, Cdc42 was localized in cell membrane and membrane protrusions (arrowheads). When CEP1 overexpression led to more protrusions in each cell, Cdc42 was also enriched in different protrusions. In contrast, when CEP1 was knocked down, Cdc42 was distributed more evenly throughout the cell membrane. Since Cdc42 needs to be in its GTP-bound active state to mediate downstream effects, we next examined whether the distribution of activated Cdc42 was influenced by the level of CEP1. To this end, we expressed a reporter GFP-wGBD (Cdc42-binding domain of N-WASP) that only binds to active Cdc42 in neural crest cells (Benink and Bement, 2005). In control and CEP1-MO cells, we observed no change in the overall level of active Cdc42, consistent with the notion that CEP1 does not have kinase activity to directly activate Cdc42. In control cells, GFP-wGBD was enriched in the leading edge, and was constantly moving between different areas of the membrane protrusions (Supplementary Movie S13). In CEP1-MO expressing cells, despite changes in cell morphology, GFP-wGBD was expressed around the cell periphery, especially at areas membrane blebs were forming (Supplementary Movie S14). This demonstrates that CEP1 influences the localization of active Cdc42 to regulate cell polarity and the formation of membrane protrusions. Conversely, when Cdc42 was overexpressed, a large portion of CEP1 was recruited to Cdc42 in cell membrane and in membrane protrusions (open arrowheads and arrowheads in Figure 8B), regardless of the orientation of the membrane. When Cdc42 was downregulated by morpholino, CEP1 was largely relocated near the nucleus (arrows), leaving little CEP1 at cell membrane. Signal intensities of Cherry-CEP1 across 16–17 cells under each condition were analyzed and plotted in Figure 8D. While CEP1 alone was localized to cell periphery as well as cell center (blue curve), Cdc42 concentrated them to cell periphery (green curve). When the Cdc42 level was reduced, CEP1 redistributed to the cell center (red curve). Gray arrows indicate that there are significant differences in CEP1 levels at cell peripheries between control and Cdc42-MO expressing cells, and at the cell centers between control and Cdc42 overexpressing cells. Given that Cdc42 is only expressed near the cell membrane, these results suggest that through physical binding, Cdc42 controls the balance of CEP1 between the two subcellular locations. To confirm that Cdc42 directly regulates CEP1 localization, we deleted the CRIB domain from CEP1, and fused the mutant construct with EGFP (EGFP-CEP1(ΔCRIB)). When expressed into neural crest cells, CEP1(ΔCRIB) is localized to the nucleus, regardless of the level of Cdc42 in the cells (Figure 8E). This is in sharp contrast to wild-type CEP1, which is largely localized to membrane protrusions when Cdc42 is overexpressed. These results indicate that CRIB domain is critical in mediating CEP1−Cdc42 interaction, and Cdc42 recruits CEP1 to membrane protrusions through direct binding. Discussion Cdc42eps co-evolved with neural crest, somite, cardiac microvasculature, and other vertebrate specific tissue or structures, thus possibly play important roles in the development of these tissue and structures. Here, we show that Cdc42ep1 (CEP1) is highly expressed in neural crest cells and plays important roles in their migration. Our study further demonstrates that CEP1 is localized to two subcellular areas, and Cdc42 directly controls the balance between these two CEP1 subpopulations. We showed that CEP1 is required for neural crest cell migration in vivo and in vitro. When CEP1 was knocked down, neural crest cells did not form lamellipodia protrusions, but instead made membrane blebs. This could result from several mechanisms. First, CEP1 may be required for the formation of lamellipodia or filopodia protrusions. It has been reported in fibroblasts that overexpression of CEP2 can induce ectopic membrane protrusions and this requires the presence of active Cdc42 as well as intact CRIB domain (Hirsch et al., 2001), demonstrating that CEP2 directly binds to and cooperates with Cdc42 in promoting protrusion formation. It is possible that CEP1 plays a similar role in interacting with Cdc42 and promoting membrane protrusions. This is supported by the result that changes in CEP1 level significantly changed the number of membrane protrusions made, especially filopodia protrusions (Figures 5C and 8C). Second, CEP1 may negatively regulate bleb formation. Membrane blebs, which grow as a result of intracellular pressure, are generated by actomyosin contractions (Charras and Paluch, 2008). They can be induced by increasing actomyosin contractility through activated RhoA/ROCK, or by increasing cortical tension by thickening actin cortex (Han et al., 2009; Bergert et al., 2012). In our experiments, when CEP1 was reduced, higher levels of pMLC and thick actin filament bundles were observed along cell periphery (Figure 5B and Supplementary Figure S3). Whether cortical tension was increased or not remains to be elucidated. In addition to increased cortical tension, blebs also associate with lower cellular adhesion. We have observed changes in the organization of focal adhesions and the distribution and magnitude of traction forces in CEP1-MO induced blebbing cells (Figure 6 and Supplementary Figure S3). Since cell adhesion can influence actin polymerization while actomyosin contraction generates tension, which can control the dynamics of focal adhesions, defect in one process may be amplified and reinforced through such feedback regulations. In our experiments, forming blebs instead of lamellipodia compromised the migration of neural crest cells. In contrast, zebrafish germ cells use membrane blebs in their migration and several cancer cells have been observed to made bleb-like protrusions when migrate in 3D matrix (Kardash et al., 2010; Lorentzen et al., 2011). It has been suggested that the dynamic shifts between actin protrusivity and actomyosin contractility, and the resulting transition between lamellipodia and blebs may be advantageous for cells to quickly adapt to local environment (Bergert et al., 2012). Since neural crest cells made different types of protrusions at different CEP1 levels, CEP1 may be involved in the mechanisms controlling the switch between the two types of cell protrusions. We observed that some CEP1 located near the nucleus. This is very similar to the observation of Liu et al. (2014) in mouse cardiac endothelial cells that CEP1 appeared as a prominent filamentous patch above the nucleus. They showed that CEP1 promotes the assembly and alignment of septin filaments and actin filaments and together they regulate actomyosin organization and persistent cell migration. Parallel actomyosin bundles have been reported previously to align with the direction of cell migration and are required for maintaining the directionality of migration (Lo et al., 2004). The actomyosin in the center of the cell may play a role in coupling the propulsive anterior and the resistive posterior to ensure that the entire cell undergoes coordinated migration (Guo and Wang, 2012). Septin has recently been considered the fourth component of the cytoskeleton and can bind to both actin filament and myosin II to promote the assembly of actomyosin (Joo et al., 2007; Mostowy and Cossart, 2012). CEPs can also bind to septin directly and stimulate the assembly of septin filaments (Joberty et al., 2001; Calvo et al., 2015; Farrugia and Calvo, 2016). It is likely that CEP1 in neural crest cells also interact with septin filaments to control the stability and the orientation of actomyosin, thus regulating the direction of cell migration. CEPs were identified through yeast two hybrid screen for binders of Rho GTPases, and they were confirmed to physically bind to Cdc42, but not to Rac1 or RhoA (Joberty et al., 1999). However, it has been unclear how Cdc42 and CEPs interact functionally. In fibroblast cells, expression of dominant negative Cdc42 abolished pseudopodia induced by CEPs, while in keratinocytes, expression of constitutive active Cdc42 or CEP2/CEP5 all stimulate stress fiber formation (Hirsch et al., 2001). However, it is also reported in fibroblast cells that CEP5 may compete with other Cdc42 effectors and results in inhibiting Cdc42 activity in cell spreading (Joberty et al., 1999). On the other hand, in cancer-associated fibroblasts and MDCK epithelial cells, Cdc42 inhibits the interaction between CEP3/5 and septin (Joberty et al., 2001; Calvo et al., 2015). Here, we showed that CEP1 has two subcellular localizations and Cdc42, rather than inhibiting CEP1 activity, physically attracts CEP1 away from perinuclear region to cell protrusions. Our results suggest that the Cdc42 level is critical for a proper balance of CEP1 at these locations to control cell morphology and behavior. Similarly, a recent report showed that Cdc42 activity regulates the correct positioning of CEP3 in cancer-associated fibroblasts (Farrugia and Calvo, 2017). Therefore, Cdc42 may play a common role in regulating different CEPs. In contrast to wild-type CEP1 that localized to membrane protrusions and perinuclear region, CEP1(ΔCRIB) is mainly localized to the nucleus in neural crest cells. Two questions remain unresolved. First is whether CEP1 has endogenous nucleus distribution. Many effector proteins for Rho GTPases (e.g. Arp2/3, WAVE1, cofilin, and formin) can act inside the nucleus and regulate the dynamics of nucleus actin and gene transcription (Huet et al., 2012). It is possible that when there is no active Cdc42 to interact with CEP1 in the cytoplasm, CEP1 can translocate into nucleus to interact with nucleus actin. Second is whether CEP1 can regulate actin organization independent of Cdc42. Since CEP1(ΔCRIB) mainly locates to the nucleus, overexpression of CEP1(ΔCRIB) failed to regulate actin organization, reflected by no hyperpigmentation as seen in wild-type CEP1 overexpression. However, Calvo et al. (2015) reported that in addition to low affinity to Cdc42, CRIB-defective CEP3 also has lower binding affinity to actin and septins. Similarly, CEP1(ΔCRIB) may also display impaired interaction with actin, septin, or other proteins, thus hindered their activity in actin regulation. Materials and methods Embryo manipulations, MOs, and RNA preparation Xenopus laevis embryos were obtained and microinjected with morpholino and RNA as previously described (Nie et al., 2009). CEP1-MO (5′-GGTTCATTGTTCCTTCTTTTTCTGA-3′) hybridizes to −18 to 7 position relative to the translational start site of Xenopus CEP1 (GenBank Accession No. NM_1114777), Cdc42-MO (5′-CTACACATTTAATTGTCTGCATGGC-3′) hybridizes to −3 to 22 position relative to the translational start site of Xenopus Cdc42 (GenBank Accession No. NM_1085899), and standard control MO (Gene tools, Philomath, OR) were used in the study. Xenopus CEP1 and Cdc42 were subcloned into pCS2 + 8NEGFP or pCS2 + 8NmCherry vectors (addgene) to generate fusion constructs. GFP-wGBD was a gift from William Bement (Addgene plasmid #26734). To generate Xenopus CEP1 with the 5′UTR for MO recognition (CEP1(UTR)) and mutation that lacks the CRIB domain (CEP1(ΔCRIB)), site-directed mutagenesis was performed using CEP1-EGFP as a template. All experimental procedures were performed according to USDA Animal Welfare Act Regulations and have been approved by Institutional Animal Care and Use Committee, in compliance of Public Health Service Policy. Red-Gal staining, in situ hybridization, immunohistochemistry, and cartilage staining Red-Gal staining, in situ hybridization, immunohistochemistry, and cartilage staining were performed as previously describe (Nie et al., 2009). Detailed procedures are provided in Supplementary Materials and methods. Primary antibodies used in the study include anti-phosphotyrosine antibody (4G10; EMD Millipore), diluted at 1:200; anti-Arp2 (Thermo PA5-19760), diluted at 1:1000; and anti-pMLC (Abcam ab2480), diluted at 1:400. FITC-conjugated secondary antibody was used at 1:1000. CNC explant culture, grafting, and microscopy CNC explants receiving different MOs or RNAs encoding fluorescent proteins were dissected from stages 13–14 embryos as previously described (Borchers et al., 2000; Alfandari et al., 2003; DeSimone et al., 2005) and cultured and imaged as previously described (Nie et al., 2009). See Supplementary Materials and methods for details. Traction force microscopy The preparation of polyacrylamide hydrogel substrates and the analysis process used to implement traction force microscopy have been described in detail previously (Yeung et al., 2005; Kovari et al., 2016) and are summarized in Supplementary Materials and methods. The Matlab code for TFM is available at https://github.com/dkovari/TFMatlab. The mean strain energy was calculated for ~30 control and CEP1-MO cells over time. The mean strain energy from samples with different phenotypes were compared with multiple comparison test using Fisher’s least significant difference (Williams and Abdi, 2010), after confirming that different samples were unequal with one-way analysis of variance (ANOVA) test. Supplementary material Supplementary material is available at Journal of Molecular Cell Biology online. Acknowledgements We would like to thank Dr Marianne Bronner (California Institute of Technology, Pasadena, USA) for her comments and discussion. We would also like to acknowledge the support from the National Institutes of Health, National Science Foundation, and Georgia Institute of Technology. Funding This work is supported by the National Institutes of Health (R00DE022796 to S.N.) and National Science Foundation (DMR-0955811 to J.E.C. and PHY-0848797 to J.E.C. and D.T.K.). Conflict of interest: none declared. References Alfandari , D. , Cousin , H. , Gaultier , A. , et al. . ( 2003 ). Integrin α5β1 supports the migration of Xenopus cranial neural crest on fibronectin . Dev. Biol. 260 , 449 – 464 . Google Scholar CrossRef Search ADS PubMed Benink , H.A. , and Bement , W.M. ( 2005 ). Concentric zones of active RhoA and Cdc42 around single cell wounds . J. Cell Biol. 168 , 429 – 439 . Google Scholar CrossRef Search ADS PubMed Bergert , M. , Chandradoss , S.D. , Desai , R.A. , et al. . ( 2012 ). Cell mechanics control rapid transitions between blebs and lamellipodia during migration . Proc. Natl Acad. Sci. USA 109 , 14434 – 14439 . Google Scholar CrossRef Search ADS Borchers , A. , Epperlein , H.H. , and Wedlich , D. ( 2000 ). An assay system to study migratory behavior of cranial neural crest cells in Xenopus . Dev. Genes Evol. 210 , 217 – 222 . Google Scholar CrossRef Search ADS PubMed Burbelo , P.D. , Drechsel , D. , and Hall , A. ( 1995 ). A conserved binding motif defines numerous candidate target proteins for both Cdc42 and Rac GTPases . J. Biol. Chem. 270 , 29071 – 29074 . Google Scholar CrossRef Search ADS PubMed Burbelo , P.D. , Snow , D.M. , Bahou , W. , et al. . ( 1999 ). MSE55, a Cdc42 effector protein, induces long cellular extensions in fibroblasts . Proc. Natl Acad. Sci. USA 96 , 9083 – 9088 . Google Scholar CrossRef Search ADS Calvo , F. , Ranftl , R. , Hooper , S. , et al. . ( 2015 ). Cdc42EP3/BORG2 and septin network enables mechano-transduction and the emergence of cancer-associated fibroblasts . Cell Rep. 13 , 2699 – 2714 . Google Scholar CrossRef Search ADS PubMed Carmona-Fontaine , C. , Matthews , H.K. , Kuriyama , S. , et al. . ( 2008 ). Contact inhibition of locomotion in vivo controls neural crest directional migration . Nature 456 , 957 – 961 . Google Scholar CrossRef Search ADS PubMed Charras , G. , and Paluch , E. ( 2008 ). Blebs lead the way: how to migrate without lamellipodia . Nat. Rev. Mol. Cell Biol. 9 , 730 – 736 . Google Scholar CrossRef Search ADS PubMed Choi , S.C. , and Han , J.K. ( 2002 ). Xenopus Cdc42 regulates convergent extension movements during gastrulation through Wnt/Ca2+ signaling pathway . Dev. Biol. 244 , 342 – 357 . Google Scholar CrossRef Search ADS PubMed DeSimone , D.W. , Davidson , L. , Marsden , M. , et al. . ( 2005 ). The Xenopus embryo as a model system for studies of cell migration . Methods Mol. Biol. 294 , 235 – 245 . Google Scholar PubMed Dupin , E. , Creuzet , S. , and Le Douarin , N.M. ( 2006 ). The contribution of the neural crest to the vertebrate body . Adv. Exp. Med. Biol. 589 , 96 – 119 . Google Scholar CrossRef Search ADS PubMed Farrugia , A.J. , and Calvo , F. ( 2016 ). The Borg family of Cdc42 effector proteins Cdc42EP1–5 . Biochem. Soc. Trans. 44 , 1709 – 1716 . Google Scholar CrossRef Search ADS PubMed Farrugia , A.J. , and Calvo , F. ( 2017 ). Cdc42 regulates Cdc42EP3 function in cancer-associated fibroblasts . Small GTPases 8 , 49 – 57 . Google Scholar CrossRef Search ADS PubMed Guo , W.H. , and Wang , Y.L. ( 2012 ). A three-component mechanism for fibroblast migration with a contractile cell body that couples a myosin II-independent propulsive anterior to a myosin II-dependent resistive tail . Mol. Biol. Cell 23 , 1657 – 1663 . Google Scholar CrossRef Search ADS PubMed Han , Y. , Eppinger , E. , Schuster , I.G. , et al. . ( 2009 ). Formin-like 1 (FMNL1) is regulated by N-terminal myristoylation and induces polarized membrane blebbing . J. Biol. Chem. 284 , 33409 – 33417 . Google Scholar CrossRef Search ADS PubMed Hirsch , D.S. , Pirone , D.M. , and Burbelo , P.D. ( 2001 ). A new family of Cdc42 effector proteins, CEPs, function in fibroblast and epithelial cell shape changes . J. Biol. Chem. 276 , 875 – 883 . Google Scholar CrossRef Search ADS PubMed Huet , G. , Skarp , K.P. , and Vartiainen , M.K. ( 2012 ). Nuclear actin levels as an important transcriptional switch . Transcription 3 , 226 – 230 . Google Scholar CrossRef Search ADS PubMed Joberty , G. , Perlungher , R.R. , and Macara , I.G. ( 1999 ). The Borgs, a new family of Cdc42 and TC10 GTPase-interacting proteins . Mol. Cell. Biol. 19 , 6585 – 6597 . Google Scholar CrossRef Search ADS PubMed Joberty , G. , Perlungher , R.R. , Sheffield , P.J. , et al. . ( 2001 ). Borg proteins control septin organization and are negatively regulated by Cdc42 . Nat. Cell Biol. 3 , 861 – 866 . Google Scholar CrossRef Search ADS PubMed Joo , E. , Surka , M.C. , and Trimble , W.S. ( 2007 ). Mammalian SEPT2 is required for scaffolding nonmuscle myosin II and its kinases . Dev. Cell 13 , 677 – 690 . Google Scholar CrossRef Search ADS PubMed Kardash , E. , Reichman-Fried , M. , Maitre , J.L. , et al. . ( 2010 ). A role for Rho GTPases and cell-cell adhesion in single-cell motility in vivo . Nat. Cell Biol. 12 , 47 – 53 . Google Scholar CrossRef Search ADS PubMed Kovari , D.T. , Wei , W.B. , Chang , P. , et al. . ( 2016 ). Frustrated phagocytic spreading of J774A-1 macrophages ends in Myosin II-dependent contraction . Biophys. J. 111 , 2698 – 2710 . Google Scholar CrossRef Search ADS PubMed Le Douarin , N. ( 1980 ). Migration and differentiation of neural crest cells . Curr. Top. Dev. Biol. 16 , 31 – 85 . Google Scholar CrossRef Search ADS PubMed Liu , Z. , Vong , Q.P. , Liu , C. , et al. . ( 2014 ). Borg5 is required for angiogenesis by regulating persistent directional migration of the cardiac microvascular endothelial cells . Mol. Biol. Cell 25 , 841 – 851 . Google Scholar CrossRef Search ADS PubMed Lo , C.M. , Buxton , D.B. , Chua , G.C. , et al. . ( 2004 ). Nonmuscle myosin IIb is involved in the guidance of fibroblast migration . Mol. Biol. Cell 15 , 982 – 989 . Google Scholar CrossRef Search ADS PubMed Lorentzen , A. , Bamber , J. , Sadok , A. , et al. . ( 2011 ). An ezrin-rich, rigid uropod-like structure directs movement of amoeboid blebbing cells . J. Cell Sci. 124 , 1256 – 1267 . Google Scholar CrossRef Search ADS PubMed Lucas , J.M. , Nikolic , I. , and Hens , M.D. ( 2002 ). cDNA cloning, sequence comparison, and developmental expression of Xenopus rac1 . Mech. Dev. 115 , 113 – 116 . Google Scholar CrossRef Search ADS PubMed Matthews , H.K. , Marchant , L. , Carmona-Fontaine , C. , et al. . ( 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 . Google Scholar CrossRef Search ADS PubMed Mostowy , S. , and Cossart , P. ( 2012 ). Septins: the fourth component of the cytoskeleton . Nat. Rev. Mol. Cell Biol. 13 , 183 – 194 . Google Scholar PubMed Nelson , K.K. , and Nelson , R.W. ( 2004 ). Cdc42 Effector Protein 2 (XCEP2) is required for normal gastrulation and contributes to cellular adhesion in Xenopus laevis . BMC Dev. Biol. 4 , 13 . Google Scholar CrossRef Search ADS PubMed Nie , S. , and Bronner , M.E. ( 2015 ). Dual developmental role of transcriptional regulator Ets1 in Xenopus cardiac neural crest vs. heart mesoderm . Cardiovasc. Res. 106 , 67 – 75 . Google Scholar CrossRef Search ADS PubMed Nobes , C.D. , and Hall , A. ( 1995 ). Rho, rac, and cdc42 GTPases regulate the assembly of multimolecular focal complexes associated with actin stress fibers, lamellipodia, and filopodia . Cell 81 , 53 – 62 . Google Scholar CrossRef Search ADS PubMed Vong , Q.P. , Liu , Z. , Yoo , J.G. , et al. . ( 2010 ). A role for borg5 during trophectoderm differentiation . Stem Cells 28 , 1030 – 1038 . Google Scholar CrossRef Search ADS PubMed Williams , L.J. , and Abdi , H. ( 2010 ). Fisher’s least significant difference (LSD) test. In: Salkind , N.J. (ed.). Encyclopedia of Research Design . Thousand Oaks, CA : SAGE Publications Inc. , 1 – 6 . Yeung , T. , Georges , P.C. , Flanagan , L.A. , et al. . ( 2005 ). Effects of substrate stiffness on cell morphology, cytoskeletal structure, and adhesion . Cell Motil. Cytoskel. 60 , 24 – 34 . Google Scholar CrossRef Search ADS Zamir , E. , Katz , B.Z. , Aota , S. , et al. . ( 1999 ). Molecular diversity of cell-matrix adhesions . J. Cell Sci. 112(Pt 11) , 1655 – 1669 . Google Scholar PubMed © The Author (2017). Published by Oxford University Press on behalf of Journal of Molecular Cell Biology, IBCB, SIBS, CAS. All rights reserved.

Journal

Journal of Molecular Cell BiologyOxford University Press

Published: Oct 6, 2017

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off