This article was accepted for inclusion in Developmental Dynamics238#6 – Special Focus on Xenopus.
Special Issue Research Article
Role of p21-activated kinase in cell polarity and directional mesendoderm migration in the Xenopus gastrula†
Article first published online: 4 JUN 2009
Copyright © 2009 Wiley-Liss, Inc.
Volume 238, Issue 7, pages 1709–1726, July 2009
How to Cite
Nagel, M., Luu, O., Bisson, N., Macanovic, B., Moss, T. and Winklbauer, R. (2009), Role of p21-activated kinase in cell polarity and directional mesendoderm migration in the Xenopus gastrula. Dev. Dyn., 238: 1709–1726. doi: 10.1002/dvdy.21985
- Issue published online: 10 JUN 2009
- Article first published online: 4 JUN 2009
- Manuscript Accepted: 17 APR 2009
- Canadian Institutes of Health Research. Grant Number: MOP-53075
- National Cancer Institute of Canada
- p21-activated kinase;
The p21 activated kinases (Paks) are prominently involved in the regulation of cell motility. Using a kinase-dead mutant of xPak1, we show that during Xenopus gastrulation, the kinase activity of Pak1 is required upstream of Cdc42 for the establishment of cell polarity in the migrating mesendoderm. Overactivation of Pak1 function by the expression of constitutively active xPak1 compromises the maintenance of cell polarity, by indirectly inhibiting RhoA function. Inhibition of cell polarization does not affect the migration of single mesendoderm cells. However, Pak1 inhibition interferes with the guidance of mesendoderm migration by directional cues residing in the extracellular matrix of the blastocoel roof, and with mesendoderm translocation in the embryo. Developmental Dynamics 238:1709–1726, 2009. © 2009 Wiley-Liss, Inc.
During gastrulation, the basic metazoan body plan is established through an extensive rearrangement of embryonic parts. Of the massive cell movements that characterize gastrulation in vertebrates, two have been well studied: convergent extension of the posterior axial mesoderm, and directional migration of the anterior mesendoderm. Both processes are driven by the oriented movement of polarized cells. In convergent extension, the mediolateral intercalation of bipolar cells contributes essentially to a change in tissue shape, leading to a lengthening and narrowing of the axis-forming tissue (Keller,2002). In this process, Wnt/PCP signaling is necessary for cell polarization (Wallingford et al.,2002), and graded activin/nodal signaling determines the orientation of cells perpendicular to the axis of elongation, by the establishment of an anteroposterior tissue polarity (Ninomiya et al.,2004).
The directional migration of the mesendoderm anterior to the converging and extending tissue has been most extensively studied in the frog, Xenopus. Here, the mesendoderm migrates across the ectodermal blastocoel roof (BCR) toward the animal pole of the embryo. Mesendoderm cells in contact with the BCR extend lamelliform protrusions (Winklbauer and Nagel,1991). Interaction with fibronectin (FN), a main component of the extracellular matrix of the BCR (Lee et al.,1984; Winklbauer and Stoltz,1995; Winklbauer,1998), is necessary for lamellipodia formation and migration (Winklbauer,1990; Winklbauer and Selchow,1992; Ramos and DeSimone,1996; Winklbauer and Keller,1996; Wacker et al.,1998). When binding to FN is blocked, cells still attach to the BCR, but do not extend locomotory processes (Winklbauer and Keller,1996).
In contrast to the bipolar axial mesoderm cells, mesendoderm cells appear unipolar. They extend lamellipodia toward the animal pole, thus underlapping their anterior neighbors and generating a distinctive “shingle arrangement” of cells (Winklbauer and Nagel,1991; Nagel and Winklbauer,1999). This uniform orientation of polarized cells, and their directional movement, depend on platelet derived growth factor (PDGF) signaling. In the gastrula, PDGF-A is expressed in the BCR, and its receptor, PDGFR-α, in the internalized mesendoderm (Ataliotis et al.,1995). Inhibition of PDGFR-α function, and inhibition or overexpression of PDGF-A randomize cell orientation and the directionality of migration (Nagel et al.,2004).
Xenopus mesendoderm cells show polarized protrusive activity when migrating as single cells in vitro on artificial FN substratum (Winklbauer and Selchow,1992; Wacker et al.,1998), indicating that they possess an intrinsic polarization independent of PDGF signaling. At the level of the control of gene expression, the homeodomain transcription factors Mix.1, Goosecoid and Siamois are necessary for the polarization of mesendoderm cells (Wacker et al.,1998; Luu et al.,2008), but the molecular mechanisms downstream of these factors are not known. Moreover, it is not clear how the polarization of protrusive activity of single cells is related to the unipolar cell morphology and to cell orientation as seen during directional migration of mesendodermal cell masses.
We show here that p21-activated kinase (Pak) function is necessary for polarized protrusive activity of single mesendoderm cells, and for the orientation and directional migration of mesendoderm cell aggregates. Paks are members of the Ste20 serine-threonine protein kinase family that were initially identified as effectors of Rho family GTPases (Manser et al.,1994). They contain a Cdc42/Rac binding site in the N-terminal region and a conserved C-terminal kinase domain. In group I Paks (Paks1–3), binding of GTP-loaded Cdc42 or Rac activates the kinase by disabling an autoinhibitory domain (Bokoch,2003, for review). However, Paks are also activated by GTPase-independent mechanisms (e.g., Lu et al.,1997; Bisson et al.,2007). Moreover, Paks can adopt kinase-independent functions as scaffold proteins, acting upstream as well as downstream of Rac/Cdc42 (Obermeier et al.,1998; Li et al.,2003; Poitras et al.,2003a; Bisson et al.,2007). The kinases are involved in the control of cytoskeletal functions, for example by targeting LIM kinase, filamin, or myosin light chain kinase. Consequently, Paks have been implicated in the control of cell motility. Other Pak functions include the control of transcription, cell survival, and proliferation (Kumar and Vadlamudi,2002; Bokoch,2003; Hofmann et al.,2004). Altered Pak1 expression is associated with tumor formation, tumor progression, and metastasis, in particular with breast, ovarian, and colorectal cancers (Kumar and Vadlamudi,2002; Schraml et al.,2003; Carter et al.,2004).
In Xenopus (Faure et al.,1997; Cau et al.,2000; Islam et al.,2000; Souopgui et al.,2002), as in mammals (Bokoch,2003), the sequences for Paks1–3 are conserved and the overall domain structure is similar. In the early embryo, xPak1 is ubiquitously expressed at the mRNA and protein level, respectively, and both maternally and zygotically (Islam et al.,2000). Likewise, xPak2 and xPak3 are expressed maternally and throughout early development (Cau et al.,2000; Souopgui et al.,2002; Faure et al.,2005). At oocyte maturation, isolated xPak1 kinase domain, but not the N-terminal domain, interferes with xPak2 function (Faure et al.,1997; Cau et al.,2000), suggesting some degree of redundancy. xPak3 antisense morpholino oligonucleotide inhibition increases proliferation and blocks differentiation in the neural plate (Souopgui et al.,2002).
Pak1 Function Is Necessary for Mesendoderm Migration on the BCR
To examine Pak1 function in the gastrula, we expressed a kinase-dead version of xPak1, KD-Pak1, that interferes with endogenous Pak1 function (Faure et al.,1997; Islam et al.,2000). This approach circumvents potential problems with maternally provided protein. Dorsal injection of KD-Pak1 mRNA affects gastrulation movements, which expresses itself for example in a shortened archenteron (Fig. 1A–C,I,J). At the mid-gastrula stage, a dose-dependent reduction of archenteron length is observed, which is maximal at 400 pg of KD-Pak1 mRNA per embryo. Coexpression of wild-type Pak1 rescues archenteron lengthening (Fig. 1K).
This gastrulation defect is not due to an inhibition of convergent extension: dorsal blastopore lip explants from KD-Pak1–expressing embryos elongate as well as control explants from uninjected gastrulae (Fig. 1L,M). Therefore, we asked whether the actual translocation of internalized anterior mesendoderm across the BCR is affected by Pak1 inhibition. To visualize the movement of the mesendoderm relative to the BCR, both tissues were labeled in register: at stage 10.5, a DiI (1,1′, di-octadecyl-3,3,3′,3′,-tetramethylindo-carbocyanine perchlorate) -soaked eyelash was inserted for a few seconds into the embryo, deep enough to reach into the mesendoderm, and embryos were fixed 2 hr later. In controls, labeled mesendoderm moved relative to the BCR label in 5 of 6 cases (Fig. 1A). In contrast, no translocation of mesendoderm directly in contact with the BCR was noted in 9 of 9 cases after KD-Pak1 injection (Fig. 1B). However, deep cells further away from the BCR moved anteriorly in 9 of 10 cases (Fig. 1C). Translocation across the BCR was rescued in 8 of 11 cases upon coinjection of KD-Pak1 and wt-Pak1 (Fig. 1D). Thus, Pak1 inhibition impedes the translocation of mesendoderm cells that are in contact with the BCR. However, movement of cells deep within the mesendodermal mass is not inhibited.
Impaired mesendoderm movement is most likely not due to changes in cell fate. Expression of the mesoderm marker Xbra (Smith et al.,1991) is similar in control and in KD-Pak1–expressing cells (Fig. 1E,G). Moreover, gsc is normally expressed in the region examined in our present study (Cho et al.,1991), and this pattern is not affected by KD-Pak1 (Fig. 1F,H). However, in the late gastrula, the position of the gsc-expressing mesoderm is altered in KD-Pak1–injected embryos, providing additional evidence for an impaired migration on the BCR (Fig. 1I,J). The uninterrupted movement of deep cells could explain how, in Pak1-inhibited embryos, the gsc-expressing domain of anterior mesoderm can spread to some extent animally on the BCR, despite the inhibition of translocation relative to the BCR. Altogether, our results suggest that Pak1 function is required in migrating cells at the mesendoderm–BCR interface.
Pak1 Affects the Morphology of Isolated Mesendoderm Cells on FN
To see whether KD-Pak1 injection affects the motility of single cells, we dissociated dorsal leading edge mesendoderm and seeded cells on FN-coated coverslips. Mesendoderm cells migrate on this substratum by extending, splitting, and retracting lamellipodia, and thus shifting constantly between spindle-like and triangular shapes. Essentially, any lamellipodium of a migrating cell can be traced back to a protrusion extended during initial cell spreading, from which it is derived by the repeated process of splitting into two daughter lamellipodia. This restriction of lamellipodia formation to an active site on the cell surface indicates a functional polarization of these cells (Winklbauer and Selchow,1992). Snap-shots of the dynamic shapes observed during migration are obtained by staining fixed cells with phalloidin-rhodamine, to visualize the actin cytoskeleton. Representative examples of differently treated cells are shown in Figure 2, and quantitative data in Figure 5G.
Confirming previous observations, elongated or triangular shapes are seen in control cells (Fig. 2A). The actin-rich processes at the tapering ends of cells can appear fragmented (left end of right cell, Fig. 2A). A typical configuration following the splitting of a protrusion, as seen in time-lapse recordings (Winklbauer and Selchow,1992), is also exhibited (top end of middle cell, Fig. 2A). Upon KD-Pak1 expression, cells extend small protrusions at multiple sites and appear stellate (Fig. 2B). At least some of these cytoplasmic processes consist of very small lamelliform protrusions (e.g., Fig. 2B, left cell process #7, right cell processes #1,4). Overexpression of wild-type xPak1 leaves cells mostly bipolar (Fig. 2C). Coexpression of KD-Pak1 and xPak1 at equal doses rescues normal cell shape (Fig. 2D). These and the corresponding quantitative data of Figure 5G suggest that Pak1 kinase activity normally functions in these cells to spatially restrict lamellipodia formation, thus polarizing the cells.
Expression of a constitutively active form of xPak1 (CA-Pak1) leads to extra lamellipodia (Fig. 2E) in approximately half of the cells (Fig. 5G). When present, extra protrusions extend from the sides of cells which nevertheless retained their elongated shapes (Fig. 2E, e.g., left cell processes #2,4, middle cell processes #2,4). Thus, two distinct polarity phenotypes are observed. With KD-Pak1, cells are stellate and lamellipodia are more numerous, but small, as if total lamellipodia mass were conserved. In contrast, with CA-Pak1, extra lamellipodia of normal size appear to be added to normally elongated cells.
Pak1 Kinase Function Is Required for Polarized Protrusive Activity on FN
Nonattached mesendoderm cells are known to be polarized with respect to protrusive activity. They extend filiform protrusion from a small active surface area only. This polarity is translated into a bipolar morphology during spreading on FN, as one lamellipodium appears at this active site, a second one opposite (Winklbauer and Selchow,1992). In contrast, nonpolarized cells from the vegetal cell mass extend protrusions in all directions before attachment, and spread irregularly on FN by forming lamellipodia randomly in different directions (Luu et al.,2008). To determine at which point Pak1 function is required for normal cell polarization, we observed the protrusive activity of KD-Pak1–expressing nonattached cells. Under these conditions, cells are strongly polarized, and protrusions form close to one another at an active site, similar to control mesendoderm cells (Fig. 3C).
Next, we examined cell spreading. We confirmed that uninjected mesendoderm cells show bipolar spreading (Fig. 3A,B). In contrast, KD-Pak1–expressing cells spread irregularly, with protrusions appearing consecutively and in random directions (Fig. 3A,B), similar to nonpolarized vegetal cells (Luu et al.,2008). Apparently, it is the establishment of a polarized morphology on FN that specifically depends on the kinase activity of Pak1.
CA-Pak1 Expression Interferes With the Maintenance of Polarized Protrusive Activity
In contrast to KD-Pak1–expressing cells, CA-Pak1–injected cells spread in a regular bipolar manner (Fig. 3A,B), but become multipolar during subsequent migration, as they are less prone to retract lamellipodia. Whereas control cells completely absorb a collapsing protrusion when moving away in the opposite direction (Fig. 4A), CA-Pak1–injected cells often remain attached to the former site of a protrusion by cytoplasmic strands, to eventually reestablish a lamellipodium (Fig. 4A). Thus, extra lamellipodia are maintained, and cells appear less polarized. This effect expresses itself also in the lifespan of protrusions. In control cells, approximately half of the lamellipodia are maintained for less than 15 min, but CA-Pak1–injected cells have a more even distribution of protrusion lifespans ranging from minutes to more than an hour (Fig. 4B).
Altogether, polarization is affected differently by KD- and CA-Pak1. With the former, the primary establishment of polarity on FN substratum is compromised, whereas the latter affects the maintenance of polarity during migration. In the following, we will refer to the KD-Pak1–induced effect as the type I polarity phenotype, and correspondingly to the CA-Pak1 effect as type II.
Pak1 Controls the Polarity of Mesendoderm Cells Upstream of Cdc42 and Rac
Cdc42 is a putative regulator of Paks, and we asked whether it initiated the effects on cell shape. In fact, dominant-negative Cdc42 (DN-Cdc42) mimicks Pak1 inhibition, generating a type I phenotype. Injected cells become stellate and extend many small protrusions (Fig. 5A,G). This would indeed be consistent with an activation of Pak1 by Cdc42. Surprisingly however, CA-Pak1, which is constitutively active independently of GTPase interaction, is not able to rescue the effect of Cdc42 inhibition. Cells coinjected with CA-Pak1 and DN-Cdc42 mRNA exhibited the multipolar shape of DN-Cdc42– or KD-Pak1–expressing cells (Fig. 5B,G), as if Cdc42 inhibition would override the consequences of Pak1 activation.
Conversely, when constitutively active Cdc42 (CA-Cdc42) was coexpressed with KD-Pak1, cells became elongated and moderately polarized (Fig. 5D), similar to CA-Pak1–expressing cells (Fig. 2E) or CA-Cdc42–expressing cells (Fig. 5C,G). It appeared as if active Cdc42 could rescue the effect of Pak1 inhibition, shifting cell morphology from the type I to the less severe type II phenotype. Taken together, this would suggest that contrary to expectation, Pak1 activates Cdc42 to achieve cell polarization. Cells coexpressing CA-Pak1 and CA-Cdc42 appear similar to controls (Fig. 5E,G), whereas coinjection of KD-Pak1 and DN-Cdc42 generates a stellate morphology, as expected (Fig. 5F,G).
Rac is also a known activator of Paks. DN-Rac–expressing cells possess small, actin-dense protrusions, but retain an elongate morphology (Fig. 6A,G). In contrast, CA-Rac leads to reduced polarization in approximately half of the cells, and to laterally extending protrusions in elongate cells (Fig. 6B,G), similar to the type II phenotype caused by CA-Pak1 expression. This suggests that overactivated Rac could impede the maintenance of polarization downstream of Pak1. Indeed, when CA-Pak1 is coexpressed with DN-Rac, cells appear “superpolarized,” with small, dense protrusions at opposite ends of extremely spindle-shaped cells (Fig. 6C,G). This could be explained by the fact that both the inhibition of Rac function and the CA-Pak1–induced Cdc42 activation promote polarity, the latter by stimulating the initial polarization of protrusive activity, the former by supporting the maintenance of polarity. Altogether, it appears that Pak1 can affect type II-related cell polarization by overactivating Cdc42 and Rac, which in turn impedes maintenance of polarization, leading to a type II polarity phenotype. Lamellipodia formation is compromised by DN-Rac, but stimulated by CA-Rac and CA-Cdc42.
Whereas Pak1 is usually considered to be an effector of Cdc42, it appears to act upstream of this GTPase in the present context. To see whether the G-protein binding ability of Pak1 is required here, a mutant form of the kinase, where the GBD site is inactive (XPak1H74,77L), was expressed. Cells appear similar to those expressing KD-Pak1 (Fig. 6D,G), suggesting that this Pak1 mutant interferes similarly with endogenous xPak1 function in polarization. Coexpression of both KD-Pak1 and XPak1H74,77L does not significantly change polarization (Fig. 6,G), although stronger polarized cells can be found (Fig. 6E). A double mutant Pak1, XPakKDΔGBD, which is both kinase-dead and GTPase-binding deficient, induces also a nonpolarized cell morphology, although lamellipodia formation is strongly promoted (Fig. 6F,G). Thus, binding of Cdc42 to Pak1 seems to be necessary for the induction of type I-related polarization.
To visualize active Cdc42/Rac, the green fluorescent protein (GFP) -tagged GBD of Pak1 was expressed at a low level (30 pg/blastomere) in mesendoderm cells. In all cells, a diffuse staining was observed. In addition, control cells showed a dense pattern of small punctae at lamellipodia (Fig. 6H) and also under the cell body (not shown), that is, over the whole substratum-contacting cell surface. These punctae were strongly reduced in KD-Pak1–expressing cells (Fig. 6I), but rescued upon coexpression of xPak1 with KD-Pak1 (Fig. 6J). These observations are consistent with Pak1 acting upstream of Cdc42 and Rac.
RhoA Activity Is Required for the Maintenance of Mesendoderm Cell Polarity
Confirming our previous findings (Luu et al.,2008), we showed that activation of RhoA is also required for the polarization of mesendodermal cells. Expression of dominant-negative RhoA (DN-RhoA) abolishes at once polarity and lamellipodia formation (Fig. 7A,J). The effect on polarity, but not that on protrusion formation, is mimicked by treatment of cells with Rho kinase inhibitor. Mesendoderm cells extend several large lamellipodia and show a type II polarity phenotype, resembling cells that express CA-Pak1 (Fig. 7B,J). To see how RhoA inhibition-induced multipolarity is related to Pak1 activation, we coexpressed DN-RhoA with a combination of factors that induces superpolarization, namely CA-Pak1 plus DN-Rac (see Fig. 6C). In this context, DN-RhoA significantly reduced polarization and prevented the extreme elongation of cell bodies (Fig. 7C,J). This is consistent with the idea that CA-Pak1/DN-Rac coexpression leads to strong RhoA activation, in addition to Cdc42 activation, and thus to superpolarization: the latter phenotype can in part be reverted by DN-RhoA.
Constitutively active RhoA (CA-RhoA) has only a minor effect on polarity (Fig. 7D,J). However, when coexpressed with CA-Rac or CA-Cdc42, it reverses the depolarizing effects of these factors (Fig. 7E,F,J; compare to Figs. 6B, 4C). This can be summarized in a possible scenario (see Fig. 11, and the Discussion section): Pak1 activates Cdc42, which then polarizes protrusive activity. At the same time, Cdc42 activates Rac, which in turn promotes lamellipodia formation. If overactivated, Rac indirectly attenuates RhoA activity, which would be required for the maintenance of polarity.
In migrating mesendoderm cells, RhoA activation requires signaling downstream of the transcription factors Goosecoid and Mix.1 (Luu et al.,2008). We confirmed this finding, and demonstrated that cells expressing the inhibitory EnRMix.1 construct show a nonpolarized type II morphology similar to that of CA-Pak1–expressing cells (Fig. 7G,J). Coexpression of CA-Pak1 and EnRMix.1 does not alter the cell phenotype (Fig. 7H,J), as expected if both CA-Pak1 and EnRMix.1 expression abolished RhoA activity. In contrast, DN-Rac substantially rescues polarity in EnRMix.1–expressing cells (Fig. 7I,J), suggesting that some residual RhoA activity is normally suppressed by Rac in EnRMix.1–expressing cells.
To visualize active RhoA (Bement et al.,2005), the GFP-tagged GBD of rhotekin (30 pg/blastomere) was expressed (Fig. 7K). In control cells, it was enriched in lamellipodia, preferentially at their proximal end. In CA-Pak1–expressing cells, this concentration of rhotekin was absent, but it was not visibly affected in KD-Pak1 cells (Fig. 7K,bottom). This is consistent with a down-regulation of RhoA activation by CA-Pak1. Because myosin contractility is a target of RhoA and Rho kinase, and could be involved in lamellipodia retraction, we also stained for nonmuscle myosin II in mesendoderm cells (Fig. 7L). As described in detail previously (Selchow and Winklbauer,1997), myosin II is enriched in the proximal part of lamellipodia in control cells (Fig. 7L, top). In CA-Pak1–expressing cells, this localized accumulation is diminished (Fig. 7L, middle), whereas it is retained in KD-Pak1 cells (Fig. 7L, bottom).
Isolated Mesendoderm Cells Migrate Independently of Pak1-Dependent Polarity
Pak1 affects migration in a variety of cell types. Surprisingly, despite their multipolar morphology, overall migration speed is not diminished in KD-Pak1–expressing cells, although cell velocities are slightly more uniform as compared to controls (Fig. 8A). A similar effect is observed with CA-Pak1 (Fig. 8C). xPak1 overexpression has only a weak stimulating effect on mesendoderm cell migration (Fig. 8B). In contrast to this insensitivity to Pak1 modulation, migration of DN-Cdc42–expressing cells is impaired (Fig. 8D). Apparently, Cdc42 has additional roles in mesendoderm cell migration beyond mediating cell polarization.
Pak1 Is Involved in the Directional Migration of Mesendoderm Cell Aggregates
In the embryo, the mesendoderm migrates as a coherent aggregate, not as single cells. Moreover, migration is directed by guidance cues located in the BCR matrix. Although apparently not relevant for the migration efficiency of isolated cells on FN, Pak1-dependent cell polarization could be important under aggregate conditions, to allow for the oriented protrusive activity that underlies directional migration (Winklbauer and Nagel,1991; Nagel et al.,2004).
Oriented migration of explants can be studied in vitro on BCR-conditioned substratum (Nagel et al.,2004). In this assay, control explants move preferentially toward the animal pole (Fig. 9A,B). The xPak1-injected explants show a weak reduction in directionality, with a quarter of explants moving vegetally instead of animally (Fig. 9B). The movement of KD-Pak1–expressing explants is completely randomized (Fig. 9B). This suggests that the kinase function of Pak1 is necessary for the guidance of mesoderm migration by substratum-borne cues. Coinjection of xPak1 with KD-Pak1 rescues directionality of migration (Fig. 9B), consistent with a permissive role of Pak1 in this process. Expression of CA-Pak does not affect directionality (Fig. 9B). Both CA-Cdc42 and DN-Cdc42 strongly affect the directionality of migration, although they do not lead to a complete randomization of movement (Fig. 9B). The migration velocities of explants vary considerably, like those of single cells, but average velocity is not significantly different between different treatments (Fig. 9C).
To observe effects of Pak inhibition on cell morphology during directional migration, we fixed embryos, and prepared the substrate-contacting side of the mesendoderm for SEM by removing the BCR. In controls, loosely packed cells underlap each other in a posterior to anterior direction (Fig. 10A), and the orientation of most lamelliform protrusions is animally biased (Fig. 10E,F). This shingle arrangement of polarized cells indicates directional, substratum-guided migration (Nagel et al.,2004). In KD-Pak1–expressing embryos, the BCR adheres strongly to the mesendoderm, as becomes apparent during its removal. Cells are tightly packed, and underlapping protrusions are large, but randomly oriented, which indicates loss of directionality (Fig. 10B,E,F). xPak1-injected cells are loosely packed, as in controls, although underlapping is less oriented (Fig. 10D,E). Coinjection of xPak1 with KD-Pak1 mRNA restores control conditions (Fig. 10C,E). Overall, these results indicate a role for Pak1 in directional migration.
We showed that Pak1 is involved in the execution of a major gastrulation movement, the migration of the dorsal mesendoderm. Its function is necessary for the polarized protrusive activity of single cells migrating on FN, and for the directional, substratum-guided migration of mesendoderm cell aggregates in vitro and in the embryo.
Control of Mesendoderm Cell Polarity by Pak1
Our experiments revealed two aspects of mesendoderm single cell polarity that are differently affected by Pak1 activity: (1) a basic polarization of protrusive activity, established during spreading on FN substratum and due to a restriction of the number of lamellipodia formed (type I phenotype-related), and (2) the maintenance of polarity during migration, in initially polarized cells, by the retraction of lamellipodia, such that the generation of new ones is balanced (type II related). The proposed control of these two polarity features by Pak1 is summarized in Figure 11.
Type I Polarity.
Nonattached cells from the vegetal mass or from the ectodermal BCR extend protrusions from all parts of their surface. At spreading, this nonpolarized, multipolar protrusive activity is transformed into a multipolar cell morphology on FN (Wacker et al.,1998; Luu et al.,2008). In contrast to all other gastrula cell types, nonattached mesendoderm cells extend protrusions from a small, active region. This polarization of protrusive activity is translated into the polarity of migrating cells during bipolar spreading on FN (Winklbauer and Selchow,1992). Our experiments show that, although nonattached cells are still polarized when Pak1 function is inhibited, protrusive activity during spreading is altered, such that many small lamellipodia are formed and maintained instead of two larger ones. Essentially, lamellipodia formation is not concentrated to two sites, but instead is spread out over the cell circumference in Pak1-compromised cells. Our results suggest that Pak1, by activating Cdc42, spatially restricts lamellipodia formation on FN.
Paks were initially identified as effectors of Cdc42 and Rac, whereas our results place Pak1 upstream of Cdc42 (Fig. 11). A similar reversal of the interaction has been observed previously. Thus, Pak1 appears to function upstream of Rac in the control of PC12 cell motility (Obermeier et al.,1998), and upstream of Cdc42 in neutrophil chemotaxis (Li et al.,2003). Also, induced EphA4 signaling in Xenopus early blastomeres involves down-regulation of Cdc42 function by Pak1, probably by sequestering active Cdc42 (Bisson et al.,2007). However, in these cases, Pak1 served as a scaffold only, with its kinase function not being required. By contrast, we found that kinase activity is in fact necessary in our experiments.
This could still be compatible with a function of Pak1 as a dynamic scaffold, whose catalytic activity is required to regulate the protein complex it helps to assemble. Such a mechanism has been proposed by ten Klooster et al. (2006). It involves Cdc42 binding to a paxillin/Git/Pix/Pak1 complex, where it activates Pak1 whose autophosphorylation then releases Pak1 and Cdc42 from the complex and allows Rac to bind, to become activated and to function in the control of the cytoskeleton. Involvement of this complex is also suggested by another line of reasoning. In our assay, the culture medium consists of a buffered salt solution without serum or growth factor supplements, and unless Pak1 were activated by a cell-intrinsic mechanism not depending on external cues (e.g., Wedlich-Soldner et al.,2003), the only possible source of external signaling should be the FN substratum. Mesendoderm cells interact with it through integrin α5β1 (Darribere et al.,1988; Ramos et al.,1996), and integrin ligation and clustering could potentially initiate activation of Pak1 in the context of the Paxillin/Git/Pix/Pak1 complex (Rosenberger and Kutsche,2006). Consistent with this, the punctate localization pattern of GFP-tagged Pak1-GBD is indeed similar to that of Paxillin in endothelial cells (Zaidel-Bar et al.,2003) and, most importantly, in dissociated Xenopus mesoderm cells (Iioka et al.,2007).
A potential problem with this model is that KD-Pak1 could lock the complex in an inactive state not to be rescued by active Cdc42, in contrast to our findings. As an alternative mechanism, Pak1 could promote activation of Cdc42 by acting on RhoGEFs, GAPs, or GDIs. For example, phosphorylation and inactivation of RhoGDI by Pak1 has been described (DerMardirossian et al.,2004). However, this does not explain how direct binding of Cdc42/Rac to Pak1 is necessary, as deduced from our observation that compromising the GTPase binding site on Pak1 blocks polarization. The observation that CA-Cdc42 rescues KD-Pak1 induced effects argues against the possibility that Pak1 acts upstream, but also downstream of Cdc42.
The localization patterns of GFP-tagged Pak1-GBD of normal or Pak-inhibited cells is consistent with Pak1-dependent Cdc42/Rac activation being uniform across the area of cell-substratum contact. In a simple model, one could assume that the widespread activation of Cdc42 raised a threshold for lamellipodia formation. Normally, the propensity for protrusion formation would exceed the threshold only at two sites during spreading on FN, one at the site of filopodia formation of the nonattached cell, and the other opposite to it. At a lowered threshold, such as when Pak1 function and Cdc42 activation are impeded, additional protrusions would form, leading to a multipolar morphology.
Type II Polarity.
The second aspect of mesendoderm cell polarity that is uncovered by manipulations of Pak1 function concerns the maintenance of a normal number of lamellipodia, and hence polarity, after its initial implementation. This requires an orderly, complete retraction of lamellipodia to balance the generation of new protrusions. Overexpression of CA-Pak1 changes the frequency of lamellipodia retraction such that on average more active, large protrusions are eventually present per cell.
This type II polarity phenotype seems to be due to a down-regulation of RhoA function. The transcription factors Mix.1 and Gsc indirectly activate RhoA in mesendoderm cells, which is necessary for polarization (Luu et al.,2008). Our experiments suggest that in addition to its effect on polarization, activated Cdc42 stimulates Rac, which in turn promotes lamellipodia formation, but apparently inhibits also RhoA function (Fig. 11). Cross-talk between Rho proteins was initially described by Nobes and Hall (1995). A similar cascade, where Cdc42 activates Rac, and both Cdc42 and Rac inhibit RhoA, has been reported by Sander et al. (1999) and also by Li et al. (2002). In our present context, this inhibition is probably indirect, as GFP-tagged Pak1-GBD and rhotekin-GFP do not obviously colocalize.
Active RhoA is concentrated at the proximal parts of lamellipodia. A major effector of RhoA is Rho kinase, and inhibition of its activity during migration mimicks the effect of CA-Pak1, EnRMix.1, or DN-RhoA. In turn, a main target of Rho kinase is myosin II, which is also enriched proximally in lamellipodia. Expression of CA-Pak1 diminishes both active RhoA and myosin II at this site, and at the same time impedes retraction of lamellipodia. Altogether, this suggests that activated RhoA, through Rho kinase, controls myosin II contractility and localization at the base of lamellipodia, which is required for the complete retraction of these protrusions.
Type I and II Polarity in Other Cell Types
The Pak1-dependent control of polarized protrusive activity seems to be a widespread phenomenon. As in mesendoderm cells, formation of many small protrusions upon expression of dominant-negative Pak1 (type I phenotype), and the extension of extra, normal-sized lamellipodia in a fraction of the cells by CA-Pak1 expression (type II phenotype), has also been observed in NIH-3T3 cells (Sells et al.,1999) or in prostate cancer cells (Even-Faitelson et al.,2005). The occurrence of this strikingly similar effect in very different cell types suggests a basic common mechanism.
Accordingly, Cdc42 which acts downstream of Pak1 in the establishment of type I-related polarity of mesendoderm cells, has also been implicated in the polarization of other cell types. In primary mouse embryonic fibroblasts, Cdc42 knockout leads to “multiple extended protrusions,” as well as to decreased cell movement and directional migration (Yang et al.,2006), as in mesendoderm cells. Likewise, in chemotactically migrating neutrophils, Cdc42 is required for the formation of a single stable protrusion (Srinivasan et al.,2003). RhoA is likewise involved in the polarization of other cell types. Thus, RhoA activation down-regulates the number of lamelliform protrusions in mesenchymal stromal cells (Jaganathan et al.,2007) and THP-1 monocytes (Worthylake and Burridge,2003), suggesting an increased polarization of cells. However, RhoA activation increases lamellipodia number in epithelial GE11 cells (Danen et al.,2005), suggesting context dependency of RhoA function in polarization.
Control of Directional Mesendoderm Migration by Pak1
Pak1 regulates translocation in various cell types, but its role in migration is strongly context dependent. Thus, Pak1 function is necessary for the migration of CHO, PC12, or HMEC-1 cells (Alahari et al.,2004; Kiosses et al.,1999), but it inhibits migration of rat or mouse embryo fibroblasts (Leisner et al.,2005). Moreover, a constitutively active form of Pak1 promotes migration in NIH 3T3 cells, but inhibits it in HMEC-1 cells (Sells et al.,1999; Kiosses et al.,1999). In mesendoderm cells, despite its inhibitory effect on polarity, interference with Pak1 function does not strongly alter migration velocity. If anything, both wtPak1 and KD-Pak1 expression increase the speed of translocation. Such an effect is also seen with NIH-3T3 (Sells et al.,1999) and endothelial cells (Master et al.,2001).
A surprising aspect of this findings is the fact that migration velocity is not linked to the polarization of protrusive activity. A lack of such a correlation had been noted previously when polarity control by homeodomain transcription factors had been studied (Luu et al.,2008). This raises the question of whether this polarity has a function, if not in single cells in vitro, then under the more complex conditions of mesendoderm aggregate directional migration. In the Xenopus gastrula, aggregates but not single mesendoderm cells are able to move in an oriented manner on the BCR or on BCR-conditioned substratum. This movement involves the orientation of apparently unipolar cells in the direction of the animal pole (Winklbauer and Nagel,1991). Restricting lamellipodia formation to a single site on the cell, or the retraction of misdirected lamellipodia could both be prerequisites for overall cell orientation.
Pak1 kinase function is indeed required for substratum-guided directional migration. In an in vitro migration assay, mesendoderm movement is randomized upon KD-Pak1 expression, and in the embryo, cells become disoriented, pointing with their protrusions in any direction and not only anteriorly. This abrogation of cell orientation in vivo could be related to the inhibition of polarized protrusive activity of single cells. CA-Pak1 affects directional migration only weakly, as does DN-RhoA (Ren et al.,2006), suggesting that maintenance of mesendoderm cell polarization by RhoA-dependent retraction of protrusions, as observed with single cells, is not essential. The relatively strong inhibition of directionality by CA-Cdc42 could be due to additional effects of this construct. On the other hand, DN-Cdc42 which inhibits single cell polarity similarly to KD-Pak1, by affecting the primary restriction of lamellipodia extension, significantly inhibits the directionality of migration. This suggests that cell polarity, expressed as an ability to restrict lamellipodia formation in single cells, may be permissive for directional migration in this system. However, the KD-Pak1 effect is stronger than that of DN-Cdc42, and Pak1 possibly serves other, Cdc42-independent functions in this process, e.g., by mediating signaling from the extracellular guidance cues.
A role for Pak1 in the chemotactic orientation of cells has been shown for neutrophils (Li et al.,2003), Dictyostelium (Chung et al.,2001; Lee et al.,2004), and for axon guidance in Drosophila (Hing et al.,1999). In this context, Pak1 acts downstream of GPCRs. A member of this receptor family, CXCR4 is expressed in the migrating mesendoderm, has the potential to act as a chemoattractant, and is involved in the control of its movement in the gastrula (Fukui et al.,2007).
In the Xenopus gastrula, PDGF-A and the PDGFRα are essential for directional migration. Pak1 has been implicated in PDGF receptor signaling in various systems. In cultured cells, PDGF can promote Pak1 activation (Yoshii et al.,1999; del Pozo et al.,2000; Howe,2001), and at least in NIH3T3 cells, this activation is anchorage-dependent and requires adhesion to FN (del Pozo et al.,2000; Howe,2001). In the Xenopus gastrula, the BCR is covered by a network of FN fibrils, and an ECM-binding form of PDGF-A is expressed in the BCR (Jones et al.,1993; Ataliotis et al.,1995). Moreover, PDGF treatment of mesendoderm stimulates its spreading on FN in vitro (Symes and Mercola,1996), and Pak1 has been suggested to mediate this response (Yoshii et al.,1999).
In the embryo, the anteriorly directed orientation of migrating cells is abolished upon inhibition of PDGF signaling, and migration on conditioned substratum in vitro is randomized (Nagel et al.,2004). Pak1 inhibition causes very similar effects, consistent with a Pak1 function in mesendoderm guidance downstream of the PDGF receptor. Moreover, PDGF-dependent directional migration is sensitive to inhibition of phosphatidylinositol 3 kinase (PI3K; Nagel et al.,2004), and PI3K has been shown to regulate Pak1 function during chemotaxis (Chung et al.,2001; Li et al.,2003). We propose that in vivo, Pak1 is activated in the layer of mesendoderm cells that is in direct contact with the FN matrix and the PDGF bound to it. Consistent with this, only the cells directly in contact with the BCR appear to require Pak1 activity for migration in vivo, whereas the deeper cells of this population, which advance by moving between neighboring mesendoderm cells, are independent of Pak1 function.
In summary, although Pak1 kinase function is not required for the translocation of single mesendoderm cells, it is needed for the directional movement of mesendodermal cell aggregates, and for the proper orientation of cells within such aggregates. Pak1-promoted cell polarization could be a permissive prerequisite for cell orientation. Alternatively, or in addition, it could be involved in the processing of PDGF-derived signaling cues that are essential for directional migration. Inhibiting Pak1 function leads also to an arrest of mesendoderm translocation relative to the BCR. It is not clear whether this is completely due to the disorientation of cells and lack of directionality of movement, or whether additional factors play a role. For example, in KD-Pak1–expressing embryos, the BCR is firmly adherent to the mesendoderm, and the normal gaps between mesendoderm cells are reduced, suggesting that adhesive interactions between mesendoderm and the BCR, and also within the mesendoderm, are probably altered. We are presently analyzing these questions, to gain a better understanding of Pak1 function in an organismal context.
Embryos and Microsurgical Operations
Xenopus laevis embryos from induced spawnings (Winklbauer,1990) were staged according to Nieuwkoop and Faber (1967). Operation techniques and buffer (Modified Barth's Solution, MBS) have been described (Winklbauer,1990; Winklbauer and Schurfeld,1999). DiI labeling was performed by soaking a mounted eyelash in DiI/sucrose solution (Ibrahim and Winklbauer,2001; modified after Gont et al.,1993) and inserting it for a few seconds into the embryo. To assay directional migration, substrate was conditioned according to Nagel and Winklbauer (1999). Stage 10 BCR explants were held against the bottom of Greiner tissue culture dishes for 2 hr. After removal of the BCR, substratum was saturated with 50 mg/ml bovine serum albumin (BSA). Displacement of mesendoderm explants (approximately 200 cells per explant) on conditioned substratum was observed by recording explant positions initially, and 1 hr later. Leading edge mesendoderm was dissociated into single cells by incubation in Ca++ and Mg++ free MBS for 20 min. To test for convergent extension, dorsal blastopore lips from control and KD-Pak1–injected stage 10 gastrulae were explanted in MBS, and cultured overnight.
Staining of the Actin Cytoskeleton, Myosin, and Active G-Proteins
Dissociated leading edge mesendoderm cells were seeded on FN-coated (200 μg/ml; bovine plasma fibronectin, Sigma-Aldrich) glass coverslips for 1 hr, and fixed in 4% formaldehyde for 20 min. 5 min after initial formaldehyde exposure, Triton-100× (0.1% final concentration) was added to the fixative. Cells were washed in 1% BSA solution, and stained with 5 μg/ml rhodamine-phalloidin for 20 min to visualize F-actin. For immunostaining of nonmuscle-myosin II, cells were incubated in rabbit anti-Xenopus laevis myosin heavy chain serum (gift from Igor David) for 30 min, washed in phosphate buffered saline (PBS)/BSA and stained with fluorescein isothiocyanate (FITC) -conjugated goat anti-rabbit antibody for 30 min. The GFP-PakGBD and pEGFP-rhotekinGBD was stained with anti-GFP rabbit IGG (Molecular Probes) for 30 min, washed in PBS/BSA and incubated in FITC-conjugated goat anti-rabbit IGG for 30 min. Coverslips were mounted on slides with SlowFade mounting solution (Molecular Probes). Images were taken on a Zeiss Axiovert Microscope equipped with epifluorescence optics. To quantitate cell polarization, spindle-shaped or triangular cells were counted as polarized, cells with more than three processes as multipolar, based on previous detailed descriptions of mesendoderm cell behavior (Winklbauer and Selchow,1992).
Measurement of Angles of Forming Protrusions
Time-lapse images of dissociated endomesodermal cells seeded on BSA-coated surface were taken at 1 picture per minute for 2 hr with ×20 magnification and Nomarsky optics. The 30° angles between 0° and 180° were copied to a transparency and the angles of the forming protrusions were then measured on the screen from approximately 15 cells per experiment. Every new protrusion was set to 0°, and the angle of the next forming protrusion was measured. The number of measured protrusions per experiment was set to 100%.
Scanning Electron Microscopy
Embryos were fixed in 2.5% formaldehyde and 2.5% glutaraldehyde in MBS, and BCRs were removed with needles. Specimens were post-fixed in 2% OsO4, dehydrated in a graded ethanol series, substitution of ethanol by hexamethyldisilazane (HMDS), dried, mounted on stabs with carbon surface, and sputter coated with gold-palladium.
In Situ Hybridization
In situ hybridization was modified after Harland (1991). Alkaline phosphatase substratum was BM Purple (Roche). For the antisense Xbra digoxigenin probe, plasmid DB30 (gift of M. Sargent, London) was linearized with BglII and used as a template for transcription with T7 RNA polymerase. Gsc probe (Cho et al.,1991; gift of H. Steinbeisser, Heidelberg) was generated by linearizing clone H7 in pBluescript SK(−) with EcoRI and transcription with T7 polymerase.
Constructs, mRNA Synthesis, and Injection
Plasmids pT7TSHA (pT7TS modified by the insertion of an HA epitope tag; Poitras et al.,2003b) containing wild-type xPak1, a kinase-dead mutant K281A of xPak1 (KD-Pak), a constitutively active xPak1 construct, L98F (CA-Pak), GTPase binding-deficient xPak1 H74,77L, and kinase-dead, GTPase binding site deleted construct, xPakKDΔGBD (Poitras et al.,2003a; Bisson et al.,2003) were linearized with XbaI and transcribed with T7 polymerase. Synthetic mRNA transcripts encoding the dominant negative (RhoN19, RacN17, Cdc42N17) or constitutively active forms (RhoV14, RacV12, Cdc42V12) of RhoA, Rac1and Cdc42 were transcribed from the SP6 promoter of pCS2-RhoN19, pCS2- RacN17, pCS2-Cdc42N17, pCS2-RhoV14, and pCS2- RacV12, and pCS2-Cdc42V12, respectively, using the mMessage mMachine kit from Ambion. pBSRN3-EnRMix.1 (construct from P. Lemaire) was linearized by sfi1 and transcribed by T3 RNA polymerase. pEGFP-rhotekinGBD was linearized with NotI and transcribed with SP6. GFP-PakGBD was constructed by fusing to enhanced GFP (EGFP; Clontech) amino acids 61–123 of xPak1 and subcloning in pT7TS-HA at EcoRI and SpeI sites. GFP-PakGBD was linearized with Xbal and transcribed with T7 polymerase. Embryos were injected at the 4-cell stage marginally into the dorsal two blastomeres with 400 pg of mRNA per blastomere, unless stated otherwise. Rho kinase inhibitor Y-27632 from Sigma was used at a concentration of 10 μM in MBS.
Time-lapse movies were made from dissociated leading edge mesendodermal cells at an interval of 1 picture per minute for 1 hr under a Zeiss Axiovert 200M Microscope, using the AxioVision software.
We thank Drs. M. Sargent, H. Steinbeisser, P. Lemaire, K. Symes, and Y. Sasai for plasmids. R.W. was funded by a Canadian Institutes of Health Research grant and T.M. received an operating grant from the National Cancer Institute of Canada with funds from the Canadian Cancer Society. N.B. received a National Cancer Institute of Canada studentship with funds from the Terry Fox Foundation. The CHUQ-HDQ Research Centre is supported by funds from the FRSQ, Quebec.
- 2004. The integrin-binding protein Nischarin regulates cell migration by inhibiting PAK. EMBO J 23: 2777–2788. , , .
- 1995. PDGF signaling is required for gastrulation of Xenopus laevis. Development 121: 3099–3110. , , , , .
- 2005. A microtubule-dependent zone of active RhoA during cleavage plane specification. J Cell Biol 170: 91–101. , , .
- 2003. The catalytic domain of xPAK1 is sufficient to induce myosin II dependent in vivo cell fragmentation independently of other apoptotic events. Dev Biol 263: 264–281. , , , , , .
- 2007. EphA4 signaling regulates blastomere adhesion in the Xenopus embryo by recruiting Pak1 to suppress Cdc42 function. Mol Biol Cell 18: 1030–1043. , , , , .
- 2003. Biology of the p21-activated kinases. Annu Rev Biochem 72: 743–781. .
- 2004. Pak-1 expression increases with progression of colorectal carcinomas to metastasis. Clin Cancer Res 10: 3448–3456. , , , , , , , , , .
- 2000. Regulation of Xenopus p21-activated kinase (X-PAK2) by Cdc42 and maturation promoting factor controls Xenopus oocyte maturation. J Biol Chem 275: 2367–2375. , , , , , .
- 1991. Molecular nature of Spemann's organizer: the role of the Xenopus homeobox gene goosecoid. Cell 67: 1111–1120. , , , .
- 2001. Control of cell polarity and chemotaxis by Akt/PKB and PI3 kinase through the regulation of PAKa. Mol Cell 7: 937–947. , , .
- 2005. Integrins control motile strategy through a Rho-cofilin pathway. J Cell Biol 169: 15–26. , , , , , , .
- 1988. The 140-kDa fibronectin receptor complex is required for mesodermal cell adhesion during gastrulation in the amphibian Pleurodeles waltlii. Dev Biol 126: 182–194. , , , .
- 2000. Adhesion to the extracellular matrix regulates the coupling of the small GTPase Rac to its effector PAK. EMBO J 19: 2008–2014. , , , , .
- 2004. Phosphorylation of RhoGDI by Pak1 mediates dissociation of Rac GTPase. Mol Cell 15: 117–127. , , .
- 2005. PAK1 regulates myosin II-B phosphorylation, filament assembly, localization and cell chemotaxis. Cell Signal 17: 1137–1148. , , .
- 1997. A member of the Ste20/PAK family of protein kinases is involved in both arrest of Xenopus oocytes at G2/prophase of the first meiotic cell cycle and in prevention of apoptosis. EMBO J 16: 5550–5561. , , , .
- 2005. Xenopus p21-activated kinase 5 regulates blastomere adhesive properties during convergent extension movements. Dev Biol 277: 472–492. , , , , , , .
- 2007. SDF-1 alpha regulates mesendodermal cell migration during frog gastrulation. Biochem Biophys Res Commun 354: 472–477. , , , , .
- 1993. Tail formation as a continuation of gastrulation: the multiple cell populations of the Xenopus tailbud derive from the late blastopore lip. Development 119: 991–1004. , , , .
- 1991. In situ hybridization: an improved whole-mount method for Xenopus embryos. Methods Cell Biol. 36: 685–695. .
- 1999. Pak functions downstream of Dock to regulate photoreceptor axon guidance in Drosophila. Cell 97: 853–63. , , , , .
- 2004. The genetics of Pak. J Cell Sci 117: 4343–4354. , , .
- 2001. Cell adhesion regulates the interaction between Nck and p21-activated kinase. J Biol Chem 276: 14541–14544. .
- 2001. Mechanisms of mesoderm internalization in the Xenopus gastrula: lessons from the ventral side. Dev Biol 240: 108–122. , .
- 2007. Wnt signalling regulates paxillin ubiquination essential for mesodermal cell motility. Nat Cell Biol 9: 813–821. , , , .
- 2000. The cytoskeletal effector xPAK1 is expressed during both ear and lateral line development in Xenopus. Int J Dev Biol 44: 245–248. , , .
- 2007. Rho inhibition induces migration of mesenchymal stromal cells. Stem Cells 25: 1966–1974. , , , , , , .
- 1993. The Xenopus platelet-derived growth factor α receptor: cDNA cloning and demonstration that mesoderm induction establishes the lineage-specific pattern of ligand and receptor gene expression. Dev Genet 14: 185–193. , , , , , .
- 2002. Shaping the vertebrate body plan by polarized embryonic cell movements. Science 298: 1950–1954. .
- 1999. A role for p21-activated kinase in endothelial cell migration. J Cell Biol 147: 831–843. , , , , .
- 2002. Emerging functions of p21-activated kinases in human cancer cells. J Cell Physiol 193: 133–144. , .
- 1984. Temporal and spatial regulation of fibronectin in early Xenopus development. Cell 36: 729–740. , , .
- 2004. Dictyostelium PAKc is Required for Proper Chemotaxis. Mol Biol Cell 15: 5456–5469. , , , , , ,
- 2005. Essential role of CIB1 in regulating PAK1 activation and cell migration. J Cell Biol 170: 465–476. , , , , .
- 2002. p21-activated kinase 1 interacts with and phosphorylates histone H3 in breast cancer cells. EMBO Rep 3: 767–773. , , , , , , , .
- 2003. Directional sensing requires Gβγ-mediated PAK1 and PIXα-dependent activation of Cdc42. Cell 114: 215–227. , , , , , , , , , , , .
- 1997. Activation of Pak by membrane localization mediated by an SH3 domain from the adaptor protein Nck. Curr Biol 7: 85–94. , , , .
- 2008. Control of gastrula cell motility by the Goosecoid/Mix.1/ Siamois network: basic patterns and paradoxical effects. Dev Dyn 237: 1307–1320. , , , , .
- 1994. A brain serine/threonine protein kinase activated by Cdc42 and Rac1. Nature 367: 40–46. , , , , .
- 2001. Dok-R plays a pivotal role in angiopoietin-1-dependent cell migration through recruitment and activation of Pak. EMBO J 20: 5919–5928. , , , , , .
- 1999. Establishment of substratum polarity in the blastocoel roof of the Xenopus embryo. Development 126: 1975–1984. , .
- 2004. Guidance of mesoderm cell migration in the Xenopus gastrula requires PDGF signaling. Development 131: 2727–2736. , , , .
- 1967. Normal table of Xenopus laevis (Daudin). Amsterdam: North-Holland. , .
- 2004. Antero-posterior tissue polarity links mesoderm convergent extension to axial patterning. Nature 430: 364–367. , , .
- 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. , .
- 1998. PAK promotes morphological changes by acting upstream of Rac. EMBO J 17: 4328–4339. , , , , , .
- 2003a. PAK interacts with NCK and MLK2 to regulate the activation of jun N-terminal kinase. FEBS Lett 543: 129–135. , , , .
- 2003b. A tissue restricted role for the Xenopus Jun N-terminal kinase kinase kinase MLK2 in cement gland and pronephric tubule differentiation. Dev Biol 254: 200–214. , , , .
- 1996. Xenopus embryonic cell adhesion to fibronectin: position-specific activation of RGD/synergy site-dependent migratory behavior at gastrulation. J Cell Biol 134: 227–240. , .
- 1996. Integrin-dependent adhesive activity is spatially controlled by inductive signals at gastrulation. Development 122: 2873–2883. , , .
- 2006. Migrating anterior mesoderm cells and intercalating trunk mesoderm cells have distinct responses to Rho and Rac during Xenopus gastrulation. Dev Dyn 235: 1090–1099. , , , , .
- 2006. AlphaPIX and betaPIX and their role in focal adhesion formation. Eur J Cell Biol 85: 265–274. , .
- 1999. Rac downregulates Rho activity: reciprocal balance between both GTPases determines cellular morphology and migratory behavior. J Cell Biol 147: 1009–1022. , , , , .
- 2003. Combined array comparative genomic hybridization and tissue microarray analysis suggest PAK1 at 11q13.5-q14 as a critical oncogene target in ovarian carcinoma. Am J Pathol 163: 985–992. , , , , , , , , , .
- 1997. Structure and cytoskeletal organization of migratory mesoderm cells from the Xenopus gastrula. Cell Motil Cytoskeleton 36: 12–29. , .
- 1999. p21-activated kinase 1 (Pak1) regulates cell motility in mammalian fibroblasts. J Cell Biol 145: 837–849. , , .
- 1991. Expression of a Xenopus homolog of Brachyury (T) is an immediate-early response to mesoderm induction. Cell 67: 79–87. , , , , .
- 2002. Xpak3 promotes cell cycle withdrawal during primary neurogenesis in Xenopus laevis. EMBO J 21: 6429–6439. , , .
- 2003. Rac and Cdc42 play distinct roles in regulating PI(3,4,5)P3 and polarity during neutrophil chemotaxis. J Cell Biol 160: 375–385. , , , , , , , .
- 1996. Embryonic mesoderm cells spread in response to platelet-derived growth factor and signaling by phophatitylinositol 3-kinase. Proc Natl Acad Sci U S A 93: 9641–9644. , .
- 2006. Targeting and activation of Rac1 are mediated by the exchange factor beta-Pix. J. Cell Biol 172: 759–769. , , , .
- 1998. Patterns and control of cell motility in the Xenopus gastrula. Development 125: 1931–1942. , , , , .
- 2002. Convergent extension: the molecular control of polarized cell movement during embryonic development. Dev Cell 2: 695–706. , , .
- 2003. Spontaneous cell polarization through actomyosin-based delivery of the Cdc42 GTPase. Science 299: 1231–1235. , , , .
- 1990. Mesoderm cell migration during Xenopus gastrulation. Dev Biol 142: 155–168. .
- 1998. Conditions for fibronectin fibril formation in the early Xenopus embryo. Dev Dyn 212: 335–345. .
- 1996. Fibronectin, mesoderm migration, and gastrulation in Xenopus. Dev Biol 148: 573–589. , .
- 1991. Directional mesoderm cell migration in the Xenopus gastrula. Dev Biol 148: 573–589. , .
- 1999. Vegetal rotation, a new gastrulation movement involved in the internalization of the mesoderm and endoderm in Xenopus. Development 126: 3703–3713. , .
- 1992. Motile behavior and protrusive activity of migratory mesoderm cells from the Xenopus gastrula. Dev Biol 150: 335–351. , .
- 1995. Fibronectin fibril growth in the extracellular matrix of the Xenopus embryo. J Cell Sci 108: 1575–1586. , .
- 2003. RhoA and ROCK promote migration by limiting membrane protrusions. J Biol Chem 278: 13578–13584. , .
- 2006. Gene targeting of Cdc42 and Cdc42GAP affirms the critical involvement of Cdc42 in filopodia induction, directed migration, and proliferation in primary mouse embryonic fibroblasts. Mol Biol Cell 17: 4675–4685. , , .
- 1999. αPIX nucleotide exchange factor is activated by interaction with phosphatidylinositol 3-kinase. Oncogene 18: 5680–5690. , , , , , , , , , .
- 2003. Early molecular events in the assembly of matrix adhesions at the leading edge of migrating cells. J Cell Sci 116: 4605–4613. , , , .