Establishment of cellular polarity and control of the orientation of the mitotic spindle during cell division are important processes in embryonic development. In the zebrafish Danio rerio, formation of the neural tube occurs by means of infolding of the planar neural plate to form the neural keel, which subsequently rounds up. During neural plate and neural tube stages, cells divide parallel to the plane of the neuroepithelium, but in the intervening neural keel/rod stage, the mitotic spindle rotates by 90 degrees and comes to lie perpendicular to the apical surface of the neuroepithelium (Geldmacher-Voss et al., 2003). During this stage, most, if not all, daughter cells are distributed, therefore, bilaterally across the midline in the early neural anlage (Kimmel and Warga, 1986; Schmitz et al., 1993; Kimmel et al., 1994; Papan and Campos-Ortega, 1994, 1997; Concha and Adams, 1998; Geldmacher-Voss et al., 2003). The mechanisms and molecules that determine the parallel orientation of the mitotic spindle in the neuroepithelial cells of the neural plate and neural tube and the reasons for the rotation of the mitotic spindle in the intermediate neural keel/rod stage are yet not understood. Of interest, cell polarity is established progressively during neurulation in zebrafish embryos, as revealed by investigations of the localization of several characteristic apical markers such as ASIP/PAR-3:EGFP, aPKC, β-Catenin, and ZO-1 (Geldmacher-Voss et al., 2003). ASIP/PAR-3:EGFP, for example, is distributed diffusely at the plasma membrane of neuroepithelial cells in the neural plate, with little or no evidence for preferential localization at apical levels. Localization at the apical pole of the cells first appears during the neural keel/rod stage and becomes quite obvious in the neural tube (Geldmacher-Voss et al., 2003).
PAR-3, which together with PAR-6 and aPKC forms the evolutionarily conserved PAR/aPKC complex (Ohno, 2001; Etienne-Manneville and Hall, 2003; Henrique and Schweisguth, 2003), has been shown to be involved in the establishment of cell polarity and the orientation of the mitotic spindle during asymmetric cell divisions in Caenorhabditis elegans and Drosophila (Horvitz and Herskowitz, 1992; Knoblich, 2001; Wodarz and Huttner, 2003; Betschinger and Knoblich, 2004; Roegiers and Jan, 2004). In C. elegans, the PAR/aPKC complex localizes to the anterior cortex in the zygote, P0, whereas PAR-1 and PAR-2 localize to the posterior cortex before the first cell division, which is asymmetric and results in a large anterior cell, AB, and a small posterior cell, P1 (Ahringer, 2003). par-3, par-6, and pkc-3 mutant embryos show defects in the polarization of the zygote and in the localization and orientation of the mitotic spindle during the initial cell division and the subsequent cleavages of the blastomeres (Kemphues et al., 1988; Cheng et al., 1995; Etemad-Moghadam et al., 1995; Doe and Bowerman, 2001). In Drosophila neuroblasts, Bazooka (the Drosophila homologue of C. elegans PAR-3) colocalizes with DmPAR-6 and DaPKC to the apical cortex before the onset of asymmetric divisions in the ventral neurogenic region, thereby recruiting Inscuteable, a key player in orienting the subsequent asymmetric neuroblast cell divisions (Kraut and Campos-Ortega, 1996; Kraut et al., 1996; Kuchinke et al., 1998; Schober et al., 1999; Wodarz et al., 1999; Kaltschmidt et al., 2000). In analogy to C. elegans, in bazooka, DmPAR-6, and DaPKC mutant Drosophila embryos, the acquisition of cell polarity is impaired severely and the mitotic spindle fails to align along the apicobasal axis of the neuroblast (Kuchinke et al., 1998; Wodarz et al., 2000; Petronczki and Knoblich, 2001). Furthermore, in mutants for the PAR/aPKC complex, the localization of cell fate determinants such as Numb (Uemura et al., 1989; Rhyu et al., 1994) and Prospero (Doe et al., 1991; Vaessin et al., 1991; Hirata et al., 1995; Knoblich et al., 1995; Spana and Doe, 1995), and their respective adaptor proteins Partner of Numb (Lu et al., 1998) and Miranda (Ikeshima-Kataoka et al., 1997; Shen et al., 1997; Schuldt et al., 1998), to the basal cell cortex during metaphase is perturbed. This disturbance in turn compromises their normally exclusive distribution to the ganglion mother cell (GMC; Schober et al., 1999; Wodarz et al., 1999, 2000; Petronczki and Knoblich, 2001).
In D. rerio, PAR-3 or ASIP (atypical PKC isotype-specific interacting protein), as the mammalian homologue of C. elegans PAR-3 was initially named (Izumi et al., 1998), is expressed in three different isoforms: ASIP/PAR-3, Pard3a, and Pard3b (Fig. 1; Geldmacher-Voss et al., 2003; Wei et al., 2004). All three isoforms exhibit the conserved modular structure typical for members of the PAR-3 family, in which three conserved regions (CRs) can be distinguished: the oligomerization domain at the N-terminus (CR1), the three PDZ domains (CR2), and the aPKC binding domain (CR3). As expected from the presence of these motifs, several studies in different organisms have confirmed that PAR-3 participates in various protein–protein interactions. Thus, the CR1 domain has been shown to self-associate and to be necessary for apical membrane localization in Drosophila follicular epithelium and in Madin–Darby canine kidney (MDCK) cells (Benton and St Johnston, 2003a; Mizuno et al., 2003). The PDZ domains have been shown to interact with PAR-6 (Joberty et al., 2000; Lin et al., 2000), the so-called junctional adhesion molecules (JAMs; Itoh et al., 2001; Ebnet et al., 2001, 2003), the adherens junction components nectin-1/-3 (Takekuni et al., 2003), and the lipid phosphatase PTEN (von Stein et al., 2005). Furthermore, coiled-coil regions in the C-terminal part of the protein have been implicated in the binding of the Rac-specific guanine nucleotide-exchange factors (GEFs) TIAM1 and STEF (Chen and Macara, 2005; Nishimura et al., 2005; Mertens et al., 2005) and the microtubule motor protein KIF3 (Nishimura et al., 2004).
Here, we show that, in zebrafish, the oligomerization domain CR1 and the PDZ domains of ASIP/PAR-3:EGFP are involved in its localization to the apical membrane in cells of the neural tube. We further show that the C-terminal portion of ASIP/PAR-3 contributes to its proper localization and that the apical localization signals that are present in the N-terminal part of ASIP/PAR-3 prevent the basolateral localization of a Numb:PAR-3 fusion protein. The parallel orientation of the mitotic spindle in the neural tube, however, is only weakly disturbed upon overexpression of various ASIP/PAR-3:EGFP constructs.
RESULTS AND DISCUSSION
We carried out a deletion analysis of ASIP/PAR-3 to determine which of the various conserved domains in the protein are required for its apical localization. For this purpose, we injected mRNA coding for different truncated versions of ASIP/PAR-3 (fused to EGFP as a reporter) into the yolk of wild-type embryos, and subsequently studied the localization of the fusion proteins in 22 to 28 hours post fertilization (hpf) old embryos in vivo by confocal laser scanning microscopy. Because the establishment of cell polarity occurs progressively, the deletion analysis was restricted to the neural tube stage, when the neuroepithelial cells are clearly polarized and all cell divisions exhibit a clearly defined and invariant orientation. It is worthwhile noting here that, in addition to the strong apical labeling observed upon injection of mRNA coding for ASIP/PAR-3:EGFP, weak cytoplasmic labeling is always seen at this stage. Injection of the mRNA into early embryos results in an almost ubiquitous but nevertheless mosaic expression of the corresponding protein. Because the cells tend to accumulate high levels of the encoded fusion protein, the cytoplasmic labeling can most simply be explained by the assumption that the amount of protein that can be anchored at the apical surface is restricted. This idea is supported by recent work in Drosophila by Harris and Pfeifer (2005), who showed that Bazooka is anchored by a saturable apical scaffold during blastoderm cellularization.
The PDZ Domains Are Necessary for the Localization of ASIP/PAR-3:EGFP to the Apical Cortex
The first PDZ domain of PAR-3 has been shown to bind to the JAM proteins of the tight junction (TJ) and to the adherens junction (AJ) components nectin-1 and nectin-3 (Itoh et al., 2001; Takekuni et al., 2003; Ebnet et al., 2004). We, therefore, hypothesized that PDZ1, and perhaps the PDZ domains 2 and 3, could mediate the apical localization of ASIP/PAR-3:EGFP, and we started our analysis by deleting each of the PDZ domains in turn. Of interest, deletion of any one of the three PDZ domains had no obvious effect on the apical localization of the corresponding fusion protein in neuroepithelial cells of the neural tube (Fig. 2A–C). All three deletion constructs localized at the apical cortex in essentially the same way as the full-length ASIP/PAR-3:EGFP protein (Fig. 3A and Supplementary Movie S1, which can be viewed at http://www.interscience.wiley.com/jpages/1058-8388/suppmat). Only when all three PDZ domains were deleted was the apical localization severely impaired (Fig. 2D), although a considerable portion of the fusion protein still localized to the apical cortex (arrow in Fig. 2D). These results show that the PDZ domains are indeed necessary for the localization of ASIP/PAR-3:EGFP to the apical cortex. However, each individual PDZ domain is dispensable. Hence, any combination of two PDZ domains is sufficient to mediate apical localization—making it likely that they act redundantly, at least in part.
Of interest, the first PDZ domain, which has itself been shown to bind to nectin-1/-3 and the junctional adhesion molecules JAM-A, JAM-B, and JAM-C (Itoh et al., 2001; Ebnet et al., 2001; Takekuni et al., 2003; Ebnet et al., 2003; Muller, 2003), is dispensable for localization to the apical cortex (Fig. 2A). In accordance with this finding, Chen and Macara (2005) have demonstrated recently that PAR-3 fragments lacking the first PDZ domain were still able to localize normally in transfected MDCK cells. JAMs are members of the CD2 subgroup of the immunoglobulin (Ig) superfamily with two extracellular Ig-like domains, a single transmembrane domain, and a cytoplasmic tail that contains a canonical PDZ binding domain (Martin-Padura et al., 1998; Ebnet et al., 2004). In a variety of mammalian epithelial and endothelial cells, JAMs are specifically enriched at tight junctions, where they colocalize with ZO-1 and the PAR/aPKC complex (Izumi et al., 1998; Ebnet et al., 2001, 2003, 2004; Suzuki et al., 2001). Nectins are Ca2+-independent, Ig-like cell–cell adhesion molecules that can bind to the actin filament-binding protein Afadin. Nectins regulate a variety of cellular functions, including adhesion, migration, and polarization. The nectin–afadin interaction, for example, is involved in the organization of E-cadherin–based AJs and claudin-based TJs in epithelial cells (Takekuni et al., 2003; Nakanishi and Takai, 2004). Both the JAM and the nectin interactions would in principle provide a means by which ASIP/PAR-3:EGFP could be localized to the apical cortex. We were surprised, therefore, to find that the first PDZ domain is dispensable for localization. However, although it has been shown that the PAR-3/JAM interaction predominantly relies on the PDZ1 domain, there is some evidence that the third PDZ domain is also capable of mediating a weak interaction with JAM-A, -B, and -C (Itoh et al., 2001; Ebnet et al., 2001, 2003). Whether this weak ability of PDZ3 to interact with JAMs (and possibly also with nectins) accounts for the proper localization to the apical cortex of ASIP/PAR-3:EGFP species lacking PDZ1 remains to be demonstrated.
Localization of ASIP/PAR-3 may also involve interaction with the GEFs TIAM1 and STEF (TIAM2; Chen and Macara, 2005; Nishimura et al., 2005). During the formation of intercellular contacts in epithelial cells, primordial adhesions (PAs), which contain components of AJs (E-cadherin, ZO1), TJs (JAM-A, occludin), and nectins, are built up into mature, belt-like AJs and TJs (Suzuki et al., 2002). It is thought that the recruitment of the PAR/aPKC complex to PAs, followed by the activation of aPKC as a result of the binding of activated Cdc42 and Rac1 to PAR-6, leads to the maturation of the TJs. Recent results indicate that TJ formation and polarity in epithelial cells are controlled by TIAM1-mediated activation of Rac and, furthermore, that association of TIAM1 with the PAR/aPKC complex is important in this context (Mertens et al., 2005). The interaction between ASIP/PAR-3 and TIAM1/STEF seems to occur mainly through coiled-coil regions in the C-terminal portion of the former (amino acid [aa] 937-1038 of rat PAR-3), but a fragment comprising PDZ1 and PDZ2 is also able to immunoprecipitate endogenous TIAM1 and STEF from rat brain lysates (Nishimura et al., 2005). Of interest, in a direct binding study, only the two PDZ domains together were able to bind to STEF; neither PDZ1 or PDZ2 alone, nor the junctional sequence between them, could bind (Nishimura et al., 2005).
Localization of ASIP/PAR-3:EGFP to the Apical Cortex Involves the CR1 Domain
That even the deletion of all three PDZ domains does not eliminate apical localization completely prompted us to analyze whether the conserved N-terminal CR1 domain is required for the process. The CR1 domain has been shown to mediate oligomerization in vitro and in vivo and to be necessary for apical membrane localization in cells of the follicular epithelium of Drosophila and in MDCK cells (Benton and St. Johnston, 2003a; Mizuno et al., 2003). As depicted in Figure 3B,C, deletion of the CR1 domain indeed impairs localization to the apical cortex. However, when compared with the full-length ASIP/PAR-3:EGFP protein, the effect on localization was rather mild: although the amount of EGFP label in the cytoplasm increased (arrowheads in Fig. 3B,C), strong labeling was still present at the apical cortex (arrows in Fig. 3B,C). Therefore, neither deletion of all three PDZ domains nor deletion of the CR1 domain alone completely abolishes apical localization. Furthermore, while deletion of the CR1 domain only weakly affects apical localization, deletion of any single PDZ domain does not affect apical localization at all. However, upon deletion of aa 1 to 363, which encompasses the CR1 domain and the first PDZ domain, apical localization is lost almost completely. Only a minor fraction of this fusion protein can still be found at the apical cortex; the vast majority of the protein is localized in the cytoplasm (Fig. 3D). Of interest, deletion of the entire N-terminal part of ASIP/PAR-3, including the CR1 domain and all three PDZ domains, enhances this phenotype, resulting in a complete loss of apical labeling and localization of the remaining C-terminal part to the cytoplasm (Fig. 3E and Supplementary Movie S2). The only difference between the ASIP-Δ3 construct (lacking aa 1–363; Fig. 3D) and the ASIP-ΔPDZ1 construct (Fig. 2A) is the presence of the N-terminal region (aa 1–273), which includes the CR1 domain. Notably, a fragment consisting of amino acids 1 to 273 alone (ASIP-Δ5:EGFP) showed a weak but obvious ability to localize apically (Fig. 3F), with the amount of apical labeling being comparable to that observed after the deletion of all three PDZ domains (Fig. 2D). This finding indicates that, in the absence of the PDZ domains, the CR1 domain (or aa 1–273), is capable of mediating apical localization, although the possibility also must be taken into account that the weak apical localization is due to CR1-mediated oligomerization of the fusion protein with the endogenous ASIP/PAR-3. In summary, neither deletion of the three PDZ domains altogether (Fig. 2D) nor deletion of the CR1 domain alone (Fig. 3B,C) can completely prevent apical localization. Furthermore, neither the PDZ domains (Fig. 3B,C) nor the CR1 domain alone (Figs. 2D, 3F) can establish the degree of apical localization observed with the full-length ASIP/PAR-3:EGFP fusion protein (Fig. 3A). Thus, apical localization requires both the CR1 domain and the PDZ domains. Based on its capability to self-associate, it seems likely that the CR1 domain fosters localization of ASIP/PAR-3:EGFP to the apical membrane by generating oligomeric complexes, whereas the PDZ domains could function in tethering these complexes to apical anchors.
Amino Acids 688 to 1127 Contribute to the Proper Localization of ASIP/PAR-3:EGFP
Despite the importance of the N-terminal part of ASIP/PAR-3 (aa 1–667) for apical localization (Fig. 3E), the C-terminal part of the protein also exercises an influence. Although apical localization was maintained upon deletion of amino acids 688 to 1127, EGFP fluorescence was distributed over the entire cell cortex, as well as being found in highly dynamic fiber-like structures within the cytoplasm (Fig. 3G and Supplementary Movie S3). The fibers possibly represent microtubules, since strong labeling of the mitotic spindle was also observed occasionally (arrowhead in Fig. 3G). This phenotype is not due to loss of the aPKC binding domain (aa 810–824), because a fusion protein lacking only this region was localized in essentially the same way as the full-length ASIP/PAR-3:EGFP protein (Fig. 3H). Of interest, localization along the entire cell cortex, which was never observed with ASIP/PAR-3:EGFP, was also frequently seen upon removal of the 15 amino acids that are specific for the ASIP/PAR-3 isoform and are missing in both the Pard3a and the Pard3b isoforms (arrowheads in Fig. 3I; compare Fig. 1). We, therefore, conclude that the localization of ASIP/PAR-3:EGFP to the apical cortex of neuroepithelial cells in the zebrafish neural tube can be thought of as a combinatorial process, which involves the CR1 oligomerization domain, the PDZ domains and the C-terminal part of the protein.
It is tempting to speculate that the C-terminal part of ASIP/PAR-3 facilitates the restriction of the protein to the apical cortex by directly or indirectly modulating the ability to bind to cytoskeletal/cortical cues at lateral and basal sites. In fact, there is growing evidence that proper positioning of PAR-3 involves the cytoskeleton. Positioning of Bazooka during cellularization in Drosophila, for example, was recently shown to be independent of DaPKC and DmPAR-6, requiring instead both the actin and the microtubule cytoskeleton, including a basal to apical transport system involving the minus-end–directed microtubule motor protein dynein (Harris and Pfeifer, 2005). In addition, Duncan et al. (2005) have shown that PAR-3 is able to associate with microtubules and that its cortical localization in mouse eggs depends on microfilaments. Furthermore, amino acids 1117 to 1231 of the ASIP/PAR-3 homologue from the rat have been shown to interact with KIF3A, a plus-end–directed microtubule motor protein (Nishimura et al., 2004). Of interest, most of the corresponding amino acids are missing in the short isoforms in zebrafish (ASIP/PAR-3 and Pard3b), whereas they are present in the long isoform Pard3a. Thus, the production of different isoforms is likely to represent a mechanism by which the properties of ASIP/PAR-3 can be modified depending on the cellular context. Indeed, many different isoforms of PAR-3 have been described, and the C-terminal part of the protein in particular seems to be highly variable (Chen and Macara, 2005; Duncan et al., 2005).
Moreover, because TIAM1/STEF is also able to bind to the C-terminal portion of PAR-3, confinement of ASIP/PAR-3:EGFP to the apical cortex could also involve interaction with this protein (see above). Notably, Mizuno et al. (2003) reported that aa 937–1024 of mouse PAR-3—in addition to the CR1 domain—are also required for the recruitment of the protein to the most apical side of the cell–cell contact region in epithelial MDCK cells. Because this region is almost completely identical with the C-terminal TIAM1/STEF binding region of rat PAR-3 (Nishimura et al., 2005) and is also highly conserved in zebrafish ASIP/PAR-3, proper apical localization could indeed require the interaction of the C-terminal portion of the protein with a zebrafish TIAM1/STEF homologue.
An alternative or additional way to restrict the localization of ASIP/PAR-3 to the apical membrane in zebrafish neuroepithelial cells would be to use an exclusionary rather than a recruitment mechanism. In Drosophila, phosphorylation of two conserved serine residues in Bazooka by DmPAR-1 creates 14-3-3 (PAR-5) binding sites. Binding of the 14-3-3 protein inhibits formation of the Bazooka/PAR-6/aPKC complex, thereby excluding Bazooka from the lateral membranes in follicular epithelial cells and from the posterior pole of developing oocytes (Benton and St Johnston, 2003b). Given that both 14-3-3 binding sites are conserved in zebrafish ASIP/PAR-3 (aa 142–147 and aa 877–884), this mechanism of mutual exclusion could also operate in zebrafish neuroepithelial cells to restrict the localization of ASIP/PAR-3 to the apical membrane. Furthermore, Benton and St Johnston (2003b) reported that the C-terminal 14-3-3 binding site plays the major role in restricting the localization of Bazooka to the apical membrane. In zebrafish, this binding site (aa 877–884) is located very close to the ASIP/PAR-3 isoform-specific 15 aa insertion (aa 848–862). Thus, the phenotype observed after deletion of this insertion (Fig. 3I) may possibly be indicative of a (partially) impaired interaction of ASIP/PAR-3 with 14-3-3 (PAR-5). However, whether lateral exclusion by means of 14-3-3 (PAR-5) indeed contributes to the apical localization of ASIP/PAR-3 remains to be investigated (e.g., by mutagenesis and/or deletion of the 14-3-3 binding sites).
Using a proteomics approach, Brajenovic et al. (2004) were able to construct a highly interconnected protein network around the putative human PAR orthologues MARK4 (PAR-1), PAR-3/PAR-6, and LKB1 (PAR-4). They confirmed most of the previously reported interactions, but they also identified more than 50 novel interactors. Therefore, direct or indirect interaction with yet unknown factors may well contribute to the proper localization of ASIP/PAR-3:EGFP in zebrafish neuroepithelial cells.
Apical Localization Signals in ASIP/PAR-3 Prevent the Basolateral Localization of a Numb:PAR-3 Fusion Protein
In a previous study, we showed that a zebrafish homologue of the Drosophila cell fate determinant Numb becomes localized to the basolateral membrane of some dividing cells during the progression from neural rod to neural tube stage (Reugels et al., 2006). Furthermore, by using Numb:EGFP deletion constructs, we were able to show that the first 196 aa, including an 11-aa insertion in the PTB domain, are necessary and sufficient for the localization.
In light of the defects seen in the localization of Numb in PAR/aPKC mutants in Drosophila, we were interested in probing the possible involvement of ASIP/PAR-3 in the corresponding process in zebrafish. The apical localization of ASIP/PAR-3:EGFP first becomes apparent during the neural keel/rod stage, whereas the basolateral localization of Numb:EGFP occurs slightly later. It is conceivable, therefore, that apical localization of ASIP/PAR-3—and, hence, the establishment of apical–basal polarity—might be a prerequisite for the basolateral localization of Numb in the neural tube. Apical–basal polarity, however, is not firmly established during neural keel/rod stage (Geldmacher-Voss et al., 2003). The apical localization of ASIP/PAR-3:EGFP, for example, is not maintained at the onset of cell division, because the fusion protein is redistributed uniformly around the cell cortex as the cells round up at prophase (Geldmacher-Voss et al., 2003). This observation, together with the mutually exclusive and complementary localization of the two proteins in the neural tube, prompted us to ask whether it might be possible to mislocalize ASIP/PAR-3:EGFP by fusing it to the first 196 aa of Numb, and whether this happening, in turn, would affect the proper establishment of cell polarity.
As shown in Figure 4, in neuroepithelial cells of the neural tube, neither the full length nor the truncated version of ASIP/PAR-3:EGFP lacking CR1 can be mislocalized to any significant extent (Fig. 4A,B; Supplementary Movie S4). Both fusion proteins still localize mainly apically (arrows in Fig. 4A,B), and only those portions of the protein that would probably otherwise be found in the cytoplasm (compare the weak cytoplasmic labeling in Fig. 3A) are mislocalized to the basolateral cortex of the cells (arrowheads in Fig. 4A,B). Only when the entire N-terminal part, including all the regions that have been shown to be necessary for apical localization (Fig. 3E), is removed can the residual C-terminal part be localized basolaterally in essentially the same way as Numb:EGFP alone (arrowhead in Fig. 4C). In this case, the apical cortex is almost completely devoid of the fusion protein (arrow in Fig. 4C). This finding shows that the apical localization signals that are present in the N-terminal part of ASIP/PAR-3 prevent the basolateral localization of the corresponding Numb:PAR-3 fusion protein. Taking into account that the asymmetric localization of ASIP/PAR-3:EGFP during neural keel/rod stage slightly precedes the basolateral localization of Numb:EGFP, a simple explanation would be that Nb(1-196):ASIP:EGFP already has been incorporated into a stable apical complex when the mechanisms that localize Numb to the basolateral cortex become active with the transition from the neural rod to the neural tube stage.
ASIP/PAR-3 and Symmetric Versus Asymmetric Cell Divisions
With regard to symmetric versus asymmetric divisions, it should be noted that ASIP/PAR-3:EGFP is most probably distributed equally to both daughter cells upon cell division (Supplementary Movie S1), because all neuroepithelial cells in the neural tube of D. rerio divide in a planar manner, i.e., with their mitotic spindles oriented parallel to the plane of the neuroepithelium (Geldmacher-Voss et al., 2003). However, we would like to mention the possibility that very subtle deviations from a strictly perpendicular cleavage plane could still result in unequal segregation of the protein. In neuroepithelial cells of the mouse, for example, Kosodo et al. (2004) showed that the prominin-1–positive apical membrane, corresponding to the so-called cadherin hole, can be segregated equally or unequally to the daughter cells upon cytokinesis, although the mitotic spindle shows a parallel orientation. Therefore, the orientation of the mitotic spindle may not necessarily be a secure criterion for distinguishing asymmetric from symmetric cell divisions.
Spindle Orientation Is Only Weakly Affected by Overexpression of Different ASIP/PAR-3:EGFP Fusion Proteins
Given the role of PAR-3 in orienting the mitotic spindle during asymmetric cell divisions in C. elegans and Drosophila, we wondered whether overexpression of the various ASIP/PAR-3 fusion proteins would affect the orientation of the spindle in the neural tube of zebrafish embryos. In addition to the analysis of the subcellular localization of the different EGFP-tagged ASIP/PAR-3 deletion constructs, we, therefore, also investigated their potential influence on the orientation of the mitotic spindle. Cell divisions in the trunk region of the zebrafish embryo during neurulation exhibit a stereotypic pattern: In the neural plate and neural tube, dividing cells are oriented parallel to the plane of the neuroepithelium, whereas in neural keel/rod the cells divide perpendicular to it, due to a 90-degree rotation of the mitotic spindle (Geldmacher-Voss et al., 2003). In a previous study, we observed 2,733 mitoses in the neural tube of embryos carrying the transgene insertions Tg(H2A.F/Z)kca37 and/or Tg(H2A.F/Z)kca66. In these embryos, 2,716 (99.4%) neuroepithelial cells divided parallel to the plane of the neuroepithelium; only 17 (0.6%) were found to divide obliquely (at angles of between 15 and 45 degrees with respect to the neuroepithelium), and perpendicular divisions were never observed (Geldmacher-Voss et al., 2003). These data were confirmed here: 187 mitoses in six Tg(H2A.F/Z)kca37 and/or Tg(H2A.F/Z)kca66 embryos were monitored, and all were found to occur parallel to the plane of the epithelium. However, neither the overexpression of the full-length ASIP/PAR-3:EGFP nor of the various deletion constructs considerably affected the planar orientation of the mitotic spindle (≤10% misoriented mitoses; Table 1; spindle orientation was deduced from the orientation of the cleavage plane in dividing neuroepithelial cells). Taking into account the finding that morpholino-mediated knockdown of ASIP/PAR-3 also has only a minor effect on spindle orientation (Geldmacher-Voss et al., 2003), it seems likely that ASIP/PAR-3 does not serve a direct or major function in orienting the mitotic spindle. Accumulating evidence from several different systems, including vertebrates (Sanada and Tsai, 2005; Zigman et al., 2005), indeed suggests that spindle orientation during central nervous system development is largely regulated by heterotrimeric G-protein signalling and Inscuteable rather than by the conserved PAR/aPKC complex. In Drosophila neuroblasts and sensory organ precursor cells (SOPs), for example, spindle orientation is mainly controlled by heterotrimeric G-protein signalling involving the Pins/Gαi complex and Inscuteable, whereas the PAR/aPKC complex is mainly involved in setting up cell polarity and regulating the localization of cell fate determinants (for review, see Betschinger and Knoblich, 2004; Bellaiche and Gotta, 2005; Wodarz, 2005). In D. rerio, planar cell polarity (PCP) signalling has been shown to be involved in the control of the orientated cell divisions in the dorsal epiblast during gastrulation (Gong et al., 2004). Whether heterotrimeric G-protein signalling or components of the PCP pathway are also involved in orienting the mitotic spindle in dividing neurepithelial cells of the neural tube remains to be determined. Further investigations will be needed to shed more light on the mechanisms and molecules required for the establishment of cellular polarity and in orienting the mitotic spindle during neurulation in zebrafish embryos.
Table 1. Orientation of Mitoses in the Neural Tube of Zebrafish Embryos After Injection of mRNA Coding for Different ASIP/PAR-3:EGFP Fusion Proteins
(400 ng/μl ± 20 ng/μl)
Zebrafish embryos were obtained from spontaneous spawnings. The embryos were staged according to Kimmel et al. (1995). For time-lapse movies, embryos were kept in a modified version of Embryo Rearing Medium (ERM; Westerfield, 1994). Embryos were anaesthetized with MS222, manually dechorionated and placed in holes cut in a thin layer of 1.7% agarose gel (in ERM) on a coverslip (see Cooper et al., 1999, for a detailed description of the procedure). As well as wild-type, we used embryos bearing the insertions Tg(H2A.F/Z)kca37 and/or Tg(H2A.F/Z)kca66 (Pauls et al., 2001).
ASIP/PAR-3:EGFP Fusion Proteins and mRNA Injections
ASIP/PAR-3 cDNA (accession no. AF465629) and the following primers were used in polymerase chain reactions (PCRs) to synthesize the ASIP/PAR-3 deletion constructs asip-Δ1:egfp to asip-Δ5:egfp:asip-Δ1:egfp (85-1127): 5′-GATGAACTAGTGATGCACGTTGGAGGAGACGACGG-3′, 5′-AAAGGAATTCGGTACCTGTCTGAAGTGGAGG-3′; asip-Δ2:egfp (271-1127): 5′-GTGAGATCGATGGATGATATTGTGAAATTGG-3′, 5′-AAAGGAATTCGGTACCTGTCTGAAGTGGAGG-3′; asip-Δ3: egfp (364-1127): 5′-CAACATCGATGAGGTCCCAGTATGAGCAG-3′, 5′-AAAGGAATTCGGTACCTGTCTGAAGTGGAGG-3′; asip-Δ4:egfp (668-1127): 5′-GGATATCGATGGAAGGCAACAAGCGAGGG-3′, 5′-AAAGGAATTCGGTACCTGTCTGAAGTGGAGG-3′; asip-Δ5:egfp (1-273): 5′-AGTGATCGATCTATGAAAGTGACGGTGTGTTTTG-3′, 5′-CAATATCGATATATCATCCAATGATCTCAC-3′.
ASIP/PAR-3 PCR fragments were digested with SpeI/EcoRI (asip-Δ1), ClaI/EcoRI (asip-Δ2 to asip-Δ4), or ClaI (asip-Δ5), and cloned into the SpeI/EcoRI, ClaI/EcoRI or ClaI-digested pCS2+/EGFP vector (Turner and Weintraub, 1994; Geldmacher-Voss et al., 2003), respectively. Construction of asip:egfp (full-length) and asip-Δ6:egfp (1-687) was described previously (Geldmacher-Voss et al., 2003). asip-ΔPDZ1:egfp and asip-ΔPDZ1-3:egfp were made by cloning the asip-Δ5 fragment into the ClaI-digested pCS2+/EGFP vector, containing asip-Δ3 or asip-Δ4, respectively.
The QuikChange site-directed mutagenesis kit from Stratagene (La Jolla, CA) was used to create specific deletions. Reactions were performed according to the manufacturer's guidelines. Starting from the full-length asip:egfp clone, four different constructs were synthesized using the following oligonucleotides: asip-Δ7:egfp (Δ810-824): 5′-CTAATGCTGAGGGGCAGACTAAGCAGTATGGAGATG-3′, 5′-CATCTCCATACTGCTTAGTCTGCCCCTCAGCATTAG-3′; asip-Δ8:egfp (Δ848-862): 5′-CAAGAGCATGGACCTAGGCTCTTCCACGCGTGATG-3′, 5′-CATCACGCGTGGAAGAGCCTAGGTCCATGCTCTTG-3′; asip-ΔPDZ2: egfp: 5′-GAAGGTTGGGCGGAGACTCACGCCCATGGGTGGGAC-3′, 5′-GTCCCACCCATGGGCGTGAGTCTCCGCCCAACCTTC-3′; asip- ΔPDZ3:egfp: 5′-GGGAGTTTCTGACATTAATGTCAACGGAAGGCAAC-3′, 5′-GTTGCCTTCCGTTGACATTAATGTCAGAAACTCCC-3′.
numb(1-197):asip:egfp was made by excising the full-length asip/par-3 sequence from the asip-pEGFP-N3 vector (Geldmacher-Voss et al., 2003) with EcoRI/XmaI, and cloning it into the EcoRI/PinAI-digested pCS2+/EGFP vector, containing numb(PTBLΔ197-680) (Reugels et al., 2006). The full-length asip:egfp clone and the asip-Δ5′:egfp (658-1127) clone (Geldmacher-Voss et al., 2003) were digested with AgeI/XbaI, and inserts of correct size were cloned into the AgeI/XbaI-digested numb(PTBLΔ197-680) vector (Reugels et al., 2006), resulting in nb(1-196):asip(254-1127):egfp and nb(1-196):asip(658-1127):egfp, respectively.
Capped RNA was synthesized in vitro using SP6 polymerase and mMESSAGE mMASCHINE or MEGAscript kits (Ambion, Austin, TX) from the constructs described above. mRNAs were injected in 5-nl aliquots (2 ng) into the yolk of wild-type zygotes.
Multilevel Confocal Time-Lapse Imaging
An LSM 410 confocal microscope (Zeiss) attached to an inverted Zeiss Diavert microscope with ×40 and ×63 immersion objectives was used to collect up to seven different vertical stacks of images (z-series) at intervals of 90 seconds, with each stack being separated by between 6 to 7 μm from the next. Z-series were transferred to a Macintosh Power PC or an Intel PC for image processing using ImageJ and Adobe Photoshop 7.0. The orientation of cell divisions in the neural tube was determined by first labeling the midline of the neural tube and defining it as 0 degrees and then drawing a line between the separating chromosomes at anaphase (perpendicular to the plane of cleavage) and projecting it to the midline. Cell divisions were classified as being parallel, oblique, and perpendicular to the plane of the neuroepithelium, based on measured angles of 0 degrees, 45 degrees, and 90 degrees (± 22.5 degrees), respectively.
J.v.T. thanks Professor Campos-Ortega for being a true teacher in the very best sense of the word. We greatly admired his lively and critical mind, his human spirit, and his sense of humor. We are indebted to Elisabeth Knust and thank her, Paul Hardy, and John Chandler for critical reading of the manuscript. We thank Iris Riedl, Christel Schenkel, and Thomas Wagner for expert technical assistance, and our colleagues for helpful discussions. This report is dedicated to the memory of José A. Campos-Ortega.