Nuclear localization of the zebrafish tight junction protein nagie oko

Abstract

The tight junctions-associated MAGUK protein nagie oko is closely related to Drosophila Stardust, mouse protein associated with lin-seven 1 (Pals1), and human MAGUK p55 subfamily member 5 (Mpp5). As a component of the evolutionarily conserved Crumbs protein complex, nagie oko is essential for the maintenance of epithelial cell polarity. Here, we show that nagie oko contains a predicted nuclear export and two conserved nuclear localization signals. We find that loss of the predicted nuclear export signal results in nuclear protein accumulation. We show that nagie oko nuclear import is redundantly controlled by the two nuclear localization signals and the evolutionarily conserved region 1 (ECR1), which links nagie oko with Par6-aPKC. Finally, deletion forms of nagie oko that lack nuclear import and export signals complement several nagie oko mutant defects in cell polarity and epithelial integrity. This finding provides an entry point to potentially novel and unknown roles of this important cell polarity regulator. Developmental Dynamics 237:83–90, 2008. © 2007 Wiley-Liss, Inc.

INTRODUCTION

The tight junction located at the apico-lateral membrane domain of vertebrate epithelial cells is an essential structure that is required as a paracellular diffusion barrier for small molecules and ions and for blocking the lateral diffusion of lipids and proteins within the membrane. The tight junction is composed of several transmembraneous components and their associated submembraneous scaffolding/adaptor proteins (for review see Shin et al.,2006). Among the associated proteins are the Crumbs-Stardust (Sdt)/Pals1/Mpp5/nagie oko (nok) and the Par6-aPKC protein complexes (Tepass et al.,1990; Tepass and Knust,1993; Kamberov et al.,2000; Hong et al.,2001; Bachmann et al.,2001; reviews in Ohno,2001; Macara,2004; Suzuki and Ohno,2006). Direct interaction between murine Pals1 and Par6 has been shown to require the N-terminal evolutionarily conserved region 1 (ECR1) and the PDZ domain of Par6 (Wang et al.,2004). In zebrafish, these proteins are required for the maintenance of cell polarity, apical junctions, and epithelial integrity (Horne-Badovinac et al.,2001; Peterson et al.,2001; Wei and Malicki,2002; Omori and Malicki,2006; Rohr et al.,2006).

Several tight junctions proteins of the membrane-associated guanylate kinase (MAGUK) family have been shown to localize to the membrane under conditions of increased confluency and to shuttle into the nucleus once confluency decreases. Both, Zonula Occludens-1 (ZO-1) and ZO-2 harbor several nuclear localization signals (NLS) and export signals (NES) that are required for this dynamic localization behavior (Gottardi et al.,1996; Riesen et al.,2002; Islas et al.,2002; Traweger et al.,2003; Betanzos et al.,2004; Jaramillo et al.,2004; Gonzalez-Mariscal et al.,2006). Currently, nuclear functions of ZO-1 and ZO-2 are not known.

Similarly, aPKC, a component of the Par6-aPKC complex, contains NLS and NES motifs (White et al.,2002) and has been shown to move into the nucleus of PC12 cells in response to nerve growth factor (NGF) (Wooten et al.,1997; Zhou et al.,1997). Within the nucleus, aPKC associates with chromatin (Wooten et al.,1997) and several nuclear target proteins have been described (Municio et al.,1995; Zhou et al.,1997). Moreover, Par3 has been shown to be involved in DNA double-strand break repair (Fang et al.,2007). The function of Par6 during nuclear localization remains to be identified (Cline and Nelson,2007).

Here, we show that nok, another MAGUK family member and a protein that is associated with Par6-aPKC, also exhibits NLS and NES motifs. We find that the nok NES motif is essential to prevent nuclear protein accumulation. Moreover, nuclear import of nok is redundantly controlled by the two NLS and the ECR1 domain, which is expected to mediate association with Par6-aPKC. Functional rescue experiments using mutant versions of nok that lack nuclear export and/or localization signals demonstrate that epithelial maintenance is not affected, which is suggestive of other unknown nuclear functions of this protein.

RESULTS

nok Contains Conserved NLS and NES Motifs

Sequence analysis of nok revealed the presence of two putative NLS motifs that are located between the ECR1 and L27 domains and between SH3 and GUK domains, respectively (Fig. 1A). Sequence comparisons with other members of the Sdt/Pals1/Mpp5/nok protein family revealed that both mouse Pals1 and human Mpp5 also harbor these two NLS motifs, which are found in a large number of nuclear proteins (Fig. 1B). Drosophila SdtA exhibits a single NLS corresponding with the N-terminal NLS1 motif of nok. Therefore, NLS motifs are an evolutionarily conserved feature of the Sdt/Pals1/Mpp5/nok protein family. In addition to NLS motifs, nok features a predicted NES within the more N-terminal of the two L27 domains (L27N) (Fig. 1C). Sequence analysis of human Mpp5 and mouse Pals1 indicates potential NES motifs between residues 124–131 (Fig. 1C).

Figure 1. Conservation of NES motifs within the Sdt/Pals1/Mpp5/nok protein family. A: nok is a MAGUK protein that contains ECR1, a bipartite L27, PDZ, SH3, and GUK domains. Potential protein–protein interactions of nok with Crumbs complex proteins and Par6-aPKC are modeled on data available for other Sdt/Pals1/Mpp5/nok protein family members (black arrows). NLS motifs are indicated in green, a single NES motif within the L27N domain is shown in red. B: Sequence alignments of the two NLS motifs are shown for Drosophila SdtA, zebrafish nok, mouse Pals1, and human Mpp5. Conserved residues are shown in red. C: Sequence alignments of the predicted NES motifs are shown for zebrafish nok, mouse Pals1, and human Mpp5. Conserved Leucines and Isoleucines are shown in red.

Functionality of the nok Putative NES Motif

First, we determined whether nok could be detected within the nucleus. Subcellular localization of endogenous nok and HisMyc-tagged wild type (wt) nok^wt protein expressed after mRNA embryo injection was characterized by fluorescence immunohistochemistry at late gastrula stages (80%-epiboly) using anti-nok and anti-Myc antibodies (Fig. 2A,B). Endogenous nok and HisMyc-tagged nok^wt protein mostly localized to junctional belts within the outer cell membrane in epithelial cells of the enveloping layer (EVL), which is the tissue that encloses the zebrafish gastrula stage embryo. Low levels of endogenous nok protein could also be detected within the nucleus (Fig. 2A).

Figure 2. The nok NES motif is required for nuclear export. Images represent reconstructions of confocal Z-stack sections imaged on late gastrula stage whole mounts. A: Endogenous nok mostly localizes to the outer cell membrane of EVL cells whereas nuclear localization is barely detectable. B: HisMyc-tagged nok^wt fusion protein detected with anti-Myc antibody, green. The wt protein localizes to the outer cell membrane in EVL cells. C: HisMyc-tagged nok^ΔNES (green) localizes to the nucleus of EVL cells and lower levels of protein are also present at the outer cell membrane. D: HisMyc-tagged nok^ΔPatj localizes to the nucleus and lower levels are present at the outer cell membrane (green). Red arrow indicates an EVL cell nucleus, which is large compared with the smaller DL cell nuclei located just underneath. Propidium iodide (PI) nuclear counterstain in red.

To characterize whether the putative NES motif is functional in shuttling nok out of the nucleus, we tested whether loss of this motif would trap nok within the nucleus. Therefore, we generated a mutation that encodes a HisMyc-tagged mutant form of nok with changes of the 8 amino acid NES motif (changes of residues L154A, L155A, L158A, and K159A) and that is referred to as nok^ΔNES. In addition, we produced a mutation of nok, which encodes a larger deletion protein lacking the L27N domain surrounding the NES motif. Previous studies have shown that the L27N domain of Pals1 targets the protein to Pals1-associated tight junction (Patj) protein, another component of the apical Crumbs protein complex (Roh et al.,2002) and we refer to this deletion protein as nok^ΔPatj. In comparison to endogenous protein (Fig. 2A) and to the overexpressed wt form of nok (Fig. 2B), both nok^ΔNES and nok^ΔPatj mutant protein localized strongly to nuclei and outer cell membranes of EVL and deep layer (DL) cells (Fig. 2C,D). Therefore, the predicted nok NES domain, which is located within the L27N domain (required for interaction with Patj), functions as a nuclear export signal.

Redundancy of nok Nuclear Import Mechanisms

Next, we analyzed whether the two putative NLS motifs are required for nok nuclear import. We first tested localization of HisMyc-tagged nok^ΔNLS protein, which is a compound mutant that lacks both NLS motifs and observed normal localization at the outer cell membrane similar to HisMyc-tagged nok^wt (data not shown). Since nuclear localization of HisMyc-tagged nok^wt is barely detectable, we tested whether loss of the two NLS motifs could suppress nuclear accumulation of protein also lacking the larger L27N domain, which includes the NES motif (we refer to this combinatorial mutant as nok^{ΔPatj NLS}). Similar to nok^ΔNES (Fig. 2C) and nok^ΔPatj mutant proteins (Fig. 3A), nok^{ΔPatj NLS} mutant protein strongly accumulated within the nucleus, which suggested that alternative routes of nuclear import independent of the two NLS motifs exist (Fig. 3B).

Figure 3. Redundant mechanisms involved in nok nuclear accumulation. Images are reconstructions of confocal Z-stack sections imaged on late gastrula stage whole mounts. HisMyc-tagged nok deletion proteins are detected with anti-Myc antibody (green). A: HisMyc-tagged nok^ΔPatj strongly localizes to nuclei of EVL and DL cells (recognizable by their size) and to outer cell membranes. B: HisMyc-tagged nok^{ΔPatj NLS} protein that lacks the L27N domain and the two nuclear import signals still localizes to the nucleus, which indicates that alternatives routes of nok nuclear import must exist. C: HisMyc-tagged nok nok^{ΔPar6 Patj} protein localizes to nuclei of EVL and DL cells indicating that direct association with Par6 is not necessary for nuclear accumulation of nok. D: In contrast, HisMyc-tagged nok^{ΔPar6 Patj NLS} protein fails to localize to nuclei of EVL and DL cells suggesting redundancy of the NLS motifs and association with Par6 in nok nuclear accumulation. E: Control non-injected embryos stained with the anti-Myc antibody.

Quantifications of total protein levels by Western blot showed higher levels of nok^ΔPatj (not shown) and nok^{ΔPatj NLS} mutant protein at 24 hr post fertilization (hpf) compared with HisMyc-tagged nok^wt protein. This finding suggests that these deletions stabilize the protein or protect it from degradation (Fig. 4). Similarly, intensity levels of immunohistochemical stainings indicated that nuclear forms of mutant protein were more prominent than the wt protein (not shown).

Figure 4. Increased levels of nuclear nok protein. Increased protein levels of HisMyc-tagged nok^{ΔPatj NLS} (expected size of 74.1 kD) compared with HisMyc-tagged nok^wt (expected size of 82.2 kD) suggest that nuclear accumulation stabilizes nok protein.

Next, we tested whether nok nuclear import depends on physical interactions with the aPKC-Par6 protein complex. Previous studies have shown that aPKC shuttles between membrane and nucleus where it associates with chromatin and phosphorylates different target proteins (Municio et al.,1995; Wooten et al.,1997; Zhou et al.,1997). Other components of this protein complex have also been shown to be present within the nucleus (Cline and Nelson,2007; Fang et al.,2007). Therefore, we analyzed whether loss of the expected Par6 binding domain ECR1 could suppress nuclear accumulation of protein that also lacked the L27N domain including the NES motif (we refer to this compound mutant form as nok^{ΔPar6 Patj}). This deletion protein strongly accumulated within the nucleus of EVL and DL cells, suggesting that alternative nuclear import routes exist that are independent of Par6-aPKC (Fig. 3C).

To test possible redundancies of the two NLS motifs and of the ECR1 domain in nuclear import mechanisms, we finally analyzed a mutant form of nok lacking the L27N domain including the NES motif, the Par6 binding site ECR1 and the two NLS motifs (this compound mutant is referred to as nok^{ΔPar6 Patj NLS}). Indeed, this nok deletion protein failed to significantly accumulate within the nucleus and rather displayed a more uniform distribution throughout the cell (Fig. 3D). This finding suggests that the ECR1 domain and the two NLS motifs are redundantly required for nok nuclear localization.

aPKC Catalytic Kinase Activity Is Required to Prevent Nuclear Accumulation

In a recent study, we showed that a catalytic kinase dead form of aPKC iota (aPKCi^KD) functions as a dominant-negative during zebrafish early heart morphogenesis and in epithelial maintenance (Rohr et al.,2006). Since overexpression of aPKCi^KD efficiently disrupts epithelial tissues, we assessed subcellular distribution of this mutant protein. Immunohistochemistry using an antibody against aPKC revealed that most endogenous protein localized to the outer cell membrane of wt epithelial EVL cells and mesenchymal DL cells at gastrula stages (Fig. 5A,B). In contrast, aPKCi^KD strongly localized to the nucleus of epithelial EVL and mesenchymal DL cells (Fig. 5D,E). Finally, an aPKCi mutant protein with an altered PB1 domain that has been shown to disrupt physical interactions with Par6 (aPKCi^D57A) mislocalized to the cytoplasm but was absent from the nucleus (Fig. 5C; Hirano et al.,2004). These findings suggest that the aPKCi kinase catalytic activity is required to prevent nuclear accumulation. Moreover, aPKCi protein that is dissociated from the tight junction and has lost its Par6 binding ability does not, per default, accumulate within the nucleus.

Figure 5. aPKC kinase catalytic activity prevents nuclear accumulation. Images are reconstructions of confocal Z-stack sections imaged on late gastrula stage whole mounts. Anti aPKC (A,B,D,E) and anti-Myc (C) stainings are false-colored. A: Endogenous aPKC localization in EVL cells is mostly at the outer cell membrane whereas (B) weak levels of wt protein are also detectable throughout the cytoplasm and nuclei of DL cells. C: HisMyc-tagged aPKCi^D57A within EVL cells localizes to the cytoplasm and is complete absent from outer cell membranes or nuclei. D: aPKCi^KD strongly localizes to the nuclei of EVL and (E) DL cells. All images are apical views onto the tissue.

nok and aPKC Are Not Sufficient to Tether Each Other Within the Nucleus

Our finding that nok may potentially interact with Par6-aPKC during nuclear accumulation prompted us to test whether nok and aPKC co-localize within the nucleus. Therefore, we first assayed whether increased nuclear levels of nok could recruit or stabilize endogenous aPKC within the nucleus. Overexpression of HisMyc-tagged nok^ΔNES protein, which strongly localized to nuclei of EVL and DL cells, did not result in increased nuclear aPKC levels (Fig. 6A,B). Therefore, nok^ΔNES apparently does not tether aPKC within the nucleus.

Figure 6. nok and aPKCi are not sufficient to tether each other within the nucleus. Images are reconstructions of confocal Z-stack sections imaged on late gastrula stage whole mounts. Myc and aPKC stainings are false-colored. A,B: Embryos expressing high levels of HisMyc-tagged nok^ΔNES were counterstained with anti aPKC. A,A': Within DL cells, high levels of nuclear HisMyc-tagged nok^ΔNES are present whereas aPKC localizes to outer cell membranes and no increased levels of nuclear aPKC are detectable (A“). B,B': Within epithelial EVL cells, HisMyc-tagged nok^ΔNES is at outer cell membranes and within the nucleus. B”: aPKC is predominantly associated with outer cell membranes. C: Embryos overexpressing HisMyc-tagged nok^wt (green) and non-tagged aPKCi^KD (red), whereas nok^wt is absent from the nucleus (C'), aPKCi^KD is nuclear. All images are apical views onto the tissue.

Conversely, we tested whether increased levels of nuclear aPKCi^KD would result in accumulation of overexpressed HisMyc-tagged nok^wt protein within the nucleus. Whereas high levels of aPKCi^KD were detected within nuclei of EVL and DL cells, no increased nuclear levels of nok^wt were observed (Fig. 6C and data not shown). These experiments demonstrate that nok and aPKC are not sufficient to tether each other within the nucleus.

nok NES and NLS Deletion Mutants Are Functional in Epithelial Maintenance and Organ Morphogenesis

To functionally characterize the role of nok NLS and NES motifs, we tested whether the loss of both NLS motifs (nok^ΔNLS), as well as, disruption of the NES motif (in nok^ΔNES and nok^{ΔPatj NLS} mutants) could partially complement the nok^s305 null mutant loss of function phenotypes, which are based on the loss of epithelial apico-basal polarity and the breakdown of epithelial integrity (Rohr et al.,2006). The phenotypes analyzed included the disruption of the mono-layered retinal pigmented epithelium (RPE) surrounding the neural retina and abnormal body curvature, which presumably corresponds with spinal cord defects (Fig. 7B). Similar to the injection of nok^wt mRNA, which completely rescued the RPE integrity defect at 30 hpf, all three mutants could partially complement RPE integrity defects (Fig. 7C–E). In contrast, whereas body form was rescued in about three quarters of nok^ΔNLS nuclear import signal mutants (Fig. 7C,E), there was no or only a slight rescue of this phenotype in nok^{ΔPatj NLS} or nok^ΔNES mutants respectively (Fig. 7D,E). Therefore, nok function in the maintenance of apico-basal polarity and of epithelial tissue integrity of the RPE does not critically depend on the presence of the NES or NLS motifs.

Figure 7. Loss of nok nuclear import and export signals does not affect epithelial maintenance. Phenotypes of (A) wt, (B) nok^s305 mutant, (C) nok^s305 mutant rescued with nok^ΔNLS mRNA or with (D) nok^{ΔPatj NLS} mRNA at 36 hpf. The straight body form is rescued in nok^ΔNLS mutants (C) but not in nok^{ΔPatj NLS} mutants (D). Insets depict details of RPE from each corresponding embryo. The cobble-stone like appearance of RPE tissue surrounding the neural retina is severely disrupted and displays wholes in (B) nok^s305 but this phenotype is partially complemented in (C) nok^ΔNLS or (D) nok^{ΔPatj NLS} mutants. E: Summary of rescue efficiencies of the different deletion mutants. Mutant phenotypes for nok^ΔNES are similar to nok^{ΔPatj NLS} (not shown).

DISCUSSION

Our study has shown that the conserved Sdt/Pals1/Mpp5/nok MAGUK proteins harbor nuclear import signals and that nok contains an NES motif, which, under normal conditions, shifts the steady state of nuclear import/export of endogenous nok protein towards efficient nuclear export. Deletion analysis of the two conserved nok NLS motifs and of the ECR1 domain, which is expected to mediate binding to Par6-aPKC (Wang et al.,2004), revealed redundant functions in nok nuclear accumulation and suggested that nok associates with Par6-aPKC during this process (Fig. 8). Indeed, Par6 has recently been shown to localize to nuclear speckles within the nucleus and to possibly function as a scaffolding protein in nuclear speckle complexes (Cline and Nelson,2007).

Figure 8. Model of nok nuclear import processes. nok protein is indicated in red. White arrows suggest different nuclear import and export routes. A: nok^wt shuttles between the outer cell membrane and nucleus. The steady state of nuclear import versus export is shifted towards efficient nuclear export. B: Loss of the NES results in nuclear accumulation of nok^ΔNES suggesting that nuclear export is affected in this mutant. C: nok^{ΔPatj NLS} protein, which lacks the L27N domain including the entire NES motif and both NLS motifs, accumulates within the nucleus suggesting that import mechanisms independent of the NLS motifs exist. D: nok nok^{ΔPar6 Patj} protein localizes to the nucleus indicating that routes of nuclear import exist that are independent of direct association with Par6 (possibly the NLS motifs). E: In contrast, nok^{ΔPar6 Patj NLS} protein accumulates within the cytoplasm and some protein is also present within the nucleus and on membranes. This observation suggests redundancy in the processes that contribute to nok nuclear accumulation.

We have characterized localization patterns of endogenous nok and aPKCi in gastrulating embryos, which revealed that these proteins are predominantly located at the outer cell membrane in epithelial EVL cells. Whereas higher levels of nuclear aPKCi protein are present within mesenchymal DL cells, nok remains largely at the outer cell membrane or throughout the cytoplasm of DL cells. The differences in aPKCi localization observed between EVL and DL cells may reflect the fact that DL cells lack apical junctions, which in turn could cause increased nuclear protein localization. Similarly, ZO-1 has been found to shuttle to the nucleus under conditions of low cell–cell interaction whereas most of the protein was at the membrane under confluent cell culture conditions (Gottardi et al.,1996; Riesen et al.,2002). Many developmental processes, including the ingression/invasion of cells during gastrulation or some aspects of organ formation require the loss of a rigid epithelial organization and a transition to a mesenchymal character of cells (Shook and Keller,2003; Thiery,2003). Whether maternal nok and aPKCi play similar roles during zebrafish gastrulation is currently not known.

We have shown that aPKCi^KD accumulates within the nucleus. Previous studies demonstrated that overexpression of aPKCi^KD has a dominant-negative effect and causes the breakdown of epithelial integrity (Suzuki et al.,2002; Rohr et al.,2006). There are several plausible explanations for aPKCi^KD nuclear accumulation. First, dissociation of apical junctions associated protein complexes may cause aPKCi^KD to accumulate within the nucleus. Alternatively, the nuclear export or the prevention of nuclear import of aPKCi may require its catalytic activity. aPKCi^D57A mutant protein, which fails to associate with Par6, localizes to the cytoplasm but not to the nucleus. This finding suggests that even a high level of aPKCi protein that dissociates from the tight junction does not, per default, accumulate within the nucleus. Therefore, we favor the latter possibility that the kinase catalytic activity prevents nuclear accumulation of aPKCi^KD.

Our findings suggest a possible link between nok and Par6-aPKC in nuclear to cytoplasmic shuttling. We found that high levels of nuclear nok^ΔNES protein did not tether endogenous aPKC to the nucleus and that, conversely, high levels of nuclear aPKCi^KD did not result in nuclear accumulation of overexpressed HisMyc-tagged nok^wt. Therefore, in wt, nok and Par6-aPKC may interact dynamically during nuclear import/export processes or in nuclear accumulation but immediately dissociate once they have entered the nucleus, which is followed by subsequent rapid nuclear export of endogenous aPKC or nok.

In functional rescue experiments, we could show that several nok deletion mutants (nok^ΔNLS and nok^ΔNES) could partially rescue nok RPE epithelial maintenance and body curvature defects, demonstrating the functionality of mutant proteins in these processes. Previous studies have shown that the L27N domain of Pals1 targets the protein to Patj, another component of the apical Crumbs protein complex (Roh et al.,2002). Similarly, direct interactions between murine Pals1 and Par6 are mediated by the ECR1 domain (Wang et al.,2004). Deletions of these domains directly impair important protein–protein interactions with proteins that have been implicated in epithelial maintenance and apico-basal polarity (Roh et al,2002; Shin et al.,2005; Michel et al.,2005; Wang et al.,2004). In this study, we did not include a detailed functional analysis of nok^ΔPatj, nok^ΔPar6, or of the corresponding compound mutants since it would be difficult to separate the phenotypes caused by the disruption of the apical tight junctions-associated protein complexes from those due to defective nuclear localization patterns (unpublished data). Phenotypic similarities of nok^ΔNES and nok^{ΔPatj NLS} mutants suggest that mutation of the NES affects the L27N domain and thereby interaction with Patj. It currently remains an unresolved issue, whether or not nok serves nuclear functions that are important in a developmental context and that are different from epithelial maintenance. Different experimental approaches will be required to address the possible nuclear roles of the Sdt/Pals1/Mpp5/nok protein family members.

EXPERIMENTAL PROCEDURES

Fish Maintenance and Stocks

Zebrafish were maintained at standard conditions (Westerfield,1994). Embryos were staged by hpf at 28.5°C (Kimmel et al.,1995). The following fish strains were used: wild type AB, and nok^s305.

DNA Constructs and Site-Directed Mutagenesis

The wt form of nok was produced by PCR amplification from a full-length cDNA template, introducing XhoI and XbaI restriction sites 5′ and 3′, respectively, and the construct was subsequently subcloned into the pCS2+HisMyc expression vector (Rohr et al.,2006). nok deletion constructs were generated by PCR based on the pCS2+HisMyc nok^wt template and blunt end ligation of the resultant PCR product. The deletion constructs generate nok deletions of the following residues: nok^ΔPar6 (Δ54–64), nok^ΔPatj (Δ152–209), nok^ΔNLS2 (Δ467–476). nok^ΔNLS1 was generated by site-directed mutagenesis of the NLS1 motif changing residues Asp88Ala and Glu93Ala, which renders the NLS motif inactive (Quick Change XL kit, Stratagene, La Jolla, CA). The nok^ΔNLS mutation is a compound mutation composed of both NLS mutations. Similarly, nok^ΔNES was generated by site-directed mutagenesis of the NES motif introducing the following exchanges that render the NES motif inactive: Leu154Ala, Leu155Ala, Leu158Ala, and Lys159Ala. Finally, compound mutations were produced by subcloning of the different single mutations. The aPKC^D57A mutation was generated by site-directed mutagenesis of pCS2+aPKC^wt. Generation of aPKC^KD has previously been described (Rohr et al.,2006). Primer sequences are available upon request.

Bioinformatics Analysis

NLS motifs were identified using the predict NLS online tool (Columbia University Bioinformatics Center; http://cubic.bioc.columbia.edu/cgi/var/nair/resonline.pl). For identification of the NES motifs, the Net NES1.1 server was used (Center for Biological Sequence Analysis, Technical University of Denmark; http://www.cbs.dtu.dk/services/NetNES/).

mRNA Synthesis and Injections

Constructs were transcribed using the SP6 MessageMachine kit (Ambion). For rescue experiments, nok^s305 embryos were injected with 50–75 pg of mRNA. For overexpression (to determine the subcellular localization patterns), 75–100 pg of mRNA were used.

Antibodies and Immunohistochemistry

Antibodies were raised against amino acids 1–200 of nok. Therefore, appropriate primers were used to amplify the cDNA encoding this portion of the protein and to subsequently clone the fragment into pET23/T7, a modified pET23a vector (Obermann et al.,1998). The His-tagged protein was expressed in E. coli BL21-CodonPlus(DE3)-RIL (Novagen), solubilized using a buffer containing 6M guanidine-HCl, and purified as described (Chumpia et al,2003). Rabbits were immunized 4 times with the purified recombinant protein according to a standard protocol (Biogenes, Berlin, Germany). Specific antibodies were affinity purified using the recombinant protein that was bound to Ni-NTA agarose beads (Sigma) essentially as described (Chumpia et al,2003). Antibody stainings were performed as previously described (Horne-Badovinac et al.,2001). The following antibodies were used: rabbit anti-aPKCζ (1:100, Santa Cruz Biotechnology), mouse anti-Myc (1:200, Invitrogen), rat anti-HA (1:100, Roche), mouse anti-His (1:100, Qiagen), goat anti-rabbit Cy5 (1:200), anti-mouse FITC (1:200), and anti-rat FITC (1:200) (Jackson ImmunoResearch). Nuclear counter-stainings were performed by RNAse treatment of embryos for 2 hr at RT followed by incubation in propidium iodide (1:1,000 dilution) for 10 min at RT. For sectioning, stained embryos were postfixed overnight at 4°C in 2%PFA, 0.3M sucrose. Embryos were embedded in 4% low melting agarose and sectioned on a Leica VT1000 Vibratome. Confocal images were obtained using the Leica TCS SP2 and Zeiss LSM 510 using 40×/1.3 oil lenses. Whole embryos were documented using the Leica MZFLIII stereomicroscope using the 1× and 10× objectives with 5–10× zoom and Leica IM50 software package. Photos were processed using Photoshop (Adobe).

Western Blot Analysis

Western blot analysis was done essentially as previously described (Anzenberger et al.,2006). Membranes were probed with mouse anti-Myc antibody (1:1,000, Invitrogen). For loading control, membranes were stripped and tested for acetylated-tubulin (mouse anti-acetylated tubulin, 1:1,000, Sigma).

Acknowledgements

We are indebted to X. Wei and J. Malicki for sharing reagents and to J. Richter, N. Cornitius, and R. Fechner for expert technical assistance. We thank Michael Bader, Ansgar Klebes, Mahendra Sonawane, and members of the Abdelilah-Seyfried lab for comments on the manuscript. This work was supported, in part, by a grant from the Helmholtz Society to S.R.