Nup358 regulates microridge length by controlling SUMOylation-dependent activity of aPKC in zebrafish epidermis

Cell polarization is a fundamental process involved in growth, development and homeostasis of multicellular organisms (Macara and Mili, 2008). The establishment and maintenance of polarity across different organisms and cell types are regulated by players that are conserved across evolution (St Johnston, 2018). The Par polarity complex, consisting of Par3, Par6 and atypical protein kinase C (aPKC), represents one such module that has well-established roles in polarity during epithelia formation, directed cell migration, establishment of the anterior-posterior axis in the C. elegans embryo and Drosophila oocyte, and axon–dendrite differentiation of neurons (Lang and Munro, 2017).

Polarized organization of the cytoskeleton, an essential aspect of cell polarity, has been shown to be governed by regulators of cell polarity including the Par3–Par6–aPKC module (Raman et al., 2018). Epithelial cells exhibit F-actin-based apical membrane protrusions that are important for absorption, secretion and mechanotransduction (Lange, 2011; Nambiar et al., 2010). aPKC is known to phosphorylate Ezrin – a membrane–actin linker – to regulate the height and number of microvilli, the apical projections of human enterocytes (Wald et al., 2008). In zebrafish, aPKC function is required to control the length of microridges, the apical projections on the epidermal cells that grow horizontally, forming long labyrinthine patterns (Raman et al., 2016). Microridges are present on various kinds of squamous epithelial cells in vertebrate lineages, including mammals, and are thought to be important for mucous retention and spread, abrasion resistance and membrane storage (Sharma et al., 2005; Sperry and Wassersug, 1976; Uehara et al., 1991). It has been reported that aPKC restricts microridge length by regulating the levels of Lgl and non-muscle myosin-II at the apical membrane domain (Raman et al., 2016). So far, it is not clear how the aPKC activity itself is regulated during the formation of apical projections in epithelial cells.

The kinase component of Par polarity complex, aPKC, phosphorylates multiple substrates such as Lgl, Numb and Par1 (Hapak et al., 2018) to achieve spatio-temporal regulation of processes that are involved in the establishment and maintenance of cell polarity. This requires a tight regulation of aPKC activity. The aPKC molecule is inhibited by intra-molecular interaction between the pseudosubstrate domain present in the N-terminal region and the C-terminal catalytic domain. Phosphorylation of specific residues on aPKC, for example, 3-phosphoinositide-dependent kinase 1 (PDK1, also known as PDPK1)-dependent phosphorylation of T410 (human PKCζ isoform) and TORC2-mediated phosphorylation of T585, is known to activate the kinase (Le Good et al., 1998; Li and Gao, 2014). Another mechanism of aPKC activation involves binding of Cdc42-GTP to Par6–aPKC (Joberty et al., 2000; Lin et al., 2000). However, whether other mechanisms of activation of aPKC also exist is not well understood.

We recently reported that aPKC can be activated by Nup358-dependent SUMOylation (Yadav et al., 2016). SUMO is a small peptide that gets covalently conjugated to specific lysine (K) residues in the target protein, which alters its activity, protein interaction partners and/or intra-cellular localization (Flotho and Melchior, 2013). SUMOylation of the target molecule takes place with a specific set of E1 (a SAE1–SAE2 heterodimer), E2 (Ubc9) and E3 ligases. The SUMO E3 ligases include members of the protein inhibitor of activated STAT (PIAS) family, Pc2 and the nucleoporin Nup358 (also known as RanBP2) (Flotho and Melchior, 2013). Nup358 has been shown to act as an E3 ligase for SUMOylation of topoisomerase II (Dawlaty et al., 2008), borealin (Klein et al., 2009) and Ran (Sakin et al., 2015). Previous studies have demonstrated that the internal repeats (IRs) present towards the C-terminal region of Nup358 possesses the SUMO E3 ligase activity (Pichler et al., 2002).

Nup358 is a nucleoporin present on the cytoplasmic side of the nuclear pore complex, and in the cytoplasm as a component of annulate lamellae (AL), which represent specific subdomains of the endoplasmic reticulum (ER) (Sahoo et al., 2017). Although Nup358 is involved in cargo-specific and receptor-specific nucleo-cytoplasmic transport, it appears to be dispensable for general transport (Wälde et al., 2012). Many non-canonical roles for Nup358 have been reported. Studies from our laboratory have revealed a role for this nucleoporin in the regulation of cell polarity in various contexts. Nup358 has been shown to be involved in microtubule dynamics, the localization of the microtubule plus-end-binding protein APC to the leading edge of cells and regulation of polarized cell migration (Joseph and Dasso, 2008; Murawala et al., 2009). It also interacts with the Par polarity complex and plays a role in the determination of axon–dendrite polarity during differentiation of isolated rat hippocampal neurons (Vyas et al., 2013). Although it is evident that Nup358 plays a role in cell polarization, the mechanism by which it functions in this process is unclear.

Given the fact that Nup358 regulates the activity of aPKC, we sought to investigate the possible in vivo function of Nup358 in formation of actin-based apical projections using zebrafish as a model system. Our analyses indicate that Nup358 is required to control the microridge length in zebrafish. Moreover, we find that Nup358 regulates Lgl-dependent microridge elongation by modulating the SUMOylation of aPKC. Our studies also suggest that Nup358 regulates aPKC activity in the cytoplasmic AL structures.

Sequence analysis revealed that the Nup358 protein is highly conserved between human and zebrafish (Fig. 1A). To investigate the role of Nup358 in the regulation of polarity in zebrafish, knockdown of Nup358 was achieved using a splice-blocking morpholino (MO) targeting the intron2–exon3 junction of Nup358 pre-mRNA (Nup358-MO). The Nup358-MO was injected into the one-cell stage zebrafish embryo, and a five-base-pair mismatch morpholino was used as control (Control-MO). The reduction in Nup358 protein level was confirmed by western blotting (Fig. 1B). RT-PCR and sequencing of the defectively spliced products showed that the intron between the exon2 and exon3 was retained in at least ∼50% of the Nup358 mRNA at 9 h post-fertilization (hpf) (Fig. 1C; Fig. S1A). Owing to the inclusion of this intron, a premature stop codon was introduced in the open reading frame (ORF) after the 47th codon, leading to possible production of a very short N-terminal peptide of Nup358.

The most prominent feature of Nup358 morphants was hydrocephaly, seen distinctly at and after 30 hpf (Fig. 1D). Almost 80% of the Nup358-MO-injected embryos showed the hydrocephaly phenotype as against none in the Control-MO-injected embryos. To rule out any off-target effects of Nup358-MO, we attempted to rescue the hydrocephaly phenotype in Nup358 morphants by expressing human Nup358 (hNup358). In vitro synthesized mRNAs encoding green fluorescent protein (GFP) or a GFP-tagged version of hNup358 was co-injected with Nup358-MO in one-cell stage embryos. Exogenous expression of hNup358 rescued the hydrocephaly phenotype caused by Nup358 knockdown, confirming the specificity of Nup358-MO used (Fig. S1B,C). Moreover, the fact that human homolog rescued the hydrocephaly phenotype in zebrafish Nup358 morphants indicate the functional conservation of this protein during evolution.

F-actin-rich microridges present on the apical surface of zebrafish outer epidermal or periderm cells have been shown to be regulated by aPKC and Lgl (Raman et al., 2016). As our previous study had demonstrated that Nup358-mediated SUMOylation activates aPKC (Yadav et al., 2016), we hypothesized that Nup358 acts upstream of aPKC and Lgl in regulating the microridge length. To test whether Nup358 functions in the regulation of microridge elongation in zebrafish, we injected Control-MO or Nup358-MO into one-cell stage embryos, and examined the microridges in peridermal cells of the head region at 33 hpf, as described previously (Raman et al., 2016). Confocal microscopy analysis followed by quantification revealed that there was a significant increase in the length of microridges in Nup358-depleted peridermal cells as compared to Control-MO-injected embryos (Fig. 1E,F). To confirm the specificity of Nup358 function in microridge regulation, we injected GFP–hNup358 mRNA along with Nup358-MO. The results showed that mRNA injection of hNup358, but not GFP, rescued the long-ridge phenotype caused by Nup358 depletion (Fig. 1F). Taken together, these results demonstrate that Nup358 is a major determinant of microridge length in zebrafish peridermal cells.

The tumor suppressor protein Lgl is known to be a strong binding partner of Non-muscle Myosin II (NMII), and it has been shown that Lgl promotes elongation of microridges through NMII (Raman et al., 2016). Zebrafish have two functionally redundant isoforms, Lgl1 and Lgl2. Since Lgl is a pro-elongation factor (Raman et al., 2016), we asked whether Lgl showed increased localization to the apical cortex and was required for microridge elongation in the absence of Nup358 function. We found that Nup358 morphants, as compared to control embryos, had enhanced localization of Lgl2 at the apical domain and microridges (Fig. S2C). To test whether increased microridge length in Nup358-deficient embryos is Lgl-dependent, MO-mediated Nup358 depletion was performed in lgl2 (penner) mutant embryos (Sonawane et al., 2005). In these mutant embryos, substantial depletion of the Lgl2 isoform was achieved by 48 hpf, as monitored by immunostaining with anti-Lgl2 antibody, which has been shown to cross-react with both the Lgl isoforms (Fig. 2A). Detailed analysis suggested that Nup358 morphants in the penner mutant background, showed significantly reduced microridge length as compared to control siblings (Fig. 2A,B). Furthermore, we achieved Lgl1 depletion using a previously characterized Lgl1-MO and investigated whether Lgl1 was also required for maintaining longer microridges in Nup358 morphants. Lgl1-MO or Nup358-MO was injected independently as well as co-injected into one-cell stage zebrafish embryos. The total amount of injected morpholino was kept constant by including appropriate control-MO along with Lgl1-MO or Nup358-MO. Microridges, in the periderm cells of head region, were imaged and analyzed at 33 hpf. In the absence of Lgl1, Nup358 morphants showed a significant reduction in the microridge length (Fig. S2A,B). Taken together, these data support the conclusion that Nup358 regulates length of microridge via Lgl.

Previous studies have indicated an important role for the aPKCλ isoform in microridge formation in developing zebrafish peridermal cells. It has been shown that in the absence of aPKCλ function, increased apical localization of Lgl – otherwise a basolateral regulator – results in elongation of microridges (Raman et al., 2016). In addition, we had earlier shown that Nup358 activates aPKC via SUMOylation (Yadav et al., 2016). Since the microridge phenotype of Nup358 depletion resembled the has (apkcλ/ι) mutant and was Lgl dependent, we asked whether Nup358 functions through modulation of aPKC SUMOylation for the regulation of microridge length.

In vitro SUMOylation experiments demonstrated that zebrafish Nup358 IR region, similar to human Nup358 IR region, possesses the SUMO E3 ligase activity (Fig. S3). Both human and zebrafish have two isoforms of aPKC, namely aPKCζ and aPKCλ (also known as aPKCι), encoded by PRKCZ and PRKCI, respectively. Previously, we had reported that human aPKCζ is post-translationally modified by SUMO, and three lysine (K) residues (K225, K284, K378) are necessary for SUMOylation (Yadav et al., 2016). Sequence analysis showed that out of the three lysine residues of human aPKCζ, two of the corresponding residues are conserved in zebrafish aPKCζ and aPKCλ (Fig. 3A). We confirmed that zebrafish aPKCλ-wild type (wt), but not the aPKCλ-K278R-K372R double mutant (SUMOmut), gets SUMOylated in mammalian cells when co-expressed with SUMO1 (Fig. S4).

We reasoned that overexpression of wild-type aPKC, but not the SUMOmut, might rescue the microridge phenotype in Nup358 morphants. To test this possibility, Nup358-MO was co-injected with HA-control or HA-tagged zebrafish aPKCλ-wt or aPKCλ-SUMOmut mRNA into one-cell stage zebrafish embryos. We used a lower amount of Nup358-MO for this experiment to achieve better rescue upon aPKCλ mRNA injection. We presumed that this partial depletion of Nup358 would also allowed retention of some SUMOylation activity that was essential for aPKC activation. Our experimental results showed that the long-ridge phenotype seen in Nup358-knockdown embryos could be partially rescued by the expression of zebrafish aPKCλ-wt but not by aPKCλ-SUMOmut (Fig. 3B,C), indicating that the regulation of microridge length by Nup358 is dependent on aPKC SUMOylation.

We found previously that endogenous aPKC is preferentially modified by SUMO1, and Nup358 is involved in the SUMOylation and activation of human aPKC (Yadav et al., 2016). We sought to investigate the role of SUMOylation and activation of aPKC in the regulation of microridge length. Initially, we examined whether aPKC gets SUMOylated in zebrafish. To find out whether aPKC is SUMOylated endogenously in a Nup358-dependent manner, the lysate prepared from Control-MO- or Nup358-MO-injected zebrafish embryos (36 hpf) were subjected to immunoprecipitation (IP) with control or aPKC antibody (recognizes both the isoforms), and the immunoprecipitates were probed for aPKC and SUMO1 by western blotting. In addition to the unmodified intact aPKC band (∼70 kDa), a major higher molecular mass aPKC band could be seen in the case of aPKC IP, which was also detected with an anti-SUMO1 antibody (Fig. 4A), suggesting that endogenous aPKC is modified by SUMO1. Interestingly, for reasons unknown, Nup358-deficient embryos showed consistently higher levels of aPKC as compared to Control-MO injected embryos (Fig. 4A, arrowhead). Despite this, quantitative analysis showed that the proportion of SUMOylated aPKC was less in the case of Nup358-MO injected embryos as compared to that of control embryos (Fig. 4B). Collectively, the data support the conclusion that endogenous aPKC is SUMOylated in zebrafish in a Nup358-dependent manner.

We had previously shown that SUMOylation of aPKC leads to an increase in its kinase activity, and expression of the C-terminal region of hNup358 (Nup358-C) that harbors the IRs involved in the SUMO E3 ligase activity was sufficient to enhance SUMOylation of endogenous aPKC in mammalian cells (Yadav et al., 2016). We reasoned that if Nup358 knockdown led to the stabilization of microridges due to reduced SUMOylation and activity of aPKC, expression of human Nup358-C should be able to rescue the phenotype. To test this, mRNAs encoding GFP-control, hNup358-C or hNup358-CΔIR (Fig. 5A) were synthesized in vitro and injected along with Nup358-MO into one-cell stage embryos. We found that the expression of the GFP–hNup358-C fragment, but not the deletion mutant that is devoid of the IR region and hence E3 ligase-defective (GFP–hNup358-CΔIR), rescued the long microridge phenotype caused by Nup358 depletion (Fig. 5B,C). Taken together, these results indicate that Nup358-mediated SUMOylation of aPKC is required for regulation of microridge length. These results also suggest that the C-terminal region of Nup358 (Nup358-C) is sufficient to function in the microridge regulation. In accordance with this, overexpression of GFP–hNup358-C resulted in shortening of microridges as compared to GFP-control expression (Fig. S2D,E).

How does Nup358 regulate SUMOylation and activation of aPKC in cells? This is particularly interesting, given that Nup358 primarily resides in the nuclear membrane as a part of NPC and aPKC is mostly present at the apical junctions of epithelial cells. Nup358, along with a subset of nucleoporins, is also found in the cytoplasm as a part of the AL structure (Sahoo et al., 2017). To analyze the distribution of Nup358 and other nucleoporins in zebrafish, we resorted to fluorescence microscopy using specific antibodies. Immunostaining of zebrafish cells with Nup358, Nup107 and mAb414 (recognizes FG-containing nucleoporins, including Nup358) antibodies, showed that the cytoplasmic structures are positive for Nup358, Nup107 and possibly other nucleoporins (Fig. 6A,B), identifying them as the AL. Furthermore, co-staining for Nup358 and aPKC indicated that ∼44% of Nup358-positive AL puncta in the cytoplasm were associated with aPKC in zebrafish cells (Fig. 6C). Taken together, the data indicate that AL might act as platforms for Nup358 to SUMOylate and activate aPKC in the cytoplasm. Based on these findings, we propose that Nup358 at the AL structures associates with, SUMOylates and activates aPKC, which in turn phosphorylates and inhibits Lgl and restricts the microridge length (Fig. 6D).

Cell polarity regulators have been shown to play a crucial role in regulation of formation and maintenance of the apical projections in epithelial cells. Among these, aPKC is known to regulate formation of microvilli in enterocytes by controlling the activation of Ezrin. In addition, aPKC prevents excessive elongation of microridges on the epidermal cells in developing zebrafish by regulating the levels of Lgl and NMII at the apical domain. However, it has remained unclear how the activation of aPKC is regulated during formation of such apical actin-based projections. Here, using microridges as a paradigm, we have unraveled SUMOylation-dependent activation of aPKC by Nup358 as an upstream regulatory step during formation of apical projections.

Previous studies from our laboratory had established a role for the nucleoporin Nup358 in cell polarity during cell migration and neuronal differentiation (Murawala et al., 2009; Vyas et al., 2013). However, whether apical-basal polarity and formation of apical projections are controlled by Nup358 remained unclear. Although Nup358 had also been shown to SUMOylate and activate the polarity kinase aPKC (Yadav et al., 2016), the mechanism by which Nup358 functions in cell polarization was not understood. Towards understanding the role of Nup358 in cell polarity, we chose to study the early development of zebrafish. We particularly focused on the formation of microridges in zebrafish peridermal cells, a process that has been shown to be regulated by aPKC-mediated inhibition of Lgl localization at the apical domain (Raman et al., 2016). The results presented here show that Nup358-mediated SUMOylation and activation of aPKC negatively regulate microridge elongation through inhibition of Lgl function, indicating that Nup358 acts as an upstream modulator of the aPKC–Lgl axis during polarization. As shown previously (Raman et al., 2016), the loss-of-function phenotype for lgl paralogues is subtle and becomes evident upon quantification suggesting redundant mechanisms involved in microridge formation. In addition, the strength of the phenotype varies depending upon parents, suggesting background dependence. Nevertheless, overexpression of Lgl2 or apically targeted Lgl1 has been shown to result in a robust increase in the microridge length, establishing its importance as a pro-elongation factor (Raman et al., 2016). It appears that Nup358-mediated SUMOylation and activation of aPKC has emerged as a mechanism to control the levels of this pro-elongation factor at the apical domain to regulate microridge length (Fig. 6D). Although these analyses of Nup358 function are based on morpholino knockdown, the fact that the microridge phenotype is rescued by injection of full-length human Nup358 or Nup358 with just SUMO ligase activity (Nup358-C) indicates the specificity of the morpholino phenotype. It is also important to note that injection of E3-ligase-deficient Nup358 does not rescue the Nup358 morphant phenotype and that increasing aPKC levels rescues the morphant phenotype, further confirming the specificity of the morpholino phenotype. Although we showed the importance of the zebrafish aPKCλ isoform in Nup358-mediated regulation of the microridges, our studies do not shed light on the specific contribution of the zebrafish aPKCζ isoform in this process. It will be also interesting to investigate the possible crosstalk or interconnection between different modes of aPKC activation such as phosphorylation, protein–protein interaction and SUMOylation and their role in the spatio-temporal regulation of aPKC function in multiple cellular contexts.

One of the fascinating questions is the location at which Nup358-mediated regulation of aPKC occurs. We identified the cytoplasmic structures called AL, which contain Nup358, as potential sites of aPKC regulation. As mentioned above, AL are specific subdomains of endoplasmic reticulum, which were recently shown to be associating with messenger ribonucleoprotein granules and playing a role in microRNA (miRNA)-dependent translation regulation of mRNAs (Sahoo et al., 2017). The finding from this study, showing the association of aPKC with Nup358-positive AL structures in zebrafish cells, points towards other unidentified processes mediated by AL. The Nup358-positive AL-associated aPKC represents a novel cytoplasmic pool in addition to the previously reported aPKC pool associated with recycling endosomes (Dhekne et al., 2014). Whether this AL-associated aPKC indicates the newly synthesized pool or the one being recycled is an interesting question. Further studies should be able to distinguish these pools and understand the mechanism of aPKC regulation at the AL.

Consistent with Nup358 acting as a SUMO E3 ligase for aPKC, there was a relative decrease in the extent of aPKC SUMOylation in Nup358 morphants (Fig. 4B). The relative decrease in aPKC SUMOylation could be an additive effect of a decrease in SUMO E3 ligase activity and increase in total amount of aPKC in Nup358-deficient cells (Fig. 4B). Under such a scenario, it is plausible that a large fraction of apical aPKC is not SUMOylated and hence inactive. Since aPKC activity is required to eliminate Lgl from the apical domain (St Johnston, 2018), this may further result in increase in Lgl localization to the apical domain and an increase in microridge length as has been shown previously (Raman et al., 2016; Fig. 6D). However, currently we do not have means to check the levels of SUMOylated and non-SUMOylated aPKC at the apical domain of the peridermal cells. How does the aPKC pool increase in the absence of Nup358 function? We have previously shown that Nup358 functions in the miRNA pathway (Sahoo et al., 2017). Therefore, the increased level of aPKC might be indicative of the impairment in the miRNA-mediated translation suppression of aPKC mRNA in Nup358-knockdown conditions. Alternatively, in addition to activation, SUMOylation of aPKC might target it for degradation by SUMO-targeted ubiquitin ligases (STUbLs) (Sriramachandran and Dohmen, 2014), and in conditions where SUMOylation is compromised, as in the case of Nup358 morphants, the protein may be stabilized. Substrates that are conjugated with multiple SUMO2 and/or SUMO3 (SUMO2/3) molecules have been often degraded by STUbLs, such as RNF4 (Tatham et al., 2008). Consistent with this, we found that zebrafish aPKC is also modified by SUMO2/3 (Fig. S5). However, further studies are required to understand the interlink between SUMO1 and SUMO2/3 modification and its effect on the stability and function of aPKC.

In summary, our study has demonstrated that Nup358 is an upstream player that regulates aPKC and Lgl to determine the microridge length in developing zebrafish epidermal cells. Nup358 exerts its effect on aPKC by SUMOylation and activation, which in turn regulates Lgl activity and the microridge length. Whether Nup358-dependent aPKC SUMOylation plays a role in other contexts of cell polarity awaits further investigation. The findings also highlight the AL as possible hubs for cytoplasmic regulatory activities.

For experiments in wild-type fish, the Tübingen (Tü) strain was used. Heterozygous fish of the Tü background carrying a mutation in the pen^t06 locus were in-crossed to get homozygous penner mutant embryos, and their siblings were used as controls (Sonawane et al., 2005). Since the strength of the phenotype varies, heterozygous pen^t06 pairs giving rise to the consistent microridge phenotype were used in this analysis. The zebrafish maintenance and experiments were performed as per the guidelines recommended by the Committee for the Purpose of Control and Supervision of Experiments on Animals (Government of India) and were approved by the institutional animal ethics committee (approval TIFR/IAEC/2017-11).

All the microinjections were performed in one-cell zebrafish embryos. An antisense splice morpholino for Nup358, designed to target the intron2–exon3 junction (5′-CATGTCTGAAGGAGAAAACACACAC-3′; Nup358-MO), and its respective five-base-pair mismatch (indicated in lowercase) control morpholino (5′-CATcTCTcAAcGAcAAAAgACACAC-3′; Control-MO) were procured from Gene Tools, LLC, and injected at a concentration of 100 µM. Lgl1 translation-blocking antisense morpholino (5′-CCGTCTGAACCTAAACTTCATCATC-3′; Lgl1 MO) (Clark et al., 2012; Hava et al., 2009) and standard control morpholino [5′-CCTCTTACCTCAGTTACAATTTATA-3′, Standard (Std) control-MO] were injected at a concentration of 250 µM. In case of Fig. S2A,B, for the ease of representation, the groups have been labeled as control-MO, Lgl1-MO, Nup358-MO and Lgl1-MO+Nup358-MO. However, to avoid any dosage-related artifacts Std Control-MO was used as filler morpholino along with Nup358-MO and five-base-pair mismatch control morpholino (Control-MO) was used as filler morpholino with Lgl1-MO. The zebrafish aPKCλ open reading frame was amplified by using 5′-GGGCTCGAGATGCCCACGCTGCGGGACAGCACCATG-3′ (forward) and 5′-CCCCCCGGGCACACACTCCTCCGCAGACATCAGCAG-3′ (reverse) primers from cDNA library prepared from wild-type Tü fish embryos. The amplicon was then cloned into T7 promoter-containing pCI-neo vector (Promega) between Xho1 and Sma1 sites. The HA-tagged zebrafish (zf)aPKCλ-SUMOmut (K278R and K372R double mutant) was generated by a PCR-based method. During RNA injections, as a control for HA-tagged or GFP-tagged proteins, HA–GFP was used. In vitro synthesis of HA–GFP-control, GFP–hNup358, GFP–hNup358-C, GFP–hNup358-CΔIR, HA–zfaPKCλ-wt and HA–zfaPKCλ-SUMOmut mRNAs was performed by using the T7 mMessage mMachine kit (Ambion; AM1344) according to the manufacturer's protocol, and the RNA was injected at concentrations ranging from 100–300 ng/µl along with the indicated morpholino. For rescue of the longer ridge phenotype caused by Nup358 knockdown with exogenously expressed HA–zfaPKCλ (Fig. 3), 50 µM morpholino was used, and 150 µM morpholino was used for experiments shown in Figs 1B and 4A.

Polyclonal antibodies against the internal repeat region (IR region) of zebrafish Nup358 was generated in rabbit and then affinity purified using the same antigen immobilized on Affi-Gel 15 (Bio-Rad Laboratories).

After injection, the embryos were grown in E3 medium at 29°C. The embryos were fixed and stained as described previously (Raman et al., 2016). Briefly, the embryos were fixed in 4% PFA (in phosphate buffered saline; PBS) overnight at 4°C followed by permeabilization in PBS with 0.8% Triton X-100 (PBT). Blocking was achieved with 10% normal goat serum (NGS) (Jackson ImmunoResearch Labs, 005-000-121) prior to incubation with primary antibody in 1% NGS overnight at 4°C. Secondary antibody incubation was performed in 1% NGS for 4 h at room temperature (RT) after washing with PBT. The embryos were then post-fixed with 4% PFA for 30 min at room temperature or overnight at 4°C followed by upgradation to 80% glycerol.

In experiments where only phalloidin staining was required, the embryos were fixed in 4% PFA overnight at 4°C, then permeabilized with PBT and incubated with 1:40 dilution of phalloidin and post-fixation step was omitted.

For co-localization of Nup358 and aPKC, 9 hpf embryos were initially fixed with 4% PFA at 4°C overnight, followed by permeabilization with methanol overnight at −20°C. Rabbit anti-zfNup358 and mouse anti-aPKC (SC-17781) antibodies were used at a dilution of 1:150 and 1:100, respectively. Co-staining with rabbit anti-Lgl antibody (1:400) (Raman et al., 2016) and Rhodamine-conjugated phalloidin (1:40, Invitrogen, R415) was performed in 48 hpf embryos.

For co-staining of Nup107 (1:25, a gift from Dr Ram Kumar Mishra, IISER, Bhopal, India) with mAb414 (1:400; BioLegend, 902901), 9 hpf embryos were fixed with 4% PFA made in PHEM buffer (60 mM PIPES, 25 mM HEPES, 10 mM EGTA, 2 mM MgCl₂, pH 6.9). Fluorescently labeled anti-mouse-IgG and anti-rabbit-IgG secondary antibodies (Invitrogen, A21202 and A21207, respectively) were used at a dilution of 1:500.

Stained embryos that were upgraded to 80% glycerol were mounted in a way such that the dorsal head epidermis was close to the coverslip. Confocal imaging was done on the Zeiss LSM 510 META confocal microscope having Plan-Apochromat 63×/1.40 Oil objective or Leica SP5 confocal microscope having 63× oil immersion HCX PL APO CS (11506188) objective. The zoom factor was either 1.5× for images taken on a Leica SP5 microscope or 3× for images taken on a Zeiss 510 META microscope. The pinhole was adjusted to 1 airy unit while imaging, keeping a line averaging of 3. The image size was 1024×1024 pixels. The imaging parameters were kept constant for any particular experiment. Imaging for the Lgl apical localization experiment (Fig. S2C) was performed at settings where the junction staining is saturated and the apical microridge decoration could be seen. For comparing the Lgl apical localization in Control-MO and Nup358-MO injected embryos, the settings were kept constant. Only apical stacks showing phalloidin-stained microridges that were in focus were chosen to make maximum intensity projections.

Differential interference contrast (DIC) images of live zebrafish embryos, at 36 hpf, were taken on a Zeiss SteREO Discovery microscope after anesthetizing the embryos in 0.04% MESAB (Sigma; A5040) for looking at the hydrocephaly and other Nup358-knockdown phenotypes. The hydrocephaly phenotype was quantified from the DIC images by measuring hind brain ventricle area by marking the edges of the ventricle (Fig. S1B,C) using ImageJ. At least six embryos per group were taken for each experiment and three independent biological repeats were performed. Box plots were made using SigmaPlot software taking all data points across three individual repeats into consideration. For microridge perimeter quantitation, the auto-local threshold tool in ImageJ was used. The number and perimeter of individual microridges were determined by using the ‘analyze particle’ command. The perimeter is mostly contributed by the length in cases of cells with microridges. Hence, for the sake of simplicity the perimeter is referred to as microridge length. Detail of the number of experiments performed, embryos and cell per group taken for quantitation are provided in Table S1. Data for all the microridge quantification experiments is plotted in the form of a bean plot that were made in R software as described previously (Raman et al., 2016). For quantifying the extent of association of Nup358 and aPKC at cytoplasmic AL structures in enveloping layer (EVL) cells of 9 hpf embryos (Fig. 6C), the number of Nup358-positive cytoplasmic structures was counted. The percentage of these structures associated with aPKC was then found using the plot profile tool in ImageJ. The quantification was performed using four to six cells per embryo and 12 embryos across three independent experiments.

For validating the knockdown of Nup358, Control-MO and Nup358-MO injected embryos at 3, 9 and 12 hpf stages were taken for RNA extraction using TRIzol (Invitrogen, 15596026). RNA isolation and cDNA synthesis using random primers was performed as previously described (Sahoo et al., 2017). By using primers flanking the morpholino target region (forward, 5′-ATGAGGAGGAGTAGAGCGGAGGTGGAGCGC-3′; reverse, 5′-CCTGAG GTTGAAGACGGCTGGATGTCCTGG-3′), the locus was amplified, and the PCR product was assessed on an agarose gel. The mis-spliced product band in Nup358-MO lane was then eluted from the gel (Promega, A9282) and sequenced for finding the defect in splicing due to Nup358-MO (Fig. 1C; Fig. S1A). As an RT-PCR control, 5′-ATCACACCTTCTACAACGAGC-3′ (forward) and 5′-CATCACCAGAGTCCATCACG-3′ (reverse) primers were used to amplify actin.

Embryos at the desired stage of development were dechorinated. Deyolking was performed with Ginzberg ringer's solution (55 mM NaCl, 1.8 mM KCl and 1.25 mM NaHCO₃). The deyolked embryos were homogenized in a micro-pestle at liquid nitrogen temperature to avoid any proteolysis during homogenization. Homogenized cells were lysed in RIPA buffer (50 mM Tris-HCl pH 8, 150 mM NaCl, 1% NP40, 0.1% SDS and 0.5% Na-deoxycholate) with 2× protease inhibitors according to a previously described protocol (Sahoo et al., 2017). For all the experiments, 4 μl of lysis buffer was used for each embryo, to ensure efficient lysis. For validating Nup358 knockdown, embryos at 24 hpf were used. For aPKC immunoprecipitation in Control-MO and Nup358-MO injected embryos, 36 hpf embryos were used. For immunoprecipitation, 200 µl of lysate was incubated with 4 µg of mouse anti-aPKC antibody (Santa Cruz Biotechnology, sc-17787) or equal amount of mouse IgG (Bangalore Genei, IGP3) along with 20 µl of Protein G–Sepharose beads (Invitrogen, 101241) for 1 h at 4°C. Proteins bound to beads were then eluted in 50 µl of SDS-PAGE loading dye, and 15 µl of the immunoprecipitate samples were loaded along with 15 µl of input sample (starting material) on 7% SDS-PAGE gels. The input sample was diluted 1:1 with SDS-PAGE loading dye prior to loading on SDS-PAGE gel. The proteins were then transferred onto PVDF membrane (Millipore, IPVH00010) using semi-dry transfer apparatus. Western blots were probed with rabbit anti-aPKC antibody (1:2000; Santa Cruz Biotechnology, sc-216) and a 1:800 dilution of rabbit anti-SUMO1 and SUMO2 antibodies (kind gifts from Dr Ram Kumar Mishra, IISER Bhopal) (Gujrati et al., 2019). Mouse anti-α-Tubulin antibody (1:5000, Sigma, T9026) was used to ensure comparable starting material. HRP-linked anti-rabbit secondary antibodies from GE (NA934V) and ECL Plus Western Detection Kit (GE Healthcare or Thermo Scientific) were used for detection of the specific proteins on the western blot following the manufacturer's protocol. The images were acquired using ImageQuant LAS 4000 (GE Healthcare). For Nup358 knockdown validation, Nup358 was detected on the western blot with the help of previously characterized rabbit anti-Nup358 antibody (1:3000) (Joseph et al., 2004), while mouse anti-vinculin antibody (1:10,000; Sigma, V9131) was used as a loading control. Anti-rabbit-IgG (1:10,000, GE, NA934V) and anti-mouse-IgG (1:10,000, GE, NA931E) secondary antibodies were used for specific detection of Nup358 and Vinculin, respectively.

Data for microridge length obtained from experiments (Table S1) represented in Figs 1E and 2A, 3A, 5B, Fig. S2A and S2D were analyzed by performing a Kruskal–Wallis test and comparisons between two groups was performed using the Dunn's method. For the quantitative data shown in Fig. 4B and Fig. S1C, the test performed and the results obtained are mentioned in the respective figure legends. Sigma plot 12.0 software was used for all the statistical analysis.

We thank Ram Kumar Mishra (IISER Bhopal, India) for SUMO1 and SUMO2 antibodies and Swati R. Gaikwad for generating the hNup358-CΔIR construct. We are grateful to Mary Dasso (NICHD, NIH, Bethesda, USA) for sharing the reagents for in vitro SUMOylation reactions. Timely help from Prateek Arora, Renuka Raman and Clyde Pinto (Sonawane lab, TIFR Mumbai, India) for microridge length quantification and analysis is acknowledged. Extensive help by Jyoti Pawar in standardizing zebrafish embryo lysis and immunoprecipitation is gratefully acknowledged. We also thank Ashwini Atre (NCCS, Pune) and Trupti Kulkarni (NCCS, Pune) for help with microscopy. We are grateful to Joseph and Sonawane lab members for scientific discussions and helpful suggestions. We thank Deepa Subramanyam (NCCS, Pune) and Vasudevan Seshadri (NCCS, Pune) for insightful discussions on the project.