Besides assembling nuclear pore complexes, the conduits of nuclear transport, many nucleoporins also contribute to chromatin organization and gene expression, with critical roles in development and pathologies. We previously reported that Nup133 and Seh1, two components of the Y-complex subassembly of the nuclear pore scaffold, are dispensable for mouse embryonic stem cell viability but required for their survival during neuroectodermal differentiation. Here, a transcriptomic analysis revealed that Nup133 regulates a subset of genes at early stages of neuroectodermal differentiation, including Lhx1 and Nup210l, which encodes a newly validated nucleoporin. These genes are also misregulated in Nup133ΔMid neuronal progenitors, in which nuclear pore basket assembly is impaired. However, a four-fold reduction of Nup133 levels, despite also affecting basket assembly, is not sufficient to alter Nup210l and Lhx1 expression. Finally, these two genes are also misregulated in Seh1-deficient neural progenitors, which only show a mild reduction in nuclear pore density. Together these data reveal a shared function of Y-complex nucleoporins in gene regulation during neuroectodermal differentiation, apparently independent of nuclear pore basket integrity.
As channels embedded in the nuclear envelope, the nuclear pore complexes (NPCs) constitute the only gateway for selective transport of macromolecules between the cytoplasm and the nucleus. These impressive structures are composed of proteins called nucleoporins (Nups), which assemble in a highly organized and modular manner (reviewed in Dultz et al., 2022). The Y-complex – also named the Nup107-160 complex – which comprises nine distinct proteins in vertebrates, is a key structural subunit of the NPC scaffold. A total of 16 copies of this complex assemble on the nuclear and cytoplasmic sides of the NPC to build up its outer rings, to which cytoplasmic filaments and the nuclear basket are anchored.
In addition to their canonical nuclear transport function, many Nups are also known to have other cellular functions, notably in cell cycle progression or as key regulators of chromatin organization and gene expression (reviewed in Buchwalter et al., 2019; Hezwani and Fahrenkrog, 2017). In line with these multiple functions, mutations in many Nups have been identified as primary causes of rare genetic diseases. Despite the presence of NPCs in all nucleated cells, most of these diseases specifically affect one or a few organs (reviewed in Jühlen and Fahrenkrog, 2018). Such tissue or cell-type specific alterations might reflect variable Nup stoichiometry at NPCs, as notably reported for several integral membrane Nups and peripheral Nups (Ori et al., 2013). For instance, increased levels of the transmembrane protein Nup210 in myoblasts and neuronal progenitors has been shown to be critical for their differentiation (D'Angelo et al., 2012). Likewise, depletion of the basket nucleoporin Nup50 reduces the differentiation efficiency of C2C12 myoblasts (Buchwalter et al., 2014). In contrast, another basket nucleoporin, Nup153, which is highly expressed in pluripotent cells and neuronal progenitors compared to differentiated cells, is required for the maintenance of their identities, notably by regulating epigenetic gene silencing and transcriptional programs (Jacinto et al., 2015; Toda et al., 2017). More recently, the Y-complex constituent Seh1, which is highly expressed in oligodendrocyte progenitor cells, has been shown to be required for oligodendrocyte differentiation and myelination by regulating the assembly of a transcription complex at the nuclear periphery (Liu et al., 2019). However, individual Y-complex Nups also contribute to earlier stages of differentiation, as underscored by the impaired neuroectodermal differentiation of Nup133−/−, Seh1−/− and Nup43−/− mouse embryonic stem cells (mESCs) (Gonzalez-Estevez et al., 2021; Lupu et al., 2008).
Earlier studies have found that the vertebrate Y-complex is, as an entity, critically required for NPC assembly both at the end of mitosis and during interphase (Doucet and Hetzer, 2010; Harel et al., 2003; Vollmer et al., 2015; Walther et al., 2003). The viability of Nup133−/−, Seh1−/− and Nup43−/− mESCs, however, indicated that the corresponding Y-complex Nups were individually largely dispensable for nuclear pore assembly in these pluripotent cells. Consistent with this, we have previously shown that mutations of Nups that form the short arm of the Y-complex, namely Nup43, Nup85 and Seh1, only lead to a mild decrease in NPC density in pluripotent mESCs (Gonzalez-Estevez et al., 2021). In contrast, pluripotent Nup133−/− mESCs feature a normal NPC density, but show specific nuclear basket defects, with half of NPCs lacking Tpr whereas Nup153 dynamics is increased (Souquet et al., 2018). How Y-complex Nups contribute to NPC assembly in differentiating mESCs had not been determined.
Because of the established implication of the basket nucleoporins Nup153 and Tpr in chromatin organization and gene regulation (Aksenova et al., 2020; Boumendil et al., 2019; Jacinto et al., 2015; Krull et al., 2010; Toda et al., 2017), we decided to investigate potential gene expression defects in Nup133−/− mESCs during neuroectodermal differentiation. Here, we show that Nup133 regulates a subset of genes, including Lhx1 and Nup210l, which are similarly misregulated in the absence of Seh1. However, Nup133 and Seh1 deficiencies display distinct NPC assembly phenotypes in neuronal progenitors, thus indicating separate roles for these proteins in NPC architecture and gene regulation in the context of mESC differentiation.
Nup133 is required for the regulation of a subset of genes during neuroectodermal differentiation
We first determined whether the impaired differentiation of Nup133−/− mESCs, initially described in cells derived from merm (behaving as Nup133−/−) blastocysts (Lupu et al., 2008), was also observed in an independent genetic background. As Lupu and colleagues, we observed altered growth and increased cell death upon neuroectodermal differentiation of Nup133−/−mESCs derived by CRISPR/Cas9-editing from HM1 mESCs (previously described in Souquet et al., 2018) (Fig. 1A,B). To evaluate the potential effect of Nup133 deficiency in gene regulation upon neuroectodermal differentiation, we assessed the mRNA levels of genes expressed in pluripotent cells [Oct4 (also known as Pou5f1) and Nanog] and in early neuronal progenitors (Sox1 and Pax6) that are considered markers for the respective states. Quantitative reverse transcription real-time PCR (RT-qPCR) analyses showed that these genes were properly repressed and activated, respectively, in Nup133−/− cells stimulated to differentiate towards neuroectoderm (Fig. 1C). This indicates that despite their impaired viability at early stages of differentiation (Fig. 1A,B) the surviving Nup133−/− cells are able to exit pluripotency and to commit towards the neuronal lineage, without overt defects in the expression of these markers.
To explore the impact of Nup133 on gene expression more broadly, we compared the transcriptome of wild-type (WT) and Nup133−/− mESCs at the pluripotent state and after 2 or 3 days of differentiation towards the neuroectodermal lineage. We therefore used WT and Nup133−/− mouse cell lines from two distinct genetic backgrounds, namely the 129/ola blastocyst-derived HM1 control cell line (Selfridge et al., 1992) and its isogenic CRISPR/Cas9-edited Nup133−/− derivatives (#14 and #19; Souquet et al., 2018), and the 129/Sv blastocyst-derived control (#1A4) and Nup133−/− mutant (merm, #319) mESC lines (Lupu et al., 2008) (Table S1). This analysis revealed that the transcriptomes of pluripotent WT and Nup133−/− mESCs were overall similar, with 43 down- and only two up-regulated genes, based on an adjusted P<0.05 and a |log2Fold Change|>2, whereas an increasing number of genes were misregulated (and notably upregulated) at day 3 of differentiation, with 57 down- and 61 up-regulated genes, using the same cutoffs (Fig. 1D; Fig. S1A).
We assayed by RT-qPCR the altered expression of a subset of these genes, filtered by criteria of differential expression [log2 fold change (FC)>2 or <−2], significance (adjusted P<0.05) and finally by average expression level (average number of reads with a log2(CPM)>1, to ensure proper detection by RT-qPCR) (Fig. S1B–E). In addition to WT (HM1) and Nup133−/− (#14) used for the initial RNA-seq experiment, these analyses were also conducted on samples from Nup133 ‘Rescue’ (denoted Rescue) cell lines generated by inserting the GFP–Nup133 transgene in Nup133−/− (#14) mESCs at the permissive Tigre locus (Zeng et al., 2008). As an additional control, we used an HM1-derived cell line that carries a transgene (OsTIR) similarly inserted in the Tigre locus [denoted WT (OsTIR)] (Fig. S2A, Table S1). In contrast to the impaired viability of Nup133−/− mESCs upon neuroectodermal differentiation, the survival of the Rescue and WT (OsTIR) cell lines were similar, confirming the functionality of the GFP–Nup133 transgene (Fig. 2A–C).
For the candidate genes localized on the short arm of the Y chromosome [Ddx3y and Eif2s3y, also located in close proximity to the loci of Uty, Uba1y, Kdm5d and Zfy (Subrini and Turner, 2021)] we observed clone-dependent expression variations (Fig. S1B). This suggests that their apparent shared misregulations, also reported in Tet1 and Tet2 mutant mESCs (Huang et al., 2014), might be in our case caused by partial loss or silencing of this genomic region, independently of Nup133 deficiency.
In contrast, we could validate the increased mRNA levels in Nup133−/− compared to WT cells for Nuggc at day 0, and of Nup210l and Lhx1 at day 3 of differentiation (Fig. S1C). We also confirmed the reduced mRNA levels in Nup133−/− compared to WT cells for Magohb and Wfikkn1 (but not Acta2) at day 3 of differentiation (Fig. S1D). Finally, reduced mRNA levels of the assayed candidate genes at the pluripotent state (day 0) were not significant due to high variability among replicates or between control cell lines (HM1 and OsTIR) (Fig. S1E). Importantly, among the validated candidate genes, Lhx1, Nup210l, Nuggc and Magohb were all efficiently restored to wild-type levels by the GFP-Nup133 transgene.
Lhx1 is a transcription factor involved in kidney and brain differentiation (Costello et al., 2015; Delay et al., 2018; McMahon et al., 2019; Shawlot et al., 1999), two organs affected in rare genetic diseases linked to Nup133 mutations, namely steroid-resistant nephrotic syndrome and Galloway–Mowat syndrome (Braun et al., 2018; Fujita et al., 2018). We further focused on this gene because of its complex misregulation in Nup133−/− cells. Indeed, although we confirmed by RT-qPCR the upregulation of Lhx1 expression at day 3 of differentiation (Fig. S1C), Lhx1 was subsequently downregulated again at later time-points (days 5 and 7 of differentiation) in Nup133−/− cells compared to WT and Rescue cells (Fig. 2D).
The other candidate gene we further characterized, Nup210l, is the differentially expressed gene (DEG) with the most significant P-value at day 3 of differentiation (Fig. 1D). It is also one of the rare DEGs whose expression had already increased at day 2 compared to WT cells (Fig. S1A). In mice, Nup210l mRNA is mainly detected in the testis and to a lesser extent in the embryonic brain (https://www.ncbi.nlm.nih.gov/gene/77595); in humans, besides testis, NUP210L expression is also detected in the prefrontal cortex neurons of rare individuals (Gusev et al., 2019). Analyses at later stages (day 5 and 7) of differentiation towards neuroectoderm showed that Nup210l was still more expressed in Nup133−/− compared to WT and Rescue cells, although a progressive increase of its expression was also observed in the latter cell lines (Fig. 2E).
As its name implies, Nup210l encodes a potential homolog of the transmembrane nucleoporin Nup210 (also known as gp210) with 43% overall amino acid identity between the two proteins (Fig. 2F). However, its putative NPC localization had never been established. To address this issue, we generated a GFP-tagged construct encompassing the minimal NPC targeting determinants previously established for gp210/Nup210 (Wozniak and Blobel, 1992), namely the Nup210L predicted signal peptide, transmembrane domain and C-terminal domain (Fig. 2F). This Nup210L mini construct when expressed at low levels in mESCs is targeted to the nuclear envelope where it colocalizes with the nucleoporin Tpr. This indicates that, like its homolog, Nup210L is indeed a nucleoporin (Fig. 2F).
The middle-domain of Nup133 is required for mESC differentiation, gene regulation and nuclear basket assembly in neuronal progenitors
Having established the requirement of Nup133 for cell viability upon differentiation and for the regulation of a subset of genes, we next aimed to determine how Nup133 contributes to these processes. We first focused on the middle domain of Nup133, which is necessary for the proper assembly of the nuclear pore basket in pluripotent mESCs (Souquet et al., 2018). We therefore integrated the pCAG-GFP-Nup133Δmid transgene in HM1-derived Nup133−/− mESCs at the Tigre locus (Fig. S2A). The GFP–Nup133Δmid protein levels in the resulting Nup133Δmid cells were comparable to those of endogenous Nup133 and of GFP–Nup133 in the Rescue cell lines throughout differentiation (Fig. 2B). Cell counts upon monolayer differentiation towards the neuroectodermal lineage showed, for the Nup133Δmid cell lines, a viability phenotype intermediate between WT and Nup133−/−, indicating that the middle domain of Nup133 is required for some, but not all, of the functions of this nucleoporin upon differentiation (Fig. 2C). In contrast, RT-qPCR analysis showed that Nup210l and Lhx1 were similarly misregulated upon neuronal differentiation in Nup133Δmid and in Nup133−/− cells (Fig. 2D,E).
The improved survival upon differentiation of Nup133Δmid compared to Nup133−/− cells enabled us to perform immunofluorescence analyses to determine whether the NPC basket assembly defects, previously observed in pluripotent mESCs lacking Nup133 or its middle domain, also occurred at the differentiated stage. Quantitative immunofluorescence analyses, performed after 5 days of differentiation, showed that the intensity of Tpr at the nuclear envelope was comparable between the WT and Rescue cell lines. In contrast, a two-fold decrease was observed in Nup133Δmid neuronal progenitors (Fig. 3A,B), a defect comparable to the one previously observed at the pluripotent state (Souquet et al., 2018). In addition, we also measured an increased Nup153 intensity at the nuclear envelope in Nup133Δmid progenitors compared to that in neuronal progenitors from WT or Rescue cell lines (Fig. 3C). This increased level of Nup153 is unlikely to solely reflect a global increase in NPC number as reported upon Tpr depletion in other cell lines (McCloskey et al., 2018), given that similar quantifications revealed a milder increase of Nup98 intensity at the nuclear envelope compared to Nup153 (Fig. 3D). Likewise, an increased accessibility of the Nup153 epitope when Tpr is absent seems unlikely, as such an effect was not previously observed in Nup133−/− mESCs at the pluripotent stage (Souquet et al., 2018). The high level of Nup153 observed might therefore reflect an increased stoichiometry of Nup153 per NPC in Nup133Δmid compared to in control-derived neuronal progenitors, possibly reflecting different stages of differentiation as previously described (Toda et al., 2017).
Nup133-dependent gene regulation and nuclear basket assembly can be uncoupled
Having identified a critical function for the middle domain of Nup133 in gene regulation, we next aimed to determine the levels of Nup133 required for this process. We therefore established Nup133-degron cell lines, which allow auxin-mediated degradation of a GFP-tagged allele of Nup133 in an OsTIR-expressing mESC line (Gonzalez-Estevez et al., 2021; see Materials and Methods and Fig. S2B).
The resulting Nup133-degron cell lines maintained normal Nup133 mRNA expression during differentiation (Fig. S3A), but actual Nup133 protein levels (without auxin treatment) were only ∼25% of that found in WT cells (Fig. 4A; Fig. S3B). This could be due to leaky OsTIR-induced degradation as previously reported (Mendoza-Ochoa et al., 2019; Yesbolatova et al., 2020), decreased stability of the tagged nucleoporin, or impaired export or translation of its mRNA. Nevertheless, these cells properly differentiated in the absence of auxin (Fig. 4C; Fig. S3D,E). As anticipated, addition of auxin to the medium throughout differentiation led to a Nup133−/−-like phenotype – normal growth at the pluripotent state but massive cell death during neuronal differentiation (Fig. 4C; Fig. S3C,D).
Importantly, the lower Nup133 levels observed in the degron cell lines in the absence of auxin was not accompanied by the altered expression of Nup210l or Lhx1 during neuronal differentiation (Fig. 4E). In contrast, more extensive depletion of Nup133 upon continuous auxin treatment mimicked the effect of Nup133 inactivation on these genes (Fig. 4E).
Quantification of the levels of the nuclear basket protein Tpr in the Nup133-degron cell lines revealed that, even in the absence of auxin, Tpr levels at the nuclear envelope were already reduced to ∼60% of the WT levels both in pluripotent cells and in neuronal progenitors (Fig. 4D; Fig. S4A). In contrast, the levels of Nup98 were not reduced in Nup133-degron cells (Fig. S4C), consistent with a largely unaltered NPC density and a specific alteration of the nuclear basket. The minor – i.e. less than 10% – increase in Nup98 intensity observed in one of the two cell lines (Nup133-degron #1) might reflect modest clonal-dependent variations of NPC density. Finally, the levels of Nup153 were very mildly increased only in the Nup133-degron #1 cells, with a similar trend in both undifferentiated and differentiated cells (Fig. S4B,D). These results indicate that a 4-fold reduction of Nup133 protein levels in untreated Nup133-degron cells is sufficient to severely impair Tpr recruitment or stabilization at nuclear pores, without major additional impact on NPC density.
Finally, although most of the GFP–mAID-Nup133 protein was already degraded after 30 min of auxin treatment in differentiated cells (Fig. 4B), a 16 to 24 h auxin treatment of these cells only led to a modest additional decrease of Tpr levels at the nuclear envelope compared to the untreated Nup133-degron cells (Fig. 4D; Fig. S4A).
Overall, these results thus demonstrate that a correct Nup133 stoichiometry is critical for nuclear basket assembly yet is not required for cell viability or gene regulation upon neuroectodermal differentiation. Taken together these data also reveal that a properly assembled nuclear basket at all NPCs is not required to regulate the expression of Nup133-target genes.
Nup210l mRNA levels rapidly increase in response to Nup133 or Seh1 depletion
We next aimed to determine whether the altered expression of Nup210l and Lhx1 was specific for Nup133 or was shared by other Y-complex constituents. In view of the requirement for Seh1 in global NPC assembly, distinct from the specific basket assembly defect of Nup133 mutant cells (Gonzalez-Estevez et al., 2021), we chose to assess its role at early stages of differentiation. Despite a very extensive death of Seh1−/− mutant cells upon differentiation (Gonzalez-Estevez et al., 2021), we could recover some mRNAs from these cells at day 3 of differentiation. As observed for Nup133−/− mESCs (Fig. 1A), Seh1−/− cells properly repressed pluripotency markers and were able to induce early differentiation markers (Fig. S5A). Importantly, mRNA levels of both Nup210l and Lhx1 were aberrantly increased in Seh1−/− cells at day 3 of differentiation as also observed in differentiating Nup133−/− cells (Fig. 5A).
The low viability of Seh1−/− cells upon differentiation (Gonzalez-Estevez et al., 2021) did not allow us to perform quantitative immunofluorescence studies at the differentiated stage. We therefore established and validated new Seh1-degron cell lines (see Materials and Methods and Fig. S2C), in which a C-terminally tagged form of Seh1 was properly and homogeneously expressed upon differentiation (Fig. 5B; Fig. S5B,C). In the resulting Seh1-degron cells, 24 h addition of auxin at day 2 or 4 of differentiation led to impaired viability of the cells (Fig. 5C). This indicates that Seh1 is not solely required at the early onset of neuronal progenitor differentiation, but also for the proper growth or viability of the progenitors themselves. Analyses of nuclear pore assembly in Seh1-degron-derived neuronal progenitors (at day 5 of differentiation) did not reveal major defects in the absence of auxin (Fig. 5D). In contrast, a 16-h treatment with auxin led to a ∼30% decrease of both Tpr and Nup98 intensities at the nuclear envelope of neuronal progenitors compared to that in the control cells (Fig. 5D; Fig. S4E). This suggests a global decrease in pore number upon Seh1 depletion, comparable to the observations previously made in pluripotent Seh1−/− and GFP-mAID-Seh1 mESCs (Gonzalez-Estevez et al., 2021). Note that unlike Tpr and Nup98, Nup153 levels were not altered in auxin-treated Seh1-degron cells, suggesting, as also observed in Nup133Δmid cells, an increased stoichiometry of Nup153 per NPC (Fig. S4F).
We also note that a 24 h auxin treatment, applied to Seh1-mAID-GFP cells at day 2 of neuronal differentiation, was sufficient to cause a significant increase in Nup210l mRNA levels (Fig. 5E). Likewise, a 24 h auxin treatment of Nup133-degron cells induced Nup210l expression (Fig. 5E). In contrast, 24 h of auxin treatment did not lead to an altered regulation of Lhx1 in Seh1-degron or Nup133-degron cells at day 3 (Fig. 5F). Together, these data indicate that Nup210l and Lhx1 are shared downstream targets of Nup133 and Seh1, with Nup210l appearing to be a gene induced early upon loss of these Y-complex Nups. On the other hand, Lhx1 seems to need a prolonged depletion of these Y-complex Nups to become misregulated, indicating that it is likely an indirect target of Y-complex Nup-dependent regulations.
In this study, we have shown in differentiating mESCs that a subset of genes is deregulated in the absence of Nup133. We found that neuronal progenitors lacking either Nup133 or just its middle domain share a defect in nuclear basket assembly and altered expression of Nup210l and Lhx1. Although this indicates a dual function for the middle domain of Nup133, this domain is large enough (416 amino acids, 17 α-helices) to possibly allow simultaneous interactions with another nucleoporin and a gene expression regulator. Alternatively, this large deletion, by shortening the length of Nup133, might also alter the head-to-tail interactions between consecutive Y-complexes, in turn affecting nuclear basket assembly or stability.
However, the basket assembly and altered gene expression phenotypes can be uncoupled. Indeed, Nup210l and Lhx1 were not misregulated in our Nup133-degron cell lines, which display a constitutive nuclear basket assembly defect, and conversely, they were both similarly misregulated in Seh1-deficient cells in which nuclear basket assembly is not specifically altered. As the untreated Nup133-degron cell lines exhibit lower Nup133 protein levels than the control cell lines, our data further argue that a limited amount of Nup133 is sufficient to keep Nup210l repressed in differentiating mESCs and to induce the proper and timely expression of Lhx1. Because this function in gene regulation is shared by Nup133 and Seh1, two physically distant members of the Y-complex, it likely involves the whole Y-complex rather than each of these two individual subunits.
The rather short lag time (below 24 h) between auxin-induced degradation of Nup133 or Seh1 and Nup210l activation suggests that there could be a direct contact between the Y-complex and the Nup210l genomic locus. Although the Y-complex is mainly visualized at NPCs to which it is stably anchored (Rabut et al., 2004), a diffuse fraction is also likely present in the nucleus, as previously described in HeLa cells (Morchoisne-Bolhy et al., 2015). Hence, Y-complex-dependent gene regulation might take place either at NPCs or ‘off-pore’. Browsing available data of LaminB1-Dam-ID tracks (Peric-Hupkes et al., 2010), we noticed that the Nup210l locus is adjacent to a lamin-associated domain (LAD) in mESCs and neural progenitor cells. A location near the nuclear periphery would be consistent with a regulation of Nup210l taking place at NPCs. In line with this hypothesis, Seh1 has been shown to recruit to the NPC the transcription factor Olig2 and the chromatin remodeler Brd7 to promote the expression of differentiation genes in oligodendrocytes (Liu et al., 2019). Additionally, Nup133 has been proposed to promote the expression of Myc in cancer cells by anchoring its super-enhancer to the NPCs (Scholz et al., 2019).
This Y-complex-mediated gene regulation may also involve epigenetic mechanisms, as reported for Nup153, which interacts with PRC1 to repress developmental genes (Jacinto et al., 2015). Along these lines, it is noteworthy that human NUP210L, initially thought to be a testis-specific gene, was found to be expressed in the prefrontal cortex neurons of some individuals. This regulation was linked to the epigenetic allele-specific activation of NUP210L, namely the deposition of the permissive histone mark H3K4me3 at its promoter (Gusev et al., 2019). In addition, another epigenetic mechanism, the DNA methylation state of NUP210L, has been linked to psychologic development disorders in individuals carrying a hemizygous 22q11.2 microdeletion (Starnawska et al., 2017). Considering its possible link with normal or pathological cognitive abilities, the mechanisms of Nup133- and Seh1-dependent Nup210l activation warrant further investigation. It is quite remarkable that Nup210l, a gene encoding a nucleoporin, is one of the earliest and most robustly upregulated genes when Y-complex-deficient mESCs are induced to differentiate towards neuronal fates. Although not sufficient to prevent the impaired differentiation of Nup133- and Seh1-deficient mESCs, Nup210L induction might be a tentative mechanism to compensate for improper Y-complex (and NPC) assembly in differentiating cells. Nup210L might, for example, help to stabilize NPC structure or to promote NPC assembly in challenging conditions. As described for its homolog, Nup210, it might also participate in gene regulation at NPCs, or contribute via its conserved large N-terminal domain to nuclear envelope or ER homeostasis (Raices et al., 2017; Gomez-Cavazos and Hetzer, 2015). Finally, Nup210L might also indirectly impact chromatin compaction, as hypothesized in the context of impaired spermiogenesis (Arafah et al., 2021). More studies are needed to elucidate its functions in normal and pathological situations.
MATERIALS AND METHODS
mESCs culture and neuroectodermal differentiation
Cell lines used in this study are listed in Table S1 and were grown as previously described (Gonzalez-Estevez et al., 2021). Briefly, mESCs were grown at 37°C and 5% CO2 on Mitomycin-C inactivated feeder cells (DR4-mouse embryonic fibroblast) plated on 0.1% gelatin (Sigma-Aldrich, St Louis, MO, USA) in serum/leukemia inhibitory factor (LIF, ESGRO, Millipore, Burlington, MA, USA)-containing stem cell medium. mESCs were used at passages below 30. Lack of contamination between the mutant cell lines was assessed by PCR on genomic DNA, proper GFP expression when pertinent and western blots analyses. Frequent DAPI staining ensured lack of major contamination by mycoplasma.
The neuroectodermal differentiation protocol used in this study was adapted from Abranches et al. (2009) and Ying et al. (2003), as previously described (Gonzalez-Estevez et al., 2021). Briefly, following trypsinization and feeder removal, mESCs were resuspended in N2B27 medium (neurobasal medium, DMEM-F12, 7.5% BSA, N2 supplement, B27 supplement, penicillin-streptomycin, L-glutamin and β-mercaptoethanol) and plated at a density of ∼0.85×104 or 3×104 cells/cm2 on gelatin-coated dishes (day 0). Medium was changed every day from day 2 onwards. To stimulate neuronal differentiation, 1 μM all-trans-retinoic acid (RA; Sigma-Aldrich) was added to the medium for 24 h on day 2.
For annexin V/propidium iodide (PI) apoptosis/viability assays, cells were trypsinized, counted, and 105 cells were centrifuged at 400 g for 3 min. Cells were resuspended in 500 µl of Annexin V-binding buffer (ab14084, Abcam) and incubated with 1 µl annexin V–Cy5 (ab14147, Abcam) and 10 µg/ml propidium iodide for 5 min at room temperature in the dark. Cells were then analyzed by flow cytometry using a CyanADP Cytomation (Beckman–Coulter), using SS (side-scatter) and FS (forward scatter) to remove debris and exclude cell doublets, and 488 nm and 635 nm excitation lasers. At least 10,000 cells were processed, and data were then analyzed using the Summit software.
To induce degradation of the GFP–mAID-Nup133 (in Nup133-degron cells) and Seh1–mAID-GFP (in Seh1-degron cells), 500 µM auxin (Sigma-Aldrich) was added to the medium (from a 280 mM stock in ethanol). The same final concentration of ethanol was added for control experiments.
RNAs were extracted at days 0, 2 and 3 of differentiation from three independent Nup133−/− clones (KO#1, HM1-derived Nup133−/− #14; KO#2, HM1-derived Nup133−/− #19; KO#3, blastocyst-derived #319 Nup133merm/merm mESCs), and from three control samples (WT#1; HM1; WT#2, HM1; WT#3, blastocyst-derived #1AA Nup133+/+) (see Table S1). Library preparation and Illumina sequencing were performed at the Ecole normale supérieure genomics core facility (Paris, France). Messenger (polyA+) RNAs were purified from 1 µg of total RNA using oligo(dT). Libraries were prepared using the strand-specific RNA-Seq library preparation TruSeq Stranded mRNA kit (Illumina). Libraries were multiplexed by 9 on 2 flowcells. Two 75 bp single-read sequencing runs were performed on a NextSeq 500 device (Illumina). A mean±s.d. number of 52.04±14.82 million reads passing Illumina quality filter was obtained for the 18 samples.
The analyses were performed using the Eoulsan pipeline version 2.0-alpha7 (Jourdren et al., 2012), including read filtering, mapping, alignment filtering, read quantification, normalization and differential analysis. Before mapping, poly N read tails were trimmed, reads ≤40 bases were removed, and reads with quality mean ≤30 were discarded. Reads were then aligned against the Mus musculus genome from Ensembl version 81 using STAR (version 2.4.0k) (Dobin et al., 2013). Alignments from reads matching more than once on the reference genome were removed using Java version of samtools (Li et al., 2009). To compute gene expression, Mus musculus GFF genome annotation version 81 from Ensembl database was used. All overlapping regions between alignments and referenced exons were counted and aggregated by genes using the HTSeq-count algorithm (Anders et al., 2015). A first analysis revealed that one of the samples (KO#2 day2) featured an abnormally high level of ribosomal transcripts; this dataset was therefore excluded from subsequent analyses.
The rest of the analysis was carried out using the bioinformatics software R [R v4.1.2 (2021.11.01)], and open access packages, as recommended from the publicly available bioinformatics course at https://diytranscriptomics.com/ (Berry et al., 2021). Mapped raw counts were transformed in counts per million (cpm) using the CPM function from the EdgeR package (v3.34.1). We filtered the genes that had a log2(CPM) above one for more than three samples, and then normalized their cpm using the trimmed mean of M-values (TMM) method (Robinson and Oshlack, 2010). The mean-variance relationship of the filtered normalized data was modeled by ‘voom’ transformation, and a linear model was fitted to the data using the lmfit function from the limma package (v3.48.3). Bayesian statistics for the chosen pair-wise comparisons (average knockout expression compared to average WT expression for each time point) were then calculated using the eBayes function from limma, and adjusted with the Benjamini–Hochberg (BH) correction. An exhaustive list of differentially expressed genes (P<0.05 and |log2FC|>1.5) was pulled-out using the decideTests function. Plots in Fig. 1D and Fig. S1A were generated in R using ggplot2 (v.3.3.5).
Transfection and CRISPR/Cas9 genome editing
mESCs were transfected as previously described (Souquet et al., 2018) using Lipofectamin 2000 (Invitrogen, Carlsbad, CA, USA) according to manufacturer instructions. To establish the Nup133-Rescue, Nup133-Δmid, Nup133-degron and Seh1-degron cell lines, 5×105 mESCs were co-transfected with 3 µg of a plasmid directing the expression of a gRNA and high fidelity (HF) Cas9 fused to mCherry and with 3 µg of DNA sequences of interest flanked by homology directed repair arms (PCR product or linearized plasmid, see Fig. S2). Plasmids are listed in Table S2, gRNAs designed using the Benchling website (https://benchling.com) are listed in Table S3, and PCR primers used to generate homology-directed repair templates are listed in Table S4. At 3 days after transfection, GFP-positive cells were FACS-sorted to select for cells expressing the tagged nucleoporin and plated on culture dishes. Individual clones were picked 6–7 days after sorting and characterized using immunofluorescence, western blot, PCR on genomic DNA and sequencing. Ploidy was assessed using chromosome spreads. Cell line characteristics are summarized in Table S1.
RNA extraction was performed using NucleoSpin RNAII isolation kit (Mascherey-Nagel) according to the manufacturer's instructions. Reverse-transcription (RT) was undertaken with the transcriptase inverse Superscript II (Invitrogen) and random hexamers (Amersham Pharmacia), using at least 150 ng of RNA per sample. Real-time quantitative PCR was performed with SybrGreen reagents (Applied Biosystems) on a LightCycler 480 instrument (Roche Life Sciences). All mRNA level results are presented as relative to the TATA-binding protein (TBP) mRNA levels. qPCR primers used in this study are listed in Table S5.
Western blots analyses
Whole-cell lysate preparations and western blot analyses were performed as previously described (Gonzalez-Estevez et al., 2021), using 4–10% SDS-PAGE gels (Mini-Protean TGX Stain free precast gels, Bio-Rad, Hercules, CA, USA) and nitrocellulose membranes (GE Healthcare). Incubations with primary antibodies were carried overnight at 4°C. Signals from HRP-conjugated secondary antibodies were detected by enhanced chemiluminescence (SuperSignal® Pico or Femto, Thermo Fisher Scientific) using ChemiDoc (Bio-Rad). Antibodies used in this study are listed in Table S6. The uncropped membranes are presented in Fig. S6.
Immunofluorescence and quantification of nucleoporin intensity at the nuclear envelope
Cells grown on glass coverslips coated with 0.1% gelatin were fixed for 20 min in 3% paraformaldehyde (VWR, Radnor, PA, USA) (resuspended in PBS and brought to pH 8.0 with NaOH), permeabilized 30 min in H-Buffer (PBS, 1% BSA, 0.2% Triton X-100 and 0.02% SDS) and incubated with the primary and secondary antibodies for 1 h at room temperature in H-Buffer, with washes in H-Buffer in-between. Antibodies used in this study are listed in Table S5. Coverslips were then incubated 5 min with DAPI (Sigma-Aldrich, 280 nM in PBS) and mounted using Vectashield (Vector, Maravai Life Sciences, San Diego, CA, USA). Images were acquired on a DMI8 microscope (Leica Microsystems), equipped with a CSU-W1 spinning-diskhead (Yokogawa, Japan) and 2 Orca-Flash 4 V2+ sCMOS cameras (Hamamatsu), using 100×1.4 NA oil objectives
Quantification of nucleoporin intensities at the nuclear envelope was performed essentially as described (Souquet et al., 2018), by mixing the cell line of interest with a reference cell line, either the WT (OsTIR) cell line or the Nup133-Rescue line, as indicated. For each field, we measured the mean intensity of 8-pixel-thick lines drawn on the nuclear rims and of a background area. After subtraction of the background, the nuclear envelope intensity value obtained for each cell was normalized to the mean value obtained for the reference cells acquired in the same field. Box plots were generated using GraphPad Prism (GraphPad Software): each box encloses 50% of the normalized values obtained, centered on the median value. The bars extend from the 5th to 95th percentiles. Values falling outside of this range are displayed as individual points. Statistical analyses were performed using unpaired nonparametric Mann–Whitney tests. ****P<0.0001, ***P<0.001, **P<0.01, *P<0.05.
We are grateful to Vedrana Andric, Alessandro Berto, Marta Boira and Salomé Neuvendel for help in mESCs culture, differentiation, cell line establishment, or western blot analyses. We thank Charlène Boumendil and Pierre Therizols for helpful discussions, and Benoit Palancade, Roger Karess, Charlène Boumendil and the Doye lab members for critical reading of the manuscript. We also acknowledge the ImagoSeine core facility of the Institut Jacques Monod for help with cell sorting, FACS analyses, and spinning disk imaging. The ImagoSeine core facility was supported by funds from the GIS-IBISA (groupement d’intérêt scientifique-Infrastructure en biologie santé et agronomie) and the France-Bioimaging (ANR-10-INBS-04) infrastructures and la Ligue contre le cancer (R03/75-79).The GenomiqueENS core facility was supported by the France Génomique national infrastructure, funded as part of the “Investissements d'Avenir” program managed by the Agence Nationale de la Recherche (contract ANR-10-INBS-0009).
Conceptualization: C.O., A.V., B.S., V.D.; Methodology: C.O., A.V., S.P., B.S., F.C., S.B., V.D.; Validation: C.O., A.V., S.P., B.S., L.J., S.B., V.D.; Formal analysis: C.O., A.V., S.P., B.S., L.J., S.B., V.D.; Investigation: C.O., A.V., S.P., B.S., F.C., L.J., S.B., V.D.; Resources: C.O., A.V., S.P., B.S., L.J., V.D.; Data curation: C.O., A.V., B.S., L.J., S.B., V.D.; Writing - original draft: C.O., L.J., V.D.; Writing - review & editing: C.O., A.V., B.S., V.D.; Visualization: C.O., A.V., S.B., V.D.; Supervision: V.D.; Project administration: V.D.; Funding acquisition: V.D., C.O., B.S.
Work in the laboratory of V.D. was supported by the Centre national de la recherche scientifique (CNRS), the Fondation pour la Recherche Médicale (FRM, Foundation for Medical Research) under grants No DEQ20150734355, Equipe FRM 2015 and EQU202003010205, Equipe FRM 2020 to V.D., and by the Labex Who Am I? (ANR-11-LABX-0071; Idex ANR-11-IDEX-0005-02). C.O. received PhD fellowships from Ecole Doctorale BioSPC, Université Paris Cité and from the Fondation pour la Recherche Médicale (fourth year), A.V. received a post-doc grant from the Labex Who Am I? (2019 post-doc call) and B.S. was supported by la Fondation ARC pour la Recherche sur le Cancer (PDF 20130606747).
The RNA-seq gene expression data and raw fastq files are available on the GEO repository under accession number GSE218080.
Peer review history
The peer review history is available online at https://journals.biologists.com/jcs/lookup/doi/10.1242/jcs.261151.reviewer-comments.pdf
The authors declare no competing or financial interests.