Nup2 requires a highly divergent partner, NupA, to fulfill functions at nuclear pore complexes and the mitotic chromatin region

Among nuclear pore proteins, Nup2 is unique because it transfers to the mitotic chromatin region to fulfill unknown functions. Analysis of Nup2 and a novel targeting partner, NupA, shows that they are required for normal anaphase and nucleokinesis. Their functions also involve an import pathway for Mad1 but apparently not general nuclear protein import.

Abstract

Chromatin and nuclear pore complexes (NPCs) undergo dramatic changes during mitosis, which in vertebrates and Aspergillus nidulans involves movement of Nup2 from NPCs to the chromatin region to fulfill unknown functions. This transition is shown to require the Cdk1 mitotic kinase and be promoted prematurely by ectopic expression of the NIMA kinase. Nup2 localizes with a copurifying partner termed NupA, a highly divergent yet essential NPC protein. NupA and Nup2 locate throughout the chromatin region during prophase but during anaphase move to surround segregating DNA. NupA function is shown to involve targeting Nup2 to its interphase and mitotic locations. Deletion of either Nup2 or NupA causes identical mitotic defects that initiate a spindle assembly checkpoint (SAC)–dependent mitotic delay and also cause defects in karyokinesis. These mitotic problems are not caused by overall defects in mitotic NPC disassembly–reassembly or general nuclear import. However, without Nup2 or NupA, although the SAC protein Mad1 locates to its mitotic locations, it fails to locate to NPCs normally in G1 after mitosis. Collectively the study provides new insight into the roles of Nup2 and NupA during mitosis and in a surveillance mechanism that regulates nucleokinesis when mitotic defects occur after SAC fulfillment.

INTRODUCTION

Nuclear pore complexes (NPCs) are large macromolecular assemblies spanning the nuclear envelope (NE) composed of multiple copies of proteins termed nucleoporins, commonly abbreviated to Nups. NPCs form channels across the NE that bridge the nucleoplasm and cytoplasm, allowing the diffusion of small molecules and active transport of larger proteins and nucleic acids (Alber et al., 2007; Stewart, 2007; D'Angelo and Hetzer, 2008). The core structure of the NPC is anchored in the NE via transmembrane Nups in combination with core subcomplexes. Additional peripheral Nups form a nuclear basket and cytoplasmic fibrils, whereas peripheral phenylalanine-glycine (FG)–repeat Nups reside within the central transport channel (Strambio-De-Castillia et al., 2010; Wente and Rout, 2010; Raices and D'Angelo, 2012; Schwartz, 2013). FG-repeat Nups of the central transport channel are intrinsically disordered and can form hydrogel-like barriers (Frey et al., 2006; Frey and Gorlich, 2007). Cargo–carrier complexes are docked and translocate through the NPC via interactions between importin or exportin carriers and the FG-repeat proteins (Bayliss et al., 2000).

During open-type mitosis when NPCs undergo disassembly, several Nups have been shown to locate to kinetochores, spindle pole bodies, and mitotic chromatin (Belgareh et al., 2001; Harel et al., 2003; Loiodice et al., 2004; Galy et al., 2006; Orjalo et al., 2006; Osmani et al., 2006a; Rasala et al., 2006; Liu et al., 2009; Xu and Powers, 2010), indicating that Nups have evolved to mediate mitotic functions away from NPCs (De Souza and Osmani, 2007, 2009; Guttinger et al., 2009; Bolhy et al., 2011). Nups including Rae1, Nup98, and components of the core Nup107/160 subcomplex regulate spindle formation and anaphase onset in Xenopus extracts, Caenorhabditis elegans, mice, and human cells (Kraemer et al., 2001; Joseph et al., 2002; Pichler et al., 2002; Arnaoutov et al., 2005; Blower et al., 2005; Jeganathan et al., 2005, 2006; Orjalo et al., 2006; Wong et al., 2006; Zuccolo et al., 2007; Platani et al., 2009; Mishra et al., 2010; Cross and Powers, 2011). In addition, the NPC nuclear basket protein Nup153 is required for early mitotic progression, as well as for the resolution of midbodies during cytokinesis in HeLa cells (Mackay et al., 2009). Nup153 depletion leads to defects in NPC basket assembly and activates an Aurora B–mediated abscission checkpoint (Mackay et al., 2010). Moreover, the spindle assembly checkpoint (SAC) proteins Mad1 and Mad2 locate to the NPC nuclear basket in mammalian cells (Campbell et al., 2001), Saccharomyces cerevisiae (Iouk et al., 2002), Aspergillus nidulans (De Souza et al., 2009), plant cells (Ding et al., 2012), and Drosophila melanogaster (Buffin et al., 2005; Katsani et al., 2008), suggesting a conserved functional relationship between the SAC and NPCs. A recent study also found that during interphase, NPCs provide a platform for generation of mitotic Mad2-inhibitory complexes previously believed to be generated exclusively from incorrectly attached mitotic kinetochores (Rodriguez-Bravo et al., 2014). In A. nidulans, the nuclear basket protein Mlp1, in addition to targeting Mad1 and Mad2 to interphase NPCs, also acts as a scaffold to locate Mad1 and Mad2 near kinetochores and the telophase spindle (De Souza et al., 2009), similar to its human and Drosophila counterparts, Tpr and Mtor. Mtor promotes the recruitment of Mad2 but not Mad1 to unattached kinetochores (Lince-Faria et al., 2009). Depletion of Tpr decreases the levels of Mad1 at kinetochores during prometaphase and leads to a defective SAC response (Lee et al., 2008; Schweizer et al., 2013).

Nup2, a mobile nucleoporin, is also part of the NPC nuclear basket. Nup2 proteins have a conserved overall domain structure divided into three domains: an N-terminal importin-α–binding domain, a domain rich in FG repeats, and a C-terminal Ran GTP–binding domain. Yeast Nup2p and the orthologous vertebrate Nup50 also contain an NPC-targeting domain, which follows the N-terminal importin-α–binding domain (Booth et al., 1999; Hood et al., 2000; Matsuura et al., 2003; Matsuura and Stewart, 2005). Budding yeast Nup2p forms a multiprotein complex with Nup60p, Kap60 (importin-α), Kap95 (importin-β), Prp20p/RCC1, and Gsp1p/RanGTP (Allen et al., 2001, 2002; Denning et al., 2001, 2002; Dilworth et al., 2001) and accelerates the rate of disassembly of Kap60p–cargo complexes (Gilchrist et al., 2002; Gilchrist and Rexach, 2003). Nup2p facilitates this process by strongly binding to importin-α (Solsbacher et al., 2000) and also participates in the nuclear export of importin-α (Booth et al., 1999; Hood et al., 2000). Nup50 (mammalian Nup2) also interacts with the importin-α/β–RanGTP complex to stimulate nuclear import (Lindsay et al., 2002). Nup2 has additionally been implicated in gene regulation (Liang and Hetzer, 2011) and binds to a set of promoters in yeast to activate gene expression (Casolari et al., 2004; Schmid et al., 2006). Nup2p also acts as a physical tether for chromatin boundary activities (Ishii et al., 2002) and plays an active role in the transition of chromatin between different transcriptional activity states (Dilworth et al., 2005). Similarly, Drosophila Nup50 interacts with transcriptionally active genes (Kalverda et al., 2010), although in this case, the physical interactions occur in the nucleoplasm rather than at NPCs. In addition, recent work shows that Nup50 is required for mouse myoblast differentiation and plays transport-independent roles in chromatin biology that occurs away from the NPC (Buchwalter et al., 2014).

A. nidulans Nup2 also has a predicted N-terminal importin-α–binding domain (amino acids [aa] 1–50), FG repeats (3 FXFGs, 3 GXFGs, and 15 FGs), and a C-terminal Ran-GTP– binding domain (aa 1285–1404; Osmani et al., 2006a) but is a significantly larger protein than its yeast and human counterparts. A. nidulans Nup2 is essential, and its deletion leads to mitotic DNA segregation defects. Moreover, A. nidulans Nup2 translocates to the main chromatin region from NPCs during mitosis (Osmani et al., 2006a), a transition conserved in rat kidney cells (Dultz et al., 2008). A study also identified Nup50 in the protein composition of mitotic chromosomes (Ohta et al., 2010). Nup2 might therefore have as-yet-unknown but conserved roles near mitotic chromatin.

In this study, by affinity purifying A. nidulans Nup2, we identify a previously uncharacterized essential A. nidulans nucleoporin we named NupA. NupA, like Nup2, translocates from interphase NPCs to locate to the mitotic chromatin region, and both play roles during mitosis that are monitored by the SAC and are also required for normal nucleokinesis. Although neither is involved in regeneration of transport-competent NPCs after mitosis, they are required for nuclear import of Mad1 during G1.

RESULTS

Nup2 is phosphorylated when it locates to the chromatin region during mitosis

Endogenous Nup2 tagged with green fluorescent protein (GFP) is functional and translocates from NPCs to the chromatin region during mitotic entry and relocates back to NPCs during mitotic exit (Osmani et al., 2006a). To determine whether Nup2 is located like Nup2-GFP, an antibody was generated and used for immunofluorescence microscopy. Similar to the GFP-tagged version, untagged Nup2 was found to locate to NPCs during interphase and the chromatin region during mitosis (Figure 1, A and B), indicating Nup2-GFP reflects the normal cell cycle– regulated locations of Nup2.

FIGURE 1:

Identification of a novel Nup via affinity purification of Nup2. (A) Western blot of A. nidulans total cell extracts from wild-type strain R153 using Nup2 antibody pretreated with or without the Nup2 antigen (Ag) as indicated. The antibody recognizes Nup2 during both G2 and mitosis (M). (B) Cells of strain R153 were fixed and labeled with the Nup2 antibody and DAPI stained to visualize DNA. Native Nup2 is at NPCs during interphase but localized near DNA during mitosis. Scale bar, ∼5 μm. (C) Coomassie-stained SDS–PAGE gel showing Nup2 copurified with a novel nucleoporin termed NupA identified by mass spectroscopy (from strain SO926). The first two samples depict affinity purifications performed for G2 and mitotic samples. The third purified mitotic sample was treated with λ phosphatase before electrophoresis. Sequence coverage obtained from mass spectroscopy: G2 sample: 72.9% for Nup2, 34.2% for NupA, 37.8% for KapB, and 36.2% for KapA; mitotic sample: 73% for Nup2, 36.7% for NupA, 34.1% for KapB, and 32.4% for KapA. (D) Cartoon depicting genomic structure of nupA from the indicated species. All contain a large intron between exons 1 and 2.

For biochemical analysis and to facilitate identification of copurifying partners via affinity purification, we generated an endogenously replaced, C-terminally S-tagged (Liu et al., 2010) version of Nup2. Both Nup2-S-Tag (146 kDa) and Nup2-GFP (171 kDa) migrate at an unexpected high molecular weight during SDS–PAGE (e.g., Nup2-S-Tag runs between the 160- and 250-kDa markers; Figure 1C), as does the endogenous protein (Figure 1A). The basis for this discrepancy in apparent size is unknown. However, in addition to the aberrant large size of Nup2 during SDS–PAGE, we also observed that mitotic Nup2 (Figure 1A) and mitotic Nup2-S-Tag (Figure 1C) appeared as a smear when compared with the more defined bands of G2 samples. In vitro phosphatase treatment caused collapse of the smear to a single band, indicating that Nup2 is a mitotic phosphoprotein (Figure 1C). Because the mitotic samples were generated after release of the temperature-sensitive nimT23 allele, the Cdc25 mitotic Cdk1–activating phosphatase in A. nidulans (Osmani et al., 1991; O'Connell et al., 1992), to a permissive temperature, the data indicate that mitotic Nup2 phosphorylation depends on mitotic Cdk1 activation (Figure 1, A and C).

A. nidulans NPCs undergo partial disassembly during mitosis, and the NIMA kinase is able to promote NPC disassembly when ectopically expressed, even during interphase (De Souza et al., 2004; De Souza and Osmani, 2007). We therefore sought to determine whether NIMA could promote the removal of Nup2 from NPCs and perhaps even trigger its transition to the chromatin region. We used a strain able to express a stable kinase-active, C-terminally truncated (NIMAΔC) version of NIMA (Pu and Osmani, 1995) under control of the regulatable alcA promoter (Waring et al., 1989). The expression of NIMAΔC was turned on in untreated cells or cells arrested in S phase by hydroxyurea (HU). The abundance of NIMAΔC and subsequent mobility shifts for endogenous NIMA and Nup2 were monitored via Western blot analysis (Figure 2A). An increase in the level of NIMAΔC occurred after 30 min of induction and accumulated further with time of induction. The increase in NIMAΔC levels correlated with mobility shifts in endogenous NIMA and Nup2, indicating that extra NIMAΔC promotes the phosphorylation of both. A similar pattern was seen after induction of NIMAΔC in the presence of HU, but in this case, induction of NIMAΔC was delayed, as were the mobility shifts of endogenous NIMA and Nup2 (Figure 2A, right).

FIGURE 2:

Induction of NIMA promotes phosphorylation and relocation of Nup2 from NPCs to the chromatin region even in S phase–arrested cells. (A) Western blot analysis of A. nidulans total cell extracts from strain SO1030 (alcA-NIMA) showing an upshift in Nup2 and endogenous NIMA upon induction of stable NIMA (NIMA-ΔC). NIMA*, endogenous NIMA at a higher exposure, showing its upshift upon induction of ectopic NIMA. (B) Cumulative timing of Nup2 transition from NPCs to the chromatin region upon induction of stable NIMA in the presence or absence of HU as indicated (wild-type strain HA377 and alcA-NIMA strain SO1030). (C) Representative images from live-cell microscopy following Nup2-GFP and NLS-DsRed with or without induction of NIMA as indicated, using strains HA377 and SO1030. Bar, ∼5 μm.

Further NIMAΔC induction experiments were completed using live-cell imaging in a strain with Nup2 tagged with GFP also expressing nuclear localization signal (NLS)–DsRed. Induction of NIMA promotes NPC disassembly and the nuclear leakage of NLS-DsRed, both normally mitotic-specific events (Suelmann et al., 1997; De Souza et al., 2004; Osmani et al., 2006a). The ability of induced NIMAΔC to promote Nup2 to the chromatin region was monitored with or without addition of HU (Figure 2B). Representative imaging in Figure 2C shows that control cells lacking inducible NIMAΔC but treated with alcA inducer were not promoted into a mitotic state. However, for cells in which NIMAΔC was induced, a mitotic state was promoted, as NLS-DsRed was released from nuclei and Nup2-GFP transferred from NPCs to the chromatin region. Quantitation of the cumulative number of cells in which Nup2-GFP located to the chromatin region revealed a delay in the transition of cells into a mitotic state with HU treatment (Figure 2B). The delay is similar to the delay we observed when induction of NIMAΔC caused mobility shifts for endogenous NIMA and Nup2 (Figure 2A). Of importance, treatment with HU caused cell cycle arrest in the control strain, as these cells failed to enter mitosis in the absence of NIMAΔC induction. Collectively the analysis indicates that Nup2 normally transitions to the chromatin region under control of mitotic-specific phosphorylation, which can be artificially promoted via increased interphase induction of NIMA.

Identification of NupA via affinity purification of Nup2

Nup2-S-tag affinity purification and mass spectrometry identified importins KapA (AN2142) and KapB (AN0906; Markina-Inarrairaegui et al., 2011) and two predicted hypothetical proteins, AN3174 and AN10382, as copurifying proteins from both G2 and mitotic cells. Further analysis indicated that AN3174 and AN10382 represent a single misannotated gene (Figure 1C, NupA). For example, pBLAST analysis using AN3174 and AN10382 identified a single gene in Aspergillus fumigatus: Afu3g13300, a predicted 140.5-kDa protein of 1299 aa. This suggests that the A. fumigatus gene could represent two genes, or, alternatively, AN3174 and AN10382 could represent a single gene. Consistent with the latter possibility, sequence analysis of A. nidulans cDNA defined a single transcript from a gene we termed nupA. nupA encodes a 768-aa protein of 84.2 kDa (Figure 1D and Supplemental Figure S1), although during SDS–PAGE, similar to Nup2, NupA ran at a higher-than-expected apparent molecular weight. This discrepancy was observed both with untagged NupA (Figure 1C) and after Western blot analysis of GFP-tagged NupA (unpublished data). Comparison between cDNA and genomic nupA identified three introns, the first being rather large (807 base pairs) for a fungal intron (Figure 1D and Supplemental Figure S1). The median length of A. nidulans introns is 59 base pairs, with 95% of introns being between 46 and 228 base pairs in length (Sibthorp et al., 2013), making the large size of the nupA intron conspicuous. To investigate whether A. fumigatus nupA contains a similar large intron, we sequenced the region spanning the predicted intron from an A. fumigatus cDNA library. This revealed that, similar to A. nidulans, nupA of A. fumigatus also contains a large (1386 base pairs) intron (Figure 1D). The size of A. fumigatus NupA is therefore 837 aa when the intron is spliced out. Given the large size of the intron, it is somewhat surprising that the open reading frame is maintained throughout the intron sequence. For nupA of Aspergillus niger and Aspergillus carbonarius, RNA-sequence data also support the existence of a large nupA intron within these species. Similar to A. fumigatus nupA, these intron sequences are devoid of in-frame stop codons (Scott Baker, Pacific Northwest National Laboratory, Richland, WA, personal communication). Of interest, tBLASTn analysis at the AspGD site using the A. fumigatus protein encoded through the intron showed that it has 90% identity in Neosartorya fischeri (score, 811 bits; E value, 0.0) and 47% identity in Aspergillus clavatus (score, 318 bits; E value 2e-86). Additional tBLASTn analysis at the National Center for Biotechnology Information detected similar sequences only in related aspergilli. This conservation suggests that the intron is subject to evolutionary selection and that it might be translated. However, using PCR to amplify across the intron sequences and gel electrophoretic analysis, we did not detect larger transcripts reflected in nupA cDNAs of A. nidulans or A. fumigatus. Using different growth conditions and developmentally staged cell types might reveal unspliced transcripts capable of encoding larger NupA isoforms.

NupA displays 46 and 49.5% identity at the amino acid level with its A. fumigatus and A. niger counterparts, respectively. This degree of similarity is low, as orthologues between A. nidulans and A. fumigatus average 68% amino acid identities (Galagan et al., 2005). This lack of sequence conservation hindered identification of orthologues of NupA outside the aspergilli and related fungi.

NupA is essential and required for normal mitosis but not short-term growth

Deletion of nupA led to the generation of diploid or heterokaryotic colonies, indicating that nupA, although poorly conserved, is essential. Using heterokaryon rescue (Osmani et al., 2006b), we identified heterokaryotic colonies that carried both ΔnupA and parental nuclei (Figure 3A). Diagnostic PCR confirmed the presence of the two genetically distinct types of nuclei (Figure 3Ab). ΔnupA mutant spores generated from the heterokaryotic colonies initially germinated and grew at a rate similar to wild-type spores but then stopped growing (Figure 3B) after several rounds of apparently defective mitosis, with nuclear DNA of the ΔnupA cells being missegregated and often clustered (Figure 3C). These phenotypes are very similar to the Δnup2 mutant (Osmani et al., 2006a), suggesting that Nup2 and NupA could have inter­dependent functions.

FIGURE 3:

NupA is essential and its deletion, similar to Nup2, does not affect initial short-term growth but causes mitotic DNA segregation defects. (A) a, Conidia from ∆nupA::pyrG transformant colonies were streaked on media supplemented with or without uridine and uracil (UU). Because nupA is essential, heterokaryons were generated after transformation that contained parental nuclei and nuclei carrying ∆nupA::pyrG. The heterokaryotic colonies produce uninucleated spores able to form colonies on +UU (parental spores can form colonies) but not on –UU (neither parental nor ∆nupA::pyrG spores can form colonies). b, Diagnostic PCR of heterokaryon SM51 showing two bands corresponding to the wild-type nupA and the ∆nupA::pyrG allele. (B) Bright-field images of germinated spores from the parental strain or heterokaryon SM51 producing wild-type and ∆nupA spores grown for 1 or 3 d at 22°C on the indicated media. Bar, ∼20 μm. (C) DAPI-stained wild-type germlings with typical one, two, four, and eight nuclei compared with ∆nupA germlings from heterokaryon SM51 of approximately the same length that had missegregated clustered DNA (arrows). Bar, ∼5 μm.

Nup2 and NupA transfer from NPCs to the chromatin region in early prophase and to the chromatin periphery during anaphase

As mentioned above, during A. nidulans mitosis, NPCs partially disassemble, and nuclear proteins, such as NLS-DsRed, disperse from nuclei and locate throughout the cell (De Souza et al., 2004; Osmani et al., 2006a; De Souza and Osmani, 2009). Dispersal of NLS-DsRed therefore acts as a marker for entry into mitosis and its nuclear reimport a marker for mitotic exit into G1. In a manner identical to Nup2, NupA-GFP was found to translocate from NPCs to concentrate to the mitotic chromatin region until cells exit mitosis (Figure 4A). Accordingly, NupA-GFP and the chromatin marker histone H1–monomeric red fluorescent protein (mRFP) localized together during mitosis (Figure 4B), and, as expected from their copurification, NupA-GFP and Nup2-ChRFP localize together at interphase NPCs and the mitotic chromatin region (Figure 4C).

FIGURE 4:

NupA localizes like Nup2 to NPCs during interphase and to the chromatin region during mitosis. (A) Images and kymograph from time-lapse microscopy (strain SM86) monitoring NupA-GFP and NLS-DsRed. NupA-GFP locates to the chromatin region during mitosis (M) as NLS-DsRed disperses from the mitotic nucleus when NPCs are partially disassembled. During exit from mitosis, NupA relocates back to NPCs as nuclear transport is reestablished. (B) NupA-GFP and Histone H1-mRFP in strain SM122 during G2, mitosis (M), and G1. Also shown are a merged image and pixel profiles for both signals through the dividing nucleus. (C) NupA-GFP and Nup2-ChRFP in strain SM88 during G2, mitosis (M), and G1. Also shown are a merged image and pixel profiles for both signals through the dividing nucleus. (D) Cartoon of the location of Nup2 and NupA (red) during mitotic phases. (E) Sequential z-stack confocal images of Nup2-GFP and NupA-GFP in combination with Histone H1-mRFP in metaphase and anaphase/telophase as indicated (strains SO572 and SM124, respectively). Nuclei that had just exited metaphase and had their DNA still condensed were classified as anaphase–telophase. Nup2 and NupA move from throughout the chromatin region to around segregating chromatin after anaphase. Bar, ∼5 μm.

By observing individual confocal slices of either Nup2-GFP or NupA-GFP and histone H1-mRFP, we found that both Nup2 and NupA locate throughout the chromatin region during metaphase but transition to the periphery of chromatin during anaphase–telophase (Figure 4, D and E, and Supplemental Figure S2, A and B). Nup2 and NupA therefore undergo a biphasic transition in relationship to chromatin during mitosis. They first move from NPCs to locate throughout the mitotic chromatin region during prophase and metaphase, and then undergo a second transition during anaphase–telophase to move to a position around the periphery of the condensed separating genomes.

NupA is required for Nup2 location at NPCs and the mitotic chromatin region

Because Nup2 and NupA display identical dynamic subcellular localizations and cause similar mitotic defects when deleted (see below), it is possible that one could be responsible for the function of the other by targeting it correctly. To investigate this possibility, we deleted Nup2 in cells carrying NupA-GFP. In the absence of Nup2, NupA-GFP was able to locate to NPCs during interphase and to the chromatin region during mitosis (Figure 5B). In contrast, when nupA was deleted, Nup2-GFP was unable to locate to NPCs or to the mitotic chromatin region (Figure 5A). NupA is therefore required for Nup2 to locate to both interphase NPCs and the mitotic chromatin region.

FIGURE 5:

NupA is required for Nup2 location to interphase NPCs and to the mitotic chromatin region. (A, B) Time-lapse images of Nup2-GFP or NupA-GFP and NLS-DsRed and a merge in ∆nupA (from heterokaryon SM79) and ∆nup2 (from heterokaryon SM92) germlings, respectively, (outlined in white) in G2, M (mitosis), and G1. Also outlined are parental spores within the same field of view that cannot germinate in the selective media used. (A) In the absence of NupA, Nup2 can locate to neither interphase NPCs nor the mitotic chromatin region. (B) Conversely, in the absence of Nup2, NupA can still locate to interphase NPCs and the mitotic chromatin region. Bar, ∼5 μm.

Deletion of Nup2 or NupA causes mitotic defects that engage the spindle assembly checkpoint and also cause defects in karyokinesis

To investigate mitotic progression without Nup2 or NupA, we used GFP-tagged tubulin to monitor the dynamics of spindle formation (Ovechkina et al., 2003). We found that the time to complete mitosis was longer and overall more variable in nup2- and nupA-deleted cells, taking 10–11 min on average to complete versus 5.5 min for wild-type cells (Figure 6, A and B, and Supplemental Figure S3). To see whether the mitotic defects caused by the absence of Nup2 and NupA are monitored by the SAC, we asked whether the mitotic delays in these mutants are dependent on a functional SAC response (Li and Murray, 1991; Musacchio and Salmon, 2007). We found that in the absence of Mad2, the mitotic delays caused by nup2 or nupA deletion were abolished, and the average time of mitosis was similar to that of wild-type controls, with the variability in the duration of mitosis no longer apparent (Figure 6, A and B, and Supplemental Figure S3). We also followed Mad1-GFP in nup2- and nupA-null mutants to monitor SAC activation, which causes the transition of Mad1-GFP from NPCs to mitotic kinetochores. As occurs in the wild type, in the mutant Mad1-GFP translocates to kinetochores during the first mitosis during spore germination. Mad1-GFP then stays associated with the kinetochore region, as defined by Ndc80-ChRFP, during the extended mitotic period (Figure 6C). Absence of Nup2 or NupA therefore causes defects in mitosis, which are monitored by the SAC.

FIGURE 6:

nup2- and nupA-deleted cells undergo delayed mitosis due to activation of the spindle assembly checkpoint. (A) Time course images of GFP-tubulin to follow mitotic spindle formation in germlings with four nuclei during G2-M-G1. ∆nup2 (from heterokaryon SM96) and ∆nupA (from heterokaryon SM98) cells display a prolonged mitosis only when the SAC is functional. Bar, ∼5 μm. (B) Bar graph representing average time spent in mitosis with mitotic spindles present for each indicated genotype. The number on each bar represents the average time in seconds. ∆nup2 and ∆nupA cells spend significantly longer time in mitosis than wild type (strain HA375) (with p < 0.01, highly significant when comparing the means). Δmad2 + ∆nup2 (from heterokaryon SM129) and Δmad2 + ∆nupA (from heterokaryon SM131) mutants show no statistical difference compared with the wild type or Δmad2 mutant (strain CDS864). (C) Time course images of wild-type and nupA- and nup2-null cells during first mitosis following Mad1-GFP in combination with the kinetochore marker Ndc80-ChRFP in strain CDS831 and spores from heterokaryon SM154 and SM152, respectively. In the first mitosis of ∆nupA and ∆nup2 nuclei, Mad1-GFP remains at kinetochores during the delayed mitosis. Bar, ∼5 μm.

In cells lacking Nup2 or NupA, although the SAC is engaged causing mitotic delay, the cells do not all subsequently complete mitosis normally, even though the SAC is apparently fulfilled to allow progression into anaphase. This was revealed by monitoring formation of daughter nuclei in ΔnupA cells using the inner nuclear membrane (INM) marker AN0162-GFP in combination with NLS-DsRed. AN0162 is an INM marker that contains a C-terminal transmembrane domain and is similar to the Schizosaccharomyces pombe INM protein Bqt4 (unpublished data and annotated at AspGD as AN0162). During karyokinesis, the NE undergoes double restrictions and abscissions to generate daughter nuclei. This process also separates the nucleolus from nuclei (Ukil et al., 2009). Twenty percent of nupA-null nuclei displayed defects during mitotic exit, and although two apparently normal G1 nuclei were initially formed, they remain associated via the NE, and karyokinesis failed, resulting in the formation of a single diploid nucleus (Figure 7, B and E). During failed nucleokinesis, AN0162 did not locate correctly around the nucleolus (compare Figure 7, A and B, arrow). Furthermore, after several rounds of mitosis, some mutant polyploid nuclei were seen to form multiple bipolar spindles during mitosis (Figure 7D), indicating that these nuclei were the products of more than one round of defective karyokinesis. Lack of Nup2 or NupA therefore compromises aspects of mitosis that are monitored by the SAC and also perturbs NE dynamics during mitotic exit, which can lead to failed karyokinesis.

FIGURE 7:

nup2- and nupA-null mutants display defects in mitotic exit. (A) Images captured during live-cell imaging of wild-type (SM116) and (B) nupA-null (ΔnupA, from heterokaryon SM127) strains during mitosis, monitoring the INM marker GFP-AN0162 and NLS-DsRed. Without NupA function, the nucleus fails to complete karyokinesis and forms one polyploid nucleus after mitosis. GFP-AN0162 does not locate to the periphery of the nucleolus after anaphase as obviously as the wild type (compare structures at arrows). Pseudocoloring represents false coloring using the thermal color scale of ImageJ. The Fire Color lookup table was applied to grayscale images to produce false-color images (pixels with a value of 0 are white, and pixels with a value of 255 are black). Bar, ∼2.5 μm. (C) Mitosis of a wild type (strain HA375) following spindle and chromatin dynamics of four mitotic nuclei. Bar, ∼5 μm. (D) A mitotic depolyploidization event in a nupA-null mutant (from heterokaryon SM98) showing multiple bipolar spindles forming in a polyploid nucleus that remains in an extended metaphase-like state before exiting mitosis. Bar, ∼5 μm. (E) Quantitation of the percentage of larger polyploid nuclei generated in wild-type (strain HA375) compared with nup2- and nupA-null cells (heterokaryons SM96 and SM98, respectively).

Nup2 and NupA are not required for assembly of transport- competent NPCs but are required for Mad1 location to nuclei in G1

Nup2 and its mammalian orthologue Nup50 facilitate importin-α/β–mediated nuclear transport (Solsbacher et al., 2000; Gilchrist et al., 2002; Lindsay et al., 2002; Matsuura and Stewart, 2005; Stewart, 2007; Makise et al., 2012). Consistent with A. nidulans Nup2 playing a similar role, it purified with importins α and β (Figure 1C). To investigate their roles in nuclear transport, we monitored NLS-DsRed to detect whether Δnup2 and ΔnupA mutant nuclei are defective in nuclear protein import. NLS-DsRed was transported into Δnup2 and ΔnupA nuclei during interphase, with little to no cytoplasmic signal apparent (Figure 8, A and B, 0′). This indicates that nuclear transport is active in the absence of Nup2 or NupA. To determine whether the rate of NLS-DsRed nuclear transport is modified without Nup2 and NupA, we followed the rate of dispersal of NLS-DsRed during mitotic entry and then its rate of import during mitotic exit. As cells entered mitosis, NLS-DsRed dispersed from nuclei and was rapidly transported back into daughter G1 nuclei during mitotic exit in wild-type as well as Δnup2 and ΔnupA nuclei (Figure 8B). Note that NLS-DsRed remained mitotically dispersed longer in the mutants than the wild type because they have prolonged SAC-mediated delayed mitosis (Figure 6, A and B, and Supplemental Figure S3). We quantified the fluorescence pixel intensity of the nuclear NLS-DsRed signal at 10-s intervals in comparison to wild type (Figure 8C). Dispersal and import rates of NLS-DsRed in the mutants were comparable to wild type, with the kinetics of both being essentially superimposable between the mutants and wild type. This suggests there are no overall defects in nuclear protein import in the absence of Nup2 or NupA. In addition, because the disassembly and reassembly of NPCs drive release and reimport of NLS-DsRed, respectively, the results further indicate that the rates of mitotic NPC disassembly and reassembly are not affected by the absence of Nup2 or NupA.

FIGURE 8:

Nup2 and NupA are dispensable for nuclear NLS-DsRed nuclear import. (A) NLS-dsRed is actively imported to the nucleoplasm in Δnup2 and ΔnupA nuclei. Bright-field and confocal images of ΔnupA or Δnup2 germinated spores from heterokaryons SM117 and SM118, respectively, and wild-type spores grown on selective media. Bar, ∼5 μm. (B) Time course images of NLS-DsRed nuclear dispersal and import during mitosis of wild-type (strain SM112), ∆nup2 (from heterokaryon SM118), and ∆nupA (from heterokaryon SM117) cells. Bar, ∼5 μm. (C) Rate of mitotic entry dispersal and mitotic exit import of NLS-DsRed in wild-type, ∆nup2, and ∆nupA cells, showing average normalized nuclear pixel intensity vs. time (N = 10; error bars ± SD). The pixel intensity of the red signal was quantified in a defined area within the nucleus every 10 s and the highest pixel intensity normalized to 100%. Time 0 represents the time at which 50% loss during mitotic entry, or gain during mitotic exit, of the normalized fluorescence intensity of the NLS-DsRed signal occurred.

Although transport of NLS-DsRed does not require Nup2 and NupA, we observed defects in the location of Mad1-GFP to nuclei after mitosis. Mad1-GFP displays distinct sequential cell cycle–regulated locations; NPCs during interphase, at kinetochores and the Mlp1-associated spindle matrix region during mitosis, transiently dispersed from nuclei during exit from mitosis and then briefly in the nucleoplasm before being reassembled back to NPCs in G1 (De Souza et al., 2009; Figure 9A). In the absence of Nup2 or NupA, live-cell imaging revealed that Mad1-GFP located normally at NPCs before mitosis and also transitioned to kinetochores upon entry into mitosis, as expected (Figures 6C and 9A). Also as in wild-type cells, Mad1-GFP dispersed from nuclei during mitotic exit. However, in the absence of Nup2, Mad1-GFP failed to accumulate back into G1 nuclei normally after completion of mitosis (Figure 9A).

FIGURE 9:

Nup2 and NupA affect the postmitotic nuclear import of Mad1. (A) Time-lapse microscopy of Mad1-GFP and the kinetochore marker Ndc80-ChRFP in wild-type (strain CDS578) and Δnup2 cells (heterokaryon SM152), showing that in the absence of Nup2, the mitotic location of Mad1 to kinetochores remains unaffected, but the nuclear import of Mad1 in G1 is impaired. Bar, ∼5 μm. (B) Quantification of the mislocalization of Mad1 in asynchronously growing wild-type, Δnup2, and ΔnupA cells.

We analyzed the locations of Mad1-GFP in populations of wild-type, Δnup2, and ΔnupA cells (Figure 9B). Because Mad1 locates to NPCs throughout most of interphase, the majority of wild-type cells contain Mad1-GFP at NPCs (Figure 9B). A smaller percentage of cells have Mad1-GFP located within nuclei at kinetochores, reflective of cells in mitosis (∼5% of the cell cycle). An even smaller percentage of cells had Mad1-GFP dispersed from nuclei (Figure 9B), when it is transiently released from nuclei during mitotic exit before nuclear transport is reestablished during G1. In the absence of Nup2 or NupA, the percentage of cells with Mad1-GFP at the kinetochores remained largely unchanged (Figure 9B), indicating that Mad1's mitotic nuclear location, when nuclear pores are opened, does not require their functions. Of note, however, the low percentage of cells with Mad1-GFP dispersed was dramatically increased to ∼50%, with a proportional decrease of cells with Mad1-GFP at NPCs (Figure 9B). Of importance, when the NPC tether for Mad1 is deleted (ΔMlp1), Mad1-GFP is effectively imported into interphase nuclei but cannot locate at NPCs (De Souza and Osmani, 2009). The fact that there is a dramatic increase in the number of cells in which Mad1-GFP remains cytoplasmic therefore indicates that without Nup2 or NupA, nuclear transport of Mad1-GFP is defective after mitosis rather than its ability to locate at NPCs. Because NLS-DsRed nuclear import is unaffected without NupA or Nup2, the nuclear import defect of Mad1 implies that they are required specifically for effective nuclear import of Mad1.

DISCUSSION

NupA is essential and was identified as a copurifying partner of Nup2 but contains no sequence characteristics of known NPC proteins and is poorly conserved even between Aspergillus species. NupA location was found to be regulated in an identical manner to Nup2 and to cause the same mitotic defects when deleted. Our analysis shows that NupA is required for the targeting of Nup2 to its interphase and mitotic locations and reveals Nup2 and NupA play interconnected roles during mitosis as well as during mitotic exit, being involved in transporting Mad1 to nuclei during G1 and also the control of karyokinesis.

Nup2 and NupA are required for aspects of mitosis as well as karyokinesis

During interphase, NupA and Nup2 locate around the nuclear periphery with NPCs, but at mitotic entry, they locate to the chromatin region within nuclei. Transition to the chromatin region depends upon mitotic activation of Cdk1, and both Nups are phosphorylated when located to the chromatin region. This adds these Nups to the growing number found to be phosphorylated in a mitotic-specific manner (Macaulay et al., 1995; Favreau et al., 1996; De Souza et al., 2004; Mansfeld et al., 2006; Glavy et al., 2007; Blethrow et al., 2008; Laurell et al., 2011). The NIMA kinase, which is required and sufficient to promote NPC disassembly, is also able to promote the disassociation of Nup2 from NPCs and its transition to the chromatin region. Because Nup2 transition to the chromatin region at mitosis is conserved, as are the mitotic roles of CDK1 and NIMA during mitotic NPC disassembly (Laurell et al., 2011), it is possible that related kinases in other species will regulate mitotic association of Nup2 to the mitotic chromatin region as occurs in A. nidulans.

The location of Nup2 and NupA first throughout the chromatin region during prophase and then around chromatin during anaphase and telophase suggests they could play roles during both early and late stages of mitosis. We show that their early mitotic roles, when they are located throughout the chromatin region, are monitored by the SAC, which is activated and engaged longer in their absence. In addition, without Nup2 or NupA function, and after apparent satisfaction of the SAC, a high percentage of mitoses failed karyokinesis, which generated polyploid nuclei. This phenotype suggests that their telophase locations around segregating genomes could also be of functional relevance, perhaps to help drive the double NE abscissions during karyokinesis. Arguing against a specific role for Nup2 and NupA in the mechanics of karyokinesis, many different mitotic defects have been shown to generate polyploid nuclei in A. nidulans (Enos and Morris, 1990; O'Connell et al., 1993; De Souza et al., 1999, 2011; De Souza and Osmani, 2011; Davies et al., 2004; Pitt et al., 2004; Prigozhina et al., 2004; Govindaraghavan et al., 2014). Because karyokinesis is aborted in response to many different mitotic defects in A. nidulans, we suggest that a regulatory mechanism is triggered, preventing completion of karyokinesis in response to postanaphase defects. We tentatively call this the karyokinesis checkpoint, although we have not formally proven that checkpoint-type regulation (Hartwell and Weinert, 1989) is in play. The karyokinesis checkpoint would prevent generation of aneuploid nuclei and generate polyploid nuclei instead. We observed polyploid nuclei generated in the absence of Nup2 or NupA form multiple bipolar spindles in a process akin to depolyploidization (Puig et al., 2008; Lee et al., 2009). Polyploid nuclei generated via the karyokinesis checkpoint can therefore contribute to cell growth and could also potentially undergo reductional mitotic division to regenerate correct ploidy, as previously seen after microtubule poison washout in A. nidulans (De Souza et al., 2011).

The karyokinesis checkpoint has similarities with the Aurora B–mediated abscission checkpoint (NoCut pathway in budding yeast), which prevents aneuploidy by causing delays in cytokinesis if chromatin becomes trapped between daughter cells (Norden et al., 2006; Chen and Doxsey, 2009; Steigemann and Gerlich, 2009; Steigemann et al., 2009). However, in the early stages of A. nidulans growth, cytokinesis is not linked to mitosis (Clutterbuck, 1970; Harris et al., 1994), and so cytokinesis cannot be a target of checkpoints involved in maintaining ploidy when mitotic DNA is slow to segregate. Instead, we propose that the final stage of NE abscission during nucleokinesis is the target of the regulatory mechanisms preventing generation of aneuploid nuclei after DNA segregation defects rather than cytokinesis in A. nidulans. In mammalian cells, delayed cytokinetic abscission can also be triggered when NPCs are not assembled correctly. This has been termed the NPC-abscission checkpoint (Mackay and Ullman, 2011). Of importance, the NPC-abscission checkpoint can be initiated by loss of Nup153 or Nup50 (Nup2 orthologue; Mackay et al., 2010). Another checkpoint, active after the SAC, has recently been reported to act upon nuclear envelope reassembly and chromosome decondensation during mitotic exit (Afonso et al., 2014). It will therefore be of interest to further investigate similarities and differences between the karyokinesis checkpoint of A. nidulans and those active in other cell types that ensure normal ploidy after defective mitosis.

Nup2 and NupA are not required for assembly of transport- competent NPCs but are required for Mad1 nuclear import

Nup2 and NupA are not required for nuclear import of NLS-DsRed, indicating that NPCs can functionally assemble without them, but they are involved in the interphase import of Mad1. The delayed nuclear import of Mad1 in the absence of Nup2 or NupA suggests they may facilitate Mad1 import. Nuclear protein import dependent upon specific Nups was previously defined for the Gag protein of the Tf1 retrotransposon of S. pombe, which requires Nup124 for its import. Nup124 is the S. pombe orthologue of Nup153 (Balasundaram et al., 1999). Of importance, sequences adjacent to the Tf1 Gag NLS are sufficient to cause the classical NLS of SV40 to become dependent upon Nup124 for its nuclear import (Dang and Levin, 2000; Kim et al., 2005; Varadarajan et al., 2005; Sistla et al., 2007). If similar functional sequences were associated with the NLSs of Mad1, this could explain the specific requirement for Nup2 and NupA for its nuclear import via the classical importin-α/β pathway.

Could NupA be a very poorly conserved functional equivalent to Nup153?

Although NupA does not contain sequence similarities to other Nups, indicating that it is a unique Nup, the results of our NupA studies reveal distinct parallels to findings from several systems for Nup153 and its orthologues, which include Nup1 and Nup60 in S. cerevisiae and Nup124 in S. pombe. They interact and purify with the same proteins, including Nup2 and importins α and β (Guan et al., 2000; Smitherman et al., 2000; Allen et al., 2001, 2002; Denning et al., 2001; Dilworth et al., 2001, 2005; Sistla et al., 2007), are required for formation of the nuclear basket complex (Denning et al., 2001; Dilworth et al., 2001; Hase and Cordes, 2003), play roles in aspects of mitosis before and after the SAC response (Mackay et al., 2009, 2010; Mackay and Ullman, 2011; Lussi et al., 2010), and are poorly conserved even between closely related species (Mans et al., 2004; Bapteste et al., 2005; Sistla et al., 2007). Given that the Nup153 equivalent in S. pombe is required for nuclear import and activity of the retrotransposon Tf1 (Varadarajan et al., 2005), one driving force promoting the rapid variance in Nup153-related proteins, potentially including NupA, could be the result of an ongoing arms race between viruses/retrotransposons and their host cells that involves Nup153-related proteins.

MATERIALS AND METHODS

General techniques

General techniques for classical genetics, media preparation, and culture of A. nidulans were as previously described (Pontecorvo et al., 1953; Osmani et al., 2006a). A. nidulans transformation was conducted as described (Osmani et al., 1987), but 10 mg/ml Vinoflow FCE (Novozymes A/S, Bagsvaerd, Denmark) was used instead of Novozyme 234 in the lytic mix. 4′,6-Diamidino-2-phenylindole (DAPI) staining of DNA was performed as described (Oakley and Osmani, 1993). Genes were endogenously tagged at their 3′ end using constructs generated by fusion PCR as described (Yang et al., 2004; Szewczyk et al., 2006) and transformed into an nkuA^ku70Δ strain (SO451) (Nayak et al., 2006). DNA was prepared from these strains as described (Osmani et al., 2006a), and diagnostic PCR using primers flanking the targeting construct was performed to verify homologous integration. Haploid strains used in this study are listed in Table 1. Protein was extracted from A. nidulans as described (Osmani et al., 2006a; Liu et al., 2010), and Western blotting analysis was performed (Osmani et al., 2006a) to confirm correct expression and the specificity of the Nup2 antibody generated. The heterokaryon rescue technique was performed as described previously (Osmani et al., 2006b), and heterokaryons used in this study are listed in Table 2.

TABLE 1:

A. nidulans haploid strains used in this study.

Strain	Genotype	Reference
R153	wA3; pyroA4	De Souza et al. (2004)
SO451	wA3; chaA1; fwA1; pyroA4; pyrG89; argB2; nkuAku70Δ::argB; SE15; nirA14	Osmani et al. (2006a)
SO926	An-nup2-S-Tag::pyrGAf; pyrG89; nimT23; pyroA4; yA2; nirA14?	This study
CDS831	ndc80-ChRFP::pyroAAf; mad1-GFP::pyroAf; ΔnkuAku70::argB; pyroA4; pyrG89; argB2; chaA1; wA3; nirA14?; fwA1?	De Souza et al. (2009)
SO572	An-nup2-GFP::pyrG; pyrG89; H1-mRFP; pabaA1; wA3	This study
CDS864	∆md2A::pyroAAf; nup49-ChRFP::pyroAAf; GFP-TubA; pyrG89; argB2; ∆nKuAKu70::argB; fwA1; chaA1; pyroA4?; nirA14?	This study
HA375	nimT23; tubA-GFP; An-H1-ChRFP::pyroAAf; pyrG89; pyroA4; argB2?; ∆nkuAku70::argB	This study
HA377	nimT23; An-Nup2-GFP::pyroAAf; ∆yA::NLS-Dsred-StuA; pyrG89; pyroA4; argB2?; ∆nkuAku70::argB	This study
SO1030	alcA-nimA-∆C S-tag::pyr4; pyrG89; An-nup2-GFP::pyroAAf; pyroA4; ∆nKuA::argB alcA; argB2?; ∆yA::nlsdsRed; nimT23; yA2.	This study
SM86	nupA-GFP::pyroA; nimT23; pyrG89; pyroA4; argB2; ∆nkuAku70::argB; ∆yA::NLS-Dsred-StuA	This study
SM88	nupA-GFP::loxp2-pyrGAf; pyrG89; nup2-ChRFP-pyroAAf; pyroA4; wA3; chaA1; fwA1	This study
SM112	pyrG89; nimT23; pyroA4; An-nup49-GFP::riboBAf; ∆nkuAku70::argB; ∆yA::NLS-DsRed-StuA; riboB2?; argB2?	This study
SM116	pyrG89; pyroA4; nirA14; pabaA1; ∆nkuAku70::argB; argB2?; ∆yA::NLS-DsRed-StuA; GFP-An0162	This study
SM122	H1-mRFP::pyrGAf; pyrG89; nupA-GFP::loxp; argB2; nirA14?	This study
SM124	H1-mRFP::pyrGAF; pyrG89; NupA GFP:: loxp; pyroA4; chA1	This study
CDS578	mad1-GFP::pyrGAf; Ndc80-ChRFP::pyrGAf; pyrG89; argB2; chaA1; fwA1; nirA14?	De Souza et al. (2009)

In some strains, we have not confirmed some nutritional or color markers that could be covered by, or be recessive to, other markers in the strain (these are designated by a question mark).

^Af Genes from A. fumigatus used for complementation of the corresponding A. nidulans nutritional mutations.

TABLE 2:

A. nidulans heterokaryons used in this study.

Strain	Genotype	Parent
SM51	wA3; fwA1; chaA1?; ∆nupA::pyrGAf; pyrG89; ΔnkuAku70::argB; argB2; pyroA4; nirA14; sE15	SO451
SM79	∆nupA::pyrGAF; An-nup2-GFP::pyroAAf; nimT23; ∆yA::NLS-Dsred-StuA; pyrG89; pyroA4; argB2?; ΔnkuAku70::argB	HA377
SM92	∆nup2::pyrGAf; nupA-GFP::pyroA; nimT23; pyrG89; pyroA4; ΔnkuAku70::argB; ∆yA::NLS-DsRed-StuA	SM86
SM96	∆nup2::pyrGAf; nimT23; GFP-tubA; An-H1-ChRFP::pyroAAf; pyrG89; pyroA4; argB2?; nkuAku70Δ::argB	HA375
SM98	∆nupA::pyrGAf; nimT23; GFP-tubA; An-H1-ChRFP::pyroAAf; pyrG89; pyroA4; argB2?; ΔnkuAku70::argB	HA375
SM117	∆nupA::pyrGAf; pyrG89; nimT23; pyroA4; An-Nup49-GFP::riboBAf; ∆nkuAku70::argB; ∆yA::NLS-DsRed-StuA; riboB2?; argB2?	SM112
SM118	∆nup2::pyrGAf; pyrG89; nimT23; pyroA4; An-nup49-GFP::riboBAf; ∆nkuAku70::argB; ∆yA::NLS-DsRed-StuA; riboB2?; argB2?	SM112
SM127	∆nupA::pyrGAf; pyrG89; pyroA4; nirA14; pabaA1; ∆nkuAku70::argB; argB2?; ∆yA::NLS-DsRed-StuA; GFP-An0162	SM116
SM129	∆nup2::pyrGAf; ∆md2A::pyroAAf; nup49-ChRFP::pyroAAf; GFP-tubA; pyrG89; argB2; ∆nKuAKu70::argB; fwA1; chaA1; pyroA4?; nirA14?	CDS864
SM131	∆nupA::pyrGAf; ∆md2A::pyroAAf; nup49-ChRFP::pyroAAf; GFP-tubA; pyrG89; argB2; ∆nKuAKu70::argB; fwA1; chaA1; pyroA4?; nirA14?	CDS864
SM152	∆nup2::pyrGAf; ndc80-ChRFP::pyroAAf; mad1-GFP::pyroAf; ΔnkuA ku70::argB; pyroA4; pyrG89; argB2; chaA1; wA3; nirA14?; fwA1?	CDS831
SM154	∆nupA::pyrGAf; ndc80-ChRFP::pyroAAf; mad1-GFP::pyroAf; ΔnkuA ku70::argB; pyroA4; pyrG89; argB2; chaA1; wA3; nirA14?; fwA1?	CDS831

The parent strain represents the strain in which the deletion or replacement of either nup2 or nupA was performed. In some strains, we have not confirmed all nutritional or color markers that could be covered by, or be recessive to, other markers in the strain (these are designated by a question mark).

^Af Genes from A. fumigatus used for complementation of the corresponding A. nidulans nutritional mutations.

Antibody generation and immunofluorescence

A rabbit polyclonal antibody was generated against the 874- to 1074-aa region of A. nidulans Nup2 cloned into Pet14b (Novagen, Merck KGaA, Darmstadt, Germany) expressed and purified from Escherichia coli strain BL21 (DE3). Purified Nup2 was used to immunize male rabbits, and affinity-purified antibodies were produced by Bethyl Laboratories (Montgomery, TX). Immunofluorescence was performed (Oakley and Osmani, 1993) with some modifications. Cells were grown on yeast extract media and fixed in 1× PHEM buffer (45 mM 1,4-piperazinediethanesulfonic acid, 45 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid, 10 mM ethylene glycol tetraacetic acid, 5 mM MgCl, pH 6.9) containing 6% paraformaldehyde (EM grade; Electron Microscopy Sciences, Hatfield, PA). Cell wall removal solution contained 10 mg/ml Vinoflow FCE (Novozymes A/S). Cells were digested for 50 min at 30°C. The antibody was used at a dilution of 1:5000.

Affinity purification and mass spectrometry

The S-tag affinity purification of endogenously S-tagged proteins was as described in Liu et al. (2009). G2 and M samples were generated by arresting cells in G2 using the nimT23 temperature-sensitive mutation, which prevents activation of mitotic Cdk1, and then releasing to permissive temperature after addition of benomyl to cause synchronous mitotic SAC arrest (Osmani et al., 1991; O'Connell et al., 1992; Lu et al., 1993). Samples were run on 10% SDS–PAGE and fixed, and protein was visualized by Bio-Safe Coomassie (Bio-Rad, Hercules, CA). Mass spectroscopy was completed at the Ohio State University Chemical Instrument Center. Phosphatase treatment was performed as described previously (De Souza et al., 2004).

Sequencing and identification of NupA

cDNA and genomic DNA of NupA were amplified from A. nidulans cDNA libraries or genomic DNA. Pfu-turbo DNA polymerase (Stratagene, Santa Clara, CA) was used for high-fidelity amplification. The resulting PCR products were cloned in pCR 2.1-TOPO vector (Invitrogen, Carlsbad, CA) and sequenced. The 5′ and 3′ rapid amplification of cDNA ends was performed following the Marathon-Ready cDNA user manual (Clontech, Mountain View, CA) to confirm the N-terminal and C-terminal ends of nupA. To define the 5′ intron in A. fumigatus nupA, we PCR amplified the region that spans the intron using an A. fumigatus cDNA library (a kind gift from Gregory S. May, University of Texas MD Anderson Cancer Center, Houston, TX), which amplified a single band of a size predicted after intron splicing that was confirmed by sequence analysis.

NIMA induction experiments

Conidia from strains HA377 (wild type for NIMA) and SO1030 (alcA-NIMA-ΔC-Stag) were inoculated overnight in 1 l of minimal medium containing glucose at 26°C. Cultures were then centrifuged, washed twice with minimal media without carbon, and shifted into minimal medium containing 0.5% ethanol to induce NIMA expression. Mycelia samples were taken every 15 min. Extracted whole-cell protein samples from mycelia were run on SDS–PAGE. Western blot analysis was performed with NIMA antibody (F68) at a dilution of 1:2000 and the Nup2 antibody generated at a dilution of 1:5000. To follow Nup2-GFP by live-cell imaging, strains HA377 and SO1030 were grown overnight at room temperature in minimal medium containing glucose. The next day, the cells were either kept on minimal medium glucose for repressing conditions or washed twice with minimum medium without carbon and supplied with minimum medium containing 0.5% ethanol to induce NIMA expression. Hydroxyurea (20 mM) was used to arrest cells at S phase. After 2 h, Nup2-GFP and NLS-DsRed were followed in these cells as described in Imaging and analysis. Graphs were plotted depicting the cumulative number of cells that had Nup2 on DNA with time.

Imaging and analysis

Fixed samples were examined using an E800 microscope (Nikon, Melville, NY) with DAPI, fluorescein isothiocyanate , or Texas Red filters (Omega Optical, Brattleboro, VT). Image capture was performed using an UltraPix digital camera (Life Science Resources, Great Shelford, United Kingdom) and analyzed by Ultraview image capture software (PerkinElmer, Waltham, MA). For live-cell imaging, spores were germinated in 35-mm glass-bottom Petri dishes (MatTek Cultureware, Ashland, MA) in 3 ml of minimal medium containing 55 mM glucose as the carbon source and 10 mM urea as the nitrogen source at room temperature (21–23°C) for 24 h or until the 4- to 8-nucleus stage was reached. Spores from heterokaryons were germinated at room temperature (21–23°C) in YG medium (yeast extract glucose: 56 mM glucose, 5 g/l yeast extract, 10 mM magnesium sulfate, 1 μg/ml p-aminobenzoic acid, 0.5 μg/ml pyrodoxine HCl, 2.5 μg/ml riboflavin HCl, 2 μg/ml nicotinic acid, 20 μg/ml choline, 20 ng/ml d-biotin, and 1 ml/l trace elements) with or without uridine or uracil complementation for 8 h (to image first mitosis) or longer to image the second or third mitosis and terminal phenotypes. Visualization of GFP/ChRFP fusion proteins was performed using an inverted microscope (Nikon) configured with an Ultraview spinning disk confocal system and controlled by Ultraview software (PerkinElmer) utilizing Nikon Plan Apo 60×/numerical aperture 1.40 oil objectives. Image analysis, kymograph generation, and pixel intensity profiles were carried out using ImageJ (National Institutes of Health, Bethesda, MD). All images are represented as maximum-intensity projections as described (Ukil et al., 2009).

Quantification of nuclear import

Cells were grown in YG with (for wild type) or without UU (for ∆nup2 and ∆nupA) at room temperature (21–23°C) overnight and NLS-DsRed images captured every 10 s. Maximum pixel intensity (MPI) of z-sections spaced at 0.8 μm was measured using ImageJ. The normalized fluorescence intensity for each time point captured was measured as

Graphs were plotted using Excel (Microsoft, Redmond, WA) as normalized mean fluorescence intensity measured versus time (8–13 nuclei were measured for each genotype).

Quantification of time in mitosis

To calculate time spent in mitosis, cells were grown on YG with (for wild type, ∆mad2) or without UU (for ∆nup2, ∆nupA, ∆nup2∆mad2, and ∆nupA∆mad2) at room temperature (21–23°C) overnight (12–16 hr), and GFP-tubulin signal was followed throughout mitosis. Images were captured every 30 s using a spinning-disk confocal microscope (details described earlier). Time in mitosis was measured starting from the time point where the spindle is visible until it disassembled. Statistical analysis was conducted using PASW statistical software (SPSS, Chicago, IL). The p values were calculated using the Student's t test, considering p < 0.05 as significantly different with 95% confidence intervals and p <0.01 as highly significant with 99% confidence.

Analysis of Mad1 dynamics without Nup2 and NupA function

For analysis of Mad1 localization in Δnup2 and ΔnupA null mutants, we used the heterokaryon rescue technique as described (Osmani et al., 2006b). Images of Δnup2 cells were captured after 24 h of growth, and the terminal phenotypes were analyzed and quantified. Δnup2::pyrG^Af and ΔnupA::pyrG^Af cells in which Mad1-GFP signals were enriched at the nuclear periphery were classified as NPC location. When the signal was uniformly distributed throughout the cell with no particular enrichment, it was classified as dispersed. For Mad1-GFP localization analysis, cells were classified as kinetochores if Mad1-GFP colocalized with the kinetochore marker Ndc80-ChRFP.

Abbreviations used:

ChRFP

mCherry variant of red fluorescent protein

DAPI

4',6-diamidino-2-phenylindole

ETOH

ethanol

HU

hydroxyurea

Kap

karyopherin

NE

nuclear envelope

NIMA

never In mitosis A

NLS

nuclear localization signal

NPC

nuclear pore complex

Nup

nucleoporin

SAC

spindle assembly checkpoint.