The Nup2 meiotic-autonomous region relieves autoinhibition of Nup60 to promote progression of meiosis and sporulation in Saccharomyces cerevisiae

During meiosis, chromosomes undergo dramatic changes in structural organization, nuclear positioning, and motion. Although the nuclear pore complex has been shown to affect genome organization and function in vegetative cells, its role in meiotic chromosome dynamics has remained largely unexplored. Recent work in the budding yeast Saccharomyces cerevisiae demonstrated that the mobile nucleoporin Nup2 is required for normal progression through meiosis I prophase and sporulation in strains where telomere-led chromosome movement has been compromised. The meiotic autonomous region (MAR), a short fragment of Nup2 responsible for its role in meiosis, was shown to localize to the nuclear envelope via Nup60 and to bind to meiotic chromosomes. To understand the relative contribution these two activities have on MAR function, we first carried out a screen for MAR mutants defective in sporulation and found that all the mutations disrupt interaction with both Nup60 and meiotic chromosomes. Moreover, nup60 mutants phenocopy nup2 mutants, exhibiting similar nuclear division kinetics, sporulation efficiencies, and genetic interactions with mutations that affect the telomere bouquet. Although full-length Nup60 requires Nup2 for function, removal of Nup60’s C-terminus allows Nup60 to bind meiotic chromosomes and promote sporulation without Nup2. In contrast, binding of the MAR to meiotic chromosomes is completely dependent on Nup60. Our findings uncover an inhibitory function for the Nup60 C-terminus and suggest that Nup60 mediates recruitment of meiotic chromosomes to the nuclear envelope, while Nup2 plays a secondary role counteracting Nup60’s autoinhibition.


Introduction
The chromosome events of meiosis, including pairing, synapsis, and crossing over between homologous chromosomes, take place in a crowded nuclear environment. It is not well understood if or how the physical location of sequences in the nucleus contribute to meiotic processes. One organizing feature of chromosomes specific to meiosis is the telomere bouquet, in which the chromosome ends cluster at the nuclear envelope by attaching to the linker of nucleoskeleton and cytoskeleton (LINC) complex (Burke 2018). The indirect attachment of telomeres to cytoskeletal motor proteins results in dramatic chromosome movements that coincide with the process of homolog pairing. Whether other chromosome regions associate with the nuclear envelope is unknown.
In budding yeast, the Ndj1 protein acts as an adapter between the telomeres and the LINC complex. Mutations in NDJ1 disrupt the telomere bouquet, abolish telomere-led chromosome movements, and lead to defects in pairing (Scherthan et al. 2007;Conrad et al. 2008;Wanat et al. 2008). Nevertheless, ndj1 mutants have nearly wild-type sporulation efficiency and spore viability, implying that yeast might have an auxiliary system that compensates for loss of the telomere bouquet.
The nucleoporin Nup2 emerged as a likely component of this potential backup mechanism from studies that uncovered a synthetic interaction between nup2 and ndj1 mutations, suggesting that the LINC complex and the nuclear pore complex (NPC) promote meiosis and sporulation by two functionally redundant pathways (Chu et al. 2017).
Nup2 is part of the nuclear basket, a fibrous structure on the nucleoplasmic face of the NPC that also contains the proteins Nup60, Nup1, Mlp1, and Mlp2 (Raices and D'Angelo 2021). Although first characterized as a component of the transport machinery, Nup2 was later found to play roles in chromatin organization, gene regulation, and DNA damage repair (Ishii et al. 2002;Dilworth et al. 2005;Kim et al. 2017;Brickner et al. 2019). Deletion analysis showed that a short fragment of Nup2 is necessary and sufficient for its meiotic function. This meiotic autonomous region (MAR) of Nup2 binds to both the nuclear envelope and meiotic chromosomes (Chu et al. 2017), suggesting that Nup2 might be helping to organize chromosomes by tethering them to the NPC. There are numerous examples in vegetative cells where Nup2 helps bring promoter regions of activated genes to the NPC (Casolari et al. recruiting meiotic chromosomes to the nuclear envelope and provides a second level of chromosome organization that complements telomere attachment to the nuclear periphery.

Strains and media
All strains in this study except for the Y2H strain are derivatives of SK1 and are listed in Supplemental information Table S1. Plasmids and primers are listed in Tables S2 and S3. Yeast media were prepared as previously described (Lui and Burgess 2009;Chu et al. 2017). Standard techniques were used for yeast manipulation. Gene knockouts and fluorescently tagged constructs were created using tailed PCR-based gene replacement and tagging techniques (Longtine et al. 1998;Sheff and Thorn 2004). Gene disruptions were confirmed by PCR, and new alleles were confirmed by both PCR and sequencing. Unmarked deletion alleles of NUP60 were created by transformation of a nup60::URA3 strain with PCR-generated DNA fragments. Transformants were selected for on plates containing 5fluoroorotic acid (Boeke et al. 1984). All alleles generated by 5-FOA selection were outcrossed twice before further analysis.

MAR mutant screen
The nup2(51-175)-GFP::CaURA3 (MAR-GFP::URA3) fragment from plasmid pSB470 was used as the template for PCR mutagenesis and contains part of the NUP2 5' UTR plus the start codon (-203 to +3), the coding sequence for Nup2 amino acids 51-175, the 2,441 bp GFP::URA3 fragment from pKT209 (Sheff and Thorn, 2004) starting 24 bp upstream of the GFP start codon, and part of the NUP2 3' UTR (+2161 to +2482). A PCR fragment from NUP2 (-203) to the GFP start codon was synthesized using Taq polymerase under standard conditions, which introduced errors at a high frequency due to the low fidelity of the enzyme. A second PCR fragment from 24 bp upstream of the GFP start codon to NUP2 (+2482) was generated using the high-fidelity polymerase Phusion (New England Biolabs). The two fragments were joined by overlap extension PCR (Horton 1989) with the Phusion polymerase and transformed into SBY6259 (MATa ho::hisG leu2::hisG ura3::hisG his4::LEU2 nup2(Δ51-175) ndj1::TRP1). The mutagenized MAR-GFP::CaURA3 fragment integrated at the NUP2 locus, replacing nup2(Δ51-175). After three days, Ura + transformants were patched onto fresh -Ura plates alongside positive and negative controls. The patches were grown for 24 hours, then replica plated onto -Ura plates overlaid with a disc of Whatman 1 paper and onto -His-Ura plates containing a mating lawn of SBY6260 (MATalpha ho::hisG leu2::hisG ura3(ΔSma-PstI) HIS4::LEU2 nup2::KanMx ndj1::KanMx). The patches on the Whatman filter were incubated at 30˚C for 24 hours, then screened for GFP fluorescence using an ImageQuant LAS 4000. Candidates that had lost GFP fluorescence were eliminated from further analysis since they were likely to contain nonsense mutations or mutations that inactivate GFP. The -His-Ura plates were incubated at 30˚C for two days, then the resulting patches of His + Ura + diploids were replica plated onto rich sporulation plates (1% potassium acetate, 0.1% peptone, 0.1% yeast extract, 0.05% glucose, 2% agar) and incubated at 30˚C for 3 days. The screen was performed in this manner rather than by transforming a diploid strain because mitochondrial loss during the transformation step led to a high frequency of false positives in trial screens. Sporulation was detected as dityrosine fluorescence as in (McKee and Kleckner 1997) but using an Alpha Innotech AlphaImager 3400 on the reflective UV setting with a SYBR green filter for detection. Sporulation-defective candidates were streaked for single colonies, patched onto -Ura plates and retested. The MAR coding region was amplified by PCR using Phusion polymerase, and the resulting fragments were commercially sequenced (Quintara Biosciences). For all further analysis, the entire MAR-GFP::URA3 fragment from each of the mutants was amplified by PCR and transformed into SBY1900 to allow construction of strains in the SBY1903 background.

Preparation of diploid colonies for sporulation assays and chromosome spreads
Glycerol stocks of haploid parent strains were patched onto YPG plates, allowed to mate for 14 hours at 30˚C, then streaked onto YPD and grown for 48 hours at 30˚C. Diploid colonies were identified based on colony and cell morphology. For experiments involving nup60 strains where the nibbled colony phenotype (Zhao et al. 2004) precluded morphology-based identification, all strains were prepared by patching haploid parents on YPG as above, then streaking the cells onto SC -His -Ura plates to select for diploids. Wild-type strains from colonies grown on SC -His -Ura plates sporulated as efficiently as those from colonies grown on YPD.

Quantitative sporulation assay
Cells were induced to enter meiosis in liquid culture according to Lui and Burgess 2009. Samples were collected after 24 hours in sporulation medium (SPM) and sonicated for 3 seconds on setting 3 using a Fisher Sonic Dismembrator 550 fitted with a microtip to disperse cell clumps. Samples were examined under white light using a Zeiss 47 30 11 9901 microscope to assess sporulation. Cells with 2 or more spores were counted as sporulated, while cells with 1 or no spores were counted as unsporulated. At least 200 cells total were counted for each sample. For all strains, the assay was performed in triplicate using cultures grown from independent colonies and repeated at least once, such that n ≥ 6. For each strain, the sporulation efficiency is reported as the mean and standard deviation for data pooled from all assays performed on that strain. Complete sporulation data are reported in Supplemental File S1.

Yeast two-hybrid assay
The coding sequence of the MAR was inserted into the Gal4 DNA-binding domain vector pGBKT7 (Clontech); the coding sequences for full length Nup60 and Nup60(188-388) were separately inserted into the Gal4 activation domain vector pACT2-2 (Arora et al. 2004). Bait and prey plasmids were cotransformed into AH109, a strain in which the HIS3 gene is driven by the GAL1 promoter. Trp + Leu + transformants were inoculated into liquid SC -Trp -Leu media and grown for 24 hours. The cultures were diluted to an OD 600 of 1.0, and 5 ul were spotted onto SC -Trp -Leu plates and SC -Trp -Leu -His plates containing 25 mM 3-aminotriazole. The plates were grown for 3 days, then imaged using an ImageQuant LAS 4000. At least three independent transformants were tested for each combination of bait and prey plasmids.

Microscopy and image analysis
All images of whole cells and of meiotic chromosome spreads were obtained with a Zeiss Axioskop epifluorescence microscope equipped with a 100x oil immersion objective and TRITC, FITC, and DAPI filter sets (Chroma). Images were collected with a Hamamatsu C4742-95 camera using Micro-Manager software (Edelstein et al. 2010) and were analyzed using Fiji/ImageJ Version 2.0.0-rc-68/1.52e (Schindelin et al. 2012).

Meiotic chromosome spreads
Cells were induced to enter meiosis using the time course protocol cited described above and were collected after 10 hours in sporulation medium. Spheroplasts were prepared as previously described (Grubb et al. 2015), then pelleted and washed twice with MES/sorbitol (0.1 M MES-NaOH, pH 6.4, 1 mM EDTA, 0.5 mM MgCl 2 , 1 M sorbitol). Chromosome spreads were prepared and stained according to (Rockmill 2009), with the following modifications. The volumes of 1X MES and FIX used were 160 ul and 440 ul, respectively. The cell mixture was evenly distributed onto 3 slides, which were incubated in a warm, humid chamber for 40 minutes before being washed with Photo-Flo. Spreads were preincubated in (PBS + 5% BSA)/fetal bovine serum prior to incubation with primary antibody and in PBS + 5% BSA prior to incubation with secondary antibody. Preincubations were for 30 minutes at room temperature in a humid chamber. Slides were dried for at least 20 minutes at room temperature and stored at -20˚C prior to staining. Antibody-stained slides were mounted in Prolong Glass with NucBlue (Thermofisher P36985) and cured for 24 hours at room temperature before imaging. Spreads were repeated for all strains at least once.

Quantitation of chromosome spread fluorescence
Pachytene nuclei were selected for imaging based on their DNA staining pattern in the DAPI channel. to the ROI. The process was repeated for TRITC channel images when applicable. The total corrected fluorescence (TCF) was calculated as TCF = 0.001 x [integrated density -(area of selected cell × mean fluorescence of background readings)]. At least 10 cells from each strain were imaged and quantified on at least two separate days (biological duplicates). Fluorescence measurements are reported in Files S2-S5. Graphing and statistical analysis was done using Prism 9 GraphPad Software.

Meiotic time course protocol
The meiotic time course was performed as described in Lui and Burgess 2009. Samples were removed from sporulation media for analysis at 2-hour intervals, starting at t=0 hours. The samples were sonicated as described above for the quantitative sporulation assay, fixed by the addition of ethanol to a final concentration of 40% by volume, and stored at -20˚C until analyzed. Samples were stained by adding an equal volume of 1 μg/mL DAPI. Cells with at least two well-defined DAPI-stained foci were counted as multinucleate. At least 200 cells were analyzed for each sample. Duplicate samples were run in parallel on each day and each time course was repeated at least one time on a separate day.
Time course data are reported in File S6.

Point mutations affect MAR function
Previous work had shown that Nup2 and Ndj1 act in functionally redundant pathways to promote meiosis and sporulation. The segment of Nup2 responsible for its function, the MAR, spans amino acids 51-175 ( Figure 1A) and has at least two activities: it localizes to the nuclear envelope via its interaction with Nup60, and it binds to meiotic chromosomes (Chu et al. 2017). To help clarify whether these activities are important for MAR function, we carried out a screen for mutations that decreased sporulation with the aim of identifying alleles that were defective for either subcellular localization or chromosome binding.
A PCR-mutagenized DNA fragment encoding MAR-GFP was integrated into an ndj1 strain at the NUP2 locus. Transformants were mated to a nup2 ndj1 strain, and the resulting diploids were screened for the ability to sporulate. Of 8700 diploids screened, 37 had reduced sporulation compared to a wild-type MAR-GFP strain. The mutant candidates were sequenced: 23 contained a unique point mutation, 12 were duplicates, and 2 contained more than 5 non-silent point mutations each and were excluded from further analysis. The 23 unique mutations affected 15 amino acid positions in the central portion of the MAR ( Figure 1B). Diploids homozygous for the MAR-GFP mutations were constructed, and sporulation efficiency was measured ( Table 1). Most of the mutants sporulated at 10-15% efficiency, or roughly the same as the null allele (GFP expressed from the NUP2 promoter). The exceptions were the mutants with the amino acid substitutions K101E and F120Y, which sporulated at approximately 50% efficiency. For these two mutants, we also measured sporulation in a mar-GFP/nup2Δ heterozygote and found that hemizygosity decreased sporulation to around 15%.

The central region of the MAR is sufficient for its function
Because the mutations all lie in the middle of the MAR, we constructed several deletions to test whether this central fragment was sufficient to provide MAR function. A 60-amino acid stretch from amino acids 72-130 corresponding to a region of homology between Nup2 and mammalian Nup50 was sufficient to support sporulation (Table 1), indicating that amino acids 51-71 and 131-175 are dispensable for MAR function. A shorter fragment confined mostly to the region affected by the point mutations (mar(89-130)) was only partially functional, but because mar(89-175) is fully functional, it seems likely that most of the MAR function is provided by the 42-amino acid stretch from 89-130.
Eleven out of the 15 affected amino acids map to conserved residues found in human and mouse Nup50 ( Figure 1C).

Most of the point mutations disrupt localization to the nuclear envelope.
We next examined the mutant strains by fluorescence microscopy and found that unlike wild-type MAR-GFP, which is localized to the nuclear envelope, most of the mutants were distributed throughout the cell ( Figure 1D). The K101E and F120Y fusions localized to the nuclear envelope, as did the mar(89-130)-GFP fusion. Because the mutants with proper localization sporulated with lower efficiency when hemizygous, we examined hemizygotes for these alleles and saw that the fusion localized to the nuclear envelope in all three cases ( Figure S1). Thus, decreasing the copy number of these three mutants does not result in improper subcellular localization, and the lower sporulation efficiency of the hemizygotes is unlikely to be due to mis-localization.

MAR mutants are defective in binding to Nup60.
Since Nup2 is known to be recruited to the nuclear envelope by Nup60 (Hood et al. 2000;Solsbacher et al. 2000), we hypothesized that the MAR mutants with aberrant localization were defective in interacting with Nup60. To test whether the MAR mutations affect interaction with Nup60, we set up a yeast two-hybrid (Y2H) assay using the MAR as bait and Nup60 as prey. We made a Gal4 activation domain fusion to full-length Nup60 and Gal4 DNA binding domain fusions to the wild-type MAR and to a subset of the MAR mutants. While the wild-type MAR interacts with Nup60, many of the mutants do not ( Figure 2). As expected, the two mutants that localize to the nuclear envelope interact with Nup60 by Y2H. However, some of the mutants that are not recruited to the nuclear envelope also interact with full length Nup60. Because the Y2H assay can detect weak or transient interactions between proteins (Mehla et al. 2017), we suspected that a defect in the interaction of these mutants with Nup60 could be masked if the MAR is able to interact with more than one part of Nup60.
Pull down assays with purified fragments of Nup2 and Nup60 have shown that Nup2 interacts with both an N-terminal fragment of Nup60 from amino acids 1-188 and an internal fragment of Nup60 from amino acids 189-388 (Denning et al. 2001). Although most of the Nup60 fragments we tested failed to interact with the wild-type MAR in the Y2H assay, we were able to detect an interaction between Nup60(188-388) and the MAR. Furthermore, all the MAR mutations disrupt this interaction, including the mutations that do not affect localization to the nuclear envelope. These results indicate that all the mutations affect interaction with Nup60 to some degree, although the defect is not always severe enough to prevent the MAR from binding to Nup60 at the nuclear envelope.

Binding of the MAR to meiotic chromosomes requires Nup60.
We next tested whether the mutant MAR-GFP fusions bind to meiotic chromosomes by staining chromosome spreads with an antibody against GFP. While wild-type MAR-GFP forms foci on meiotic chromosomes, the mutants were absent from the spreads ( Figure 3A). Quantitation of the fluorescent signal showed that binding of the mutants was comparable to that of GFP ( Figure 3B). Therefore, the mutants appear to be defective in interacting with both Nup60 and meiotic chromosomes, suggesting that the MAR might require Nup60 to bind meiotic chromosomes.
To test whether the MAR can bind to meiotic chromosomes in the absence of Nup60, we prepared chromosome spreads from NUP60 and nup60Δ strains expressing wild-type MAR-GFP and stained the spreads with antibodies against GFP ( Figure 3C). Binding of MAR-GFP in the nup60Δ strain was reduced to the background levels exhibited by GFP alone ( Figure 3D). These results demonstrate that the MAR requires Nup60 to bind to meiotic chromosomes.

Loss of Nup60 blocks sporulation in strains with telomere bouquet defects.
The observation that Nup60 was required for binding of the MAR to meiotic chromosomes led us to speculate that Nup60 might be playing a larger role in promoting sporulation than we initially suspected.
To explore this possibility, we began by asking whether nup60Δ, like nup2Δ, blocks sporulation in strains lacking NDJ1 or CSM4. Both nup60Δ and nup60Δ ndj1Δ strains grow poorly compared to wild-type strains; however, only the double mutant has a significant sporulation defect ( Figure 4A and B, row 1), indicating that the growth defect alone does not result in poor sporulation and that nup60Δ is synthetic with ndj1Δ. Likewise, the nup60Δ csm4Δ strain exhibits a synthetic sporulation defect, albeit to a lesser degree than the nup2Δ csm4Δ double mutant (Table 2). We predicted that nup60 and nup2 should not have a synthetic phenotype if the two proteins are acting in concert to promote sporulation.
However, because the nup60Δ nup2Δ double mutant either grows extremely poorly or is inviable depending on the strain background (Dilworth et al. 2001), epistasis analysis required an allele of NUP60 that affected sporulation but not growth.

A Nup60 N-terminal fragment is sufficient for sporulation function.
We constructed a set of NUP60 truncations with the aim of finding an allele defective for sporulation alone. The N-terminus of Nup60 contains two helical regions that contact the nuclear envelope and the NPC core, while the C-terminus contains FxF repeats that bind nuclear transport receptors (Denning et al. 2001;Mészáros et al. 2015). We hypothesized that deleting the central portion of Nup60 would create the desired allele by removing a Nup2 interaction domain while leaving regions involved in pore structure and nuclear transport intact. Instead, we found that the strain missing the central domain of Nup60 (Δ189-388) grows as poorly as the nup60Δ mutant, while another expressing only the central portion of Nup60 (189-388) grows as well as wild type (Figure 4B and D). In addition, we tested growth on hydroxyurea (HU) plates, since nup60Δ strains are known to be hypersensitive to DNA-damaging agents (Bermejo et al. 2011). The Nup60(189-388) fragment was necessary and sufficient for wild-type growth in the presence of HU ( Figure S2). We next tested how the NUP60 deletions affect sporulation and found that only full-length Nup60 and the N-terminal fragment from amino acids 1-388 complement the sporulation defect ( Figure   4B). Although the middle fragment from 189-388 fully rescues the growth defect, it complements the sporulation defect only slightly. Collectively, these results demonstrate that the middle region of Nup60 is necessary but not sufficient for sporulation, while the C-terminal region is completely dispensable.

NUP60 and NUP2 are in the same genetic pathway
The ability of the nup60(189-388) allele to complement growth but not sporulation allowed us to test for a synthetic interaction between nup2 and nup60. We found that nup60(189-388) on its own, like nup2, has little effect on sporulation and that the nup2 nup60 double mutant also sporulates with wild-type efficiency ( Table 2). In addition, the nup60(189-388) nup2 ndj1 triple mutant sporulates with approximately the same efficiency as either the nup60(189-388) ndj1 or nup2 ndj1 double mutant.
Thus, nup60(189-388) and nup2Δ do not have a synthetic sporulation phenotype. We also measured the kinetics of nuclear division in a meiotic time course of wild-type, nup2Δ, nup60 , and nup2Δ nup60(189-388) strains. Again, nup60(189-388) phenocopies nup2Δ, causing an approximately onehour delay in the first meiotic division, while the double mutant exhibits a delay similar to that seen in the single mutants ( Figure 5A). In addition, we compared the kinetics of nuclear division in wild-type, ndj1Δ, nup60(189-388) and ndj1Δ nup60(189-388) strains. The ndj1Δ nup60(189-388) double mutant takes significantly longer to undergo the first meiotic division than either single mutant ( Figure 5B).
These results demonstrate that NUP60 acts in the same pathway as NUP2 but in a different pathway from NDJ1. Strains that showed a delay in the first meiotic division also exhibited a delay in spore formation ( Figures 5C, D). Oddly, in the nup2 ndj1 and nup60 ndj1 strains, the percentage of multinucleate cells is significantly higher than the percentage of sporulated cells at the 24-hour time point. When we examined the DAPI stained cells under white light, we noticed that many of the multinucleate cells in these double mutants had not formed visible spores ( Figure 5E). In contrast, when we examined the wild-type strain, all multinucleate entities examined had sporulated.

Full-length Nup60-GFP is required to recruit the MAR to the NE.
To determine which part or parts of Nup60 can recruit the MAR to the nuclear envelope, we tagged the Nup60 fragments with GFP. We first ascertained that the GFP-tagged alleles behave like their untagged counterparts in terms of growth and sporulation ( Figure S4). We next examined strains expressing each of the tagged Nup60 fragments by fluorescence microscopy and found that all the Nup60-GFP fusions localized to the nuclear envelope to some degree ( Figure 4C, middle column).
However, fusions lacking the N-terminal 188 amino acids were also aberrantly present in the nucleoplasm and cytoplasm. These results are consistent with published observations that helical domains in the N terminus of Nup60 are important for contacting the nuclear membrane and the core of the NPC (Mészáros et al. 2015).
When we examined the localization of MAR-mCherry that was co-expressed in each of the Nup60-GFP strains, we found that MAR-mCherry is properly localized at the nuclear envelope only in the strain expressing full-length Nup60-GFP ( Figure 4C, right column). When the N-terminus, middle, or C-terminus of Nup60 is deleted, MAR-mCherry appears largely throughout the cell. Curiously, Nup60(1-388)-GFP supports sporulation despite being unable to recruit the MAR to the nuclear envelope.

Full-length Nup60-GFP is required to recruit the MAR to meiotic chromosomes.
We next examined binding of the Nup60 fragments to meiotic chromosomes spreads. We found that the only Nup60 fragment that supports sporulation (Nup60(1-388)) binds to meiotic chromosomes while the Nup60 fragments that do not support sporulation do not bind to meiotic chromosomes ( Figure 6A and B, middle column). These results show that the N-terminal fragment from amino acids 1-388 is sufficient for meiotic chromosome binding and demonstrate that there is a correlation between Nup60's chromosome-binding activity and its ability to promote sporulation.
In spreads of strains expressing full-length Nup60-GFP, both Nup60-GFP and MAR-mCherry are present on the chromosomes and appear to be co-localized with one another. In contrast, in spreads with Nup60 fragments that do not bind to the chromosomes, MAR-mCherry is absent ( Figure   6A and C, right column). Finally, although Nup60(1-388)-GFP is present on meiotic chromosomes, it does not appear to recruit MAR-mCherry. These results indicate that the C-terminus of Nup60 is dispensable for the binding of Nup60 to meiotic chromosomes but required for binding to the MAR.
Since Nup60(1-388) is sufficient for sporulation but does not appear to interact with the MAR at the nuclear envelope or on meiotic chromosomes, it seemed likely that Nup60(1-388) does not require Nup2 to function.
While the NUP60 strain sporulates poorly, the nup60(1-388) and nup60(1-388)-GFP diploids both sporulate nearly as well as a strain that is completely wild type (Table 3). Oddly, we found that the full length Nup60-GFP fusion also restored sporulation. The nearly wild-type level of sporulation observed with the NUP60-GFP nup2 ndj1 strain is in sharp contrast to the poor sporulation of the nup2 ndj1 strain in which Nup60 is untagged. To test whether the GFP tag itself was conferring a new function onto fulllength Nup60, we also measured sporulation in nup2 ndj1 strains where Nup60 was tagged with mCherry or the myc antigen. Since NUP60-mCherry and NUP60-13myc also suppress the ndj1 nup2 sporulation defect, it seems more likely that the tags are somehow interfering with the C-terminus of Nup60, causing the full-length fusions to behave like the truncated alleles. In addition, we found that suppression of the nup2 ndj1 sporulation defect is dominant, indicating that the presence of untagged, full-length Nup60 does not prevent the tagged or truncated versions from functioning in the absence of Nup2.

The Nup60-GFP fusion binds to meiotic chromosomes in the absence of Nup2.
Although Nup60-GFP is able to recruit MAR-mCherry to both the nuclear envelope and meiotic chromosome spreads, the efficient sporulation of the NUP60-GFP ndj1 nup2 strain suggested that the Nup60-GFP fusion does not actually require the presence of the MAR to bind to meiotic chromosomes.
We prepared meiotic spreads from NUP60-GFP and NUP60-GFP nup2 strains and found that Nup60-GFP was bound to the chromosomes in both cases (Figure 7). Therefore, Nup60-GFP is capable of binding to meiotic chromosomes without Nup2.

Discussion
Previously, we showed that a mutation in NUP2 produces a synthetic sporulation phenotype when combined with a mutation in either NDJ1 or CSM4, two genes required for the association of meiotic telomeres with the LINC complex. These results suggested that the NPC and the LINC complex contribute to sporulation through functionally redundant but distinct pathways, perhaps by organizing meiotic chromosomes in the nucleus. The work presented here builds on those earlier findings to show that mutations in NUP60 also give a synthetic sporulation defect with ndj1∆ and csm4∆ mutations.
Moreover, while binding of the Nup2 MAR to both the nuclear envelope and meiotic chromosomes is completely dependent on Nup60, the N-terminus of Nup60 is capable of binding chromosomes and promoting sporulation in the absence of Nup2. This is significant because Nup2 is thought to be only transiently associated with the NPC, while Nup60's interaction with the core NPC is more stable. If Nup60 rather than Nup2 is responsible for promoting sporulation, that would suggest that the association of chromatin with the nuclear envelope is important for sporulation.
Normally, both Nup60 and the Nup2 MAR are required for sporulation in ndj1 mutants: it is only when the C-terminus of Nup60 is deleted that Nup2 is no longer needed. These results indicate that the sporulation function of Nup60 can be attributed to the N-terminal domain, that the C-terminus somehow represses this function, and that Nup2 counteracts that inhibition (Figure 8). One simple explanation for how autoinhibition of Nup60 might occur is that the protein interacts with itself and that binding of the MAR disrupts this interaction. AlphaFold modeling of Nup60's structure (Jumper et al. 2021) predicts that the Nup60 C-terminus could potentially mask a large part of the N-terminal domain ( Figure S5B).
Although the confidence score is very poor for most of the Nup60(389-539) fragment, and FxF repeat regions are thought to be largely unstructured (Denning et al. 2003), the helix at the extreme Cterminus is predicted with high confidence and may be responsible for mediating an interaction between Nup60's N-and C-terminal domains.
Another possibility is that an inhibitory protein binds to the Nup60 C-terminus either directly or as cargo of the transport machinery. Fusion of a fluorescent protein (GFP, mCherry) or an epitope tag (13myc) to Nup60's C-terminus mimics the effect of a C-terminal deletion, as any of these modifications disrupts Nup60's autoinhibition. A parallel phenomenon has been observed with Nup1, where adding a large protein tag or deleting the C-terminal 36 amino acids interferes with Nup1's ability to recruit damaged telomeres to the NPC but does not affect any of Nup1's other known functions (Aguilera et al. 2020). Like Nup60, Nup1 is considered to be a functional homolog of mammalian Nup153, partly due to the presence of a conserved importin-α docking site at the extreme C-termini of Nup1 and Nup153, which in Nup1 is known to bind importin-β (Kap95) (Pyhtila and Rexach 2003;Sistla et al. 2007). While the predicted helix at the end of Nup60 is not homologous to the Nup1/Nup153 docking site, importinα/β heterodimers bind to Nup60 via its C-terminus, and this binding can prevent Nup60 from interacting with Nup2 (Denning et al. 2001).
If the main role of Nup2 in promoting sporulation is to counteract the inhibitory effect of the Nup60 C-terminus, then it is perhaps not surprising that only a short segment of Nup2 is required for function and that the mutations identified by our screen all seem to affect the ability of the MAR to bind to Nup60. In principle, inhibition could be achieved simply by masking a small surface on Nup60. While a crystal structure of full-length Nup2 is not available, AlphaFold predicts that Nup2 contains two helices that span residues 87-132 (Jumper et al. 2021) and coincide almost exactly with the functional region of the MAR (residues 89-130) ( Figure S5C). When we map the MAR mutations identified by our screen onto these modeled helices, we find that many of the mutations lie on the side of the helices facing a groove in the Nup2 protein ( Figure S5D), potentially defining a surface of the MAR that interacts with Nup60. Most of the remaining mutations map to the interface between the two helices and possibly result in partial unfolding of the MAR.
We identified three MAR mutants (K101E, F120Y, and mar(89-130)) that localize properly to the nuclear envelope but do not bind to meiotic chromosomes or promote sporulation. Despite their proper subcellular localization, there is evidence that these mutations also alter the MAR's interaction with Nup60. First, the three mutants do not interact with the central fragment of Nup60 by Y2H. Second, the sporulation phenotype is severe in hemizygous (mar/Δ) strains but mild in mar/mar homozygotes, suggesting that the defect can be overcome by increasing the amount of mutant MAR protein. Finally, K101E and F120Y lie on the same face of the MAR helices as most of the mutations that disrupt localization to the nuclear envelope. One possibility is that K101E, F120Y and mar(89-130) bind to Nup60 but not in a way that counteracts the inhibitory effect of the C-terminus; as a result, the mutant Because C-terminally tagged Nup60 can bind to meiotic chromosomes without Nup2, it is tempting to speculate that the function of the C-terminus is to inhibit chromosome binding and that Nup2 overcomes that inhibition. However, it is important to point out that we do not yet know whether untagged Nup60 binds to meiotic chromosomes in the absence of Nup2. A Nup60 antibody is not commercially available, and we have been unable to construct a tagged version of Nup60 that behaves like the untagged protein despite numerous attempts (Table 3, Figure S6). We believe that the chromosome binding exhibited by Nup60-GFP fusion reflects the intrinsic chromosome binding activity of untagged Nup60, as Nup60-mCherry and Nup60-3HA also bind to meiotic chromosomes ( Figure   S7).
We note that the results presented here differ from our previously published work in two respects. First, the sporulation efficiencies are slightly lower for wild type (~90% vs. 100%) and higher for ndj1 nup2 (10% vs. 1%). Second, whereas we previously saw almost no meiotic division (i.e., separation of chromosome masses at meiosis I) in the ndj1 nup2 double mutant, we now see that a significant proportion of cells (~ 40%) eventually execute both meiotic divisions. We see these postmeiotic cells both in newly constructed strains and in strains that were used in Chu et al., 2017. We do not yet have a full explanation for the disparity. However, the new observations do not substantially alter the basic conclusion that loss of Nup2 causes a meiotic delay that is made more severe by the absence of Ndj1. The formation of multinucleate cells in ndj1 nup2 strains is also consistent with the previous observation that the double mutant undergoes two rounds of spindle-pole body duplication, suggesting that meiotic progression is taking place (Chu et al., 2017).
Although a significant fraction of nup2 ndj1 and nup60 ndj1 cells execute both meiotic divisions, many of these cells fail to form visible tetrads. This implies that defects in the NPC basket affect both meiosis and the sporulation process itself. One possibility is that an extended delay in meiosis I somehow interferes with proper sporulation, perhaps by interfering with NPC assembly following meiotic rejuvenation. During the second division of budding yeast meiosis, rDNA circles, nucleolar proteins, and core nucleoporins of the mother cell are sequestered away from the four gamete nuclei into a fifth vesicle that is ultimately excluded from the spores (Fuchs and Loidl 2004;King et al. 2019).
This process eliminates age-damaged nuclear components but leaves the spore nuclear membranes bereft of intact NPCs. The basket nucleoporins are the only NPC subunits that segregate with the chromatin. While neither Nup2 nor Nup60 has been directly implicated in NPC assembly, their metazoan homologs Nup50 and Nup153 have been shown to seed interphase NPC assembly in vitro (Schwartz et al. 2015;Vollmer et al. 2015). Because the yeast nuclear membrane remains intact throughout mitosis and meiosis, post-meiotic NPC assembly most likely resembles interphase NPC assembly. Thus, it is plausible that retention of the basket is mediated through Nup60's association with the segregating chromosomes and that disruption of the basket leaves the newly formed nuclei unable to properly repopulate their membranes with NPCs. It is unclear, however, why an effect on sporulation would only manifest when the cells are also ndj1 or csm4.
The mechanism by which Nup60 promotes meiosis and sporulation once it has bound Nup2 is still an open question. Nup136, the functional homolog of Nup153 in Arabidopsis, helps remove meiotic interlocks by promoting chromosome movement (Martinez-Garcia et al. 2018). While nup2 mutants do not show any obvious defect in rapid prophase movement (Chu et al, 2017), it is possible that Nup2/Nup60 influences chromosome movement in more subtle ways. Nup60 may also be acting indirectly via effects on gene expression. Nup60 represses transcription of CLN2 by recruiting the gene to the nuclear envelope (Kumar et al. 2018). Interestingly, deacetylation of a lysine residue in the Cterminus of Nup60 prevents it from binding to the CLN2 promoter. Alternatively, Nup2/Nup60 might be acting through Hop1 or Pch1. Nup2 facilitates removal of Hop1 from interstitial regions of chromosomes, allowing double strand break potential to persist near the chromosome ends and helping to ensure that short chromosomes form a crossover (Subramanian et al. 2019). Nup2 has also been shown to promote association of the checkpoint protein Pch2 with meiotic chromosomes and to control the distribution of Pch2 between the nucleus and the cytoplasm (Herruzo et al. 2021). It will be interesting to see if Nup2's effect on either Hop1 or Pch2 requires Nup60. Finally, several landmarks of meiotic prophase are delayed in the nup2 mutant, including homolog pairing, double-strand break repair, and synapsis, and this delay is much worse in an ndj1 mutant background (Chu et al, 2017). On the other hand, double-strand breaks and the products of recombination (both crossovers and noncrossovers) are slightly elevated in the nup2 mutant. It would not be surprising if nup60 mutants exhibited these same phenotypes.
The NPC was initially defined by its role in nucleocytoplasmic trafficking (Wente and Rout 2010) but also helps to maintain the functional organization of chromatin within the nucleus. This organization influences a variety of DNA transactions including gene expression and DNA repair. Our work has uncovered a mechanism linking the NPC basket proteins Nup60 and Nup2 to sporulation, with Nup60 playing the major role and a 42-aa region of Nup2 acting to alter Nup60's activity. Nup60's C-terminus appears to inhibit the N-terminus, and binding of Nup2 relieves this inhibition. Whether the interaction between Nup2 and Nup60 is regulatory in nature and whether it is preserved in other processes involving Nup2/Nup60 binding to chromatin remains to be determined.

Data availability
Strains and plasmids are available upon request. All supplemental figures, tables, and files and raw image data will be deposited in the Dryad repository upon acceptance for publication; doi TBD.
Supplemental information contains supplemental figures, tables, and methods; Supplemental File S1 contains sporulation data; File S2 contains fluorescence measurements for Figure 3AB; File S3 contains fluorescence measurements for Figure 3CD; File S4 contains fluorescence measurements for Figure 6; File S5 contains fluorescence measurements for Figure 8; File S6 contains time course data.
We thank members of the Burgess lab for stimulating discussions and Carly Cheung and Daniel Chu for technical support with plasmids and yeast strains. We are grateful to Shannon Owens and Amy MacQueen for advice on the chromosome spread procedure. This work was funded by NIH R01 GM701195 to SMB.       using a two-tailed t-test.

Figure 8
Cartoon representation of Nup60 auto-inhibition. Left: In the absence of Nup2, Nup60 is inactive due to inhibition by its own C-terminus.