A ubiquitin–proteasome pathway degrades the inner nuclear membrane protein Bqt4 to maintain nuclear membrane homeostasis

The eukaryotic nucleus is enclosed by the nuclear envelope (NE), which separates the nucleoplasm from the cytoplasm. The NE comprises the inner and outer nuclear membranes (INM and ONM, respectively). The ONM is continuous with the endoplasmic reticulum (ER). The INM is involved in the organization and regulation of chromosomes by modulating chromatin potential for gene expression, DNA replication, DNA damage repair and genome stability (Mekhail and Moazed, 2010; Hirano et al., 2020a; Pawar and Kutay, 2021). Excess amounts of INM proteins result in their aberrant accumulation at the NE, often causing nuclear deformation and malfunctioning (Gonzalez et al., 2012). The aberrant accumulation of INM proteins has been linked to certain mammalian diseases. For example, accumulation of the INM protein SUN1 was found to be pathogenic in human laminopathies such as Emery–Dreifuss muscular dystrophy and Hutchinson–Gilford progeria syndrome (Chen et al., 2012). Therefore, the amount of INM proteins must be regulated. However, the mechanisms that regulate levels of resident INM proteins to maintain INM homeostasis remain poorly understood.

Mechanisms for ER membrane homeostasis have been well studied: ER membrane proteins are regulated through the ER-associated degradation (ERAD) system, which is involved in degrading misfolded or mislocalized ER proteins as a means of protein quality control (Vembar and Brodsky, 2008; Ruggiano et al., 2014; Christianson and Carvalho, 2022). ERAD involves the ubiquitin–proteasome system (UPS), in which polyubiquitin chains are covalently attached to a substrate protein, leading to its recognition and degradation by the 26S proteasome complex. Ubiquitination reactions involve three different classes of enzymes: E1 ubiquitin-activating enzymes, E2 ubiquitin-conjugating enzymes and E3 ubiquitin ligases, such as Hrd1 and Doa10 (human MARCHF6 or TEB4). INM-associated degradation (INMAD) has been proposed as an extension of ERAD to maintain INM homeostasis, which also depends on the UPS (Foresti et al., 2014; Khmelinskii et al., 2014; Koch and Yu, 2019; Smoyer and Jaspersen, 2019; Mannino and Lusk, 2022). It has been shown that in the budding yeast Saccharomyces cerevisiae, ERAD and INMAD require additional machinery for the degradation of integral membrane proteins, involving Cdc48 (also known as p97 or VCP in mammals), an ATPase associated with diverse cellular activities (AAA-ATPase), which removes ubiquitinated proteins from the membrane before transferring them to the proteasome for degradation (Ruggiano et al., 2014; Christianson and Carvalho, 2022; Koch and Yu, 2019; Smoyer and Jaspersen, 2019; Mannino and Lusk, 2022).

In the fission yeast Schizosaccharomyces pombe, several INM proteins have been studied for their roles in the organization and regulation of chromosome stability, including Ima1, Lem2, Man1, Bqt3 and Bqt4 (Chikashige et al., 2006, 2009; Hiraoka et al., 2011; Steglich et al., 2012; Gonzalez et al., 2012; Tange et al., 2016; Hirano et al., 2020b). In particular, we found that Bqt4 was degraded in the absence of Bqt3 (Chikashige et al., 2009). Bqt3 is a transmembrane protein composed of seven transmembrane domains (TMDs) along its entire length. Bqt4, which interacts with Bqt3, is predicted to be a tail-anchored protein. Immunoelectron microscopy analysis has shown that Bqt3 and Bqt4 localize at the INM (Chikashige et al., 2009). Bqt3 and Bqt4 are required for anchoring telomeres to the NE during vegetative growth and clustering telomeres to the spindle pole body through the meiosis-specific telomere-binding proteins Bqt1 and Bqt2 (Chikashige et al., 2006, 2009). Importantly, Bqt4 interacts with Lem2 (Hirano et al., 2018) and the depletion of Bqt4 is synthetic lethal with the depletion of Lem2 (Tange et al., 2016; Kinugasa et al., 2019). Therefore, Bqt4 might play an important role in cell viability. However, the non-telomeric vital roles of Bqt4 and the mechanisms underlying its degradation remain unknown. In this study, we identified the factors required for Bqt4 degradation through an INM-associated degradation mechanism. We also demonstrated that excess amounts of Bqt4 resulted in impaired nuclear morphology and growth defects, indicating the importance of regulated degradation of Bqt4 for membrane homeostasis.

Because the UPS accounts for the selective degradation of many proteins in the nucleus (Enam et al., 2018; Franić et al., 2021), we first tested the requirement of the proteasome for Bqt4 degradation. We treated cells expressing Bqt4 tagged N-terminally with GFP (GFP–Bqt4) under its own promoter with bortezomib (BZ), a chemical proteasome inhibitor effective in S. pombe (Takeda et al., 2011), and assessed the fluorescence levels of GFP–Bqt4 at 4 h after drug addition in the presence or absence of Bqt3. As a result of the treatment, GFP–Bqt4 fluorescence in the nuclei of both bqt3⁺ and bqt3Δ cells was elevated (Fig. 1A). To confirm this result, we measured protein levels by western blotting, which consistently showed an increase in GFP–Bqt4 protein levels in these strains upon proteasomal inhibition by BZ (Fig. 1B). Notably, the GFP–Bqt4 signal was restored upon proteasome inhibition in the absence of Bqt3 and also elevated in the presence of Bqt3. This suggests that Bqt4 is constitutively turned over by the proteasome system and its association with Bqt3 antagonizes this process. We sought to confirm this phenomenon by introducing temperature-sensitive mts2-1 and mts3-1 proteasome mutations (Gordon et al., 1993; Gordon et al., 1996). Using both fluorescence microscopy and western blotting, we found that GFP–Bqt4 levels were increased in bqt3⁺ and bqt3Δ cells carrying these proteasome mutations at a restrictive temperature of 36°C, compared to GFP–Bqt4 levels in the corresponding wild-type (WT) strains (Fig. 1C,D; Fig. S1). These results indicate that Bqt4 degradation proceeds via the proteasome system.

As proteins are ubiquitinated as a requirement for recognition by the proteasome, we examined whether Bqt4 was modified by polyubiquitination. We immunoprecipitated GFP–Bqt4 under denaturing conditions and detected associated ubiquitin by immunoblotting. Ubiquitinated forms of Bqt4 were detected in bqt3Δ cells and were enriched in both bqt3⁺ and bqt3Δ cells when proteasomal activity was inhibited (Fig. 1E), suggesting that Bqt4 was targeted to the proteasome by polyubiquitin modification.

Bqt4 is degraded via the UPS in the absence of Bqt3, suggesting that Bqt3 plays a role in stabilizing Bqt4 by protecting Bqt4 from degradation. To assess whether Bqt3 stabilizes Bqt4, we determined the degradation kinetics of GFP–Bqt4 in the presence or absence of Bqt3 by a cycloheximide chase assay followed by western blotting. The results showed that GFP–Bqt4 degraded faster in the absence of Bqt3 than in the presence of Bqt3 (Fig. 1F). This indicates that Bqt4 is protected from proteasome-dependent degradation by Bqt3. Thus, Bqt4 is maintained at the INM by associating with Bqt3, and is polyubiquitinated and degraded quickly once it dissociates from Bqt3.

Vacuolar autophagy is another pathway through which cells degrade proteins. Thus, we next examined whether Bqt4 degradation is regulated by the vacuolar autophagy pathway by analyzing GFP–Bqt4 levels in bqt3Δ cells lacking the isp6⁺ and atg1⁺ genes. Isp6 is a serine protease necessary for vacuolar function and Atg1 is a serine/threonine protein kinase that is the upstream initiator of the autophagy machinery. Deletion of these two genes has been shown to completely impair the vacuolar autophagy pathway (Kohda et al., 2007). We found that the deletion of isp6 and atg1 had no detectable effect on the quantity of GFP–Bqt4 in the absence of Bqt3, suggesting that Bqt4 degradation does not involve the vacuolar autophagy system (Fig. 1G). Collectively, we conclude that the Bqt4 protein is degraded by the UPS without the involvement of the vacuolar autophagy pathway.

Next, we determined whether Bqt4 was degraded in the nucleus by the nuclear proteasome. In S. pombe, proper targeting of proteasomes to the nucleus requires the tethering factor Cut8 (Tatebe and Yanagida, 2000). In cut8 mutants, proteasomes fail to localize to the nucleus and are mostly cytoplasmic (Tatebe and Yanagida, 2000). Because nuclear protein degradation depends on the localization of proteasomes in the nucleus (Takeda and Yanagida, 2005), we examined the levels of GFP–Bqt4 in bqt3⁺ and bqt3Δ cells carrying the temperature-sensitive cut8-563 mutation. We found that GFP–Bqt4 degradation was suppressed in cut8-563 cells shifted to the nonpermissive temperature of 36°C (Fig. 2A,B). Taken together, our results strongly suggest that Bqt4 is degraded mainly by the proteasome within the nucleus.

We next sought to identify the specific E3 ubiquitin ligase that directly recognizes Bqt4. As Bqt4 is a protein associated with the INM, we investigated whether membrane-localized Hrd1 and Doa10 E3 ligases participate in the Bqt4 degradation pathway. Deletion of doa10⁺ partially restored GFP–Bqt4 levels in the absence of Bqt3, whereas deletion of hrd1⁺ did not (Fig. 3A,B, comparing hrd1Δ to doa10Δ). These results are consistent with previous studies that showed that Doa10 partially localizes to the INM and is involved in the degradation of certain nuclear and INM substrates, whereas Hrd1 is found exclusively in the ER (Deng and Hochstrasser, 2006; Boban et al., 2014). Additionally, the double deletion of doa10⁺ and hrd1⁺ showed no obvious synthetic effect on GFP–Bqt4 restoration compared with the deletion of doa10⁺ alone (Fig. 3A,B). We also measured the amount of GFP–Bqt4 in mutants lacking Ubc6 and Ubc7, which are E2 ubiquitin-conjugating enzymes associated with Doa10 (Swanson et al., 2001). We found that GFP–Bqt4 levels increased to some extent in ubc6Δ and ubc7Δ single and ubc6Δ ubc7Δ double mutants (Fig. 3C,D). These findings indicate that the Doa10 complex contributed to the degradation of Bqt4 in the INM.

Although degradation of Bqt4 depended on Doa10, the level of GFP–Bqt4 was only partially restored in the doa10Δ mutant (Fig. 3A,B). Therefore, we attempted to identify other redundant E3 ligases involved in Bqt4 degradation. We constructed mutants of E3 ligases and their components that have been suggested to localize to or function in the ER and nucleus, namely meu34Δ, dsc1Δ, ubr1Δ, ubr11Δ, san1Δ, hul5Δ, ptr1-1, and cut9-665 (PomBase, https://www.pombase.org; Matsuyama et al., 2006, Nielsen et al., 2014; Enam et al., 2018; Franić et al., 2021), and observed GFP–Bqt4 fluorescence intensity in these mutants. The mutants tested showed no detectable increase in fluorescence, except for the hul5Δ mutant (Fig. S2). hul5Δ cells showed a slight increase in the fluorescence levels of GFP–Bqt4, but showed no increase in GFP–Bqt4 levels by western blotting (Fig. S3). Instead, a small degraded fragment was observed (Fig. S3), suggesting that the increase in the GFP signal in hul5Δ cells reflects this degraded product. Therefore, Doa10 is the only E3 ligase identified to date that ubiquitinates Bqt4.

We have previously shown that Bqt4 localizes exclusively to the nucleoplasmic side of the INM, and this localization requires its C-terminal helix (Chikashige et al., 2009). However, as this C-terminal helix was predicted to be a weak candidate to be a TMD (Fig. S4), we examined whether Bqt4 is an integral or peripheral membrane protein to understand how Bqt4 is removed from the INM before degradation. Therefore, we performed membrane protein extraction experiments. In brief, the membrane fraction was isolated from the cell extract, treated under various conditions, and then separated into pellet and supernatant fractions by ultracentrifugation, followed by western blotting: a protein was expected to enter the pellet fraction if it remained in the lumen or associated with the membrane after the treatment. GFP–ADEL (Pidoux and Armstrong, 1992) and Ish1–GFP were used as controls for luminal and integral membrane proteins, respectively. As shown in Fig. 4A, all these proteins were extracted from the membrane by treatment with Triton X-100 and 0.5 M NaCl. However, similar to Ish1, Bqt4 did not dissociate from the membrane by sodium carbonate (pH 11.5) or urea, which can strip peripheral membrane proteins. In contrast, GFP–ADEL was released from the membrane following sodium carbonate treatment, as expected for luminal proteins. Thus, Bqt4 is likely to behave as a tail-anchored protein integrally embedded in the INM in the presence of Bqt3.

To further characterize the properties of the TMD of Bqt4, we examined whether Bqt4 behaves as an integral membrane protein in the absence of Bqt3 using a membrane extraction assay. As most of the Bqt4 protein is degraded in bqt3Δ cells, we used bqt3Δ doa10Δ cells to prevent the degradation. Bqt4 was extracted from the membrane by detergent, but not by sodium carbonate or urea (Fig. 4B), as was the case in the bqt3⁺ condition, indicating that the TMD of Bqt4 was embedded in the INM independently of Bqt3.

The ATPase Cdc48 has been shown to facilitate the degradation of ubiquitinated ERAD membrane substrates by extracting them from the ER membrane in S. cerevisiae cells (Ye et al., 2001; Jarosch et al., 2002; Rabinovich et al., 2002; Wolf and Stolz, 2012). Therefore, we investigated whether Cdc48 is necessary for Bqt4 degradation. To this end, we generated a cdc48 mutant bearing mutations in both the D1 (P267L) and D2 (A550T) ATPase domains based on the temperature-sensitive cdc48-6 allele in S. cerevisiae (Schuberth and Buchberger, 2005; Ruggiano et al., 2016) (Fig. 5A). The cdc48 mutant (cdc48-P267L/A550T) exhibited cold-sensitive growth defects (Fig. 5B), suggesting that these mutations impaired Cdc48 activity in S. pombe. Inactivation of cdc48-P267L/A550T by shifting to a nonpermissive temperature of 16°C elevated the level of GFP–Bqt4 in bqt3⁺ and bqt3Δ cells, as observed by fluorescence microscopy (Fig. 5C) and western blotting (Fig. 5D); however, it was degraded at 33°C (Fig. S5), indicating that Cdc48 was required for Bqt4 degradation.

We previously showed that Bqt4 degradation depends on its C-terminal TMD (residues 414–432) (Chikashige et al., 2009). To identify the minimal region of Bqt4 that is sufficient for its degradation, we analyzed truncated C-terminal fragments based on the structure of Bqt4 predicted by AlphaFold2 (Fig. 6A; Fig. S6) (https://alphafold.ebi.ac.uk/entry/O60158). The Bqt4 fragment containing the helix domain and the adjacent intrinsically disordered sequence (Bqt4C^369–432) reproduced the behavior of full-length Bqt4 (Fig. 1A), that is, localization to the NE and responses to BZ, both in the presence and absence of Bqt3 (Fig. 6A, Bqt4C^369–432). The helix domain alone (Bqt4C^394–432) showed weaker localization to the NE and was somewhat diffused to the membrane compartments in the cytoplasm; nevertheless, this fragment was degraded in the absence of Bqt3 and its levels increased with BZ treatment (Fig. 6A, Bqt4C^394–432). The TMD alone (Bqt4C^414–432) showed even weaker localization at the NE with diffusion to the cytoplasm, but showed the same responses, that is, degradation in the absence of Bqt3 and increased levels upon BZ treatment (Fig. 6A, Bqt4C^414–432). These results indicate that the C-terminal fragment containing residues 369–432 was necessary and sufficient for the NE localization of Bqt4, and its truncation successively reduced NE localization. These results also indicate that the C-terminal TMD of Bqt4 is sufficient for its proteasome-mediated degradation.

To further narrow down the TMD of Bqt4 for proteasome-dependent degradation and Bqt3-dependent protection against degradation, we generated constructs containing a shorter TMD with the last seven residues truncated (Bqt4C^369–425). The fragment Bqt4C^369–425 lost its dependency on Bqt3 for degradation and exhibited similar behaviors in the presence or absence of Bqt3 (Fig. 6A, Bqt4C^369–425): this fragment showed increased signal levels in the presence of BZ, indicating that it is also a substrate of proteasomal degradation in a Bqt3-independent manner. As the level of Bqt4C^369–425 were not restored by doa10 deletion, the fragment is likely degraded by a pathway distinct from that of Doa10 (Fig. S7). To further characterize the reason why these fragments were degraded independent of Bqt3, we examined the interaction between these fragments and Bqt3 as well as the membrane association state of these fragments. The results of the yeast-two-hybrid assay showed that the fragments lacking the last seven residues did not bind to Bqt3 (Fig. 6B). In addition, Bqt4C^369–425 was partially extracted by sodium carbonate (Fig. 6C), suggesting that the last seven residues assisted the membrane-embedding ability of the TMD. Thus, we concluded that the entire TMD (414–432) is the minimal region of the Bqt3-binding site and is a degron for proteasome-dependent degradation that is protected by Bqt3 (Fig. 6D).

We observed that inhibition of proteasomal degradation caused striking NE expansion and deformation, with irregular protrusions and invaginations, especially in bqt3⁺ cells, in which Bqt4 levels were excessively high (Fig. 1A,C). We hypothesized that high levels of Bqt4 in these cells were responsible for this abnormal NE morphology. To test this possibility, we inhibited proteasomal degradation in the presence or absence of endogenous Bqt4 and visualized the NE using Ish1–GFP as an NE marker, and then nuclear deformation was assessed by measuring the roundness of the nucleus. Only when Bqt4 was present, a strong deformation of the NE was observed upon proteasomal inhibition (see bqt4⁺+BZ; Fig. 7A), suggesting that Bqt4 was an essential factor causing this phenomenon. To ascertain whether this abnormal phenotype was caused by an accumulation of Bqt4, we overexpressed GFP–Bqt4 under the nmt1 promoter using the chemical compound YAM2 to control the expression level. YAM2 suppresses nmt1 promoter activity depending on its concentration (Nakamura et al., 2011). Overexpression of GFP–Bqt4 reproduced the nuclear-deformed morphology (Fig. 7B), and NE proliferation and deformation became more striking in cells that expressed higher levels of GFP–Bqt4 (Fig. 7C,D), indicating that the abnormal NE morphology depended on the amount of Bqt4. Taken together, these results support our hypothesis that Bqt4 is the only protein for which its accumulation upon inhibition of degradation induces aberrant nuclear morphology. Finally, we investigated whether an excess of GFP–Bqt4 was toxic. We found that overproduction of GFP–Bqt4 caused a growth defect in both bqt3⁺ and bqt3Δ strains compared to cells expressing endogenous levels of Bqt4 (Fig. 7E). Altogether, these observations suggest that Bqt4 plays an important role in regulating NE morphology, and a nuclear proteasomal degradation pathway controlling the quantity of Bqt4 is crucial for nuclear membrane homeostasis.

In this study, we identified the molecular factors required for the degradation of the INM integral protein Bqt4 in S. pombe (Fig. 8). Bqt4 is ubiquitinated and degraded by the nuclear proteasome via an INMAD mechanism involving the Doa10 E3 ligase, which possibly recognizes the C-terminal TMD of Bqt4 as a degron. In addition, the AAA-ATPase Cdc48 is required for Bqt4 degradation. We also demonstrated that the accumulation of undegraded Bqt4 in the INM caused strong deformation of the NE and resulted in growth defects. Bqt4 escapes degradation by interacting with Bqt3 through its C-terminal TMD. Excess amounts of Bqt4 that were not bound to Bqt3 were subjected to INMAD-mediated degradation to avoid the accumulation of Bqt4 at the NE. These findings contribute to our understanding of the significance of INM protein degradation in INM homeostasis.

Studies on mammalian cells have suggested that integral INM proteins are largely stable (Toyama et al., 2013; Buchwalter et al., 2019); however, it has been difficult to study their turnover. In the budding yeast S. cerevisiae, most of the INMAD substrates characterized thus far are not native INM proteins, but rather nuclear soluble proteins (Swanson et al., 2001; Ravid et al., 2006) and ER membrane proteins that are unassembled or mislocalized to the INM (Foresti et al., 2014; Khmelinskii et al., 2014; Natarajan et al., 2020). Thus, our study on Bqt4 as a native INM degradation substrate provides an opportunity to obtain further insight into the mechanisms governing INM protein degradation. In addition to several previous studies that reported a handful of native INM degradation substrates (Boban et al., 2014; Pantazopoulou et al., 2016; Smoyer et al., 2019; Koch et al., 2019), our results on the degradation of the INM protein Bqt4 demonstrate that INMAD is responsible not only for maintaining INM identity by removing foreign proteins but also for maintaining the proteostasis of INM integral components.

An important question is whether INMAD machinery specifically recognizes its substrates. Although a degradation signal has yet to be identified for integral INM substrates of INM-localized E3 ubiquitin ligases, such as Doa10 and the Asi complex, we identified that Bqt4 has an intramembrane degron. This suggests that INMAD does not simply detect and degrade misfolded or mislocalized proteins in the INM, but instead specifically recognizes a substrate to regulate its protein levels.

In S. cerevisiae, the Doa10 E3 ubiquitin ligase cooperates with the Asi complex in INMAD (Koch and Yu, 2019; Smoyer and Jaspersen, 2019; Mannino and Lusk, 2022). As homologs of Asi proteins are absent in S. pombe as in other species, INMAD mechanisms might be diverse among organisms. However, increasing evidence has shown that the degradation of INM proteins in plants and mammals also employs the UPS and Cdc48/p97 (Tsai et al., 2016; Huang et al., 2020; Krshnan et al., 2022), suggesting a general mechanism for INM protein degradation.

In S. cerevisiae, the Doa10 E3 ligase can passively diffuse from the ER to the INM and functions in both ERAD and INMAD (Deng and Hochstrasser, 2006). In particular, Doa10 has been implicated in mediating the degradation of the INM protein Asi2, which is a component of INMAD (Boban et al., 2014). In S. pombe, Bqt4 degradation also depends on Doa10, indicating that Bqt4 is subject to regulation by INMAD. In addition, our data indicate that S. pombe might possess additional unidentified redundant E3 ligases cooperating with Doa10 at the INM. Because Bqt4 is supposed to be recognized by these E3 ligases at its transmembrane degron, it is expected that the unidentified E3 ligases also localize to the INM [‘X(E3)’ in Fig. 8]. Similar situations in which unidentified E3 ligases degrade an INM substrate have been reported in S. cerevisiae (Pantazopoulou et al., 2016). Similarly, INM-localized E3 ligases have yet to be identified in higher eukaryotes. We speculate that the uncharacterized E3 ligases functioning at the INM in S. pombe might operate with E2 ubiquitin-conjugating enzymes other than Ubc6 and Ubc7, because degradation of Bqt4 exhibits limited dependence on these enzymes. Thus, it would be of great interest to identify and characterize other membrane-associated E3 ligases and E2 conjugating enzymes that mediate protein degradation at the INM.

The Cdc48 complex plays a central role in the ERAD degradation of ER integral membrane proteins by exerting a pulling force on the ubiquitinated membrane protein via its ATPase activity (Braun et al., 2002; Rabinovich et al., 2002; Wolf and Stolz, 2012; Bodnar and Rapoport, 2017). Cdc48 has also been implicated in the degradation of INM proteins Asi1 and Msp3 in S. cerevisiae (Pantazopoulou et al., 2016; Koch et al., 2019). In S. pombe, Cdc48 mutations at its ATPase active center impair Bqt4 degradation (Fig. 5), indicating that ERAD and/or INMAD degradation involves the Cdc48 complex; however, it remains unclear whether Cdc48 extracts Bqt4 from the INM.

The INM protein Bqt4 has been suggested to have non-telomeric functions that are currently unknown but are vital for cell growth, as the double deletion of Bqt4 and Lem2 results in synthetic lethality (Tange et al., 2016; Hirano et al., 2018; Kinugasa et al., 2019; Hirano et al., 2023b). In cells lacking both Bqt4 and Lem2, NE rupture occurs in association with leakage of nuclear proteins into the cytoplasm, suggesting that Bqt4, in cooperation with Lem2, plays an essential role in maintaining nuclear membrane integrity (Kinugasa et al., 2019). The nuclear membrane rupture phenotype can be rescued by the increased expression of Elo2, a fatty acid elongase that catalyzes the synthesis of very-long-chain fatty acids (Kinugasa et al., 2019), and Tlc4, a ceramide synthase homolog which is localized in the Golgi (Hirano et al., 2023b). In addition, the depletion of Bqt4 and Lem2 leads to a reduction in the quantity of t20:0/24:0 phytoceramide, the most abundant ceramide species in S. pombe, and this reduction can be prevented by the overexpression of these proteins (Kinugasa et al., 2019; Hirano et al., 2023b). These findings strongly suggest that Bqt4 is involved in nuclear membrane lipid metabolism. Thus, we propose that the abnormal accumulation of Bqt4 at the INM might result in excessive lipid synthesis, ultimately leading to nuclear membrane expansion and deformation. One possibility is that Bqt4 promotes the synthesis of phospholipids, a major component of the nuclear membrane, as Bqt4 binds to an enzyme involved in lipid synthesis (Hirano et al., 2023a). Several lines of evidence suggest that increased phospholipid synthesis from phosphatidic acid induces striking NE membrane deformation (Santos-Rosa et al., 2005; Barbosa et al., 2015). The exact mechanisms leading to the deformed nuclear morphology upon the accumulation of Bqt4 remain to be elucidated. Further investigations will advance our understanding of the role of INM protein degradation in the maintenance of nuclear membrane homeostasis.

All S. pombe strains used in this study are listed in Table S1. Unless otherwise specified, all S. pombe strains were cultured in yeast extract with supplements (YES) medium. Malt extract medium (ME) was used to induce sporulation. G418 disulfate (Nacalai Tesque, Japan), hygromycin B (Wako, Japan), nourseothricin sulfate (Werner BioAgents, Germany) and blasticidin S (Wako) were used at concentrations of 100, 200, 100 and 100 μg/ml, respectively, to select strains bearing the respective selection marker.

All plasmids were constructed using the NEBuilder system (New England BioLabs, MA, USA) and enzymatic digestion and ligation (TaKaRa Bio, Japan) systems, according to the manufacturer's instructions. Gene disruption and GFP tagging were performed using the direct chromosomal integration method (Wach, 1996; Bähler et al., 1998). The integration cassettes were amplified using a two-step PCR. In the first PCR, fragments (∼500 bp genomic regions upstream and downstream of the gene of interest) were amplified from the WT S. pombe genome using KOD One DNA polymerase (TaKaRa Bio). The first PCR products were used as primers in the second PCR step to amplify a cassette for integration, which contained a selection marker. For gene disruption, pFA6-kanMX6 (Bähler et al., 1998), pCR2.1-hph (Sato et al., 2005), pCR2.1-bsd (a gift from Dr Takeshi Sakuno, Osaka University, Japan) and pCR2.1-nat (Sato et al., 2005) were used to generate the integration cassettes. For GFP tagging, pFA6-GFP(S65T)-kanMX6 (Bähler et al., 1998) was used to generate the GFP-containing integration cassettes. The cassettes were then transformed into S. pombe cells and the desired integrant strains were selected on YES plates containing the appropriate selection agent. Correct disruption and integration were confirmed by genomic PCR at the 5′ and 3′ ends using KOD FX Neo (TaKaRa Bio). We confirmed that GFP–Bqt4 is functional, as it suppressed the synthetic lethality of bqt4 and lem2 double deletion (Tange et al., 2016) and anchored telomeres to the NE (Chikashige et al., 2009).

All fluorescence microscopy images were obtained from live cells. Fluorescence microscopy data of intracellular GFP fusion proteins were collected using a DeltaVision microscope system (GE Healthcare, USA) equipped with a pco.edge 4.2 sCMOS camera (PCO, Germany) and a 60× PlanApo N OSC oil-immersion objective lens (numerical aperture=1.4, Olympus, Japan). Cells were precultured from fresh colonies and cultured in YES medium at 30°C to reach the logarithmic growth phase (5–10×10⁶ cells/ml). The cells were then attached to glass-bottomed culture dishes coated with soybean lectin (Sigma-Aldrich), covered with Edinburgh minimal medium with glutamate (EMMG) medium (Moreno et al., 1991), and observed at specified temperatures. To control Bqt4 expression level under the nmt1 promoter, cells were precultured in EMMG medium supplemented with 2 μM thiamine overnight and washed three times with EMMG medium; the cells were inoculated in EMMG medium supplemented with 0–2 μM YAM2 (cat. no. Z2895181411, Enamine, Kyiv, Ukraine), and cultured overnight until the cell concentration reached mid-log phase.

An identical fluorescence intensity range was applied to all images presented in each panel using the Fiji software (Schindelin et al., 2012). To quantify the fluorescence intensity in the nucleus, we used Fiji. First, we generated the sum projection image from the raw z-stack images and then subtracted the background signal from the entire image by measuring the mean intensity of a region where no cells were present. Second, the nuclear region was selected manually or semi-automatically. For semi-automatic selection, individual nuclei were selected (original image), and the selected image was blurred using Gaussian blur (blurred image). An appropriate sigma value was applied depending on the image contrast. An edge-enhanced image was generated by subtracting the blurred image from the original image. A threshold was applied to the nuclear region using the appropriate minimum and maximum intensity ranges. Third, the threshold region was applied to the original image and the total fluorescence intensity of the region was quantified. The nuclear intensities obtained from bqt3⁺ cells were averaged and the values were used to calculate the relative fluorescence intensity. The relative fluorescence intensities were plotted, and the results are presented as the mean±standard deviation.

For the roundness index measurement, after selecting individual nuclei, a threshold was applied to the nuclear membrane signal using the appropriate fluorescence intensity ranges. The nuclear region was filled in, and the roundness index was measured using the built-in function in Fiji.

Cells were precultured in YES medium at 30°C overnight, inoculated in fresh YES medium, and then cultured until the cell concentration reached mid-log phase (5×10⁶–10×10⁶ cells/ml). Approximately 5×10⁷ cells were harvested by centrifugation at 2000 g at 4°C, and the cell pellet was washed with 1 ml ice-cold water containing 1 mM phenylmethylsulfonyl fluoride (PMSF). The cells were lysed in 150 μl of ice-cold 1.85 M NaOH containing 7.5% β-mercaptoethanol (β-ME) on ice for 15 min, followed by the addition of 150 μl of ice-cold 55% (v/v) trichloroacetic acid (TCA) to precipitate the proteins, incubation on ice for 15 min, and centrifugation (10 min, 14,000 g, 4°C). The protein pellet was then washed twice with ice-cold acetone and dissolved in sample solubilization buffer [50 mM Tris-HCl (pH 8.0), 1% SDS, 1× protease inhibitor cocktail (P8215, Sigma-Aldrich), 1 mM PMSF and 1 mM BZ (LC Laboratories, USA)]. Total protein concentrations were determined using the BCA assay (Thermo Fisher Scientific) according to the manufacturer's protocol. The samples were further diluted in 2× Laemmli SDS sample buffer to a concentration of 1 μg/μl. After the addition of dithiothreitol to a concentration of 50 mM, the samples were heated at 95°C for 10 min before immunoblotting. Identical amounts of protein (10 μg) were separated by SDS–PAGE and transferred to a polyvinylidene difluoride membrane. The membranes were blocked for 1 h at room temperature (approximately 26–28°C) in blocking solution [PBST (10 mM Na₂HPO₄, 137 mM NaCl, 2.7 mM KCl, 1.76 mM KH₂PO₄ and 0.1% Tween 20) containing 5% skim milk (Nacalai Tesque)]. To detect the GFP fusion proteins and β-actin as a loading control, the membranes were probed with anti-GFP rabbit polyclonal antibody (cat. no. 600-401215, Rockland Immunochemicals, USA; 1:2000 dilution) or anti-β-actin mouse monoclonal antibody (cat. no. ab8224, Abcam, UK; 1:2000) in blocking solution overnight at 4°C. After three rounds of washing with PBST, the membranes were incubated with appropriate horseradish peroxidase-conjugated secondary antibodies (cat. nos. NA931 and NA934, Cytiva, USA; 1:2000) in a blocking solution at room temperature for 2 h. After three rounds of washing with PBST, the immunoreactive bands were detected using a chemiluminescence system (ImmunoStar LD, Wako) on a ChemiDoc imaging system (Bio-Rad, USA).

Cells were grown in YES medium until early mid-log phase (2–5×10⁶ cells/ml), as described above. The culture was then divided into two groups, and the proteasome inhibitor BZ dissolved in dimethyl sulfoxide (DMSO) was added to one culture to a final concentration of 1 mM; the same amount of DMSO was added to the other culture. The cells were incubated at 26°C for 4 h before harvesting for fluorescence microscopy and immunoblotting.

For the temperature-sensitive mutant experiments, cells harboring temperature-sensitive or cold-sensitive mutations were grown in YES medium until the early mid-log phase (2–5×10⁶ cells/ml) at 26°C and 33°C, respectively. The culture temperature was shifted up to 36°C or down to 16°C, and then the cells were incubated for 4 h and 15 h, respectively, before harvesting.

The ubiquitination detection assay was performed as described previously (Boban et al., 2014) with modifications. Cells were grown in YES medium at 30°C until the mid-log phase (5–10×10⁶ cells/ml), and approximately 5×10⁸ cells were harvested by centrifugation at 2000 g at 4°C. Where indicated, the proteasome inhibitor BZ was added at a final concentration of 1 mM for 4 h before the cells were collected. The same volume of DMSO was added to the control samples. The cells were washed with cold water containing 1 mM PMSF and lysed in 250 μl of 1.85 N NaOH solution containing 7.5% β-ME for 15 min on ice. Subsequently, the proteins were precipitated by adding 250 μl of 55% (v/v) TCA, followed by incubation for 15 min on ice. The proteins were pelleted by centrifugation (10 min, 14,000 g, 4°C) and washed with 500 μl of 50 mM Tris before being dissolved in 200 μl denaturation solution [25 mM Tris-HCl pH 7.4, 2% SDS, 2 mM ethylenediaminetetraacetic acid (EDTA), 1× protease inhibitor cocktail, 1 mM PMSF, 1 mM BZ, 10 μM PR-619 (SML0430, Sigma-Aldrich) and 50 mM 2-chloroacetamide (032-09762, Fuji Film Wako, Japan)]. Total protein concentrations were determined using the BCA assay (Thermo Fisher Scientific) according to the manufacturer's protocol. The samples were incubated at 65°C for 10 min and clarified by centrifugation (5 min, 15,000 g, room temperature). Equal amounts of protein lysates were diluted fivefold with cold immunoprecipitation dilution buffer (50 mM Tris-HCl pH 7.5, 1.2% Triton X-100, 2 mM EDTA, 100 mM NaCl, 1×protease inhibitor cocktail, 1 mM PMSF, 1 mM BZ, 10 μM PR-619 and 50 mM 2-chloroacetamide), incubated with gentle rotation at 4°C for 1 h, and cleared by centrifugation at 14,000 g for 10 min. A portion (20 μl) of the sample (2% input) was mixed with 2× Laemmli SDS sample buffer (20 μl) and frozen until analysis. The remaining sample was mixed with antibody-coupled beads prepared from 2 μg anti-GFP rabbit polyclonal antibody (cat. no. 600-401215, Rockland Immunochemicals) and 50 μl protein-A-coupled Dynabeads (Life Technologies, USA), according to the manufacturer's instructions. The samples were immunoprecipitated at 4°C for 3 h with gentle rotation, and then pelleted using a magnet. The pellets were then washed as follows: twice with washing solution A (50 mM Tris-HCl pH 7.4, 2 mM EDTA, 100 mM NaCl, 1% Triton X-100 and 0.4% SDS), once with washing solution B (50 mM Tris-HCl pH 7.4, 2 mM EDTA, 350 mM NaCl and 1% Triton X-100) including a 10 min incubation at 4°C with gentle rotation, and finally once with washing solution A. To elute the immunoprecipitated protein, 50 μl of 1× Laemmli SDS sample buffer was added to the pellet, and the mixture was incubated for 10 min at 65°C. Ten microliters of each sample were separated on SDS-PAGE gels (8%). For the analysis of immunoprecipitated GFP–Bqt4, the eluate was diluted 1:20 with 1× Laemmli SDS sample buffer. Immunoblotting was performed using an anti-GFP mouse monoclonal antibody (JL8, TaKaRa Bio; 1:1000) and an anti-ubiquitin mouse monoclonal antibody (P4D1, Santa Cruz Biotechnology, USA; 1:1000). Protein levels in the input samples were assessed using an anti-β-actin antibody (cat. no. ab8224, Abcam; 1:2000).

The bqt3⁺ or bqt3Δ cells expressing GFP or GFP–Bqt4 were grown in YES medium to the mid-log phase (approximately 5×10⁶ cells/ml). After adding final 200 μg/ml cycloheximide (C7698, Sigma-Aldrich), the cells were cultured for 0, 5, 10, 30 and 60 min for protein degradation. Immediately after culturing, stop solution (2 mM PMSF with 5× protease inhibitor cocktail) was added to the cells, and the cells were placed on ice for 5 min. The cells were harvested by centrifugation (8000 g for 3 min) and subjected to western blotting as described above.

The membrane association assay was performed according to a previously published protocol (Habeck et al., 2015) with slight modifications. S. pombe cells expressing GFP–Bqt4 and control proteins (Ish1–GFP and GFP–ADEL) were grown to the mid-log phase and mixed at an appropriate ratio before harvesting by centrifugation at 2000 g at 4°C. The harvested cells (∼5×10⁸ cells) were washed with cold water containing 1 mM PMSF and resuspended in 1 ml cold fractionation buffer (FB; 0.7 M sorbitol, 50 mM Tris-HCl pH 7.4, 150 mM NaCl, 1× protease inhibitor cocktail, 1 mM PMSF and 1 mM BZ). Resuspended cells were lysed using a bead shocker (2700 rpm for 10 cycles of 60 s on and 60 s off; Yasui Kikai, Japan), followed by washing the beads three times with 200 μl FB. The cell extracts were combined, and cell debris and intact nuclei were removed by centrifugation (2000 g for 5 min at 4°C). The membrane fraction was isolated by ultracentrifugation at 100,000 g (in a Beckman TLA-100.4 rotor) for 1 h at 4°C. The membrane was thoroughly resuspended in FB and divided into four separate aliquots that were subjected to one of the following treatments: (1) FB alone; (2) 0.1 M Na₂CO₃, pH 11.5; (3) 2.5 M urea; or (4) 1% Triton X-100 and 0.5 M NaCl. After incubation for 1 h on ice with occasional mixing, the samples were separated into pellet and supernatant fractions by ultracentrifugation at 100,000 g for 45 min at 4°C. The pellets were washed once with appropriately supplemented FB by centrifugation (100,000 g for 45 min at 4°C) before being resuspended in FB. Proteins were then precipitated from the pellet and supernatant fractions by adding TCA to a final concentration of 20% (v/v). After 30 min of incubation on ice, the proteins were pelleted by centrifugation (16,000 g for 10 min at 4°C), washed twice with ice-cold acetone, and dissolved in 1× Laemmli SDS sample buffer. Samples were incubated for 10 min at 95°C before immunoblotting.

The yeast two-hybrid assay was performed as described in our previous paper, with slight modifications (Asakawa et al., 2022). GFP–Bqt4 fragments and Bqt3 were expressed from pGADT7 (cat. no. 630442, Clontech) and pGBKT7 (cat. no. 630443, Clontech) plasmids in the budding yeast strain (AH109). To examine protein interactions, suspensions of the transformants were spotted onto synthetic defined medium plates with dropout supplements: Leu/−Trp (630417, TaKaRa Bio) or −Ade/−His/−Leu/−Trp (630428, TaKaRa Bio). The strains used in this assay are listed in Table S1.

Cells expressing GFP–Bqt4 under the control (bqt4p) or nmt1 promoter (nmt1p) were pre-cultured in YES medium overnight. Fiv-fold dilution series of the cells (starting concentration was 8.0×10⁵ cells/ml) were spotted on EMMG medium plates supplemented with or without 10 μM thiamine, and the plates were then incubated at 30°C for 4 days. For the temperature-sensitivity growth assay, WT and Cdc48 mutant cells were spotted in a fivefold dilution series and incubated at 16, 20, 26, 30, 33 and 37°C for 3–6 days.

We thank Chizuru Ohtsuki and Yoshino Kubota for their technical assistance. We also thank Prof. Mitsuhiro Yanagida (Okinawa Institute of Science and Technology) and the National BioResource Project Yeast in Japan for providing the S. pombe strains.