F-box proteins Pof3 and Pof1 regulate Wee1 degradation and mitotic entry in fission yeast

The key mitotic inducer cyclin-dependent kinase 1 (Cdk1) controls the G₂/M transition and is highly conserved in all eukaryotic cells (Nurse, 1990). Wee1 kinase acts as a major negative regulator of Cdk1 by phosphorylating tyrosine 15 on Cdk1. At the onset of mitosis, Wee1 is inactivated by both phosphorylation and degradation in human somatic cells (Watanabe et al., 1995). Wee1 degradation has also been observed in the budding yeast Saccharomyces cerevisiae and in Xenopus (Michael and Newport, 1998; Sia et al., 1998). In the fission yeast Schizosaccharomyces pombe, Wee1 activity is downregulated mainly by protein phosphorylation, most likely by kinases Cdk1, Cdr1/Nim1 and Cdr2 (Aligue et al., 1997; Coleman et al., 1993; Guzman-Vendrell et al., 2015; Kanoh and Russell, 1998; Parker et al., 1993; Tang et al., 1993; Wu and Russell, 1993). However, the specific phosphorylation sites in fission yeast Wee1 have never been carefully mapped or fully confirmed in vivo. On the other hand, although oscillation of fission yeast Wee1 protein levels during the cell cycle and a decrease at mitosis have been observed (Aligue et al., 1997), the possible proteolytic degradation of Wee1 has not been extensively studied because of the low expression levels of Wee1.

In S. cerevisiae, Xenopus oocytes and human somatic cells, Wee1 is degraded after ubiquitylation by the SCF (Skp1/Cul1/F-box) complex at the onset of M phase (Ayad et al., 2003; Kaiser et al., 1998; Michael and Newport, 1998; Petroski and Deshaies, 2005; Watanabe et al., 2004). SCF belongs to the largest family of multicomponent E3 ubiquitin ligases and consists of four subunits: the scaffold Cul1, the RING domain protein Rbx1/Roc1/Hrt1, the adaptor Skp1 and a substrate-binding F-box protein (Petroski and Deshaies, 2005). All F-box proteins carry a loosely conserved 50 amino acid motif (known as the F-box) and are involved in binding to Skp1 and recognizing diverse targets (Bai et al., 1996; Patton et al., 1998). Studies in yeasts and mammals have revealed the existence of multiple F-box proteins in each organism, for example, 69 in human and 21 in budding yeast (Lee et al., 2011; Willems et al., 2004). The entire S. pombe genome encodes 18 F-box proteins (named Pofs and Pops) (Hermand, 2006; Lehmann et al., 2004). Currently, most of the S. pombe F-box proteins remain uncharacterized, and none of them have been assigned a function in regulating Wee1 stability, although three F-box proteins (Pop1, Pop2 and Pof3) have been shown to play crucial roles in cell cycle progression and coordination (Katayama et al., 2002; Toda et al., 1999).

So far, the best-characterized F-box proteins involved in ubiquitylation of Wee1 are human β-TrCP (BTRCP) and Tome-1 (CDCA3) when incorporated in SCF complexes (Ayad et al., 2003; Watanabe et al., 2004); Tome-1 probably has the stronger effect on promoting efficient mitotic entry (Smith et al., 2007). In the present study, we examined the potential contribution of 17 out of 18 S. pombe F-box proteins to control of Wee1 stability, using genomically tagged Wee1–Renilla luciferase as the reporter. We found that both Pof3 and Pof1 controlled Wee1 stability. We could also detect the in vivo interaction between Pof3 or Pof1 and Wee1 proteins, suggesting that redundant SCF ubiquitin ligase activity mediated by at least two F-box proteins regulates Wee1 degradation and mitotic entry in fission yeast, which is similar to Wee1 stability control in human somatic cells. Our genetic analyses demonstrated that Pof3 might play the major role in timely Wee1 degradation, similar to human Tome-1 protein.

It has been shown that the level of fission yeast wee1⁺ mRNA does not fluctuate during the cell cycle, but that the protein level of Wee1 oscillates, with its peak at S and G₂ phases and a significant drop at M phase (Aligue et al., 1997). We previously demonstrated that Wee1 protein in S. pombe is more stabilized in proteasome mutant mts3-1 or slightly stabilized in SCF component mutant skp1-A7 (Yu et al., 2013), which strongly suggests that the SCF and proteasome mediate Wee1 proteolysis. Consistently, we observed that Wee1 detected by western blotting using a newly developed antibody (anti-SpWee1) (Lucena et al., 2017) was also slightly stabilized in mts3-1 and skp1-A7 mutants (Fig. S1A). Furthermore, a characteristic smear pattern was detected in mts3-1 cells expressing 6His-myc-ubiquitin when cell extract was blotted with anti-Wee1 antibodies, demonstrating that Wee1 is indeed ubiquitylated in vivo (Fig. S1B). It is noteworthy that the abundance of ubiquitylated Wee1 was not apparently decreased in mts3-1 skp1-A7 double mutant cells (Fig. S1B, lane 3), most probably because of the low penetrance of the skp1-A7 mutation.

We then examined which F-box proteins are responsible for Wee1 recognition and subsequent proteolysis in fission yeast. We chose 17 available F-box protein deletion or temperature-sensitive mutants and introduced Renilla luciferase-tagged Wee1 and firefly luciferase-tagged Ade4 into these mutants as reporters (Fig. 1A). This strategy was employed to monitor the abundance of Wee1 protein because it has been used reliably in previous studies (Keifenheim et al., 2017; Yu et al., 2013). Strikingly, our quantitative measurements indicated that Wee1 was significantly stabilized in the pof3Δ and pof1-6 mutants, but not in the other F-box protein deletion mutants examined (Fig. 1B). Our immunoblotting results for these mutants using anti-Wee1 antibodies largely confirmed the previous data, although the stability of Wee1 protein was not significantly increased in pof1-6 or pof3Δ mutants (Fig. S1C). We postulated that the changes in Wee1 protein levels in pof1-6 and pof3Δ mutants had been minimized or lost during the relatively ‘long’ procedures of SDS-PAGE and immunoblotting, in contrast to the easy and rapid procedures of luciferase measurement. In addition, we also found that the double mutant pof3Δ pof1-6 was synthetically sick at 32°C (Fig. 1C), and that pof3Δ and pof1-6 had an additive effect on the Wee1 protein levels at both 25°C and 37°C (Fig. 1B) and on cell size at 25°C (Fig. 1D,E). By contrast, at 37°C, the small size of pof1-6 and pof3Δ pof1-6 mutant cells was completely disproportionate to their dramatic increase in Wee1 levels (Fig. 1D,E), which was most probably the result of Pof1 function(s) in processes unrelated to cell cycle.

In S. pombe, the size at which cells enter mitosis is quite sensitive to the gene dose of wee1⁺ and cdc25⁺; thus, cell length at division is commonly employed as an easy measure of mitotic timing (Nurse, 1975; Russell and Nurse, 1986, 1987). The average cell length at division in pof3Δ cells was reported to be 24.2±4.1 μm at 26°C versus around 14 μm in wild-type cells, which was interpreted as a result of activation of the DNA damage checkpoint (Katayama et al., 2002). The cell length at division in the pof3Δ wee1-50 double mutant (22.1±3.7 μm at 26°C and 15.7±2.7 μm at 36°C) prompted the authors to consider the G₂ delay as Wee1-independent (Katayama et al., 2002). Based on our result of significantly increased Wee1 in pof3Δ (Fig. 1B), we assumed that the intermediate cell length at division in the pof3Δ wee1-50 double mutant does not necessarily reflect the independence of wee1⁺. Rather, the elongated phenotype of the single pof3Δ mutant is possibly derived from a combination of checkpoint activation and increased stability of Wee1. The lack of tight coupling between Wee1 level and cell length was observed in cells expressing probably fourfold Wee1 protein levels with cell length of 28 μm at division (Russell and Nurse, 1987) and was more apparent for the pof1 mutant, which displayed acute growth arrest with small cell size (about 8.5 μm at 36°C) (Harrison et al., 2005), similar to the strong wee phenotype. Our direct measurement of Wee1 protein levels uncovered a dramatic increase in Wee1 levels in pof1-6 cells (Fig. 1B), which was completely disproportionate to its small size. Thus, the pof1-6 mutant is one extreme example in which the cell size does not tightly couple with Wee1 protein level; this is most probably a result of poor growth caused by a large number of cellular defects upon loss of pof1⁺ gene function.

Because skp1-A7 mutation also caused elevated Wee1 protein levels (Yu et al., 2013), Skp1 might cooperate with Pof1 and Pof3 in the same function. Supporting this assumption, we observed extreme synthetic growth defects in double mutants of skp1-A7 combined with pof3Δ or pof1-6, but not with the other F-box protein mutants examined (Fig. 1F and data not shown). These data indicate that at least two fission yeast F-box proteins (Pof1 and Pof3) are possibly involved in protein level control of Wee1 at mitotic entry.

As in vertebrates, the fission yeast SCF ubiquitin ligases consist of multiple conserved core complex subunits Skp1, Pcu1 (Cul1) and Rbx1. Each SCF complex also contains an F-box protein as the variable receptor subunit. Previous studies have confirmed that both Pof1 and Pof3 can bind SCF core components Skp1 and Pcu1 (Harrison et al., 2005; Katayama et al., 2002). If Wee1 were a substrate for SCF^Pof3 or SCF^Pof1 (an SCF complex containing Pof3 or Pof1, respectively, as the F-box component), then Pof3 and Pof1 would bind Wee1. To verify this, immunoprecipitation was performed in strains carrying 3HA–Wee1, Pof3–13myc or Pof1–13myc alone or doubly in the background of mts3-1 mutants, which inactivate proteasomes (Gordon et al., 1996) and should slow down the degradation of ubiquitylated Wee1. Our results confirmed the interactions between Wee1 and Pof3 or Pof1 (Fig. 2). These interactions were specific because Wee1 did not precipitate Pof10–13myc in similarly treated cells (Fig. S2), which is consistent with our finding that deletion of pof10⁺ does not affect Wee1 protein levels (Fig. 1B, Fig. S1C).

A previous study showed that overexpression of wee1⁺ is lethal (Russell and Nurse, 1987). Although pof3Δ and pof1-6 mutants are defective in downregulating Wee1 levels, their growth is not seriously inhibited at ambient temperatures (Harrison et al., 2005; Katayama et al., 2002). Therefore, we expected that further elevated Wee1 in pof3Δ and pof1-6 mutants would lead to lethality as a result of strong G₂ arrest. To test this possibility, we first constructed a series of yeast strains expressing different levels of Wee1 under promoters of various strengths, including nmt promoter variants (Basi et al., 1993; Maundrell, 1990) and constitutively expressing adh promoters, in which P_adh11, P_adh21 and P_adh81 are weaker versions of P_adh1 in the order P_adh1>P_adh11>P_adh21 >P_adh81 (Chen et al., 2017; Kawashima et al., 2007; Tada et al., 2011; Yokobayashi and Watanabe, 2005). We obtained strains carrying genomically integrated wee1⁺ driven under different promoters, except with adh11 and adh1 promoters (Fig. S3A). We assumed that constant overexpression of Wee1 under adh11 or adh1 promoters was lethal, because wee1⁺ expressed from the strongest P_nmt1 (about 20 times the Wee1 level in wild-type cells; Fig. S3D) permanently inhibited colony formation after induction and caused dramatically elongated cell morphology in liquid cultures (Fig. S3A–C). We found that increased wee1⁺ expression correlated closely with increased cell size at division (Fig. S3B,C).

We introduced P_adh21-wee1⁺ into skp1-A7, pof3Δ and pof1-6 mutant backgrounds. Interestingly, skp1-A7 P_adh21-wee1⁺ and pof3Δ P_adh21-wee1⁺ mutants were much sicker than skp1-A7, pof3Δ or P_adh21-wee1⁺ alone (Fig. 3A) and showed strengthened cell elongation phenotypes (Fig. 3C,D), but the pof1-6 mutant was less sensitive to P_adh21-wee1⁺ (Fig. 3A). However, the growth fitness of pof1-6 mutants became sensitive to P_nmt1-wee1⁺ (Fig. 3B), although its cell elongation phenotype was not as strong as in skp1-A7 or pof3Δ backgrounds (Fig. 3C,D and Fig. S4). These genetic analyses support our conclusion that SCF^Pof3 and SCF^Pof1 are responsible for cell cycle-dependent degradation of Wee1, in which SCF^Pof3 might play a major role. We assumed that the additive effect on growth of increased Wee1 levels in pof1-6 and pof3Δ mutants combined with wee1⁺ overexpression does not necessarily only result from delay in G₂/M transition. It is fairly possible that other defects caused by pof1-6 or pof3Δ mutations contribute to a reduction in growth rate; the detailed mechanisms definitely require further investigation.

It is well accepted that Wee1 acts antagonistically to Cdc25 in activation of Cdc2; thus, cells defective in Cdc25 function should also be sensitive to even slight overproduction of Wee1. We confirmed this idea by crossing cdc25-22 with either P_adh81-wee1⁺ or P_adh21-wee1⁺, which resulted in the synthetic sick or lethal phenotype, respectively (Fig. 4A). Interestingly, skp1-A7 and pof3Δ combined with the cdc25-22 mutation were also much sicker, whereas the pof1-6 mutant was not so sensitive to the presence of the cdc25-22 mutation (Fig. 4B).

In conclusion, it is very likely that the mechanism whereby redundant SCF ubiquitin ligase activities regulate Wee1 levels through incorporating multiple F-box proteins is conserved in yeasts and humans, and even throughout eukaryotes. Supporting this idea, it has been shown that the human F-box protein Tome-1 has a stronger effect on mitotic entry than β-TrCP (Ayad et al., 2003; Smith et al., 2007). Similarly, we have provided genetic evidence that Pof3 is more important than the fission yeast β-TrCP homologue Pof1 in degrading Wee1. Although human Tome-1 and fission yeast Pof3 lack apparent sequence similarity or homology (C.Q. and Q.-w.J., unpublished observation), the possibility that Pof3 and Tome-1 are functionally similar and related is very interesting.

Escherichia coli host strain DH5α was used to propagate plasmids in standard lysogeny broth (LB). Standard molecular biology methods were employed, and enzymes were used as recommended by the suppliers (New England Biolabs, Fermentas, Transgen Biotech and TIANGEN Biotech).

Schizosaccharomyces pombe strains used and created in this study are listed in Table S1. Standard fission yeast techniques were used throughout for cell culture and genetic crosses (Forsburg and Rhind, 2006). Cells were grown in supplemented yeast extract (YE5S) or Pombe glutamate (PMG) medium containing 4 g/l sodium glutamate as nitrogen source. Appropriate nutritional supplements were used at 175 mg/l. Thiamine (5 µg/ml) was used to repress expression from the nmt promoters. G418 disulfate (Sigma-Aldrich), hygromycin B (Sangon Biotech) or nourseothiricin (clonNAT; Werner BioAgents) was used at a final concentration of 100 μg/ml where appropriate. Serial-dilution drop tests for growth were performed as previously described (Yu et al., 2013).

To generate strains expressing wee1⁺ under nmt promoters at the leu1⁺ locus, the coding sequence of wee1⁺ was amplified from yeast genomic DNA using the polymerase chain reaction (PCR) and subcloned into the vector pJK148-P_nmt1::leu1⁺ (Chen et al., 2017) using SalI/BamHI sites, which resulted in the construction of the final vector pJK148-P_nmt1-wee1⁺::leu1⁺. The plasmids of pJK148-P_nmt41-wee1⁺::leu1⁺ and pJK148-P_nmt41-wee1⁺::leu1⁺ with weaker nmt promoters were constructed by Quikgene site-directed mutagenesis as previously described (Chen et al., 2017).

To construct strains expressing wee1⁺ under different strength adh promoters at the lys1⁺ locus, the PCR-amplified coding sequence of wee1⁺ was subcloned into a series of pUC119-based vectors carrying sequences corresponding to different strength adh promoters (Chen et al., 2017) using SalI/BamHI sites. This procedure generated four vectors: pUC119-P_adh81-wee1⁺::hphMX6::lys1*, pUC119-P_adh21-wee1⁺::hphMX6::lys1*, pUC119-P_adh11-wee1⁺::hphMX6::lys1* and pUC119-P_adh1-wee1⁺::hphMX6::lys1*.

Wild-type ubiquitin fused with both 6His and myc (i.e. 6×His-myc-ubiquitin) was expressed under the control of nmt1 promoter (P_nmt1) at the leu1⁺ locus. To construct the strains, a pREP1-6His-myc-ubiquitin vector was originally constructed in which a sequence corresponding to myc (AACGGTGAACAAAAGCTAATCTCCGAGGAAGACTTGGGATCC) was inserted into pREP1-6His-ubiquitin (a kind gift from Kathy Gould) behind 6His using Quikgene site-directed mutagenesis (Mao et al., 2011). Then, the PstI-SacI fragment containing nmt1 promoter (P_nmt1), 6His-myc-ubiquitin and nmt terminator (T_nmt) sequences from the above vector was cloned into the pJK148 vector (Keeney and Boeke, 1994) using PstI/SacI sites. This process resulted in the construction of pJK148-P_nmt1-6His-myc-ubiquitin::leu1⁺.

For integration of wee1⁺- or 6His-myc-ubiquitin-expressing vectors at either leu1-32 (for pJK148-based vectors) or lys1⁺ (for pUC119-based vectors) sites in the genome, each plasmid DNA (∼1 µg) was digested and linearized with NruI (for pJK148-based vectors) or ApaI (for pUC119-based vectors) before being transformed into yeast using a lithium acetate method (Bähler et al., 1998; Keeney and Boeke, 1994). Transformants were selected on YE5S plates containing hygromycin B (for pUC119-based vectors), or PMG plates plus thiamine but lacking leucine (for pJK148-based vectors). Correct integration was verified by a colony PCR method.

Yeast strains carrying both ade4-firefly luciferase::kanMX6 and wee1-Renilla luciferase::natMX6 were used. Exponential phase cultures were pre-grown in liquid YE5S at 30°C (for wild-type and most mutants), 25°C or shifted from 25°C to 37°C for 4 h (for skp1-A7 and pof1-6 mutants). Cell lysates were prepared and luciferase activities were measured using the Dual-Reporter Assay System (Promega) as described previously (Yu et al., 2013).

Samples were collected and processed for western blotting as described previously (Lucena et al., 2017). For detection, rabbit polyclonal anti-Wee1 (Lucena et al., 2017) was used as the primary antibody (1:1000 dilution). Cdc2 was detected using rabbit polyclonal anti-PSTAIRE (Santa Cruz Biotechnology, sc-53) as loading control (1:1000 dilution). For co-immunoprecipitation experiments, whole-cell lysates were prepared and subjected to immunoprecipitation with rat anti-hemagglutinin (HA) monoclonal antibodies (clone 3F10; Roche, Mannheim, Germany) as described previously (Li et al., 2010).

To measure cell length at division, the yeast strains were grown either in PMG medium with or without thiamine or in YE5S medium to an A₆₀₀ of <0.5. Cells with clear septa under a DIC microscope were selected for measurement. A minimum of 100 septated cells were scored for each mutant. Photomicrographs were obtained using a Nikon 80i fluorescence microscope coupled to a cooled CCD camera (Hamamatsu, ORCA-ER). Image processing and analysis was carried out using Element software (Nikon) and Adobe Photoshop.

All experiments were repeated at least three times with similar results. To determine the statistical significance of our data, two-tailed Student's t-tests were performed; P<0.05 was considered statistically significant.

We thank Drs Kathy Gould, Li-lin Du, Takashi Toda, Dieter Wolf and Yoshinori Watanabe for providing the yeast strains and plasmids. We are grateful to Dr Nicholas Rhind for his advice on some experiments.