The fission yeast interphase spindle pole body (SPB) is a bipartite structure in which a bulky cytoplasmic domain is separated from a nuclear component by the nuclear envelope. During mitosis, the SPB is incorporated into a fenestra that forms within the envelope during mitotic commitment. Closure of this fenestra during anaphase B/mitotic exit returns the cytoplasmic component to the cytoplasmic face of an intact interphase nuclear envelope. Here we show that Brr6 is transiently recruited to SPBs at both SPB insertion and extrusion. Brr6 is required for both SPB insertion and nuclear envelope integrity during anaphase B/mitotic exit. Genetic interactions with apq12 and defective sterol assimilation suggest that Brr6 may alter envelope composition at SPBs to promote SPB insertion and extrusion. The restriction of the Brr6 domain to eukaryotes that use a polar fenestra in an otherwise closed mitosis suggests a conserved role in fenestration to enable a single microtubule organizing center to nucleate both cytoplasmic and nuclear microtubules on opposing sides of the nuclear envelope.
Many eukaryotes rely upon two microtubule organizing centers (MTOCs) to nucleate the antiparallel microtubule arrays of the mitotic spindle. Despite functional conservation, structure can vary dramatically (Heath, 1980). This structural variation is often accompanied by significant variation in the behavior of the nuclear envelope during mitosis (Kubai, 1976; Heath, 1980). In most higher eukaryotes, the nuclear envelope fragments to enable radial microtubule arrays to capture one set of chromosomes. In syncytial systems, complete nuclear envelope breakdown could be catastrophic, as it could facilitate chromosome exchange between neighboring spindles. A partial or complete membrane barrier is therefore retained in many of these systems. The evolutionary paths taken by fungi and protists means that conservation of nuclear envelope integrity throughout mitosis in a “closed mitosis” is a common feature of microbial cell division (Kubai, 1976; Heath, 1980).
The permanent separation of cytoplasm and nucleoplasm in closed mitoses presents significant challenges if cells possess an MTOC that executes cytoplasmic functions alongside genome segregation. Many fungi, including the yeasts Schizosaccharomyces pombe and Saccharomyces cerevisiae, solve this problem by integrating planar MTOCs into the nuclear envelope so that the microtubules nucleated from the two faces can function in different compartments (McCully and Robinow, 1971; Byers and Goetsch, 1975; Heath, 1980; Ding et al., 1997). In budding yeast spindle pole body (SPB), duplication is coupled with integration of the new SPB into the membrane alongside the old. Both SPBs then reside within pores in the nuclear envelope for the remainder of the cell cycle (Byers and Goetsch, 1975; Adams and Kilmartin, 2000). In contrast, integration of the fission yeast SPB is transient and restricted to mitosis (McCully and Robinow, 1971; Ding et al., 1997).
Electron microscopy of budding yeast SPBs in G1 phase of the cell cycle reveals a triple layered structure with a “half bridge” that extends from the central plaque along the cytoplasmic face of the nuclear envelope and ends in an amorphous “satellite” structure (Byers and Goetsch, 1975; Adams and Kilmartin, 2000). The new SPB forms by elaboration of the satellite to form a duplication plaque on the cytoplasmic side of the nuclear envelope. The completely formed duplication plaque is then inserted into a pore in the nuclear envelope next to the integrated old SPB. Defects in the half bridge components Kar1, Cdc31, Mps3, and Sfi1 block half bridge formation, and hence duplication plaque formation (Adams and Kilmartin, 2000; Kilmartin, 2003; Jaspersen et al., 2006). Mps3 contains a Sad1/Unc84 (SUN) domain (Jaspersen et al., 2006). As SUN domains cooperate with Klarsicht, ANC-1, and Syne homology (KASH) domain proteins to establish connections between nuclear and cytoplasmic molecules across the nuclear envelope (Razafsky and Hodzic, 2009), Mps3 may act alongside the membrane protein Kar1 to tether half bridge components to the nuclear envelope. The membrane protein Mps2 forms a complex with Nbp1 and the Kar1-interacting protein Bbp1 that sits at the central plaque/half bridge (Schramm et al., 2000; Araki et al., 2006). This complex cooperates with the integral membrane protein Ndc1 to promote the integration of the duplication plaque into the nuclear envelope (Araki et al., 2006).
The fission yeast interphase SPB comprises a cytoplasmic component that is separated from a nuclear component by the nuclear envelope (McCully and Robinow, 1971; Ding et al., 1997). A structure reminiscent of the S. cerevisiae half bridge/bridge extends from the cytoplasmic component over the surface of the nuclear envelope. Fine striations through the envelope connect this cytoplasmic component to a nuclear component that contains γ-tubulin and recruits centromeres to the SPB (Funabiki et al., 1993; Ding et al., 1997; Kniola et al., 2001). The SUN domain protein Sad1 and the KASH proteins Kms1 and Kms2 appear to mediate the association of centromeres with the SPB and thus the cytoplasmic microtubules (Goto et al., 2001; King et al., 2008). The ability to differentiate between the old and new SPBs with a time-sensitive fluorescent protein suggest that SPB duplication is conservative, with a new SPB forming de novo alongside the old (Grallert et al., 2004). Upon commitment to mitosis, the membrane separating the two SPB components disperses and the old and new SPBs insert into the resulting fenestra (Ding et al., 1997). Such polar fenestration is a feature shared by a subset of eukaryotes that use closed mitosis (Kubai, 1976; Heath, 1980). The local removal of the nucleoplasm/cytoplasm barrier in S. pombe is transient and rapid, as there is no detectable leakage of the nucleoplasm into the cytoplasm (Tallada et al., 2009). The two SPBs then become active, each nucleating microtubules to generate each half of the spindle (Ding et al., 1993). During anaphase B, the membrane grows back between the two components to recreate the interphase partitioning of SPB components (Ding et al., 1997).
Although less is known about SPB integration in fission yeast, Sad1 is a clear homologue of Mps3, Alm1, Mlp2, and Cdc31, and Sfi1 bridge component orthologues are also required for spindle formation (Hagan and Yanagida, 1995; Flory et al., 2002; Kilmartin, 2003; Paoletti et al., 2003; Niepel et al., 2005; Jaspersen et al., 2006). The membrane-spanning Ndc1 orthologue Cut11 is recruited to the SPB during mitosis, where it is required for the integration of the new SPB into the nuclear envelope (West et al., 1998; Tallada et al., 2009).
There are striking parallels between the integration of SPBs and nuclear pore complexes (NPCs) into the nuclear envelope such that the two systems appear to compete for assembly factors (Witkin et al., 2010). Mps2 and Mps3 associate with the nuclear periphery in addition to the SPB, Cdc31 modulates mRNA export, and the nuclear pore complex (NPC) component Mlp2 binds to SPBs and participates in SPB assembly (Muñoz-Centeno et al., 1999; Jaspersen et al., 2002; Fischer et al., 2004; Niepel et al., 2005). Most notably, Ndc1 is a component of both NPCs and SPBs (Chial et al., 1998). In NPCs, Ndc1 interacts with Pom34 and Pom152 in the membrane ring surrounding NPCs, where these molecules perform redundant functions in NPC assembly (Wozniak et al., 1994; Chial et al., 1998; Miao et al., 2006; Alber et al., 2007; Onischenko et al., 2009). Importantly, reducing translation of Pom34 mimics mutation of Ndc1 in rescuing the otherwise lethal SPB insertion arising from removal of Mps2 (Sezen et al., 2009).
We describe the recruitment of fission yeast Brr6 (Saitoh et al., 2005; Lo Presti et al., 2007) to SPBs during mitotic commitment and anaphase B/mitotic exit. We show that it is required for SPB insertion and envelope integrity during SPB extrusion. Phenotypic and genetic interactions with apq12 that mirror those reported for Saccharomyces cerevisiae (Hodge et al., 2010) suggest that it modifies the lipid composition of the envelope to facilitate SPB insertion/extrusion. We discuss the evolutionary restriction of the “Brr6 domain” to a specific subset of fungi and protists that exploit a polar fenestra to enable an MTOC to organize both cytoplasmic and spindle microtubules (Kubai, 1976; Heath, 1980).
S. pombe Brr6 is required for spindle formation
Screening for mutations that formed monopolar spindles upon exposure to 4% DMSO for one generation (Broek et al., 1991; Poloni and Simanis, 2002) identified a glycine–to–aspartic acid change at position 145 of brr6+ (brr6.ds1; Fig. 1, A and B; for comparative images of wild-type cells, see Fig. S1 A and Fig. 2 A). brr6+ is an essential gene (Lo Presti et al., 2007). Inoculating spores isolated from a diploid strain in which the brr6+ ORF had been replaced with the kanR marker into medium containing geneticin to select against growth of brr6+ spores revealed monopolar rather than bipolar spindles in their first division of 88% of mitoses (unpublished data). Repressing brr6+ transcription in a haploid brr6.Δ background also blocked mitotic progression with monopolar spindles (Fig. S2, B and C).
brr6.ds1 was not a penetrant allele (Fig. 1 B). We therefore generated tighter temperature-sensitive mutations, the tightest of which, brr6.ts8, gave no binucleate cells after a shift to the restrictive temperature from 25°C to 36°C (unpublished data). brr6.ts8 behaved as a recessive mutation in a heterozygous diploid (Fig. S2 D). Monitoring spindle formation in an asynchronous brr6.ts8 culture shifted from 25°C to 36°C by immunofluorescence established that, after a wave of monopolar spindles, cells with condensed chromosomes devoid of any microtubule staining appeared 40 min after shift to the restrictive temperature (Fig. 1, C–E). To ensure that the lack of tubulin staining within these cells was not an artifact of defective processing, we mixed brr6.ts8 cells that had been held at 36°C for 2 h with spg1.B8 cdc7.A20 cells that had been incubated at 36°C for 4 h, and processed the mixed culture for immunofluorescence microscopy. Microtubules were present in the long multinucleated spg1.B8 cdc7.A20 cells (Fig. 1 E, asterisks) but not in the shorter mononucleated brr6.ts8 cells within the same sample (Fig. 1 E, arrows), which indicates that brr6.ts8 cells were unable to maintain a mitotic spindle after prolonged incubation at the restrictive temperature.
Compromising Brr6 blocks SPB insertion into the nuclear envelope
To determine the basis for the brr6− monopolar spindle phenotype, brr6.ts8 cells were processed for electron microscopy. Size-selected G2 cells were used for the shift to 36°C to maximize the frequency of mitotic cells when sampling took place between 110 and 150 min later. Serial sections indicated that SPB duplication was normal; however, the cytoplasmic component of one SPB failed to insert into the nuclear envelope and appeared to drift away from it (Fig. 3 and Fig. S3, for comparative images of an inserted pro-metaphase wild-type SPB, see Fig. S1 B). There was a striking concordance in the distance between the nonfunctional SPB and the nuclear envelope in all 12 cells examined (thick brackets in Fig. 3, B and J; and Fig. S3, D and E). Furthermore, some residual osmophilic material persisted on the outer face of the nuclear envelope (outer NE) below this SPB in the same, or subsequent section (thin brackets Fig. 3, C and I; and Fig. S3, C and O). Thus, this nonfunctional SPB may retain some connection with the nuclear envelope. In all cases, the second SPB inserted into the nuclear envelope and nucleated an extensive microtubule array; however, the insertion of this active SPB was defective, as it was associated with a large hole in the nuclear envelope in 11 out of 12 complete sets of serial sections of brr6.ts8 cells in which both SPBs could clearly be identified (the exception is shown in Fig. S3, M–Q). In contrast, the nuclear envelope was intact in all 15 mitotic wild-type control cells (Fig. S1 B and not depicted).
The new SPB fails to insert into the envelope of brr6.ts8 cells at 36°C
As the old and new SPBs differ in their potential to recruit elements of the septum initiation network (SIN) and nucleate microtubules in cut12.1 and cut11.1 mutations (Bridge et al., 1998; Sohrmann et al., 1998; West et al., 1998; Grallert et al., 2004; Tallada et al., 2009), we asked whether SPB age accounted for the heterogeneity in SPB functions during insertion in brr6 mutants. pcp1.RFP atb2.GFP brr6.ts8 and pcp1.RFP atb2.GFP brr6+ cells were isolated by filtration and resuspended in medium lacking nitrogen at 25°C to arrest cell cycle progression in G1 for 16 h. This arrest gave sufficient time for the slow folding RFP to fold into an actively fluorescent configuration. Addition of a nitrogen source as the temperature was increased to 36°C induced cell cycle reentry whereupon a new SPB was synthesized. The RFP within the new SPB did not have time to fold into a fluorescent configuration before division so that the old SPB, but not the new one, emitted a red signal (Fig. S1 C, i). The red signal, in all 17 brr6.ts8 cells in which a single spot of Pcp1.RFP fluorescence could be seen, presided at the base of a microtubule array (Fig. S1 C, ii–iv), which indicates that it was the old SPB that was active.
Brr6 is not required for spindle formation after SPB insertion
To ask whether Brr6 was required for spindle function after the insertion of the SPB into the nuclear envelope, we used the cold-sensitive nda3.KM311 β-tubulin mutation. Incubation of nda3.KM311 cells at 20°C arrests mitotic progression in prophase after the SPBs have inserted into the nuclear envelope (Kanbe et al., 1990), after which a temperature shift between 30°C and 37°C restores tubulin function, generating a synchronized mitosis (Uemura et al., 1987). For this approach to really test Brr6 function, we had to be confident that Brr6.ts8 protein was inactivated immediately upon temperature shift of brr6.ts8. The stress response invoked when a brr6.ts8 culture is shifted from 25°C to 36°C transiently blocks mitotic commitment at precisely the point at which we wish to study Brr6 requirement in the nda3.KM311 experiment (Fig. 1 D). We therefore monitored spindle integrity as wee1.50 brr6.ts8 cells were shifted from 25°C to 36°C because Wee1 inactivation overrides this stress response, driving all cells into mitosis (Nurse, 1975). All spindles formed over the 20 min after the heat shift were monopolar (Fig. 4 A), which indicates that Brr6.ts8 function was indeed immediately inactivated by the temperature shift. We therefore shifted brr6.ts8 nda3.KM311 cells from 20°C to 36°C and observed normal segregation of chromosomes on bipolar spindles (Fig. 4 B). Furthermore, cell cycle progression through anaphase B was unaffected, as the recruitment of SIN kinase Cdc7 to the new SPB in anaphase B (Grallert et al., 2004; Sohrmann et al., 1998) was not perturbed (Fig. 4 C). We conclude that Brr6 is not required for chromosome segregation after the SPB has inserted into the nuclear envelope.
Extensive genetic interactions between brr6 and genes encoding SPB components
brr6.ds1 and brr6.ts8 mutants exhibited significant genetic interactions with mutations in SPB components (Fig. S4 A). The inclusion of brr6.ts8 blocked growth of sad1.2, cut12.1, and kms2.Δ mutants at temperatures at which either single mutant was viable. In contrast, perturbation of Brr6 function with brr6.ts8 (or brr6.ds1; unpublished data) reduced the requirement for the Ndc1 homologue Cut11. Conversely, the inclusion of a temperature-sensitive allele of cdc31+, cdc31.1, or deletion of kms1+ alleviated the cold sensitivity of the brr6.ds1 mutation. Monitoring spindle integrity in cut11.1 brr6.ts8 mutants 3 h after a shift to the semipermissive temperatures of 28°C and 30°C confirmed the results of the spot test analysis, which showed that compromised Brr6 function alleviates the impact of defective Cut11 function (Fig. S4 B). The recruitment of Cut12, Sad1, or Cut11 to SPBs was unaffected by the brr6.ts8 mutation (Fig. S4 C; Hagan and Yanagida, 1995; Bridge et al., 1998; West et al., 1998).
Brr6 is recruited to SPBs during SPB insertion and extrusion
Brr6 is a member of a protein family that contains the Pfam motif PF10104 (Finn et al., 2008; http://pfam.sanger.ac.uk//family/PF10104#tabview=tab6; Brr6_like_C_C), which has members in fungi and protists, but not higher eukaryotes (Lo Presti et al., 2007). Brr6 family members contain two regions predicted to span membranes that are separated by a conserved region that contains, among other conserved residues, four cysteine residues (Fig. 2; brr6.ts8 is a mutation of the conserved cysteine at 192 to glycine [Fig. 2 A]). Immunofluorescence with antibodies to the nonconserved domains of Brr6 (Fig. 5 A) revealed a signal around the nuclear periphery throughout the cell cycle and an association with the mitotic SPB (Fig. 5, C–E; not depicted). The SPB fluorescence was initially a single dot, indicating that Brr6 was recruited at the earliest stages of insertion of the adjacent two SPBs (Ding et al., 1997). This dot staining broadened to give two foci that were barely resolved from one another and that often resided within the circle defined by the staining at the nuclear periphery, after which the SPB staining was at the nuclear periphery (Fig. 5, C and E). The signal persisted at both SPBs until midway through anaphase, whereupon it is diminished asymmetrically. Labeling the new SPB with Cdc7.HA (Sohrmann et al., 1998; Grallert et al., 2004) established that the loss of Brr6 from the SPB was stochastic and did not correlate with SPB age (Fig. 6 A). There was a striking variation in the intensity of SPB staining. Early mitotic and mid-anaphase SPBs had stronger signals than metaphase or late mitotic cells (Fig. 5, C and D; cells are numbered to indicate descending intensity levels).
Although we were unable to tag the carboxyl terminus of Brr6 with GFP (unpublished data), fusion of the EGFP sequences to the amino terminus generated a functional fusion protein that showed very weak association with cytoplasmic membranes, strong enrichment around the nuclear periphery, and very strong associations with mitotic SPBs. Fluctuations in the intensity of this SPB association of Brr6.NEGFP confirmed that SPB recruitment of Brr6 was transient, first during mitotic commitment and then during anaphase B/mitotic exit (Fig. 6 B). The lack of recruitment between these two phases was graphically illustrated in the greatly extended anaphase B that follows transient arrest at the G2/M boundary of cdc25.22 cells (Fig. 6 C; Hagan et al., 1990). Thus, at times when the SPB enters or leaves the nuclear envelope, Brr6 association with the SPB increased. Western blot analysis of extracts from cells synchronized with respect to cell cycle progression by centrifugal elutriation with anti-Brr6 antibodies established that the changes in Brr6 localization were not accompanied by marked alteration of Brr6 levels or migration in SDS-PAGE (Fig. 5 B).
The re-recruitment of Brr6 to SPBs in anaphase B suggested that it may play a role in coordinating membrane dynamics with SPB extrusion. We therefore addressed the impact of brr6.ts8 upon nuclear integrity with a GFP–β-Gal NLS fusion protein (Yoshida and Sazer, 2004). As this NLS-tagged fusion protein is constantly imported into the nucleus, cytoplasmic fluorescence is negligible throughout the closed mitosis of S. pombe (Fig. 7, A and C; and Videos 1 and 3). Any breach of nuclear integrity at any point in the cell cycle is therefore easily detected as the fluorescent signal disperses throughout the cell (Yoshida and Sazer 2004; Gonzalez et al., 2009; Tallada et al., 2009). As expected from the electron microscopic analysis of the brr6.ts8 phenotype (Figs. 3 and S3), the GFP signal flooded out of brr6.ts8 nuclei, as nuclear integrity was compromised during mitotic commitment (Fig. 7 B, asterisks; and Video 2).
To address the integrity of the nuclear envelope during anaphase as the SPB is extruded from the nuclear envelope once more, we introduced the GFP–β-Gal NLS fusion protein into the nda3.KM311 arrest release approach (Fig. 4) to monitor nuclear envelope integrity during anaphase, when Brr6 function is compromised by temperature inactivation of the brr6.ts8 gene product. As before, cells were arrested in prophase at 20°C before warming to 36°C to inactivate Brr6 function. Although GFP–β-Gal NLS was retained within the envelope of control nda3.KM311 brr6+ cells (Fig. 7 C and Video 3), it transiently flooded out of nda3.KM311 brr6.ts8 nuclei during anaphase B/mitotic exit, which indicates that the integrity of the nuclear envelope at this time was compromised by the brr6.ts8 mutation (Fig. 7 D, asterisks; and Video 4). If this disruption of envelope integrity arose from errors in SPB/membrane topology during extrusion, the fidelity of SPB insertion in the subsequent mitosis could be compromised, even when Brr6 function is restored for the duration of the intervening G2 phase. We therefore monitored cell viability and spindle integrity in brr6+ nda3.KM311 and brr6.ts8 nda3.KM311 cells after transient incubation at 36°C (50 min) before a return to the brr6.ts8 permissive temperature of 26°C (Fig. 8 A). Control cultures were warmed directly from 19°C to 26°C (Fig. 8 B). The colony-forming ability of the brr6.ts8 mutant cells was reduced if they underwent division at 36°C, but not if they did not experience Brr6 inactivation during mitosis (Fig. 8 C). Importantly, mitotic defects were apparent in the subsequent mitosis at the permissive temperature, when brr6.ts8 prophase cells had been transiently raised to 36°C to undergo mitosis, and at the restrictive temperature before being returned to the permissive temperature for the duration of interphase (Fig. 8 D). In contrast, no mitotic defects were seen if cells had not undergone mitosis at the restrictive temperature, but merely shifted back to the permissive temperature of 26°C immediately after the prophase arrest (Fig. 8 E). These data suggest that recruitment of Brr6 to anaphase B SPBs facilitates the resolution of SPB/nuclear envelope topology during SPB extrusion.
Brr6 does not associate with nuclear pores
The association of Brr6 with the nuclear periphery throughout the cell cycle prompted us to ask whether this distribution reflected an association of Brr6 with nuclear pores; however, we did not detect a marked registration between the signals for Brr6.NEGFP and the NPC marker Nup107.cherry (Fig. 9, A and B). Furthermore, imaging either the Nup107.GFP or a Cut11.GFP signal in mixed brr6+/brr6.ts8 cultures at the restrictive temperature did not reveal any alterations in the distribution or quantity of nuclear pores in S. pombe brr6.ts8 mutants after 1 h at the restrictive temperature (Fig. 9, C and D).
Cooperation between Apq12 and Brr6
apq12 + was isolated as a multicopy suppressor of brr6.ds1 (Fig. S5 A). The intimate genetic and functional relationship between budding yeast Apq12 and Brr6 (Hodge et al., 2010) prompted us to assess the relationship between the fission yeast homologues. Although apq12+ is not an essential gene (Decottignies et al., 2003), the addition of 4% DMSO to apq12.Δ cells did compromise spindle formation (Fig. S5 B), which suggests that it may indeed act in partnership with Brr6. This view was consolidated by the synthetic lethality arising from combining apq12.Δ cells with either brr6.ds1 or brr6.ts8 (Fig. S5 C and not depicted), and the striking concordance of the genetic relationships between apq12.Δ and mutations in SPB components and those displayed between the same SPB mutants and brr6.ts8/brr6.ds1 (Fig. S5 D and not depicted). Apq12.tdTom did not associate with SPBs at any point in the cell cycle but exhibited a more general membrane reminiscent of the distribution of the endoplasmic reticulum (Fig. S5 E; Broughton et al., 1997).
The sensitivity of S. cerevisiae brr6 and apq12 mutants to reagents that increase membrane fluidity led to the conclusion that they modulate lipid homeostasis to facilitate the assembly of the outer components of the nuclear pore on the nuclear envelope (Hodge et al., 2010). Strikingly, S. pombe brr6 and apq12 mutants also had a marked impact upon the tolerance of these agents (Fig. S5 F). Moreover, mutation of either brr6 or apq12 compromised the incorporation of NBD cholesterol into membranes (Fig. 7 E), which is indicative of an impact upon sterol metabolism.
We show that Brr6 is required for the correct insertion of the fission yeast SPB during mitotic commitment. It is also required for nuclear envelope integrity during anaphase B/mitotic exit. This latter requirement coincides with the time at which the interphase separation of the nuclear and cytoplasmic SPB components is reestablished (Ding et al., 1997). Consistently, Brr6 is transiently recruited to the SPB at mitotic commitment and again during anaphase B.
Ablation of Brr6 function with the brr6.ts8 mutation blocked the dispersal of the nuclear envelope at the new SPB during mitotic commitment (Fig. 10). The retention of osmophilic material on the outer face of the envelope beneath the inactive cytoplasmic SPB component and the consistency in the distance separating it from the envelope both suggest that the new SPB retains a degree of attachment to the nuclear envelope. Integration of the active, old, SPB into the envelope was accompanied by the formation of a large hole next to it, indicating that this integration event was also defective. These data suggest that, unlike Ndc1/Cut11 that anchors the SPB within the nuclear envelope (West et al., 1998), Brr6 coordinates the SPB insertion cycle with the generation of the pore within the envelope.
The loss of nuclear integrity during mitotic commitment in cut11 mutants could indicate that Cut11 anchorage must be tightly coordinated with the Brr6-driven integration. However, nuclear integrity is also compromised during mitotic commitment of cut12.1 mutants. In contrast to Cut11 and Brr6, the sequence and relationship between cut12 and cdc25 (for example, the ability to suppress the nuclear integrity defect by elevation of Cdc25 levels) suggest that Cut12 regulates SPB activation and mitotic commitment rather than providing a physical interface with the envelope (Tallada et al., 2009). Thus, loss of envelope integrity may reflect a generic “Achilles heel” of spindle formation in fission yeast that arises from the need to transiently generate a hole large enough to accept the SPBs. In this context, it is notable that compromised nuclear envelope integrity has not been reported for Ndc1, or any budding yeast SPB insertion mutants (Adams and Kilmartin, 2000). Thus, the route of wholesale fenestration to accept both SPBs simultaneously may place more of a strain upon a system than one in which a second SPB is “pulled” into the envelope via interactions with its embedded sister.
The failure to integrate the new SPB while an adjacent, active, old SPB borders a hole in the envelope also mirrors the consequences of mutating the core SPB component Cut12 and the Ndc1 orthologue Cut11 (West et al., 1998; Tallada et al., 2009). As Cut11 and Brr6 recruitment to the SPB are both confined to mitosis, the functional asymmetry between the SPBs most likely reflects an inherent difference between the two SPBs by themselves rather than a differential impact of the mutation in these molecules upon SPB structure. This relationship between SPB age and nuclear envelope dynamics extends to more global organization. Reducing membrane incorporation into the mitotic envelope results in asymmetric nuclear separation; a larger nucleus separates from a smaller one (Gonzalez et al., 2009). Fascinatingly, the large nucleus is associated with the old SPB when the membrane deficiency arises from mutation of the RAN pathway; however, there is no correlation with SPB age when the membrane deficiency arises from a simple reduction in membrane biogenesis (Gonzalez et al., 2009). Furthermore, SPB age dictates function at other cell cycle stages; the molecules of the SIN are recruited to the new SPB in anaphase B, Kms2 associates with one of the two interphase SPBs in ima1.Δ cells, and only one of the two SPBs expands after overproduction of the MPF inhibitor Rum1 (Sohrmann et al., 1998; Grallert et al., 2004; Uzawa et al., 2004; King et al., 2008). Thus, there are many inherent functional differences between the old and new SPBs; the old SPB appears to be more robust, as it tolerates deficiencies in constituent or interacting molecules that render the new SPB defective. The molecular basis for these distinctions between the old and new SPBs remains to be established.
Fission yeast Brr6 is named after the budding yeast homologue BRR6 (Bad Response to Refrigeration). BRR6 and BRL1 (BRR6 related 1) are essential components of the nuclear periphery that are required for efficient nuclear transport (de Bruyn Kops and Guthrie, 2001; Saitoh et al., 2005). Loss of either can be complemented by S. pombe or Pneumocystis carinii brr6+ (Saitoh et al., 2005; Lo Presti et al., 2007). BRL1 and BRR6 are thought to have arisen from a genome duplication event that is specific to the Saccharomycotina subphylum of the Ascomycota. As brr6+ genes outside this subphylum are represented once in their genomes, it has been proposed that Brr6 function was originally executed by a single isoform in budding yeast ancestors before the duplication event (Lo Presti et al., 2007).
S. cerevisiae Brl1 was identified as a multicopy suppressor of a mutation in the XPO1/exportin gene CRM1 (Saitoh et al., 2005). brl1 mutants are compromised for protein export, and Brl1 associates with the NPC outer ring component Nup145 (Alber et al., 2007) in two-hybrid assays (Ito et al., 2001). However, Brl1 does not behave as an integral NPC component; it is not found in NPC complexes (Alber et al., 2007), and no Brl1/Brr6 homologues exist in higher eukaryotes whose NPCs mirror yeast NPCs in composition and structure (Lo Presti et al., 2007). Furthermore, its distribution around the budding yeast nuclear periphery is not altered by the nup133.Δ mutation that clusters nuclear pores (Saitoh et al., 2005). The transport defects exhibited by brl1 mutants may therefore arise from altered nuclear envelope organization, as NPC distribution is disrupted in such cells (Saitoh et al., 2005). Similarly, although S. pombe Brr6 neither associates with NPCs nor impacts upon NPC number or distribution, defects in the ability to sequester a β-Gal NLS fusion protein indicate that S. pombe Brr6 function does have an impact on nuclear envelope physiology (unpublished data).
Altered assimilation of sterols and sensitivity to agents that alter membrane fluidity in both S. pombe and S. cerevisiae brr6 mutants suggests that a principle function for Brr6-related molecules is to alter membrane composition. In the case of fission yeast Brr6, its recruitment to SPBs during SPB integration and extrusion suggests that the physical association of Brr6 with the SPB at this time facilitates changes in SPB association with the envelope because it physically alters the lipid composition of the nuclear envelope in the neighborhood of the SPB at these times. Strikingly, defects in both fission and budding yeast affect events on the cytoplasmic face of the nuclear envelope (new SPB integration in S. pombe and assembly of NPC components in S. cerevisiae; Hodge et al., 2010), raising the question as to whether Brr6-related molecules specifically modify this outer membrane?
The membrane-spanning Pfam motif PF10104, which defines a Brr6 molecule, is found in fungi and protists that generate a polar fenestra in the nuclear envelope to enable an MTOC to organize both cytoplasmic and nuclear microtubules within a closed mitosis (Heath, 1980); e.g., Dictyostelium discoideum (Moens, 1976). We believe that it is highly significant that PF10104 is not found in microbes that do not need a fenestra to form within the envelope, either because their anastral spindles are organized by an intranuclear MTOC as in Trypanosoma brucei or no MTOC at all, as in Tetrahymena (Fujiu and Numata, 1999; Ogbadoyi et al., 2000). This striking distribution of PF10104 and the clear demonstration that fission yeast Brr6 mediates SPB insertion and exit lead us to propose that Brr6-related molecules modulate similar interactions between MTOCs and the nuclear envelope in other species.
Although we propose a conserved function for Brr6-related proteins, elaborations around a core PF10104 function could account for the diversity of solutions that have been adopted to cope with the challenges of segregating genomes within an enclosed nuclear envelope (Kubai, 1976; Heath, 1980). For example, it is possible that the need to retain the SPB within the envelope throughout the cell cycle (Adams and Kilmartin, 2000) has set different hierarchies for SPB incorporation in budding and fission yeasts. As discussed earlier, the insertion event in S. cerevisiae in which the duplication plaque is pulled into the envelope alongside its integrated parental SPB differs considerably from insertion into a fenestra within the S. pombe nuclear envelope. Maybe these distinct topological demands explain the absence from S. pombe of the Nbp1 molecule that associates with Ndc1 and Mps2 to drive the insertion of the S. cerevisiae duplication plaque (Araki et al., 2006). In this context, it would be interesting to determine the contributions of Brl1 and Brr6 to SPB integration in strains in which a modified SESA network facilitates life without Mps2 (Sezen et al., 2009).
There is currently no link between Brr6 or Brl1 and SPB insertion in S. cerevisiae. Perhaps this reflects a need to simultaneously mutate both loci to reveal an SPB phenotype. However, the functional data that links Brr6 with NPC assembly (Hodge et al., 2010) does challenge our proposal that the primary requirement for the PF10104 domain is to generate polar fenestrae. It may be important to consider the generic evolutionary conservation of NPC architecture and the restriction of PF10104 distribution; although NPC architecture is conserved, many eukaryotes do not need a PF10104 protein for NPC assembly. So why does S. cerevisiae need one? Perhaps this requirement reflects a coevolutionary adaptation to the coexistence of NPCs and PF10104 domains; i.e., although PF10104 domain proteins are not required for the assembly of a generic eukaryotic NPC, if one is present, NPC components adapt to the alterations in membrane composition they promote, leading to an “addiction” to PF10104-driven modifications for NPC assembly. Thus, if a PF10104 molecule is modifying membranes, it is understandable that resident membrane proteins would adapt to accommodate, or take advantage of, the modified membrane properties.
In summary, we believe that the characterization of S. pombe Brr6 presented here and the evolutionary distribution of the PF10104 domain suggests that the Brr6 family enables a significant subset of lower eukaryotes to deal with one of the significant distinctions between many higher and lower eukaryotes: genome segregation within a closed mitosis.
Materials and methods
Cell culture and strains
Strains are listed in Table I. Cells were cultured in YE5S or appropriately supplemented Edinburgh Minimal Medium number 2 (EMM2; Moreno et al., 1991). Size selection for cell cycle synchronization used an elutriator rotor (JE-5.0; Beckman Coulter; Mitchison and Carter, 1975). The screen of Broek et al. (1991) was modified by the inclusion of cut12.NEGFP to identify conditional mutations that condensed chromosomes and exhibited defects in SPB separation in the presence of 4% DMSO (Craven et al., 1998; Poloni and Simanis, 2002). The pURB1 and pURB2 libraries were used to clone plasmids that complement the brr6.ds1 mutation according to standard procedures (Moreno et al., 1991; Barbet et al., 1992).
Isolation of apq12+ and brr6+
Transformation of IH2586 with the pURB2 library (Barbet et al., 1992) generated 3 × 105 transformants. Plasmid rescue, retransformation, and subcloning established that two of the transformants are viable because the apq12+ gene on the plasmids they contained suppressed the DMSO sensitivity of brr6.ds1. Mitotic haploidization with lys1/ura1 leu1/his3 and ade6 markers (Kohli et al., 1977) established that the sdd1.ds1 mutation resided on chromosome I. After initial long-range mapping in a swi5.158 background (Schmidt, 1993), crosses between sdd1.ds1 and cdc24.M38 produced no recombinants in 13 tetrads. Subcloning cosmid SPAC8F11 (Wellcome Trust Sanger Institute) and Tn1000 transposition analysis (Morgan et al., 1996) established that sdd1+ was allelic to brr6+. The SMART RACE cDNA kit (Takara Bio Inc.) established that the 5′ untranslated region (UTR) and 3′ UTR for brr6+ were 257 nt and 232 nt, respectively.
Mutagenesis of S. pombe loci
PCR technologies were used to delete brr6+ and tag the nup107+ and apq12+ loci (Bähler et al., 1998). The temperature-sensitive brr6.ts8 and N-terminally tagged brr6.NEGFP and brr6.N3Pk alleles were generated by a marker switch approach (MacIver et al., 2003): The ADH1 terminator and kanMX6 marker cassette were integrated immediately after the brr6 ORF stop codon (IH3412; Bähler et al., 1998). This cassette was switched with a fragment in which ura4+ was inserted downstream of an altered brr6 ORF (IH3587). To create conditional alleles, the second fragment (the 3.7-kb genomic PstI–HindIII fragment from cosmid SPAC8F11) was cloned into the pUC18 to generate pUCBrr6. A BglII site was then introduced 10 nt downstream from the brr6+ ORF into which the BamHI fragment from pREP42PkN (Craven et al., 1998) was inserted to generate pUCBrr6.ura4. Mutagenic PCR using pUCBrr6.ura4 as a template generated the fragment for transformation (MacIver et al., 2003). To generate alleles in which 3Pk or EGFP were inserted at the initiator ATG, the ATG of pUCBrr6 was converted to an NdeI site, creating pUCBrr6-5NdeI. The SacI–PstI fragment of pUCBrr6-5NdeI was inserted into pGEM/T (Promega) in which the NdeI site had been destroyed generating pGEMBrr6-5NdeI. The 3Pk and EGFP NdeI cassettes (Craven et al., 1998) were inserted into the NdeI site of pGEMBrr6-5NdeI. The introduction of a BglII and the BamHI fragment from pREP42PkN as for pUCBrr6.ura4 generated pGEM-PkNBrr6.ura4 and pGEM-EGFPNBrr6.ura4. Fragments spanning from 352 bp upstream of brr6 to 189 bp downstream of ura4+ were transformed into strain IH3412. PCR analysis of strains that switched the kanR marker for ura4+ identified strains bearing the tagged alleles. The SacI–PstI fragment from pGEM-PkNBrr6 was cloned into the pINTA (Petersen et al., 2001) for integration at the leu1 locus to create IH3508.
DNA fragments corresponding to codons of 1–79 (Brr6-Nt) and 253–297 (Brr6-Ct) of Brr6 were cloned into pGEX-6P-1 (GE Healthcare). GST fusion proteins were expressed in Escherichia coli BL21(DE3)RIL and purified on glutathione Sepharose 4B. A mix of GST-Brr6-Nt and GST-Brr6-Ct were used to generate and affinity purify separate rabbit polyclonal antibodies (Scottish Antibody Production Unit).
Indirect immunofluorescence of S. pombe to stain microtubules (using TAT1; a generous gift from K. Gull, Oxford University, England, UK) and Sad1 has been described previously (Hagan and Yanagida, 1995). For Brr6 and Cdc7.HA staining, cells were processed after fixation in 3.6% formaldehyde for 10 min. A DeltaVision system (Applied Precision) fitted with a 100× 1.45 NA α-plan Fluor objective lens (Carl Zeiss) was used for live cell imaging in conjunction with SoftWorX and Imaris (Bitplane) software. A UApo TIRFM 150×/1.45 NA lens (Olympus) was used with a microscope (Deltavision) for imaging the registry of Nup107.cherry and Brr6.NEGFP signals. The cells were mounted in a Bioptechs chamber. Bioptechs controls regulated the temperature of the chamber and objective while the stage was maintained in a temperature-controlled environmental chamber (Solent Scientific). Fig. 6 B, Fig. 9, Fig. S4 C, and Fig. S5 E show maximal projections of 14 sections, 0.2 µm between the slices. Fig. 7 (A–D) and Videos 1–4 show maximal projections of 14 sections, 0.3 µm between the slices. Images were taken every 3 min.
Cells were prepared for electron microscopy using high-pressure freezing and freeze substitution as described previously (Murray, 2008; Tallada et al., 2009). Wild-type brr6+ cells were filtered and fixed from mid-log phase asynchronous cultures. brr6.ts8 cells were synchronized with respect to cell cycle progression by centrifugal elutriation (Mitchison and Carter, 1975) using an elutriator rotor (JE-5.0; Beckman Coulter). After filtration onto 0.45-µm membrane filters (Millipore), cells were loaded into interlocking brass hats (Swiss Precision Instruments) and fixed by high-pressure freezing with a high-pressure freezer (HPM010; Bal-Tec). Freeze substitution into 2% OsO4 + 0.1% uranyl acetate in anhydrous acetone was conducted using an automatic freeze substitution chamber unit (AFS; Leica) at 90°C for 72 h with a 5°C h−1 slope to raise the temperature to −20°C, at which point cells were held for 2 h before a final increase to 4°C for 4 h at a rate of 5°C h−1. After infiltration with Spurr’s resin, blocks were trimmed and serial sections were prepared with a diamond knife, mounted on Formvar/carbon-coated single-slot grids, stained with Reynolds’s solution for 5 min, and imaged on a transmission electron microscope (model 1220; JEOL) at 80 kV. 12 and 15 complete serial thin section images of SPBs were analyzed for brr6.ts8 and wild-type cells, respectively.
Online supplemental material
Fig. S1 shows wild-type controls for anti-tubulin/Sad1 immunofluorescence and electron microscopy alongside Pcp1.RFP/Atb1.GFP images of old pole activation in brr6.ts8. Fig. S2 shows the phenotype arising from complete loss of Brr6 and that brr6.ts8 is a recessive allele. Fig. S3 shows further electron micrographs of brr6.ts8 cells at the restrictive temperature. Fig. S4 shows the genetic interactions between brr6 mutant alleles and strains harboring mutations in SPB components. Fig. S5 shows the characterization of apq12.Δ phenotype and genetic interactions with SPB components and brr6 alleles. Fig. 5 also shows the sensitivity of brr6 alleles and apq12.Δ to reagents that alter membrane fluidity. Video 1 shows the maintenance of nuclear integrity during mitotic commitment of wild-type cells, whereas Video 2 shows that nuclear integrity is transiently breached during mitotic commitment of brr6.ts8 cells. Video 3 shows that the integrity of the nucleus is maintained during anaphase of nda3.KM311 brr6+ cells, whereas Video 4 shows that this integrity is breached, sometimes repeatedly, toward the end of anaphase B/mitotic exit of brr6.ts8 cells.
We thank Osami Niwa (Kazusa DNA Research Institute, Japan), Valerie Doye (Institut Curie, France), and Katherine Sheldrick for strains; Stephen Murray and Sandra Rutherford for technical assistance; and Steve Bagley, the Paterson advanced imaging facility for technical support, and Keith Gull (Oxford University, UK) for the TAT1 antibody.
This work was supported by Cancer Research UK (CRUK) grant C147/A6058.
Tiina Tamm and Agnes Grallert contributed equally to this study.
T. Tamm’s present address is Institute of Molecular and Cell Biology, University of Tartu, Tartu 51010, Estonia.
F.E. Stevens’ present address is Electron Microscope Unit, University of Sydney, New South Wales 2006, Australia.