The transmission of nuclear pore complexes to daughter cells requires a cytoplasmic pool of Nsp1

Asymmetric cell divisions are critical for cell fate determination during embryogenesis, organogenesis, and differentiation (Neumüller and Knoblich, 2009). Since the budding yeast Saccharomyces cerevisiae undergoes an asymmetric division, it is an effective model for identifying factors that are actively segregated along the polarity axis and the underlying molecular mechanisms responsible for their segregation (Pruyne et al., 2004). In yeast, the two type V myosin motors Myo2 and Myo4 deliver organelles, secretory vesicles, and mRNAs to the daughter cell (Pruyne et al., 2004; Chung and Takizawa, 2010; Eves et al., 2012). Myo2 also plays a role in nuclear migration by guiding spindle microtubules along actin cables in concert with a complex of proteins at the plus ends of microtubules, including Kar9 and Bim1 (Korinek et al., 2000; Miller et al., 2000; Yin et al., 2000). A redundant nuclear positioning pathway requires the dynein–dynactin complex (Eshel et al., 1993; Li et al., 1993; Grava et al., 2006). Recent data also implicates the exocyst complex in anchoring ER tubules that extend from the mother nuclear envelope (NE) to the bud cortex in maintaining nuclear position at the bud neck (Kirchenbauer and Liakopoulos, 2013). Further, the ubiquitylation of a component of the nuclear pore complex (NPC) was shown to function in nuclear migration through the recruitment of dynein light chain to the NE (Hayakawa et al., 2012). The latter process reflects several connections uncovered between NPCs and the cytoskeleton (Stelter et al., 2007; Splinter et al., 2010; Bolhy et al., 2011; Steinberg et al., 2012).

NPCs are massive protein assemblies embedded in the NE that control the flux of molecules between the nucleus and cytoplasm. Each NPC is composed of ∼30 individual protomers termed nucleoporins (nups; Rout et al., 2000; Cronshaw et al., 2002) found in distinct subcomplexes (Siniossoglou et al., 1996, 2000; Grandi et al., 1997; Marelli et al., 1998; Belgareh et al., 2001; Vasu et al., 2001; Alber et al., 2007a,b; Onischenko et al., 2009). These subcomplexes are thought to form modular building blocks that contribute to the formation of the concentric membrane, inner and outer ring complexes that surround a central transport channel (Alber et al., 2007a,b). The channel itself is rich in unstructured nups like Nsp1/Nup62 that contain repetitive peptide motifs of Phe-Gly (FG-nups; Alber et al., 2007a,b). Nsp1 helps form two subcomplexes at the NPC composed of Nup49, Nup57, and Nic96, or Nup82 and Nup159 (Nehrbass et al., 1990; Mutvei et al., 1992; Grandi et al., 1995; Schlaich et al., 1997; Bailer et al., 2000, 2001).

Because transport through NPCs is essential for cell life, there are likely mechanisms to ensure that NPC numbers can accommodate cell type–specific nuclear transport loads. We understand little about mechanisms that control NPC number. Lymphocyte stimulation results in a doubling of NPC number, which suggests that external inputs can up-regulate the NPC assembly pathway (Maul et al., 1972). Further, the S-phase doubling of NPCs observed in cell culture suggests that NPC assembly is linked to the cell cycle (Maul et al., 1972), perhaps through cyclin-dependent kinases (Maeshima et al., 2010). Mutations in nups important for NPC assembly can also impact differentiation programs (Lupu et al., 2008; de Jong-Curtain et al., 2009; D’Angelo et al., 2012). These studies cumulatively suggest that NPCs themselves might be important for cell fate determination and underscore the importance of identifying mechanisms that control NPC number.

One way to modulate NPC number is to regulate the de novo assembly of NPCs, which occurs by postmitotic and interphase mechanisms (Doucet et al., 2010). During de novo NPC assembly in interphase, the membrane and inner ring complexes assemble at the NE first and might directly contribute to fusion of the inner and outer nuclear membranes (Makio et al., 2009; Onischenko et al., 2009; Doucet et al., 2010; Dultz and Ellenberg, 2010; Fichtman et al., 2010; Vollmer et al., 2012). Other conserved ER and inner nuclear membrane proteins might also act at these early steps (Dawson et al., 2009; Chadrin et al., 2010; Talamas and Hetzer, 2011; Yewdell et al., 2011), which are followed by the recruitment of the outer ring complex, Nic96, and the FG-nups (Zabel et al., 1996; Doucet et al., 2010).

Since budding yeast NEs/NPCs remain intact during mitosis, determining NPC number relies on both de novo assembly and mechanisms that segregate NPCs between mother and daughter cells. One study suggests that mother NPCs are restricted from being inherited by the daughter, placing a burden on de novo NPC assembly to support daughter viability (Shcheprova et al., 2008). Since the known NPC assembly mechanism is relatively slow compared to yeast mitosis (Winey et al., 1997; Dultz and Ellenberg, 2010), there might be a faster daughter-specific NPC assembly mechanism that has yet to be uncovered. It is also plausible that there is a mechanism that ensures NPC inheritance. Such mechanisms exist for several organelles including mitochondria (Itoh et al., 2002), vacuoles (Hill et al., 1996), peroxisomes (Hoepfner et al., 2001; Fagarasanu et al., 2006), cortical ER (Du et al., 2001; Estrada et al., 2003) and late Golgi (Rossanese et al., 2001). Consistent with the idea that NPCs are also actively transmitted to daughter cells, the use of a tandem fluorescent “timer” protein showed a bias of “old” nups in daughters (Khmelinskii et al., 2012).

In an effort to identify putative inheritance and/or de novo NPC assembly mechanisms needed to ensure that adequate numbers of NPCs are present in daughter cells, we investigated the distribution of newly synthesized and assembled pools of representative components of distinct nup subcomplexes through mitotic divisions in budding yeast. We identified a newly synthesized pool of the nup Nsp1 that accumulates between S and G2 phase and is directed into the bud by interactions with ER in a pathway that requires Myo2. Inactivation of newly synthesized Nsp1 results in a dramatic loss of NPC inheritance by daughter cells. Our data are consistent with a model in which Nsp1 is a critical player in a pathway governing NPC number in daughter cells, which ensures their viability.

To investigate how budding yeast nups are segregated between mother and daughter cells, we visualized both newly synthesized and existing pools of representative members of the membrane (Ndc1) and inner (Nup170) and outer ring (Nup85) complexes, in addition to the FG-nup Nsp1 (Fig. 1 A) using a recombination induced tag exchange (RITE) approach (Fig. 1 B; Verzijlbergen et al., 2010). Nup-RITE fusions are expressed from their endogenous gene loci, and tag exchange between mCherry and GFP genes is initiated by an estradiol-induced activation of a constitutively expressed Cre recombinase–estradiol-binding domain (EBD) fusion (Fig. 1 B; Logie and Stewart, 1995). Unlike photoconvertible proteins like Dendra2, which mature slowly (t_1/2 of maturation ∼1.5 h; Zhang et al., 2007), the RITE approach leverages the fast maturation of GFP (t_1/2 of maturation <15 min; Iizuka et al., 2011) to visualize the production of “new” nups. Further, the “old” mCherry signal did not increase after the first appearance of “new” green fluorescence, which supports the finding that we can accurately assess the distribution of both old and new versions of these nups through anaphase (Fig. S1).

Consistent with published data (Khmelinskii et al., 2010), we observed that the “old” mCherry nups segregated with the NE (Fig. 1, C–F, bottom panels), which suggests that NPCs are inherited by daughter cells. We simultaneously visualized the appearance and segregation of the “new” GFP nups. Since the genetic switch is not synchronous between cells, we focused on those in which green fluorescence appeared late in the cell cycle to enable the visualization of NPC protomers rather than newly assembled NPCs. GFP fusions of Nup85, Ndc1, and Nup170 segregated equivalently to their “old” counterparts, as reflected in similar daughter/mother total nuclear fluorescence ratios (tf_d/tf_m) of red and green proteins (Fig. 1, C1, D1, and E1). To normalize for the amount of NE in the daughter and mother, we divided the tf_d/tf_m ratio for a given daughter–mother pair by a daughter/mother ratio of nuclear surface area (sa_d/sa_m). In general, this number hovers around unity for both the newly synthesized and old nups, which indicates a strong correspondence between NE surface area and the number of NPCs (Fig. 1, C2, D2, and E2). These data suggest that for the membrane and inner and outer ring complexes (the scaffold of the NPC), there is little bias in the enrichment of these NPC protomers in either the mother or daughter. Thus, we find that NPCs are segregated concomitantly with the NE between mother and daughter cells and that asymmetry in the NPC assembly pathway is minimal for these components.

In contrast to the scaffold nups, the FG-nup Nsp1-GFP did not display a symmetrical distribution. First, although low levels of NE fluorescence could be visualized in the mother and daughter, a discrete Nsp1-GFP focus was visualized in the daughter at the cell cortex (Fig. 1 F, arrowheads). As anaphase progressed, the focus moved from the cortex and integrated into the NE (Fig. 1 F and Video 1). Consistent with these observations, the calculated tf_d/tf_m of the GFP fluorescence remained at ∼0.2 for the first three stages of anaphase, but by cytokinesis the ratio reached ∼0.7, as exhibited by old and new ratios of all nups (Fig. 1, F1). To rule out that this pool of Nsp1 resulted from an aggregation or instability of the GFP-tagged Nsp1, we confirmed that the cytoplasmic Nsp1-GFP accumulation could be reproduced with monomeric GFP (Fig. S2 A), and there was no change in protein levels or stability of Nsp1-GFP compared to untagged Nsp1 by Western blotting (Fig. S2 B). Further, as shown in Fig. S2 C, we could observe bud-localized foci of untagged Nsp1 by immunofluorescence using specific anti-Nsp1 antibodies. In addition, we affinity purified Nsp1-GFP and confirmed that all established interacting partners including Nup49, Nup57, Nic96, Nup159, Dyn2, and Nup82 (Nehrbass et al., 1990; Grandi et al., 1993, 1995; Zabel et al., 1996; Schlaich et al., 1997; Bailer et al., 2001; Stelter et al., 2007) were specifically enriched compared to controls (Fig. S2 D). Last, the Nsp1 fusion proteins generated in this study support viability (Fig. 2 B). Collectively, these data support the finding that the Nsp1-GFP foci represent a previously undiscovered pool of Nsp1 localized to daughter cells before completion of anaphase.

To directly query the function of this bud-localized pool of Nsp1, we needed an experimental system capable of rapid and specific abrogation of the function of a newly synthesized cytoplasmic pool of nups (which we term nup_CYT) without affecting the pool assembled into NPCs (nup_NPC). We adopted the “anchor-away” system (Haruki et al., 2008), which exploits the conditional high-affinity dimerization of the FRB and FKBP12 protein domains in the presence of the drug rapamycin (Fig. 2 A). By expressing nups as an FRB fusion, we selectively recruit their newly synthesized nup_CYT pools to an abundant plasma membrane trap (Pma1-FKBP12; Fig. 2 A). Growth of Nup-FRB fusion–containing strains on medium with carrier alone (DMSO) showed that the FRB moiety does not appreciably influence their fitness (Fig. 2 B), whereas the addition of rapamycin led to growth arrest, which is consistent with the trapping of essential proteins (Figs. 2 B and S3). Further, plasmid-expressed NSP1 and NIC96 were able to suppress rapamycin-induced growth inhibition, ensuring that nup trapping did not act as a dominant-negative (Fig. 2 B). As expected, we observed an accumulation of all tested Nup-FRB-GFP fusions at the plasma membrane after incubation in the presence of rapamycin (Fig. 2 C).

To test whether our trap was specific to nup_CYT over nup_NPC, we incubated strains expressing Nup-FRB-GFP fusions in the presence of rapamycin and cycloheximide (to inhibit protein synthesis). Under these conditions, the Nup-FRB-GFP plasma membrane pool was completely absent, and nup_NPC was unaffected (Fig. 2 C). These data are consistent with the interpretation that the plasma membrane pool is made up exclusively of Nup-FRB-GFP_CYT. Thus, this system is capable of specifically inhibiting nup_CYT without affecting nup_NPC.

Since the Nsp1_CYT foci are newly synthesized and appear to be integrated into the NE at the end of anaphase (Fig. 1 F), we reasoned that they might represent NPC assembly intermediates. We therefore examined how trapping of Nsp1-FRB-GFP_CYT impacted the relative levels of NPCs in mother and daughter cells after anaphase completion. Because there is little turnover of nups_NPC (Fig. 2 C), we assume that Nup-FRB-GFP_NPC represents NPCs; we localized Nup170-mCherry in the same cells as an independent NPC marker. Nsp1-FRB-GFP–expressing cells were grown to mid-log phase, incubated in the presence of rapamycin or carrier alone (DMSO), and imaged over one to two cell cycles (Fig. 3 A). Strikingly, the trapping of Nsp1-FRB-GFP_CYT resulted in a dramatic reduction in the levels of Nsp1-FRB-GFP_NPC in daughter cells after mitosis, while also affecting (although to a lesser extent) the daughter levels of Nup170-mCherry (Fig. 3B). When expressed as a tf_d/tf_m ratio, we observed a 72% reduction in daughter levels of Nsp1-FRB-GFP_NPC from an average tf_d/tf_m of 0.57 in DMSO-treated cells compared to 0.16 in trapped cells (Fig. 3 C). This change in NPC levels could not be explained by the relatively modest (34%) change in nuclear size observed upon trapping Nsp1-FRB-GFP_CYT (Fig. S4). Similar experiments were performed with additional members of the Nsp1 complex including Nup49-FRB-GFP, Nup57-FRB-GFP, and Nic96-FRB-GFP. In these cells, the tf_d/tf_m ratios were also significantly reduced in the presence of rapamycin (Fig. 3 C). Because we did not observe a restriction of NPC inheritance under wild-type conditions (Fig. 1), these data suggest that there is a specific inhibition of the transmission of NPCs upon sequestering newly synthesized, cytoplasmic components of the Nsp1 complex.

To further test if trapping of the Nsp1 complex leads to the specific loss of the transmission of NPCs to daughter cells, we localized an assembled and stable component of the NPC, Nup133, tagged with the photoconvertible Dendra fluorescent protein in an Nsp1-FRB–containing strain. We irreversibly photoconverted the majority of Dendra from its green to red form in G1/S-phase cells such that any remaining Nup133 protomers would complete assembly into NPCs before anaphase (Fig. 3 D; “photoconversion”). We then tracked the red fluorescent (assembled/old) NPCs through the cell cycle in the presence of either DMSO or rapamycin. Strikingly, rapamycin treatment resulted in a severe loss of the inheritance of old Nup133-Dendra (Fig. 3 D), which is reflected in the change of tf_d/tf_m ratios from 0.56 to 0.15 (Fig. 3 E). These data support the conclusion that NPCs are not inherited when Nsp1_CYT is inhibited.

Interestingly, we observed a weak accumulation of the green “new” Nup133-Dendra in daughter cells (Fig. 3 D) shown by the difference in the mean red and green tf_d/tf_m ratios (0.15 and 0.44, respectively; Fig. 3 E), which suggests that NPC assembly continues after Nsp1_CYT is trapped. To test this, we monitored the levels of Nsp1-FRB-GFP_NPC and Nup170-mCherry in cells progressing from G1 to G2/M. In DMSO-treated cells, both GFP and mCherry fluorescence increased with similar kinetics, which we interpret to represent de novo NPC assembly (Fig. 3 F). In contrast, in rapamycin-treated cells, Nsp1-FRB-GFP levels remain constant while Nup170-mCherry increases, perhaps reflecting the accumulation of an NPC assembly intermediate (Fig. 3 F). Consistent with this idea, the tf_d/tf_m ratios of Nup170-mCherry were significantly higher than the tf_d/tf_m ratios of their Nsp1 complex FRB-GFP_NPC partners (Fig. 3 C). Thus, the Nup170-mCherry–containing NPC assembly intermediates might either be transmitted or assembled in the daughter when the Nsp1 complex is trapped.

To test if the mechanism of NPC transmission required de novo NPC assembly, we generated a strain expressing the essential inner ring nup Nup192 as a FRB-GFP fusion (Fig. S3). Because the inner ring complex is thought to assemble upstream of the Nsp1 complex (Kosova et al., 1999; Gomez-Ospina et al., 2000; Makio et al., 2009; Onischenko et al., 2009), trapping of Nup192 allowed us to ask whether inhibition of NPC assembly influences the transmission of NPCs. As shown in Fig. 4 A, under rapamycin and DMSO conditions the relative levels of Nup192-FRB-GFP_NPC between mother and daughter cells did not change (tf_d/tf_m ratios of ∼0.6; Fig. 4 B), which suggests that compromised NPC assembly does not influence the mechanism of NPC transmission. Similar results were obtained upon trapping the outer ring component Nup120 (Fig. 4, A and B).

If there is a mechanism to specifically drive NPC transmission to daughter cells, we hypothesized that inhibition of this pathway would lead to daughter-specific phenotypes distinct from those in NPC assembly mutants. To test this idea, we grew Nsp1-FRB-GFP– and Nup192-FRB-GFP–containing cells in the presence of rapamycin. In both cases, the cells ceased to grow after ∼24 h and were vacuolated, suggesting a loss of viability (Fig. 4 C). Interestingly, while the initial Nsp1-FRB-GFP mother cell (time 0) was able to give rise to multiple progeny (mean of 3.2; Fig. 4 D), almost all of its daughters failed to bud, which indicates a daughter-specific growth arrest. In contrast, daughter cells from the Nup192-FRB-GFP mother continued to bud with a frequency reflecting the order by which they were born; the first daughter budded an average of 2.3 times, whereas the fifth daughter never budded (Fig. 4 D). These data are consistent with the interpretation that the transmission of NPCs to daughters results in their dilution after each division, leading to a loss of viability of all cells. We suggest that the loss of viability of the Nsp1-GFP-FRB mother cell is not due to the retention of NPCs, rather it is likely the result of the assembly of NPCs lacking Nsp1, which would ultimately compromise nuclear–cytoplasmic compartmentalization.

To unravel the mechanism of NPC transmission, we performed experiments to understand the production, dynamics, and localization of Nsp1_CYT foci. We first investigated whether Nsp1_CYT production is cell cycle stage dependent. We localized Nsp1-GFP after blocking cell cycle progression in G1, S, or G2 using α-factor, hydroxyurea, and SWE1 overexpression, respectively (Fig. 5 A). In G1-arrested cells, no Nsp1-GFP accumulated within cytoplasmic structures. In contrast, upon S and G2 arrests, Nsp1-GFP could be visualized within cytoplasmic foci (Fig. 5 A, arrowheads). Consistent with our RITE experiments, these foci were specific for the Nsp1 complex, with Nsp1, Nup82, Nup49, and Nup57 appearing in a cytoplasmic pool in ∼80% of S-phase–arrested cells (Fig. 5 B), whereas analogous structures were not observed with Nup170-GFP or Pom34-GFP (Fig. 5, A and B). We observed Nup85-GFP in cytoplasmic foci in 16% of cells, but our inability to similarly localize either membrane (Pom34) or inner ring (Nup170) nups argues against the idea that these cytoplasmic foci are annulate lamellae (Kessel, 1992).

Interestingly, the distribution of the Nsp1_CYT foci was biased toward the daughter bud (Fig. 5 C). This bud bias was most evident in cells arrested in G2 in which the polarized/apical growth pathway is hyperactivated (Sia et al., 1998; McMillan et al., 1999); under these conditions the number of foci increased with bud size, with greater numbers found in the bud (Fig. 5 C). These data suggest an active targeting or retention mechanism for Nsp1_CYT in daughter cells. To further establish that this bud pool is distinct from Nsp1_NPC, we arrested cells containing the Nsp1-RITE cassette in G2 in the presence of estradiol. In these cells, the first “new” Nsp1-GFP signal appeared as a focus in the daughter bud (Fig. 5 D, arrowheads; and Video 2). As we prolonged the arrest, additional Nsp1-GFP_CYT foci accumulated in the bud concomitantly with a NE pool. We did not observe any “old” Nsp1-mCherry in the cytoplasm, nor did we detect a similar bud-localized accumulation of “new” or “old” Nup85 (Fig. 5 D and Video 3). These data further support the existence of two pools of Nsp1: one that is assembled into NPCs (Nsp1_NPC) and another that is actively targeted and/or retained in the daughter cell (Nsp1_CYT).

We felt that understanding the role of Nsp1 in NPC transmission required defining how Nsp1_CYT is localized to the bud. We therefore tested colocalization of Nsp1-GFP_CYT with components of the polarity apparatus and organelles actively transported to daughters, like ER. For these experiments, we delayed cells before anaphase using a temperature-sensitive cdc6-1 allele (Liang et al., 1995). By arresting cdc6-1 cells in G2/M at the nonpermissive temperature and then imaging them under permissive conditions, we can exert consistent control over the production of the Nsp1-GFP_CYT foci. Using this strategy, we failed to observe colocalization between Nsp1-GFP_CYT and several bud-directed factors including Kar9 (Fig. 6 A), Myo2, Abp140 (Yang and Pon, 2002), Abp1 (Wesp et al., 1997; Goode et al., 2001; Fig. S5), or Bud6 (unpublished data; Amberg et al., 2005). We did observe a transient colocalization of Nsp1-GFP_CYT with the exocyst subunit, Exo70 (Fig. 6 B). Interestingly, mammalian Nsp1/Nup62 interacts with Exo70 at the leading edge of migrating cells (Hubert et al., 2009; Béaslas et al., 2012), raising the possibility that this interaction is conserved, even if transient.

In contrast to the fleeting Nsp1-Exo70 colocalization, we observed a stable association between Nsp1-GFP_CYT foci and ER (visualized with HDEL-DsRed; Fig. 6 C). For example, in Fig. 6 D, an Nsp1-GFP_CYT focus (arrowhead) remains associated with an ER tubule that extends from the mother NE to the daughter cortex for >95 s (Video 4). Moreover, we colocalized Nsp1-GFP_CYT with a bolus of ER originating from the mother NE and traveling to the daughter cortex over 15 min (Fig. S5 B). Because ER morphology in yeast requires an intact actin cytoskeleton (Prinz et al., 2000; Fehrenbacher et al., 2002), we tested the association of Nsp1_CYT with ER after disrupting actin filaments by treatment with latrunculin A. Under this condition, we observed a reduction in tubular ER in the bud (Fig. 6 E), and the majority of Nsp1-GFP_CYT foci localized to the cell cortex, presumably with cortical ER (Fig. 6, E and F). No effect on ER structure, or Nsp1_CYT distribution, was observed after disrupting the spindle with nocodazole (Fig. 6, E and F).

The association of Nsp1_CYT with ER was further confirmed by immuno-EM using anti-Nsp1 antibody staining followed by gold-labeled secondary antibodies on ultrathin sections. The anti-Nsp1 antibody was highly specific, as quantification of gold particles at the nucleus showed that >93% were at the NE (Fig. 7 A). Further, NPCs were often decorated with two or more gold particles (Fig. 7 B). Outside the nucleus, we observed gold particles in the cytoplasm and on membranes (Fig. 7 C, arrowheads). We assessed that 40% of the gold particles were associated with membranes that could be morphologically identified as either internal ER (Fig. 7, A and D) or cortical ER (Fig. 7, E and F). Collectively, the combination of our fluorescence and EM data support the conclusion that Nsp1_CYT associates with ER, which likely plays a role in its localization to the bud.

The association of Nsp1_CYT with ER extending from the mother NE and the potential interaction with Exo70 prompted us to investigate whether Nsp1’s role in NPC transmission might function alongside a recently described nuclear positioning pathway (Kirchenbauer and Liakopoulos, 2013). In this pathway, ER tubules connect the mother NE to the daughter cortex through the exocyst complex (of which Exo70 is a member; Fig. 8 A). This interaction is thought to help drive the formation of NP; dynamic bud extensions of the mother NE are proposed to help maintain nuclear position at the bud neck (Kirchenbauer and Liakopoulos, 2013; Fig. 8 A). We first tested whether the Nsp1_CYT foci might be physically connected to NP by time-lapse imaging of cdc6-1 cells. After release from the cdc6-1 arrest, NP are observed as the mother NE vectorially extends into the bud and dynamically retracts throughout the ∼1 h time lapse shown in Fig. 8 B (see Video 5 and also Videos 6 and 7). Remarkably, NP move toward a bud-localized Nsp1-GFP_CYT focus (Fig. 8 B, arrowheads), appear to “kiss” it, and retract to the mother. This occurs multiple times before the NE absorbs the focus at the end of the time lapse. NP in strains expressing Nup170-GFP do not extend substantially beyond the bud neck, and the NE signal is continuous between mother and bud portions (Fig. 8 B, right; and Video 8).

The interaction between Nsp1_CYT and NP prompted us to investigate whether trapping of Nsp1-FRB_CYT might, in addition to affecting NPC transmission, impact nuclear positioning at the bud neck. To test this idea, we treated cells expressing either Nsp1-FRB-GFP or Nup192-FRB-GFP with rapamycin and monitored nuclear position in small or unbudded cells as they moved through the cell cycle. Hmg1-mCherry was used to visualize the NE (Fig. 8 C). As expected, under DMSO-treated conditions, 94% of nuclei moved to the bud neck before anaphase and 98% of nuclei were properly positioned when Nup192-FRB-GFP was trapped (Fig. 8, C and D). In contrast, in 25% of rapamycin-treated Nsp1-FRB-GFP cells, the nuclei were found at a distal location to the neck, which supports the conclusion that the inhibition of Nsp1_CYT impacted the ability of cells to position their nuclei correctly (Fig. 8, C and D).

Our observations that trapping of Nsp1_CYT impacts NPC inheritance and nuclear positioning suggested that these two processes might be coupled. To explore this idea, we investigated Nsp1_CYT distribution and NPC inheritance in strains with mutations in the canonical myosin- and dynein-dependent nuclear positioning pathways. We tested the distribution of Nsp1-GFP by arresting strains containing temperature-sensitive alleles of MYO2 or knockouts of dynein heavy (DYN1) and light (DYN2) chain in S phase (Fig. 9 A). Remarkably, in the myo2 strains, we observed a substantial redistribution of the Nsp1-GFP_CYT foci from a predominantly bud-biased distribution to one that was unbiased or biased to the mother (Fig. 9, A and B). In contrast, Nsp1-GFP_CYT was not affected in dyn1Δ, dyn2Δ, and myo4Δ strains (Fig. 9, A and B). These data predict that NPC transmission requires Myo2 function. To test this idea, we calculated the mean fluorescence intensity (mfi) of a scaffold component of the NPC (Nup85-GFP) at the NE as an indirect measure of NPC density and expressed this as a daughter/mother ratio (mfi_d/mfi_m). Consistent with an active mechanism to deliver NPCs to the daughter, in wild-type cells the average mfi_d/mfi_m ratio was >1 (1.20), which suggests a higher density of NPCs in daughter cells compared to mother cells after anaphase completion (Fig. 9, C and D). These results were mirrored in dyn1Δ, dyn2Δ, and myo4Δ cells. In contrast, in myo2-14 cells the mean mfi_d/mfi_m was 0.94 (Fig. 9 D), a significant difference (P < 0.0001) from wild-type cells. Collectively, these data provide support to the conclusion that NPC transmission to the daughter cell specifically depends on the Myo2-dependent localization of Nsp1.

We have uncovered a mechanism that controls the inheritance of NPCs in budding yeast based on a bud-directed pool of the Nsp1 complex. Such a mechanism helps explain why daughter cells have a higher density of NPCs (Fig. 9, C and D) and are enriched in “old” nups (Khmelinskii et al., 2012). We conclude that inheriting, as opposed to assembling, NPCs provides the cell the most efficient means to ensure daughter viability. Consistent with this idea, newly synthesized components of the scaffold of the NPC show no preference for the mother or daughter side of the anaphase NE, which argues against a daughter-specific NPC assembly mechanism (Fig. 1). Furthermore, there is no detectable bias in the NPC assembly reaction, as inhibiting assembly by the conditional trapping of either Nup120 or Nup192 did not affect the relative levels of NPCs between mother and daughter cells (Fig. 4, A and B). Indeed, after inhibiting NPC assembly, we and others (Zabel et al., 1996; Makio et al., 2009) show that daughter cells are capable of additional rounds of division, whereas restriction of NPC transmission results in a daughter-specific growth arrest (Fig. 4, C and D). We suggest that an actively controlled NPC inheritance mechanism might be most relevant when NPC assembly rates cannot keep up with rapid cell divisions. While the mechanism described here likely varies throughout eukaryotes, the principle that it is more efficient to ensure that components of the NPC are reused as opposed to resynthesized/assembled is likely universal and might help to explain why nups are so stable (Daigle et al., 2001; D’Angelo et al., 2009; Savas et al., 2012).

NPC transmission requires the production of an ER-bound pool of Nsp1_CYT synthesized before anaphase, likely between S and G2 phase (Fig. 5). Collectively, our data argue that Nsp1_CYT acts autonomously from NPCs to carry out this function. First, “old” Nsp1_NPC rarely, if ever, exchanges into Nsp1_CYT even over many hours of G2 arrest (Fig. 5 D). Consistent with this observation, Nsp1-FRB_NPC cannot be trapped at the plasma membrane (Fig. 2). Further, depleting Nsp1_CYT has a highly penetrant and rapid effect on daughter viability, whereas a general inhibition of NPC assembly (which would not be predicted to affect Nsp1_CYT or NPC transmission; Fig. 4 A) manifests gradually in both mother and daughter cells over several cell cycles (Fig. 4, C and D). Last, the reduction in NPC density in myo2 daughters more likely reflects a loss of bud targeting of the Nsp1 complex rather than NPC malfunction as nups localize normally to NPCs in these strains (Fig. 9).

There are two general models for how Nsp1_CYT might function in promoting the inheritance of NPCs: it helps NPCs either overcome a mother cell retention mechanism similar to that proposed for mitochondria (Yang et al., 1999) or a diffusional barrier at the bud neck (Shcheprova et al., 2008). We favor the model presented in Fig. 10 that invokes the existence of a diffusion barrier since a retention mechanism would be predicted to restrict NPC mobility, which has not been observed (Bucci and Wente, 1997; Belgareh and Doye, 1997). Further, since diffusion barriers would be more likely to impede the passage of large complexes such as NPCs, it predicts that smaller proteins like Hmg1 are able to access daughter cells (Fig. 8 C), as well as putative NPC assembly intermediates (Fig. 3).

In our model, Nsp1_CYT moves into the bud by interacting with ER targeted to the daughter (Fig. 10). Interestingly, while the canonical ER inheritance pathway uses Myo4 (Estrada et al., 2003), neither Nsp1_CYT distribution nor NPC inheritance are impacted in myo4Δ strains (Fig. 9). These data suggest that ER bound by Nsp1_CYT may be transmitted to the daughter in a mechanism mediated by the exocyst complex, which subsequently contributes to the formation of NP (Fig. 10; Kirchenbauer and Liakopoulos, 2013). A functional relationship between NP and Nsp1_CYT is consistent with our observation that NP dynamics are directed toward Nsp1_CYT (Fig. 8 B). Moreover, both the formation of NP and the localization of Nsp1_CYT depend on an intact actin cytoskeleton and Myo2 (Fig. 6, E and F; and 9, A and B) and are required for nuclear positioning at the bud neck (Fig. 8 D). Last, our data point to a potentially conserved interaction between Nsp1 and Exo70, first observed in mammalian tissue culture cells (Fig. 6 B; Hubert et al., 2009).

We propose that the passage of Nsp1_CYT through the bud neck licenses NPC transmission during anaphase by either directly (or through an unidentified factor) overcoming the diffusion barrier (Fig. 10). In the absence of Nsp1_CYT and/or Myo2 function, the barrier remains intact and NPCs are not able to enter the daughter, leading to a loss of daughter viability (Fig. 10). At this point we can only speculate as to the molecular composition of the bud neck barrier and the signaling factors and/or forces that modulate its function. We envision a mechanism analogous to the recently described ER stress surveillance pathway in which ER stress is signaled through the MAP kinase Slt2 to regulate septin function and impede cortical ER inheritance (Babour et al., 2010). Future experiments will be focused on identifying the internal and/or external inputs that impact Nsp1_CYT function that might modulate NPC number.

All yeast strains used in this study and their derivation are listed in Table S1. ABY530 and ABY534 were gifts from A. Bretscher (Cornell University, Ithaca, NY). YB0044 and PCCPL397 were gifts from B. Stillman (Cold Spring Harbor Laboratory, Cold Spring Harbor, NY), and D. Lew (Duke University, Durham, NC), respectively. Most experiments were performed on yeast grown at 30°C, with the exception of those containing temperature-sensitive alleles (myo2-14, myo2-20, and cdc6-1), which were grown at RT or at 34°C as described in the figure legends. All strains were grown in YP (1% yeast extract and 2% peptone) with 2% dextrose (YPD), 2% raffinose (YPR), or 2% galactose (YPG). Standard yeast manipulations including transformation, tetrad dissection, and PCR-based integration were performed as described in Amberg et al. (2005).

All plasmids and their derivation are listed in Table S2. To generate PLPMR2, the GFP and mRFP genes encoded in pKV015 (a gift from F. van Leeuwen, Netherlands Cancer Institute, Amsterdam, Netherlands; Verzijlbergen et al., 2010) were sequentially replaced by mCherry and GFP genes, respectively. ORFs encoding NSP1 and NIC96 with their cognate promoters and 3′ UTRs were inserted into pRS416 (Sikorski and Hieter, 1989) to generate PLPC19 and PLPC20. pKW2329 and pKW1358 were gifts from K. Weis (University of California, Berkeley, Berkeley, CA).

To arrest cells in different phases of the cell cycle, strains were grown to mid-log phase. For G1 arrests, α-factor (Keck Biotechnology Resource Laboratory) was added to a final concentration of 50 µg/ml for 2 h. For S-phase arrest, hydroxyurea (Sigma-Aldrich) was added to the growth medium at a final concentration of 0.2 M for 3 h. G2 arrests were achieved by overexpressing the Swe1 kinase as described in Booher et al. (1993). Strains containing SWE1 behind the control of the GAL1 promoter were grown in YPR to mid-log phase and were shifted to YPG for 3 h. In all cases, cell cycle arrests were confirmed visually by microscopy.

For experiments using strains containing the cdc6-1 allele, cells were grown at RT to mid-log phase and arrested in G2/M by shifting to the nonpermissive temperature of 34°C for 3 h. Cells were then placed on an agarose pad and imaged at RT.

For all imaging experiments, cells were grown to mid-log phase and immobilized on a 1.4% agarose pad containing complete synthetic medium (CSM) with 2% glucose and sealed with VALAP (1:1:1 Vaseline/lanolin/paraffin). All the microscopy experiments were carried out on a wide-field deconvolution microscope (DeltaVision; Applied Precision/GE Healthcare) equipped with a 100×, 1.40 NA objective lens and solid state illumination. The images were acquired using a charge-coupled device (CCD) camera (CoolSNAP HQ²; Photometrics) or an Evolve EMCCD camera (Photometrics). Temperature control was achieved through the enclosure of the microscope within an environmental chamber. In all cases, a z series of images (0.2–0.5-μm sections) were acquired and further processed as described under “Image processing and analysis.” Time-lapse series were performed with different time intervals as described in the figure legends.

Cells grown to mid-log phase were collected by centrifugation and fixed in 2% PFA in PBS for 20 min at RT. Cells were subsequently washed in PBS and cell walls were permeabilized by an incubation for 30 min in sorbitol citrate buffer (0.1 M K₂HPO₄, 0.04 M citric acid, 1.2 M sorbitol, and 0.5 mM MgCl₂, pH 6.5) containing 0.02% zymolyase 100T (MP Biomedicals). Cells were washed in sorbitol citrate and transferred to a slide pre-coated with 0.1% poly-l-lysine (Electron Microscopy Sciences). The slide was incubated for 4 min in methanol (−20°C), followed by 4 min in acetone (−20°C). Slides were air dried. The processed cells were then blocked for 20 min with PBS containing 1% BSA and 0.1% Triton X-100 before incubation with an anti-Nsp1 primary antibody (32D6; Abcam) followed by a Alexa Fluor 594–conjugated secondary antibody (Invitrogen).

In all images shown, a z series was deconvolved using the iterative algorithm in softWoRx (version 5.5; Applied Precision). In some cases (indicated in the figure legends), maximum-intensity projections were also generated by softWoRx. Subsequent image processing and analysis was performed using Fiji/ImageJ (Schindelin et al., 2012). To measure tf in daughter and mother cells (tf_d and tf_m), a z stack was sum-projected and background subtracted, and the integrated density of a region of interest (ROI) encompassing the mother or daughter NE was calculated. As an indirect measure of NPC density, we calculated the mfi over an ROI that surrounded only the NE in mother or daughter cells of one middle z plane. To calculate nuclear volume, we measured the ferret diameter of the nucleus through a middle z section. An approximate nuclear volume was calculated by considering that the nucleus was a sphere using the equation V = 4/3 πr³. All plots and statistical analysis were performed using Prism software (GraphPad Software).

Nup-RITE–containing strains were grown to mid-log phase in YPD or YPR in the presence of 300 µg/ml hygromycin (Invitrogen) to maintain selection for mCherry expression (Verzijlbergen et al., 2010). Cells were synchronized in G1 with α-factor, washed in YPD or YPR, and released for 1 h in YPD or YPG containing 1 µM estradiol before imaging on an agarose pad.

For plate-based growth assays, cells were grown overnight and plated in 10-fold serial dilutions on YPD plates containing DMSO or rapamycin (EMD Millipore) at a concentration of 1 µg/ml. Growth was assessed after 48 h at 30°C. For microscopy, mid-log phase cells expressing Nup-FRB-GFP fusions were grown in YPD and placed onto agarose pads containing DMSO, rapamycin (10 µg/ml), or rapamycin and cycloheximide (10 µg/ml; VWR International). Cells expressing Nup133-2xDendra were exposed to seven 0.5-s pulses of UV light to photoconvert Dendra from its green to red fluorescent form. In all experiments, cells that were unbudded or small budded were imaged through at least one cell cycle. For long-term viability/growth assays, cells were transferred to a microfluidic plate (Y04C/CellASIC) within the ONIX microfluidic platform (CellASIC) inserted onto the microscope stage. YPD containing DMSO or rapamycin (10 µg/ml) was perfused into the microfluidic plate at a pressure of 2 psi for 24 h.

For SILAC labeling, lys2Δ (PCCPL314) and lys2Δ NSP1-GFP (PCCPL322) yeast cells were grown in CSM-Lys. Lysine (Sigma-Aldrich) or “heavy” [¹³C₆/¹⁵N₂]l-lysine (Cambridge Isotope labs) was added to a final concentration of 30 mg/liter. 100 OD of cells were harvested by centrifugation and resuspended in 500 µl of lysis buffer (150 mM KOAc, 20 mM Hepes, pH 7.4, 10% glycerol, and complete protease inhibitor cocktail [Roche]). 500 µl of zirconia beads (0.1 mm in diameter; BioSpec Products Inc.) were added and cells were lysed using a bead mill (Retsch) running at 30 Hz for 5 min. The beads were removed by centrifugation, and Triton X-100 was added to a final concentration of 1%. After a 30-min incubation at 4°C, the lysate was clarified by a 10-min centrifugation at 1,000 g. Equivalent amounts of PCCPL329 and PCCPL322 lysates were incubated (separately) with GFP-Trap agarose beads (Allele Biotechnology) for 30 min at 4°C. After washing, the beads were combined, and 100 µl of 8 M urea and 50 mM Tris-HCl, pH 7.4, was added. Bound proteins were reduced, alkylated, and digested with the endoproteinase LysC directly on beads. The resulting peptide mixture was removed from the beads, acidified with trifluoroacetic acid, and desalted using stop-and-go extraction tips (StageTips; Rappsilber et al., 2003), which are pipette tips packed with C18 solid phase chromatography resin (3M). Peptides were washed with 0.1% formic acid, eluted with 60 µl of 80% acetonitrile and 0.1% formic acid, and reduced to 5 µl in a speed vacuum centrifuge (Eppendorf). The peptide mixture was subjected to reversed phase chromatography on an Easy nLC system (Thermo Fisher Scientific) using a 50-cm column (New Objective) with an inner diameter of 75 µm, packed in-house with 3 µm C18 resin (Dr. Maisch GmbH). Peptides were eluted with an acetonitrile gradient (5–30%) and directly electrosprayed into a mass spectrometer (LTQ Orbitrap Velos; Thermo Fisher Scientific; Aguilar et al., 2010). Data were acquired with Xcalibur software (version 2.1; Thermo Fisher Scientific) and analyzed with MaxQuant (version 1.2.2.5; Cox and Mann, 2008; Cox et al., 2011) as described previously (Fröhlich et al., 2013). In brief, peak lists were searched against a local database for S. cerevisiae (obtained from Saccharomyces Genome Database). Maximum allowed mass deviation for tandem mass spectrometry (MS/MS) peaks and missed cleavages were 20 ppm and 3, respectively. Maximum false discovery rates were 0.01 both on the peptide and on the protein levels. Minimum required peptide length was six residues. Proteins with at least two peptides were considered identified. All calculations and plots were performed with the R software package.

Exponentially growing cells were collected by centrifugation and lysed in 0.1 M NaOH for 5 min at RT. The pellet was resuspended in SDS-PAGE sample buffer, and proteins were separated on SDS-PAGE gels and transferred to a nitrocellulose membrane (Bio-Rad Laboratories). We used anti-Nsp1 antibodies (32D6; Abcam) followed by HRP-conjugated secondary antibodies and ECL for detection.

Cells were fixed in PBS containing 4% PFA and 0.1% glutaraldehyde for 1 h at RT and embedded in 10% gelatin. They were subsequently infiltrated with 2.3 M sucrose at 4°C overnight and frozen in liquid nitrogen. Thin sections (∼60 nm) were cut with a cryo-ultramicrotome (EM FC6; Leica). Immunolabeling was carried out with the anti-Nsp1 primary antibody for 30 min followed by 10 nm of gold-conjugated secondary antibody for an additional 30 min. Grids were visualized with a transmission electron microscope (Tecnai Biotwin; FEI) at 80 kV. Images were taken using a CCD camera (Morada) fitted with iTEM software (both from Olympus).

Fig. S1 demonstrates that Nup-mCherry levels do not increase after RITE is induced. Fig. S2 presents multiple experiments confirming the functionality of the Nsp1-GFP protein. Fig. S3 shows the growth of strains containing Nup-FRB-GFP alleles. Fig. S4 is the quantification of the change of daughter nuclear size upon trapping of Nsp1-FRB-GFP. Fig. S5 examines colocalization between Nsp1-GFP_CYT and several components of the polarity apparatus and ER. In Video 1, the production and NE integration of Nsp1-GFP_CYT is shown during mitosis. Video 2 is a RITE experiment performed under G2 arrest showing the specific bud localization of only Nsp1-GFP and not Nsp1-mCherry. Video 3 shows the distribution of green and red versions of Nup85-RITE during G2 arrest. In Video 4, the interaction between an Nsp1-GFP_CYT focus with an ER tubule is highlighted. Videos 5–7 demonstrate the physical connection between the Nsp1-GFP_CYT foci and NP. Video 8 shows the absence of a cytoplasmic pool of Nup170-GFP during NP formation. Online supplemental information is available.

We are indebted to the generous support from the G. Harold and Leila Y. Mathers Foundation. Thank you to Xinran Liu and Morven Graham for their electron microscopy expertise, T. Walther for use of his mass spectrometer, and C. Vogel for image analysis. We are grateful for the reagents sent by F. van Leeuwen, B. Stillman, L. Weisman, A. Bretscher, D. Lew, and K. Weis in addition to helpful criticisms of the manuscript from M.C. King.

Funding was also provided by the National Institutes of Health, grant R01GM105672 to C.P. Lusk.