Unexplored Cdc42 functions at the budding yeast nucleus suggested by subcellular localization

ABSTRACT

In budding yeast, the Rho-family GTPase Cdc42 has several functions that depend on its subcellular localization and the cell cycle stage. During bud formation, Cdc42 localizes to the plasma membrane at the bud tip and bud neck where it carries out functions in actin polymerization, spindle positioning, and exocytosis to ensure proper polarity development. Recent live-cell imaging analysis revealed a novel localization of Cdc42 to a discrete intracellular focus associated with the vacuole and nuclear envelope. The discovery of this novel Cdc42 localization led to the identification of a new function in ESCRT-mediated nuclear envelope sealing. However, other aspects of this intracellular localization and its functional implications were not explored. Here, we further characterize the Cdc42 focus and present several novel observations that suggest possible additional Cdc42 functions at the nucleus, including nucleus-vacuole junction formation, nuclear envelope tethering, nuclear migration, and nucleopodia formation.

KEYWORDS:

• CDC42
• vacuole
• nucleus-vacuole junction
• nucleoporins
• nucleopodia

INTRODUCTION

Cdc42 is a Rho-family GTPase that is a key regulator of polarity development in essentially all eukaryotic cell types [1]. It has well-established functions in actin polymerization [2,3], spindle positioning [4–6], and exocytosis [7,8]. In budding yeast, Cdc42 regulates cytoskeletal rearrangement, membrane remodelling, and septin organization to ensure proper polarized cell growth [9]. Cdc42 was thought to function exclusively at the plasma membrane in budding yeast until a recent study provided evidence for a novel Cdc42 function at the nucleus, where the small GTPase functions with ESCRT-III proteins to seal holes in the nuclear envelope [10].

The discovery of the Cdc42 nuclear envelope sealing function was engendered by an internal fluorescent protein fusion allele of CDC42 that allowed live-cell imaging of Cdc42 in budding yeast [11]. While the aforementioned study focused on establishing a previously unrecognized function of Cdc42 in nuclear envelope sealing, it did not address other novel aspects of Cdc42 localization and the functional implications. Here, we present several novel live-cell imaging observations that suggest possible additional Cdc42 functions at the nucleus.

RESULTS

Cdc42 appears to associate with a membrane-bound compartment that segregates into the bud with the vacuole

We previously reported the novel subcellular localization of Cdc42 as a peri-vacuolar focus that sometimes overlaps with the nuclear ER [10]. In this previous study, while we elucidated a novel function for Cdc42, we did not fully investigate the nature of the discrete Cdc42 focus. To determine whether this Cdc42 spot is associated with a membrane-bound body, we first stained cells endogenously expressing Cdc42-mCherry^SW with MDY-64, a green lipophilic dye that stains the plasma membrane and intracellular membranes in fungi [12]. We found that MDY-64 stained the plasma membrane, vacuole membrane, and the Cdc42 spot, suggesting that the Cdc42 focus is associated with a membrane-bound compartment (Figure 1a,

Figure 1. Fluorescence microscopy reveals Cdc42 spot is within a membrane-bound compartment that segregates with the vacuole. (a) Fluorescence micrograph of 5 minute pulse of 10 µM MDY-64 (green) followed by immediate imaging of cells endogenously expressing Cdc42-mCherry^SW (magenta). (b) Montage from a movie of a cell endogenously expressing Cdc42mCherry^SW (magenta) and Vph1-GFP (green). (c) Fluorescence micrograph of wildtype cells (top) or vac17∆ cells (bottom) expressing Cdc42-mCherry^SW (magenta) and Vph1-GFP (green). (d) Quantification of wildtype or vac17∆ cells containing a Cdc42 spot in the mother cell or bud of budding cells. Wildtype: spots in mother 28.8 ± 1.7% (n = 842); vac17∆: spots in mother 15.3 ± 3.1% (n = 1139) . Time in minutes. All micrographs are maximum intensity projections of 2.5 µm at 0.5 µm steps. All scale bars 5 µm

Supplemental Figure S1a). The high intensity of the MDY-64 staining of the Cdc42 spot relative to the vacuole is likely because the body is a putative early compartment in the endosomal pathway. Because of the short 5 minute pulse (with no chase), at early timepoints, the concentration of the dye at the Cdc42 body should be higher relative to the concentrations of the dye at the vacuole, which is a terminal point for internalized dye. At later timepoints, the vacuole will eventually become saturated with the dye, and all membrane-bound bodies should have similar intensities of the dye.

Both FM4-64 and MDY-64 belong to a class of lipophilic styryl dyes that stain the plasma membrane, vacuole, and other membrane-bound intracellular organelles in budding yeast. The first publication describing FM4-64 contains several figures that show a spot on the vacuole stained by FM4-64 that resembles the Cdc42 spot we report [13]. The authors do not comment on this persistent peri-vacuolar spot, which apparently is not pathological as it is present in the wild-type strain, at normal growth conditions. Thus, it appears that this spot has been observed before, but may have been overlooked or dismissed as an artefact. In any case, this present study notes this finding that suggests that the Cdc42 spot is associated with a membrane-bound compartment.

While our imaging of fixed cells suggested that the spot is associated with a membrane-bound compartment, live-cell microscopy revealed that this spot is tightly associated with the vacuole. We had previously shown that the spot colocalizes with the vacuole at all times throughout the cell cycle [10]. Here, we further analysed the dynamic localization of the spot relative to the vacuole in cells expressing Vph1-GFP to mark the vacuolar membrane and Cdc42-mCherry^SW [14]. We observed that the spot segregates into the bud with the vacuole in budding cells, suggesting that it derives its movement from the same mechanisms that drive vacuole inheritance (Figure 1b, Movie 1, Supplemental Figure S2a, Supplemental Movie 1). To determine whether the Cdc42 spot drives vacuole segregation or if the Cdc42 spot is simply a passenger on segregating vacuoles, we examined the dynamics of Cdc42 spots in vac17∆ cells. vac17∆ cells fail to segregate vacuoles to daughter cells because vacuole inheritance involves the myosin-to-vacuole adapter protein Vac17p; vac17∆ budding cells thus contain buds devoid of vacuoles (Figure 1c) [15,16]. We reasoned that if the Cdc42 spot drives vacuole segregation, migration of the Cdc42 spot into the bud might occur even if vacuole inheritance is disrupted and might result in budded cells containing a Cdc42 spot in the bud but no vacuoles. vac17∆ cells expressing Cdc42-mCherry^SW displayed a significant decrease in the population of cells containing a Cdc42 spot, wherein 15.3 ± 3.1% (n = 1139) of vac17∆ cells had a Cdc42 spot compared to 28.8 ± 1.7% (n = 842) in a wildtype background. Interestingly, we were unable to find any budded cells containing a Cdc42 spot in the bud (Figure 1d), suggesting that the Cdc42 spot resides on a vacuole and is likely migrating into the bud along with the vacuole. Because vac17∆ cells eventually develop vacuoles through de novo vacuole biogenesis, we reasoned that the Cdc42 spots observed in mother cells formed in this way [17]. Together, these findings suggest that the Cdc42 spot marks an intracellular membrane-bound compartment that is tightly associated with the vacuole.

Cdc42 is at the site of nucleus-vacuole junction genesis

We had previously reported that in addition to invariably associating with the vacuole, the Cdc42 spot also overlaps with the nuclear ER in a cell-cycle-dependent manner. To test the possibility that the Cdc42 spot is at the nucleus-vacuole junction, we examined strains expressing Cdc42-mCherry^SW and Nvj1-GFP, which marks the interface between the nuclear and vacuolar membranes in budding yeast [18]. Similar to our observation of Cdc42 spot association with the nuclear ER, we found that Cdc42 overlaps with Nvj1 in a cell-cycle dependent manner. Cdc42 and Nvj1 colocalized in unbudded but not budded cells (Figure 2a, Supplemental Figure S1b). Live-cell imaging revealed that the colocalization of Cdc42 and Nvj1 in an unbudded cell is sustained throughout the duration of a 90 minute movie (Figure 2b, Movie 2, Supplemental Figure S1c). In budding cells, however, it appears that Cdc42 and Nvj1 never overlap in the mother cell, even transiently. Instead, they dynamically maintain their separation throughout an entire budding event

Figure 2. Cdc42 localizes to the NVJ in a cell-cycle dependent manner. (a) Still from a live-cell fluorescence microscopy movie of an unbudded cell (top) and a budded cell (bottom) endogenously expressing Cdc42-mCherry^SW and Nvj1-GFP (green). Representative images of unbudded cells displaying overlapping Cdc42 spot and Nvj1 signal (88.5 ± 14.1%, n = 31) and budded cells displaying no overlap of Cdc42 spot and Nvj1 signal (89.4 ± 10.4%, n = 55) (b) Montage of a movie of an unbudded cell endogenously expressing Cdc42-mCherry^SW (magenta) and Nvj1-GFP (green). (c) Montage of a movie of a budding cell endogenously expressing Cdc42-mCherrySW (magenta) and Nvj1-GFP (green) with Cdc42 spot in mother. (d) Montage from a movie of a cell endogenously expressing Cdc42-mCherry^SW (magenta) and Nvj1-GFP (green) with Cdc42 spot in the bud. Yellow arrow indicates the first frame where the Nvj1-GFP signal is observed. Representative images of cells with bud containing overlapping Cdc42 spot and Nvj1 signal (100 ± 0%, n = 8). All micrographs are maximum intensity projections of 2.5 µm at 0.5 µm steps. All scale bars 5um. All time in minutes

(Figure 2c, Movie 3).

Because Cdc42 migrates into the bud with the vacuole, and vacuole inheritance precedes nuclear inheritance [19], we were curious whether there is a spatial or temporal relationship between the Cdc42 spot and the formation of a new nucleus-vacuole junction in the bud. Live-cell imaging of a large-budded cell revealed that the bud contains a Cdc42 spot, but not an Nvj1-GFP signal, presumably because the nucleus has not yet migrated into the bud (Figures 2D, 0:00–0:06, Movie 4). Later in the cell cycle, we observed the appearance of an Nvj1-GFP signal in the bud at the Cdc42 spot, suggesting that the Cdc42 spot is at the site of nucleus-vacuole junction genesis (Figures 2D, 0:09, Movie 4, Supplemental Figure S1d). Together, these live-cell imaging observations suggest that the intracellular Cdc42 spot is the site of NVJ formation and remains associated with the NVJ until the start of the next cell cycle, when the Cdc42 spot presumably dissociates from the nucleus, and therefore the nucleus-vacuole junction.

Cdc42 directly interacts with Nup59 and overlaps with cytoplasmic Nups

We established that the Cdc42 spot colocalizes with the vacuole 100% of the time and overlaps with the nuclear envelope and nucleus-vacuole junction in a cell-cycle-dependent manner. We next sought to identify protein binding partners that might mediate the cell-cycle-dependent nuclear localization. In an effort to identify candidate cell-cycle dependent Cdc42-binding partners associated with the nuclear envelope or nucleus, we searched the sequences of nuclear proteins for motifs and domains found in proteins known to interact with Cdc42 or other Rho GTPases. We found that the nuclear pore complex component Nup59 contains a putative Cdc42- and Rac- Interactive Binding (CRIB) motif (Figure 3a), a 15–16 residue sequence that directly binds to the Switch I and II regions of Cdc42 and Rac GTPases [20,21]. The nucleoporins (Nups) are a group of ~30 proteins that form the nuclear pore complex, a large channel that spans the nuclear envelope to facilitate the exchange of components between the cytoplasm and nucleoplasm [22,23]. Nup59 is a soluble FG nucleoporin which has been mapped to the central core of the nuclear pore complex [22,23]. To determine whether Nup59 directly interacts with Cdc42, we assayed for direct binding using a GST-pulldown assay using purified proteins. We found that a GST-fusion of full-length Nup59 (GST-Nup59^FL)

Figure 3. Cdc42 interacts with nucleoporins in vivo and in vitro. (a) Multiple sequence alignment of CRIB motifs from several proteins and a similar sequence in Nup59 (black boxed sequence). Pink shaded letters denote sequence identity. Residues in unshaded pink boxes are similar to conserved residues. (b) GST-pulldown assay using purified GST-fusion proteins as prey proteins immobilized on glutathione agarose and purified soluble His₆-Cdc42 as bait. Presence of prey protein was detected by immunoblot against His₆ epitope tag. GST-Bem1 serves as a positive control for Cdc42 interaction and GST-Nsp1 serves as an additional negative control to GST alone. (c) Fluorescence micrographs of cells endogenously expressing Cdc42-mCherry^SW (magenta) and Nup59-GFP (green). Representative image of cell displaying overlapping Cdc42 spot and Nup59 signal (83.7%, n = 43). (d) Fluorescence micrographs of cells endogenously expressing Cdc42-mCherry^SW (magenta) and Nup159-GFP (green, top) or Nup82-GFP (green, bottom). Yellow arrows indicate enrichment of Nups on the nuclear envelope. Representative images of cell displaying overlapping Cdc42 spot and Nup159 signal (80%, n = 30) and overlapping Cdc42 spot and Nup82 signal (71.6 ± 1.6%, n = 55). (e) Fluorescence micrographs of a small budded cell (top) and large-budded cell (bottom) endogenously expressing Cdc42-mCherry^SW (magenta) and Nsp1-GFP (green). Yellow arrows indicate cytoplasmic puncta or NE enrichment of Nsp1-GFP. White arrow indicates Nsp1_CYT. Representative image of cell displaying overlapping Cdc42 spot and Nsp1 signal (72.8 ± 4.5%, n = 48). (f) Fluorescence micrographs of cells endogenously expressing Cdc42-mCherry^SW (magenta) and Nup188-GFP (green) in nup133∆ background. Top cell is unbudded. Bottom cell is budded. All micrographs are maximum intensity projections of 2.5 µm at 0.5 µm steps. All scale bars 5 µm

robustly pulled down Cdc42 (Figure 3b). We next tested whether the 50 residue portion of Nup59 containing the putative CRIB motif (GST-Nup59^CRIB) is sufficient for the full-length protein’s direct interaction with Cdc42 and found that it is not (Supplemental Figure S3a). Furthermore, the CRIB domain does not appear to be required for the interaction with Cdc42 as an alanine mutation of the putative Nup59 CRIB motif did not disrupt the Cdc42 association (Supplemental Figure S3b). Thus, although Nup59 contains a putative CRIB motif, it appears that Nup59 interacts with Cdc42 in a region outside of the identified CRIB motif.

To look for evidence that the direct interaction between Cdc42 and Nup59 we detected in vitro is reflected in vivo, we examined by live-cell fluorescence microscopy yeast endogenously expressing Cdc42-mCherry^SW and Nup59-GFP. We noted overlapping fluorescence signals between Cdc42 and Nup59 at the nuclear envelope (Figure 3c). Because live-cell fluorescence imaging and 3D projections of confocal images (Supplemental Movie 2) suggest that the Cdc42 spot is on the cytoplasmic face of the nuclear envelope, we also examined nucleoporins on the cytoplasmic face of the nuclear envelope. We performed fluorescence microscopy on cells endogenously expressing Cdc42-mCherry^SW and various cytoplasmic Nups C-terminally tagged with GFP. We observed colocalization of Cdc42-mCherry^SW with Nup159-GFP, Nup82-GFP, and Nsp1-GFP (Figure 3d, E, Supplemental Figure S1e-h). Interestingly, we observed that the Cdc42 spot often overlaps in signal with apparent aggregates and punctae of Nup82 and Nup159 (Figure 3d, Supplemental Figure S1e, f). We also observed a similar localization pattern for Cdc42 and Nsp1, a Nup that notably forms cytoplasmic punctae referred to as Nsp1_CYT, an NPC (Nuclear Pore Complex) subcomplex whose inheritance by the bud is required for the transmission of nuclear pore complexes from mother to bud during cell division [24,25]. While we did observe Nsp1_CYT foci, the Cdc42 spot did not colocalize with this Nsp1 subcomplex, but instead localized to enrichments of Nsp1 on the nuclear envelope, as well as with cytoplasmic Nsp1 punctae that were retained in the mother cell, and therefore appear to be distinct from Nsp1_CYT (Figure 3e, Supplemental Figure S1g, h).

Because nucleoporin aggregates are indications of aberrant NPC assembly, we wanted to determine if the Cdc42 spot localizes to newly assembled NPCs or is involved in their biogenesis [26–30]. We examined Cdc42 localization in nup133∆ cells, a genetic background in which nucleoporins form morphologically distinct clusters consistent with a block in the assembly of all NPCs [26]. Cellular components involved in NPC biogenesis will unambiguously colocalize with the halted NPC assemblies which manifest as distinct clusters, providing a microscopy assay to determine whether or not a protein of interest is likely involved in NPC biogenesis [31]. We did not see any correlation between the location of the Cdc42 spot and the location of NPC clusters (visualized with Nup188-GFP) that would suggest Cdc42’s involvement in NPC biogenesis (figure 3f). Thus, the enriched nucleoporins we observe near or overlapping with the Cdc42 spot may represent novel structures whose functions are yet to be uncovered. In summary, our biochemical experiments show that the predicted CRIB motif of Nup59 directly binds to Cdc42 and our live-cell fluorescence microscopy shows that, though Cdc42 is not likely to be involved in NPC assembly, the Cdc42 spot frequently appears proximal to enrichments, aggregates, and punctae of various Nups in vivo.

The Cdc42 spot comigrates with an inherited Ndc1 focus

Nup59 is a soluble nucleoporin that is a member of a nucleoporin network composed of Ndc1, Nup53, Nup157, and Nup170. This subcomplex functions in NPC biogenesis and maintenance of NPC structural integrity [25]. Ndc1 is an essential integral transmembrane nucleoporin that localizes to both the NPC and the spindle pole body (SPB), a structure that spans the double membrane of the nuclear envelope [32]. Because Ndc1 directly binds to Nup59²⁵, which directly binds to Cdc42 (Figure 3b), we wondered whether Cdc42 has any localization relationship with Ndc1 in vivo. We observed by live-cell microscopy that the Cdc42 spot and Ndc1-GFP overlap on the nuclear envelope (Figure 4a, Movie 5).

Figure 4. A Cdc42 spot localizes to the distal end of nucleopodia. (a) Montage of a movie of a cell endogenously expressing Cdc42-mCherry^SW (magenta) and Ndc1-GFP (green). The contrast was enhanced in the asterisked micrograph to more clearly show the nuclear envelope tether (yellow arrowhead). White arrowheads indicate Ndc1-GFP foci at the spindle pole bodies. Cdc42 spot overlap with Ndc1 signal observed 90.5 ± 6.7%, n = 53 (b) Micrographs of cells endogenously expressing Cdc42-mCherry^SW (magenta) and GFP-HDEL (green) at three hours of hydroxyurea treatment (top panels) or after a one-hour release following three-hour hydroxyurea treatment (bottom panels). (c) Micrograph of cell endogenously expressing Cdc42-mCherry^SW (magenta) and GFP-HDEL (green) containing Cdc42 spot in the bud proximal to nucleopodium (yellow arrowhead). White arrow indicates Cdc42 spot. Observed 100% of the time, n = 3. (d) Micrograph of cdc23-1 cells endogenously expressing Cdc42-mCherry^SW (magenta) and Nsp1-GFP (green) grown at the restrictive temperature of 37°C for three hours (top) or after one-hour of growth at the permissive temperature of 25°C for one hour following three hours of growth at the restrictive temperature of 37°C (bottom). White arrowheads indicate Cdc42 spots. (e) Micrograph of cdc23-1 cells endogenously expressing Cdc42-mCherry^SW (magenta) and Nsp1-GFP (green) containing Cdc42 spot in the bud (white arrowhead) grown at the restrictive temperature of 37°C for three hours, then released into the permissive temperature of 25°C for one hour. Yellow arrowhead indicates nucleopodium. All micrographs are maximum intensity projections of 2.5 µm at 0.5 µm steps. All scale bars 5 µm

Interestingly, in a budding cell undergoing transport of the Cdc42 spot to the bud, we observed what appeared to be a tether of Ndc1-GFP-decorated nuclear envelope trailing behind the Cdc42 spot as it migrated into the bud. Intriguingly, once in the bud, the Cdc42 spot was met by the inherited Ndc1 punctum associated with the SPB. Thus, the Cdc42 spot briefly overlaps with the SPB at the end of nuclear migration, where it appears to be a landmark for the inherited SPB. This very rare observation shows that Cdc42 colocalizes with Ndc1 in vivo, specifically the Ndc1 focus that is inherited by the bud, suggesting that nucleoporins might provide a link between the Cdc42 spot and the nuclear envelope.

The Cdc42 spot localizes to the distal tip of nucleopodia

We previously reported that the intracellular Cdc42 spot localizes to the nuclear ER in the mother cell of a dividing cell in a cell-cycle dependent manner [10]. Here, we report that the nuclear envelope is ephemerally connected to the Cdc42 spot during migration of the nuclear envelope into the bud (Figure 4a, Movie 5). Together, these observations suggest that there is a relationship between the Cdc42 spot and the nuclear envelope during nuclear migration. We therefore determined whether the Cdc42 spot associates with the nuclear envelope during transmission into the bud.

Because the emerging bud is small, it is often difficult to resolve the spatial organization of the segregated organelles within the emerging bud. To unambiguously assess the spatial distribution of both the Cdc42 spot and the nuclear envelope within the bud, we uncoupled mitosis from polarized growth to induce a nuclear migration delay, such that nuclear migration occurs when the bud is larger than normal. In cells arrested in S phase or mitosis with drugs such as Hydroxyurea [33], or with temperature-sensitive alleles of cell cycle genes such as cdc23-1 [34], the uncoupling of polarized growth and nuclear division results in mononucleated, large-budded cells. We reasoned that upon release into permissive conditions, when nuclear division resumes, the larger bud will allow us to more easily observe the position of the nucleus within the bud during nuclear migration.

Cells endogenously expressing Cdc42-mCherry^SW and GFP-HDEL were treated with 0.4 M hydroxyurea to induce arrest as large-budded cells with a single nucleus juxtaposed with the mother-side of the bud neck (Figure 4b, top). One hour after release into fresh imaging media free of hydroxyurea, the cells began to resume nuclear division, with many exhibiting peanut-shaped nuclei spanning the budneck (Figure 4b, bottom). Interestingly, we observed a handful of cells harbouring a single nucleus adjacent to the mother bud neck with ER membrane emanating from the mother nucleus into the bud (Figure 4c, Supplemental Figure S2b). We suspect this ER protrusion to be a nucleopodium, a term for nuclear protrusions devoid of DNA that are thought to play a role in spindle positioning and nuclear migration in budding yeast [35]. Among the few cells displaying a putative nucleopodium, we observed rare cells (n = 3) containing a Cdc42 spot in the bud which appeared to localize to the tip of the putative nucleopodium, which appeared to be sampling the bud space (Figure 4c, Supplemental Figure S2b). Live-cell imaging revealed that the putative nucleopodium is very dynamic and tracks with the Cdc42 spot in the bud (Movie 6, Supplemental Movie 3).

To determine if the ER protrusions we observed are nucleopodia, we performed a similar cell-cycle arrest experiment using cells expressing the nucleoporin Nsp1-GFP so that we can distinguish the nuclear envelope from other ER membranes. Cells endogenously expressing Cdc42-mCherry^SW and Nsp1-GFP in the metaphase-arrest cdc23-1 mutant background were grown at the restrictive temperature of 37°C for 3 hours, then released into the permissive temperature of 25°C and visualized by live-cell microscopy. Like the hydroxyurea-arrested cells, we observed cdc23-1 mutants to arrest as mononucleated, large-budded cells (Figure 4d, top). Upon release into permissive growth conditions, cells slowly resumed nuclear division, displaying peanut-shaped nuclei that spanned the bud neck (Figure 4d, bottom). Similar to the HU-arrest/release cells, we also observed the formation of nucleopodia extensions into the bud in several cdc23-1 cells upon release into permissive growth conditions (Figure 4e). Interestingly, many of the Cdc42 spots were very broad and diffuse (Figure 4d). In cells that contained a discrete Cdc42 focus in the bud, we observed instances where the Cdc42 spot overlapped with the nucleopodia extending into the bud (Figure 4e). Together with the hydroxyurea experiments, the cdc23-1 cell-cycle arrest experiments suggest that the Cdc42 spot in the bud might be associated with the nuclear envelope.

DISCUSSION

Here, we report several novel live-cell imaging observations that hint at new Cdc42 functions at the nucleus.

Light microscopy analysis of MDY-64 staining suggests that the Cdc42 body contains membrane that originated from the plasma membrane. Our previous studies found that the spot localizes to the vertices of fusing vacuoles and that Cdc42 genetically and physically interacts with several members of the ESCRT machinery [10]. Together with the MDY-64 analysis from this study, these cumulative observations support the idea that this body may be an endosomal compartment that fuses with the vacuole.

Live-cell imaging of the Cdc42 spot and the nucleus-vacuole junction revealed a surprising observation that suggests possible new Cdc42 functions at the nucleus. We previously found that the Cdc42 spot localizes to both the vacuole and nucleus, implying that Cdc42 is at the nucleus-vacuole junction. Our present study established that the Cdc42 spot colocalizes with Nvj1, a nucleus-vacuole junction protein involved in the formation and maintenance of the inter-organelle contact site [18]. As expected, we found that Cdc42 colocalizes with Nvj1 in a cell-cycle dependent manner, mirroring the cell-cycle dependent localization of the Cdc42 spot to the nuclear envelope we previously reported [10] (Figure 2a – D). Interestingly, we observed that the genesis of the nascent nucleus-vacuole junction in the bud occurs at the Cdc42 spot (Figure 2d). This observation suggests that the Cdc42 spot may be involved in formation of the nucleus-vacuole junction. This interesting finding warrants further investigation into a possible novel Cdc42 function in formation of the contact site between the nucleus and vacuole in budding yeast.

Additional clues to possible new Cdc42 functions at the nucleus emerged from live-cell image analysis of Cdc42 and various nucleopore complex proteins. We found that Cdc42 interacts with Nup59 and Ndc1, members of the Nup157-Nup170-Nup53-Nup59-Ndc1 interaction network essential for NPC biogenesis [25,36–40]. Additionally, we observed the Cdc42 spot near punctae containing many of the nups on the cytoplasmic face of the nuclear envelope, including Nup82, Nup159, and Nsp1 (Figure 3d, E, Supplemental information), which are suggestive of possible NPC assembly defects [26–30]. When we examined whether the Cdc42 spot localized to clusters of halted NPC assemblies in the nup133∆ genetic background, which produces NPC assembly defects, however, we found no overlapping fluorescence signal between Cdc42 and NPC clusters (figure 3f), suggesting that the Cdc42 spot is not involved in NPC biogenesis.

While it is unlikely that Cdc42 is involved in NPC biogenesis, Cdc42 might interact with the Nup157-Nup170-Nup53-Nup59-Ndc1 interaction network to establish a connection with the nuclear envelope. We report several observations that suggest that the Cdc42 spot is physically connected to the nuclear envelope. Thus, the Nup157-Nup170-Nup53-Nup59-Ndc1 interaction network may provide a physical link between Cdc42 and the nuclear envelope. In cells arrested in mitosis, we observed a sustained connection between the distal tip of the nucleopodium extending from the mother nucleus and the Cdc42 spot in the bud (Figure 4c, E, Supplemental Figure S2b). However, in cells dividing normally, we observed what appeared to be the untethering of the nuclear envelope (visualized by Ndc1-GFP) from the Cdc42 spot as it dissociated from the nucleus to segregate into the bud (Figure 4a). Interestingly, we found that towards the end of nuclear migration, the Cdc42 spot colocalizes specifically with the segregated Ndc1 punctum, a structure that has previously been shown to be associated with the SPB [32,41]. The Mitotic Exit Network (MEN) also associates with the spindle pole body that is segregated into the bud, suggesting a possible association with Cdc42 [42,43,44]. These combined observations suggest that the Cdc42 spot is reversibly associated with the nuclear envelope, possibly through the Nup157-Nup170-Nup53-Nup59-Ndc1 interaction network, and that the connection between the spot and the nucleus may be cell-cycle-dependent.

In sum, here we shared several novel observations that emerged from live-cell imaging and biochemical analyses. The intriguing observations we report here warrant functional investigation into possible Cdc42 roles in MVB function, formation of the nucleus-vacuole junction, and nuclear envelope tethering. Adding to the already previously described function of the Cdc42 body in nuclear envelope sealing, the intracellular Cdc42 spot appears to have several functions within the cell depending on cell cycle stage (Figure 5). As a molecular switch, Cdc42 appears poised to regulate any of several functions depending on cell cycle stage, which aligns with our observations of cell-cycle-dependence on the subcellular localization of the Cdc42 spot.

Figure 5. Schematic depicting Cdc42 spot throughout cell cycle progression

(1) Throughout the cell cycle, the Cdc42 spot (red) remains associated with the vacuole (orange) and functions at the nucleus (green) to seal holes formed in the nuclear envelope (blue) after ER (blue) tubule fission or nuclear envelope fission. At start, Cdc42 marks the site of bud emergence on the plasma membrane. (2) Early in the cell cycle, the ER is inherited by the bud and Cdc42 is localized to the expanding bud cortex. In budding cells, the intracellular Cdc42 spot can be observed proximal to, but not colocalized with, the nucleus vacuole junction (orange/blue dashed). (3) Bud inheritance of the Cdc42 spot occurs during vacuole transmission as the spot is tightly associated with the vacuole membrane. As the Cdc42 spot migrates to the bud, the nuclear envelope can be seen tethered to the Cdc42 spot. (4) During spindle elongation and concomitant nuclear migration, the Cdc42 spot in the bud appears to be a landmark for the migrating spindle pole body (blue spot). (5) After nuclear migration, the inherited spindle pole body is observed to colocalize with the Cdc42 spot in the bud after spindle elongation. The Cdc42 spot in the bud is at the site of nascent nucleus-vacuole-junction formation. Cdc42 relocalizes from the bud cortex to the bud neck as cytokinesis marks the end of the cell cycle. (6) New daughter cells exhibit the Cdc42 spot colocalized with the nucleus-vacuole-junction.

Author contributions

M.S. Lu and D.G. Drubin conceived of the experiments. M.S. Lu generated the reagents, performed the experiments, and analysed the data. M.S. Lu and D.G. Drubin wrote the manuscript. D.G. Drubin secured funding.

Acknowledgments

This work was supported by the National Institutes of Health grant no. R35GM118149 to D.G. Drubin. The authors declare no competing financial interests.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Supplementary material

Supplemental data for this article can be accessed here

Additional information

Funding

This work was supported by the National Institute of General Medical Sciences [R35GM118149].