The microtubule-associated protein She1 coordinates directional spindle positioning by spatially restricting dynein activity Free

By transporting various cargoes along microtubules, the dynein and kinesin families of molecular motors play important roles in many cellular processes, including coordinating the spatially and temporally appropriate positions of membrane-bound vesicles, organelles and the mitotic spindle. In addition to being affected by various motor-specific accessories and regulators, microtubule motors are regulated by a family of microtubule-associated proteins (MAPs). For instance, the neuronal MAP tau, well known for its implication in Alzheimer's disease (Iqbal et al., 2010), inhibits kinesin-1-based transport of vesicles and organelles in neurons in vivo, and induces detachment and pausing behavior of kinesin-1 and dynein in vitro (Dixit et al., 2008; Seitz et al., 2002; Tan et al., 2019; Trinczek et al., 1999). In addition to tau, several other MAPs, including MAP4 and MAP9 affect dynein transport functions in vivo and in vitro (Monroy et al., 2020; Samora et al., 2011; Semenova et al., 2014). Although the mechanisms by which MAPs affect in vitro motor motility are beginning to be understood (Ecklund et al., 2017; Monroy et al., 2018, 2020; Tan et al., 2019), how their activities result in the appropriate positioning of the various dynein and kinesin cargoes in cells is unclear.

Sensitive to high expression-1 (She1) is a yeast-specific MAP that has been shown to potently affect dynein motility in vitro and dynein-mediated spindle positioning in vivo (Markus et al., 2012b, 2011; Woodruff et al., 2009). Although the precise mechanism by which She1 promotes appropriate in vivo dynein activity is unclear, our recent in vitro data determined that She1 affects dynein motility through simultaneous interactions with the dynein microtubule-binding domain (MTBD) and the microtubule (Ecklund et al., 2017). In addition to reducing dynein velocity, She1 reduces the dynein–microtubule dissociation rate (Ecklund et al., 2017), suggesting that it might promote dynein–microtubule interaction in cells and potentially affect dynein force production. Given that dynein functions in concert with the cortical receptor Num1 (potentially orthologous to NuMA in humans) (Greenberg et al., 2018; Heil-Chapdelaine et al., 2000) and the dynactin complex in cells (dynein activator) (Adames and Cooper, 2000; Geiser et al., 1997; Lee et al., 2003; Moore et al., 2008), it is unclear whether these prior findings with dynein alone apply to active Num1–dynein–dynactin complexes. Thus, a clear picture of how this MAP affects dynein-mediated spindle positioning is lacking.

She1 is one of several known effectors of budding yeast dynein activity, which also include Ndl1 (Nde1 in humans) (Li et al., 2005), Bik1 (CLIP170 in humans) (Sheeman et al., 2003) and Pac1 (LIS1 in humans) (Lee et al., 2003). In contrast to higher eukaryotes, the only known function for dynein in budding yeast is to position the nucleus with the enclosed mitotic spindle at the future site of cytokinesis (Eshel et al., 1993; Li et al., 1993), the narrow neck between the mother and daughter cells. Dynein performs this activity from cortical Num1 sites, to which it is delivered by a multi-step ‘offloading’ mechanism (Lee et al., 2005, 2003; Markus and Lee, 2011; Sheeman et al., 2003). Specifically, (1) dynein indirectly associates with the plus ends of dynamic microtubules in a Pac1- and Bik1-dependent manner (Lammers and Markus, 2015; Lee et al., 2003; Markus et al., 2009; Sheeman et al., 2003); (2) the dynactin complex is recruited to plus end-bound dynein (Moore et al., 2008); and (3) upon encountering cortical Num1, dynein and dynactin are offloaded and activated to translocate the nucleus and spindle (Lammers and Markus, 2015; Markus and Lee, 2011). Although the offloading process appears to be biased towards the daughter cell (Markus and Lee, 2011) – the likely result of regulated asymmetric recruitment to microtubule plus ends by Kip2, various kinases and B-type cyclins (Chen et al., 2019; Grava et al., 2006) – it is unclear whether this is sufficient to promote directionally tuned dynein activity in cells, or whether other molecules play a role in this process.

Our recent studies discovered that She1 is a key effector that polarizes dynein-mediated spindle positioning. Specifically, we found that deletion of She1 compromises dynein-mediated spindle translocation events that lead to the spindle crossing the bud neck, and results in a higher prevalence of anaphase onset within the mother cells, quite distal from the bud neck (Markus et al., 2012b). How She1 performs this function, however, is unknown. In this study, we employ a combination of in vitro and in vivo methods to determine the precise basis by which She1 affects dynein activity. In particular, we find that our previous in vitro data indicating a role for She1 in reducing dynein velocity does not likely reflect the basis by which it reduces in-cell dynein activity. Rather, we find that She1 supports dynein-mediated spindle positioning by attenuating the initiation of cortical dynein–dynactin-mediated spindle movements predominantly in the mother cell, and that this is potentially a consequence of She1 reducing the plus end localization of dynein–dynactin complexes in this compartment. We find that this activity requires She1 interactions with the microtubule-bound conformation of the dynein microtubule-binding domain as well as microtubules, partially reconciling our in vivo and in vitro data. Although unclear, our findings suggest that She1 may affect dynein localization and activity by promoting the autoinhibited ‘phi’ conformational state of dynein. Finally, we find that the likely basis for mother cell-specific inhibition of dynein activity is the asymmetric binding of She1 to astral microtubules within this compartment. In summary, our work describes how an asymmetrically localized MAP can spatially tune the activity of a molecular motor to coordinate appropriate cargo transport.

We previously found that She1 is important for polarizing dynein-mediated spindle movements towards the daughter cell (Markus et al., 2012b). Consistent with its importance in spindle positioning, single time-point images acquired of cells grown at low temperatures (16°C; to exacerbate dynamic microtubule-mediated processes) revealed a mild spindle positioning defect (Ecklund et al., 2017). To gain additional insight into the role of She1 in dynein-mediated spindle positioning and cell cycle progression, we performed time lapse imaging of cells over the course of several cell cycles using a microfluidics-based platform (CellAsic ONIX, Millipore Sigma), in which cells are imaged at 30°C in the presence of constantly replenished nutrients. We imaged cells expressing Spc110–Venus, NLS–3mCherry and mTurquoise2–Tub1 to visualize spindle pole bodies (SPBs), the nucleus and microtubules, respectively. Consistent with the low-temperature single time-point assay (Ecklund et al., 2017), 17.7% of she1Δ cells (compared to only 2.5% of SHE1 cells) exhibited mispositioned spindles at anaphase onset, confirming the importance of She1 in spindle positioning (Fig. 1A,B).

Given that She1 is important for proper spindle positioning, we wondered whether she1Δ cells exhibited any cell cycle delays. Cells with mispositioned spindles trigger the spindle positioning checkpoint (SPC), which delays cytokinesis and mitotic exit by preventing activation of the mitotic exit network (MEN) (Bardin et al., 2000; Bloecher et al., 2000; Caydasi et al., 2010; Daum et al., 2000; Pereira et al., 2000; Yeh et al., 1995). We assessed several aspects of cell cycle progression by measuring the time intervals between various temporal landmarks (Fig. S1). This revealed an 11.0% increase in mitotic duration in she1Δ cells (Fig. 1C; see Fig. S1 and Materials and Methods) that was largely a consequence of a delay between anaphase onset and cytokinesis (Fig. 1E). This is likely reflective of the previously determined dynein-independent role for She1 in promoting spindle disassembly and thus mitotic exit (Woodruff et al., 2010). We observed no significant delay between SPB duplication and anaphase onset due to She1 deletion (Fig. 1D), suggesting that the timing of spindle assembly is unperturbed upon She1 deletion. Separate analysis of cells with mispositioned spindles and those with properly positioned spindles revealed a 13.6% delay between anaphase onset and cytokinesis in SHE1 cells (note the very small dataset due to the low prevalence of mispositioned spindles in wild-type cells), as would be expected given the presence of an intact SPC (Fig. 1F) (Bardin et al., 2000; Bloecher et al., 2000; Daum et al., 2000; Pereira et al., 2000). Although she1Δ cells exhibit no such delay, the daughter-bound SPB of the mispositioned spindles in almost all such cases immediately translocated into the daughter cell following anaphase, which is sufficient to silence the SPC, and promote mitotic exit.

We sought to clarify the role of She1 in dynein-mediated spindle positioning. In light of our previous finding that She1 enhances dynein–microtubule binding affinity (Ecklund et al., 2017), we hypothesized that She1 promotes daughter cell-directed dynein activity by enhancing its force generation capacity (by reducing dynein–microtubule dissociation). In this model, She1 assists dynein in generating sufficient force to pull the nucleus (with an enclosed spindle) through the narrow neck (about half the width of the nucleus; see Fig. 3A,B). If true, we predicted that loss of She1 would lead to scenarios where dynein-mediated spindle movements into the neck would be more prone to failure. To determine whether this is the case, we imaged dynein-mediated spindle movements in cells arrested in a metaphase-like state (‘preanaphase’) by treatment with the DNA synthesis inhibitor hydroxyurea (HU), which permits observations of many dynein-mediated spindle movements. To ensure that all spindle movements are dynein dependent, we performed these and all similar subsequent experiments in cells deleted for KAR9, a key member of an actomyosin-mediated spindle orientation pathway (Markus et al., 2012a; Miller and Rose, 1998), in which preanaphase spindles exhibit an oscillatory movement along the mother–daughter axis (Yeh et al., 2000). Dynein-mediated spindle movements that were directed toward the bud neck were scored based on whether they resulted in a successful or unsuccessful crossing of the neck (Fig. 1G). Whereas 61.3% of such spindle movements resulted in a neck cross in wild-type cells, this value was reduced to 27.1% in she1Δ cells (Fig. 1H), lending support to an assisted force model for She1 function.

To determine whether She1 affects dynein force generation, we employed optical trapping with an artificially dimerized dynein motor domain fragment that is affected by She1 in motility assays (Ecklund et al., 2017; Markus et al., 2012b) (Fig. 2A). The motility of single dynein motor-coated beads captured in an optical trap was measured in the absence or presence of 2 or 5 nM recombinant She1–HaloTag^TMR (She1–TMR), concentrations that approximate cellular levels (Markus et al., 2012b). As expected, pre-stall bead velocity decreased significantly in the presence of She1–TMR (Fig. 2B,C). Although there was no appreciable change in dynein stall force in the presence of She1, we observed a 4.6-fold increase in the mean stall time (Fig. 2B,D,E), indicating that She1 induces a persistent force generating state for dynein in vitro.

Our results indicate that She1 may improve the ability of dynein to remain bound to microtubules in the presence of an opposing force, which could potentially be provided by the nucleus being squeezed through the bud neck during nuclear migration. To test this in vivo, we employed live-cell imaging with cells expressing Nup133–3mCherry (to visualize the nucleus) and GFP–Tub1. HU-arrested cells were imaged, and we measured several parameters of dynein-mediated nuclear translocation through the bud neck, including the frequencies with which the nuclei cross the neck. Specifically, those events that were directed toward the bud neck were scored based on whether they resulted in a successful or unsuccessful insertion of the nucleus into the narrow neck (see Fig. 3C). In addition to performing this assessment in otherwise wild-type cells, we also employed cells deleted for BNI1, a formin that when deleted results in a widening of the bud neck (Gladfelter et al., 2005; Moore et al., 2009) but not the nucleus (Fig. 3B). We reasoned that if the neck provides a barrier over which dynein-mediated nuclear migration must overcome, and that She1 assists dynein in performing this function, then widening the neck may rescue she1Δ phenotypes. Although we noted a small, but statistically insignificant increase in the success rate for such events as a consequence of BNI1 deletion, the presence or absence of She1 has no apparent effect on this phenomenon (Fig. 3D). However, deletion of She1 causes a small reduction in the velocity of these movements, while combined deletion of She1 and Bni1 increases this value to levels greater than she1Δ and SHE1 cells (Fig. 3E). We also noted a small, statistically insignificant, decrease in the translocation distance in she1Δ cells, and a significant increase in this value for she1Δ bni1Δ cells (Fig. 3F), with no apparent change in the duration of these movements (Fig. 3G). These data suggest that She1 may indeed support dynein during these movements, and that widening the neck helps overcome a barrier to this translocation. However, given the lack of change in the translocation success rate, the basis by which She1 promotes dynein-mediated spindle and nuclear position remains unclear.

Previous studies have suggested that She1 reduces dynein activity in cells (Markus et al., 2012b, 2011), which, if compartmentally specified (i.e. to the mother cell), could have the capacity to polarize dynein-mediated spindle movements (i.e. to the daughter cell). Specifically, cells lacking She1 exhibit faster and more frequent dynein-mediated spindle movements, which are apparent from spindle movements coincident with astral microtubules sliding along the cell cortex (Bergman et al., 2012; Markus et al., 2012b, 2011; Woodruff et al., 2009). We sought to determine the basis by which She1 polarizes dynein activity by performing a highly detailed assessment of dynein-mediated spindle movements. HU-arrested cells were imaged, and dynein-mediated spindle movements were manually curated from automated spindle tracking data. We quantified various parameters of spindle movements from these data, including velocity, displacement per event and relative degree of activity (i.e. the number of events per minute, and the total dynein-mediated spindle displacement, cell⁻¹ min⁻¹). As expected, this analysis revealed that loss of She1 led to increases in spindle velocity, displacement and overall dynein activity (Fig. 4A–E).

To gain additional insight into the role of She1 in spindle positioning, we employed a custom written code that identifies the relative position of the spindle over time with respect to the boundaries of the cells (i.e. mother versus daughter, and with respect to the bud neck or cell cortex; to account for differences in cell lengths, the lengths of all mother and daughter cells were set to 1). This revealed that loss of She1 led to: (1) the spindle spending a greater fraction of time within the mother cell, (2) an increased fraction of time during which the spindle resides within close proximity (≤1 µm) of the cell cortex, and (3) the spindle residing more distal from the bud neck (Fig. 4F–H). For example, whereas the spindle exhibits only a partial bias toward wild-type mother cells, with a mean position within very close proximity of the bud neck, the spindle spent a significantly larger fraction of time within she1Δ mother cells more distal from the neck (Fig. 4G,H).

Given previous data indicating a key role for Kar9 in polarizing the yeast microtubule network (Cepeda-García et al., 2010; Hotz et al., 2012; Liakopoulos et al., 2003; Moore et al., 2006; Pereira et al., 2001), we wanted to determine whether any of these phenotypes are a consequence of the lack of Kar9. Analysis of Kar9-expressing cells revealed a very similar pattern for all spindle position metrics noted above in kar9Δ cells (Fig. S2E–G), indicating that She1 polarizes dynein activity in a manner that is independent of Kar9. Interestingly, whereas She1 reduces spindle displacement values and overall dynein activity in KAR9 cells (Fig. S2B–D) – albeit to a somewhat lesser extent – the presence of She1 had no effect on spindle velocity in these cells (Fig. S2A), indicating that this specific metric is a consequence of combined deletion of She1 and Kar9, and does not account for the polarized spindle movements promoted by She1. These observations suggest that She1 and Kar9 play complementary roles in affecting dynein motility in cells. However, given the fact that the spindle position metrics were statistically indistinguishable in KAR9 SHE1 and kar9Δ SHE1 cells (Fig. S2F,G; P≥0.2118), and KAR9 she1Δ and kar9Δ she1Δ cells (P≥0.3557), our results indicate that She1 polarizes dynein activity in a manner that is distinct and independent of Kar9 function.

Unlike the analyses described in Fig. 4A–E and Fig. S2A–D, which focus exclusively on dynein-dependent events, our tracking data (Fig. 4F–H and Fig. S2E–G) include all positions in which the spindle resides. To determine to what extent these positions are due to dynein, we repeated our analysis on cells deleted for the dynein heavy chain (dyn1Δ). As expected, these cells exhibit a large reduction in spindle movements that were only minimally increased upon further deletion of She1 (Fig. S2H). These data reveal that dynein and She1 are both required to promote spindle translocation into the daughter cell, as spindles in both dyn1Δ and dyn1Δ she1Δ cells spent a very low fraction of time in this compartment (Fig. 4G). We also noted that the spindle resides within close proximity of the cell cortex for only a small fraction of time in these mutants, indicating that this phenomenon is also dynein dependent (Fig. 4F). Finally, although deletion of DYN1 leads to an increased distance of the spindle from the bud neck, this value was slightly increased in dyn1Δ she1Δ cells, suggesting that She1 may play a dynein-independent role in promoting spindle/neck proximity (Fig. 4H). Taken together, these data indicate that She1 attenuates dynein-mediated spindle movements, and ensures that the spindle remains distal from the cell cortex, and within close proximity of the neck. Given the biased residence of the spindle within she1Δ mother cells, these data suggest that She1 may specifically reduce dynein activity within the mother cell.

Our data indicate that She1 affects at least two aspects of dynein activity: (1) the quality of dynein motility (e.g. it reduces the velocity and displacement of dynein-mediated spindle movements, although the former only in the absence of KAR9), and (2) initiation of dynein motility (e.g. it reduces the frequency of such events). Although our previous reconstitution experiments could potentially account for the reduced dynein-mediated spindle velocities observed in kar9Δ cells, none of our in vitro data revealed a She1-mediated reduction in microtubule binding by dynein, raising the question of how She1 may inhibit initiation of dynein-mediated spindle translocation events. To gain insight into the mechanism underlying this phenomenon, we employed our spindle dynamics assays in combination with a series of mutants.

In addition to binding to microtubules in the cytoplasm (Woodruff et al., 2009), She1 localizes prominently to spindle microtubules in the nucleus (Wong et al., 2007), where it has been proposed to play several roles [e.g. in promoting spindle stability and disassembly, and in kinetochore function (Wong et al., 2007; Woodruff et al., 2010; Zhu et al., 2017)]. Thus, it is conceivable that loss of She1 impacts spindle movements from within the nucleus. For instance, dynein-mediated outward forces could be impacted by the loss of inward-directed forces provided by such spindle-stabilizing factors. To determine whether it is the nuclear or cytoplasmic pool of She1 that is responsible for affecting the various spindle motility parameters, we added a strong nuclear localization signal to endogenous She1. We confirmed that She1^NLS is expressed at levels similar to that of wild-type She1 (Fig. S3A), is significantly enriched in the nucleus (Fig. S3B), and binds to yeast microtubules with very similar affinity to wild-type She1 (Fig. S3C–F). Of interest, we noted that the addition of the NLS eliminates the growth arrest phenotype associated with overexpressed She1 (Fig. S3G), indicating that the cytoplasmic pool – and not the nuclear pool – of She1 is causative of the sensitive-to-high-expression (SHE) phenotype (Espinet et al., 1995).

Analysis of spindle movements in she1^NLS cells revealed a phenotypic signature that was almost identical to that in she1Δ cells. This was apparent in both the quality of spindle movements (Fig. 4A–C), the relative extent of dynein activity (Fig. 4D,E), and the spindle positioning metrics, all of which mirror that of she1Δ cells (Fig. 4F–H). Thus, She1 affects both the quality and quantity of dynein-mediated spindle movements from within the cytoplasm.

As noted above, our previous work demonstrated that She1 must be bound to microtubules to affect dynein motility in vitro (Ecklund et al., 2017). To determine whether this is the case in cells, and to assess the contribution of microtubule-bound She1 to the initiation of dynein-mediated spindle movements, we employed cells expressing a mutant α-tubulin (Tub1^G437R) to which She1 binds to a significantly reduced extent (∼55.7% reduction) (Denarier et al., 2021). Consistent with the notion that She1 must bind microtubules to affect dynein motility, tub1^G437R cells exhibit increased spindle velocity and displacement values that roughly scale with the relative reduction in She1–microtubule binding in these cells (Fig. 4B,C). Moreover, tub1^G437R cells exhibit an increase in the extent of dynein activity (Fig. 4A,D,E), mean spindle position distal from the bud neck (Fig. 4H), and increased fractions of time during which the spindle resides in close proximity of the cell cortex and within the mother cell (Fig. 4F,G). Further deletion of She1 in tub1^G437R cells leads to more robust phenotypes, consistent with the partial ability of She1 to bind to microtubules in these cells. We noted that some of the dynein activity metrics were higher in tub1^G437R she1Δ cells than in TUB1 she1Δ cells, suggesting the mutant tubulin may affect other aspects of microtubule function that impact the dynein pathway. This may be a consequence of microtubule length differences between TUB1 and tub1^G437R cells, the latter of which exhibit a larger fraction of long microtubules (Denarier et al., 2021). Microtubule length differences have been found to correlate with increased dynein activity in cells (Estrem et al., 2017). Taken together, these data suggest that astral microtubule-bound She1 reduces the quality and quantity of dynein-mediated spindle movements. Given the increased residence of the spindle in the mother cell in both tub1^G437R and she1^NLS cells, these data indicate that astral microtubule-bound She1 reduces dynein activity specifically in the mother cell.

In addition to directly impacting dynein motility, She1 has also been implicated in affecting the interaction between dynein and dynactin, a multi-subunit complex that is required for dynein activity in cells (Markus and Lee, 2011; Markus et al., 2011; Woodruff et al., 2009). Although the mechanism by which She1 performs this activity and the impact of this regulation on cellular dynein function are unclear, we wondered whether some of our observations are a consequence of She1 affecting dynein–dynactin binding. Previous studies have found that human dynactin promotes a motility-competent configuration of dynein, and also stimulates microtubule-binding of the dynein-dynactin complex (McKenney et al., 2016; Zhang et al., 2017), the latter of which is due in large part to the N-terminal cytoskeleton-associated protein-glycine-rich (CAP-gly) microtubule-binding domain on the dynactin subunit p150^Glued (also known as DCTN1; Nip100 in budding yeast) (McKenney et al., 2016). Thus, enhanced dynein–dynactin binding could lead to an increase in the initiation of dynein-mediated spindle movements. Given that deletion of any of the dynactin subunits in yeast severely compromises dynein activity (Moore et al., 2008), we chose to employ a Nip100^ΔCAP-gly mutant that has been shown to have only minor effects on dynein activity (Moore et al., 2009).

With the exception of a small reduction in the time the spindle spends within the daughter cell (Fig. S4G), our analysis of nip100^ΔCAP-glykar9Δ cells revealed no significant impact on dynein function (Fig. S4A–H). These data contrast with prior observations (Moore et al., 2009), and instead suggest that this CAP-gly domain plays a very minor role in dynein function. Moreover, additional deletion of She1 leads to a phenotypic signature largely indistinguishable from that of NIP100 she1Δ cells. Although we cannot rule out the possibility that She1 affects dynein activity in a dynactin-dependent manner, these data indicate that the dynactin MTBD is not involved in this mode of regulation.

We sought to determine the importance of the She1–dynein MTBD interaction on the various aspects of spindle migration. To this end, we employed a chimeric dynein mutant (Dyn1^mMTBD; in which the MTBD is replaced by the corresponding mouse sequence) that exhibits reduced affinity for, and sensitivity to, She1 (Ecklund et al., 2017). Given that the dyn1^mMTBD allele possesses a C-terminal 3YFP tag (in contrast to the untagged DYN1 employed above), we compared spindle dynamics of this mutant to cells expressing Dyn1–3YFP. Although Dyn1–3YFP supports normal spindle positioning (Lee et al., 2005), we noted that these cells exhibit somewhat reduced spindle velocity values (Fig. S2A). Moreover, although deletion of She1 led to similar alterations in most spindle migration parameters in DYN1-3YFP cells, many of these changes were somewhat attenuated (Fig. S2B–G), suggesting that Dyn1–3YFP is less susceptible to She1-mediated inhibition.

Unexpectedly, we noted a she1Δ-mediated increase in the quality of dynein-mediated spindle motility in dyn1^mMTBD-3YFP cells (Fig. 5A–C,F), suggesting that in-cell modulation of dynein motility by She1 does not occur through binding to the dynein MTBD. However, the she1Δ-dependent increase in the extent of dynein activity in dyn1^mMTBD cells was significantly reduced compared to wild-type cells, indicating that inhibition of dynein activity by She1 occurs in a manner that requires its interaction with the dynein MTBD (Fig. 5D,E).

Our prior in vitro work revealed that She1 exhibits higher affinity for the nucleotide-free state of the dynein motor domain (Ecklund et al., 2017). Given the interaction of She1 with the dynein MTBD, and its preferential binding to the high microtubule affinity state of the MTBD (Ecklund et al., 2017), this indicates that She1 exhibits higher affinity for dynein when it is bound to microtubules during a processive run (Nishida et al., 2020; Uchimura et al., 2015). We also found that Dyn1^mMTBD exhibits a lower affinity for microtubules than the wild-type motor (Ecklund et al., 2017), suggesting that the basis for the partial insensitivity of this mutant to She1 may be a consequence of conformational differences between the two MTBDs. To determine whether this is the case, we introduced two point mutations into Dyn1^mMTBD (mouse residues E3289K and E3378K, which are in the coiled-coil and MTBD, respectively) that have been reported to increase dynein–microtubule binding affinity (Redwine et al., 2012). This mutant (dyn1^mMTBD-HA, for high affinity) would be expected to spend a greater fraction of time in a microtubule-bound conformation – to which She1 would be predicted to preferentially bind – and would potentially restore She1 sensitivity to the chimeric mutant. Consistent with this notion, the she1Δ-mediated increase in the number and extent of dynein-mediated spindle movement events was completely restored in dyn1^mMTBD-HA she1Δ cells (Fig. 5D,E). Analysis of tracking data revealed an increased fraction of time during which the spindle resides within the mother cell, and a mean position more distal from the bud neck as a consequence of She1 deletion in dyn1^mMTBD-HA, but not dyn1^mMTBD cells (Fig. 5G,H). These data indicate that She1-mediated inhibition of dynein activity indeed requires a microtubule-bound conformation of the MTBD. In light of the varying extents to which She1 deletion affects the dynein activity parameters for each of the mutants, these data further support the notion that She1-mediated inhibition of the initiation of dynein-mediated spindle movements, and not its effect on the quality of movement, is the key determinant that dictates the ability of She1 to promote bud neck proximity, and daughter cell-directed spindle movements.

Recent data indicate that yeast dynein adopts an autoinhibited conformational state referred to as the ‘phi’ particle (Amos, 1989; Marzo et al., 2020). This conformation restricts dynein activity in cells by precluding its interactions with cargo adaptors and dynactin (Zhang et al., 2017). Given the ability of She1 to inhibit the dynein–dynactin interaction in cells (Markus et al., 2011; Woodruff et al., 2009), we wondered whether She1 may function in part by promoting the phi conformation. To this end, we employed a dynein mutant that is compromised in its ability to adopt this state, Dyn1^D2868K (Marzo et al., 2020). In stark contrast to all other dynein mutants tested, cells expressing Dyn1^D2868K exhibited motility and activity metrics that were reduced upon She1 deletion (Fig. 5B–E), indicating that rather than inhibit this mutant, She1 in fact promotes its activity. However, we noted that the baseline levels of activity in dyn1^D2868K SHE1 cells were similar to those in DYN1 she1Δ cells (Fig. 5D,E), indicating that dyn1^D2868K phenocopies wild-type dynein in she1Δ cells.

In spite of these observations, the spindle tracking data in dyn1^D2868K SHE1 and dyn1^D2868K she1Δ cells revealed a pattern very similar to cells expressing wild-type Dyn1 (Fig. 5F–H), raising the question of whether She1-mediated inhibition of the initiation of dynein-mediated spindle movements is the only determinant that dictates the ability of She1 to promote bud neck proximity and daughter cell-directed spindle movements. Moreover, these data suggest that She1 may in fact promote microtubule-binding of active cortical Num1–dynein–dynactin complexes in a manner that is reflective of our single-molecule data (Ecklund et al., 2017; Markus et al., 2012b), our optical trapping data (Fig. 2) and our nuclear/neck translocation data, in which She1 appears to support dynein-mediated translocation of the nucleus through the bud neck (Fig. 3E).

We next sought to determine whether She1 directly affects the initiation of dynein-mediated spindle movements in a compartment-specific manner. Astral microtubule plus ends make contacts with the cell cortex as they dynamically sample the mother and daughter cells, with a subset of these ‘cortical contacts’ resulting in dynein-mediated spindle movements (‘productive events’; Fig. 6A). Initiation of a productive event is thought to occur coincidently with offloading and activation of dynein–dynactin complexes at Num1 sites (Lammers and Markus, 2015; Markus and Lee, 2011). We wondered whether the increased frequency of dynein-mediated spindle movements in she1Δ cells is a consequence of more astral microtubule contacts with the cell cortex or more of such contacts transitioning into a productive event. To address this, we separately counted each in mother and daughter cells.

This revealed that the number of cortical contacts was significantly greater in mother than in daughter cells for almost all yeast strains (except for dyn1^D2868K; Fig. 6B), which is surprising in light of previous observations that the daughter-oriented SPB nucleates more and longer microtubules (Estrem et al., 2017; Lengefeld et al., 2018; Vogel et al., 2001; Vogel and Snyder, 2000). Although loss of She1 had very little impact on the total number of cortical contacts in most mother and daughter cells (Fig. 6B), it did cause an increase in the ratio of mother:daughter cortical contacts compared to isogenic parent strains in wild-type, nip100^ΔCAP-gly and dyn1^D2868K cells (1.27 in SHE1 versus 1.93 in she1Δ; 1.91 in nip100^ΔCAP-gly versus 3.42 in nip100^ΔCAP-glyshe1Δ; and, 0.98 in dyn1^D2868K versus 1.81 in dyn1^D2868K), suggesting that She1 may regulate microtubule dynamics in a compartment-specific manner, which may partially account for the mother cell bias in spindle residencies. However, this ratio was largely unchanged with respect to wild-type in she1^NLS, tub1^G437R and tub1^G437R she1Δ cells (mother:daughter ratios of 1.45, 1.35 and 1.35, respectively), indicating that relative differences in cortical contacts do not account for the mother cell bias in these mutants.

In spite of the small change in the number of microtubule-cortex encounters in she1Δ cells, the number of productive events was significantly greater in she1Δ mother, but not she1Δ daughter cells, compared to wild-type cells (Fig. 6C). Although the same was true for most of the other mutants that exhibit a mother cell bias of spindle residencies (i.e. she1^NLS, tub1^G437R she1Δ, dyn1^mMTBD-HA she1Δ and dyn1^D2868K she1Δ), it was not true for tub1^G437R cells, which exhibit a greater enrichment of mother cell spindle residencies and a mean spindle position further from the neck than she1^NLS cells (see Fig. 4G,H), thus raising the question of whether the absolute number of productive events alone accounts for the spindle position phenotype. Moreover, although dyn1^D2868K she1Δ cells did not exhibit a significant increase in productive events in the mother cell, the mother:daughter ratio of these events was increased due to a reduction in events noted in the daughter cell. Unlike other mutants, this change appears to mirror the changes in cortical contacts noted as a consequence of She1 deletion, likely accounting for the mother cell bias in this mutant.

Plotting the relative fraction of cortical contacts that transition to a productive event in each compartment revealed that She1 dampens the initiation of dynein-mediated spindle migration events specifically in mother cells (Fig. 6D). Specifically, whereas an approximately equal fraction of cortical contacts in wild-type mother and daughter cells result in a productive event, these values were increased to 72.0% in she1Δ mother cells, but only 59.6% in she1Δ daughter cells. We observed a very similar pattern of disproportionate increases in the fraction of productive events in the mother cells for most of those mutants that exhibit a biased enrichment of spindle residencies in the mother cell, including she1^NLS, tub1^G437R, tub1^G437R she1Δ and dyn1^mMTBD-HA she1Δ cells. Exceptions include nip100^ΔCAP-glyshe1Δ cells, which exhibited a broad distribution of values, and an overall increase in the mean activity in both mother and daughter cells upon She1 deletion (Fig. S4K), and dyn1^D2868K cells. As noted in our activity measurements described above (Fig. 5D), dyn1^D2868K SHE1 cells phenocopy DYN1 she1Δ cells in their fraction of productive events (Fig. 6D), suggesting that preventing dynein autoinhibition renders it refractory to She1-mediated inhibition. Notably, dyn1^mMTBD cells, which exhibit no she1Δ-mediated shift in spindle residencies, also show no she1Δ-dependent increase in dynein activity in mother or daughter cells.

Taken together, these data indicate that astral microtubule-bound She1 dampens the initiation of dynein-mediated spindle movement events specifically in the mother cell in a manner that requires an interaction with the microtubule-bound conformation of the dynein MTBD. Although this activity likely accounts for the biased spindle residence in mother cells upon loss of She1, data from nip100^ΔCAP-gly and dyn1^D2868K cells raise the possibility that microtubule dynamics also plays a role in the mother cell bias. Since the probability of a dynein-mediated spindle movement is the product of the number of astral microtubule–cortical contacts and the fraction of productive events, a change in the number of cortical contacts likely affects the mean spindle position.

Given that She1 reduces the likelihood of a cortical contact transitioning to a productive event specifically in the mother cell, we wondered whether She1 affects the levels of dynein or dynactin at microtubule plus ends or the cell cortex in a compartment-specific manner (Fig. 7A). Although mother and daughter kar9Δ SHE1 cells exhibit a similar number of cortical dynein foci, this balance was shifted significantly toward mother cells in those lacking SHE1 (Fig. 7B,C), suggesting that the increased productive events in she1Δ mother cells is a consequence of enhanced offloading.

Consistent with previous observations (Grava et al., 2006), we noted that daughter cells exhibit a higher frequency and intensity of plus end-associated dynein foci than mother cells, in spite of the lack of KAR9 in these cells (Fig. 7D; P≤0.0006). We also noted a higher frequency of dynactin (Jnm1–3mCherry) at plus ends in daughter than mother cells (Fig. 7E; P=0.0002). Interestingly, the frequency and intensity of dynein and dynactin foci were much higher in she1Δ mother cells (P<0.0001), but only slightly higher in she1Δ daughter cells (P≥0.0324), indicating that She1 precludes association of dynein and dynactin with plus ends in mother cells to a greater extent than in daughter cells. Notably, dyn1^mMTBD SHE1 cells exhibit levels of dynein and dynactin very similar to DYN1 she1Δ cells, which were only slightly increased upon deletion of SHE1, suggesting that She1 affects their plus end association in a manner that requires an interaction with the dynein MTBD. These data indicate that She1 may attenuate the initiation of dynein-mediated spindle movements in mother cells by preventing association of dynein and dynactin with plus ends, which in turn reduces their offloading and activation at cortical sites in this compartment.

We next wondered whether She1 preferentially localizes to astral microtubules in the mother cell. To circumvent the difficulty in visualizing endogenous levels of She1–3GFP (due to low cellular concentrations), we overexpressed She1–3GFP using the galactose-inducible promoter GAL1p. To correlate She1 localization patterns to our spindle dynamics data, we arrested GAL1p:SHE1-3GFP cells with HU. Live-cell imaging revealed a ∼4-fold greater fraction of mother cells with apparent She1–3GFP along astral microtubules with respect to daughter cells (Fig. 8A,B; note only those cells with astral microtubules apparent in both compartments were assessed). Fluorescence intensity measurements confirmed its preferential localization to astral microtubules in mother cells (Fig. 8C), and that this was not a consequence of differences in microtubule polymer mass (e.g. microtubule bundling; Fig. 8D).

To determine whether this biased localization is a consequence of asymmetric enrichment within the cytoplasm of the mother cell or preferential microtubule-binding in this compartment, we measured the fluorescence intensity values for regions within the cytoplasm (i.e. regions excluding microtubules) of untreated cells and those treated with the microtubule depolymerizing agent nocodazole (see Fig. S5A). This analysis revealed a small enrichment of She1–3GFP in the untreated mother cells (Fig. 8E) that was somewhat reduced in nocodazole-treated cells. Given the large difference in microtubule-binding between the two compartments compared to these small differences in cytoplasmic intensities, these data suggest that the biased localization of She1 to mother cell-associated astral microtubules is largely achieved through differential microtubule binding activity in the two compartments.

As an alternative method to assess She1 localization, we exploited the fact that She1 overexpression leads to a relocalization of dynein from microtubule plus ends to along the length of astral microtubules as a consequence of the interaction between microtubule-bound She1 and the dynein MTBD (Ecklund et al., 2017; Markus et al., 2012b). Assessment of dynein localization in She1-overexpressing cells revealed many examples of cells with Dyn1–3YFP bound along the length of astral microtubules within the mother, but not the daughter cell (Fig. 8F; Fig. S5C).

We noted that overexpressed She1–3GFP forms diffuse ‘clouds’ of fluorescence in the vicinity of microtubules (Fig. 8A, blue arrowhead and dashed circle), suggesting that She1, and the microtubules on which it is observed, is enclosed in a membrane-bound compartment, such as the nucleus. To determine whether the She1 localization pattern described above is cytoplasmic or nuclear, we overexpressed Kip2, which leads to a substantial increase in the length of astral microtubules, and assessed the localization pattern of both She1–3GFP and Dyn1–3YFP, the latter of which is exclusively cytoplasmic (note the absence of Dyn1 from spindle regions in Fig 8F and Fig. S5C). Although we cannot rule out that some of the diffuse She1 fluorescence is nuclear, this revealed that most of the ‘clouds’ of She1–3GFP are in fact within the cytoplasm (Fig. 8G and Fig. S5D; note the presence of She1 and dynein on long astral microtubules within the ‘clouds’). Taken together, our data reveal that She1 preferentially localizes along astral microtubules within mother cells.

Our work sheds light on the means by which She1 affects dynein activity in cells, and how it polarizes dynein-mediated spindle movements toward the daughter cell. We find that astral microtubule-bound She1 precludes the initiation of dynein-mediated spindle movements in mother cells by preventing association of dynein and dynactin with microtubule plus ends in this compartment. We also find that this requires interactions between She1 and the microtubule-bound conformation of the dynein MTBD. This activity is compartmentally specified, likely as a consequence of preferential binding of She1 to astral microtubules within mother cells. Although our prior in vitro work revealed the She1–dynein MTBD interaction, and demonstrated the importance of this interaction for velocity reduction in single molecule assays with purified dynein alone (Ecklund et al., 2017), the nature of in-cell inhibition of dynein activity appears to occur through an entirely distinct mechanism that is unrelated to the ability of She1 to reduce the velocity of dynein, or reduce its microtubule-dissociation rate in vitro. However, our new findings suggest that She1 may in fact promote microtubule interactions by active cortical Num1–dynein–dynactin complexes, similar to its effect on dynein alone in optical trapping assays. Thus, the in-cell activity of She1 appears to involve two distinct activities: (1) it precludes the initiation of dynein-mediated spindle translocation events in cells (by blocking the interaction of dynein with Pac1 and Bik1, and/or the dynein–dynactin interaction, thus limiting plus end and cortical localization), and (2) it promotes dynein–microtubule encounters subsequent to offloading and activation.

How does She1 prevent the initiation of dynein-mediated spindle movements? We find that She1 only reduces dynein events when it can adopt the phi state (see dyn1^D2868K data), and when it is in the microtubule-bound conformational state for which She1 exhibits high affinity (compare dyn1^mMTBD to dyn1^mMTBD-HA cells) (Ecklund et al., 2017). Importantly, dynein does not contact the plus end directly, but rather interacts with it indirectly via Bik1 (Lammers and Markus, 2015; Sheeman et al., 2003). Thus, dynein is in a microtubule-unbound state prior to offloading to the cortex. However, we recently proposed a model whereby dynein makes direct contact with the microtubule during the offloading process, and that this contact is required to disengage dynein from the plus end-targeting machinery, thus permitting spindle translocation (Lammers and Markus, 2015). If true, this would be the first moment in the activation process when the microtubule-bound dynein MTBD encounters She1. Thus, we posit that it is at this moment when She1 somehow causes dynein to adopt the phi state (see below), which consequently breaks its contacts with Pac1 and/or dynactin (Marzo et al., 2020; Zhang et al., 2017), thus terminating the offloading event. This model is further supported by our prior observation that deletion of She1 leads to enhanced offloading of dynein to cortical sites (Markus and Lee, 2011).

How might She1–dynein MTBD binding promote the phi conformation? Although such a mechanism has not yet been described, previous studies have identified long-range allosteric communication from the MTBD to the AAA ring via the coiled-coil stalk (Kon et al., 2009; Niekamp et al., 2019; Uchimura et al., 2015), where contact points have been identified that stabilize the phi state (Marzo et al., 2020; Zhang et al., 2017). In support of such a model, we find that a dynein mutant that is less able to adopt this conformation (Dyn1^D2868K) is somewhat refractory to She1-mediated inhibition (and in fact relies on She1 for maximal in-cell activity). The same is true for Dyn1 with a C-terminal 3YFP tag, addition of which has been shown to reduce the propensity of dynein to adopt the phi conformation (Torisawa et al., 2014). Moreover, we find that She1 may prevent offloading of dynein to the mother cell cortex by reducing its plus end association, which is mediated by Pac1 (Lee et al., 2003; Sheeman et al., 2003). Our recent studies have found that Pac1 exhibits higher affinity for dynein when it is in a non-phi state (Marzo et al., 2020). Thus, reduced plus end binding may also be a consequence of She1 promoting the phi conformation.

How is She1 asymmetrically localized to the mother cell? In light of our data that She1 asymmetry is likely imparted by differential microtubule-binding affinities in the mother and daughter cells, and the likely role of phosphorylation in modulating the She1-microtubule affinity (Markus et al., 2012b), it is possible that a daughter cell-localized kinase phosphorylates She1 to reduce its microtubule binding in this compartment. Previous studies have identified the mitotic kinase Aurora B (Ipl1) and the mitogen-activated protein kinase Hog1 as candidates that phosphorylate She1 (Markus et al., 2012b; Pigula et al., 2014; Woodruff et al., 2010). At least one upstream activator of Hog1 is enriched in the daughter cell (Sho1) (Raitt et al., 2000), suggesting this kinase may be key in modulating She1 asymmetry.

Although She1 is the only known MAP that has the capacity to polarize dynein-mediated cargo transport, recent studies have revealed several MAPs in higher eukaryotes that impact motor-mediated cargo transport. For instance, MAP4 and MAP9 affect dynein–dynactin transport functions in higher eukaryotes (Monroy et al., 2020; Samora et al., 2011; Semenova et al., 2014). Of note, depletion of MAP4 leads to hyperactive cortical dynein activity and pronounced spindle movements, much like deletion of She1 (Samora et al., 2011). The mechanism by which MAP4 functions is unclear; however, MAP9 precludes the interaction between microtubules and dynein–dynactin complexes by blocking dynactin–microtubule binding (Monroy et al., 2020). Although there is limited sequence homology among the MAP family (restricted to the microtubule-binding domains), many of them are enriched with regions of intrinsic disorder [95%, 94% and 96% of MAP9, MAP4 and She1, respectively, are predicted to be disordered according to MetaDisorder (Kozlowski and Bujnicki, 2012)]. Thus, although sequence alignments reveal no clear She1 homolog in higher eukaryotes, MAP4 and/or MAP9 may have evolved from She1, or a common ancestor.

For expression and purification of She1–NLS^SV40–HaloTag, we introduced sequence encoding the nuclear localization sequence (NLS) from SV40 large T antigen (hereafter referred to as ‘NLS^SV40’) between She1 and the C-terminal HaloTag. Briefly, the C-terminal HaloTag sequence was PCR amplified from pProEX-HTb-TEV:SHE1-HALO (Markus et al., 2012b) using a forward primer that includes a sequence that encodes a short linker (Gly-Ser-Gly-Ser) followed by the NLS^SV40 (Ala-Ala-Ala-Pro-Lys-Lys-Lys-Arg-Lys-Val-Gly). This PCR product was assembled into pProEX-HTb-TEV:SHE1-HALO digested with BamHI and NotI (which excises the HaloTag) using Gibson assembly, yielding pProEX-HTb-TEV:SHE1-NLS^SV40-HALO. Note this sequence (linker and NLS^SV40) is identical to the one introduced at the 3′ end of the SHE1 locus to generate she1^NLS.

Strains are derived from either YEF473A or W303, and are available upon request (listed in Table S1). We transformed yeast strains using the lithium acetate method (Knop et al., 1999). Engineered yeast strains (e.g. she1^NLS and dyn1m^MTBD-HA) were constructed by PCR product-mediated transformation (Longtine et al., 1998) or by mating followed by tetrad dissection. Proper tagging and mutagenesis was confirmed by PCR and/or by sequencing. Fluorescent tubulin-expressing yeast strains were generated using plasmids and strategies described previously (Markus et al., 2015; Song and Lee, 2001). Yeast synthetic defined (SD) complete medium was obtained from Sunrise Science Products (San Diego, CA).

Assessment of cell cycle progression and spindle positioning (see Fig. 1 and Fig. S1) was performed by imaging cells in the CellAsic ONIX system using microfluidic cassettes designed for haploid yeast cells (Y04C; MilliporeSigma). In brief, after an overnight growth in SD complete medium supplemented with 2% glucose at 30°C, cells were diluted 50-fold into the cell inlet well of the microfluidic cassette, which was primed with SD complete medium prior to addition of cells (per manufacturer's instructions). Pressure was maintained at 7.0 psi throughout the imaging period to ensure a constant replenishment of medium into the cassette, which was set to 30°C. Note that temperature is not directly monitored by the CellAsic system, and is thus affected by ambient room temperature. To account for temperature differences between experiments, wild-type and mutant cells were imaged simultaneously by introducing respective cells into adjacent imaging chambers of the microfluidics cassette. Z-stacks (7 steps with 0.5 µm spacing) from multiple XY coordinates (10 for replicate 1, and 5 for replicate 2) were acquired for 10 h at 90 s intervals. Data were normalized [i=(x-minimum value from control)/(maximum value from control−minimum value from control), where i is the normalized value, and x is the non-normalized value] to account for differences between independent replicates due to differences in imaging conditions (likely due to fluctuations in room temperature).

For spindle dynamics and nuclear translocation assays, mid-log phase cells were arrested with 200 mM hydroxyurea (HU) for 2.5 h in SD complete medium supplemented with 2% glucose, and then applied to slide-mounted agarose ‘pads’ consisting of 1.7% agarose dissolved in SD complete medium supplemented with 2% glucose and 200 mM HU for confocal fluorescence microscopy. Full Z-stacks (19 planes with 0.2 µm spacing) of GFP-labeled microtubules (GFP–Tub1; for Figs 1G, 4, 5 and 6), and/or NLS-3mCherry (for Fig. 3) were acquired every 10 s for 15 min on a stage pre-warmed to 28°C. To image She1-3GFP in HU-arrested cells (for Fig. 8A–E), mid-log phase cells cultured in SD complete medium supplemented with 2% glucose were pelleted (400 g for 1 min), and then resuspended in SD complete medium supplemented with 2% galactose (to activate the GAL1 promoter). After 30 min, the cells were pelleted again, and resuspended in SD complete medium supplemented with 2% galactose and 200 mM HU for 2.5 h followed by applying them to slide-mounted agarose pads for confocal microscopy (for a total galactose induction time of 3 h). To image She1 or Dyn1 in She1- and/or Kip2-overexpressing cells (in non-HU arrested cells), cells were cultured in galactose-containing medium for 2.5 to 3 h prior to imaging. Z-stacks (7 steps with 0.6 µm spacing) of each respective channel were acquired every 20 s for 1 min. Note that to prevent bleed-through of mTurquoise2 fluorescence in the GFP channel (i.e. in cells expressing mTurquoise2–Tub1 and She1–GFP, such as in Fig. 8G) the mTurquoise2/CFP channel was acquired using a 445 laser and a CFP filter, while the GFP channel was acquired using a 514 nm laser and a YFP filter, which permits visualization of GFP, but not mTurquoise2.

Images were collected on a Nikon Ti-E microscope equipped with a 1.49 NA 100× TIRF objective, a Ti-S-E motorized stage, piezo Z-control (Physik Instrumente), an iXon DU888 cooled EM-CCD camera (Andor), a stage-top incubation system (Okolab), and a spinning disc confocal scanner unit (CSUX1; Yokogawa). 445 nm, 488 nm, 514 nm, 561 nm and 594 nm lasers housed in an LU-NV laser unit equipped with AOTF control (Nikon) were used to excite mTurquoise2, GFP, Venus, mRuby2 and mCherry, respectively, which were used with emission filters mounted in a filter wheel (ET480/40 m for mTurquoise2, ET525/50 M for GFP, ET520/40 M for Venus, and ET632/60 m for mRuby2 and mCherry; Chroma). The microscope was controlled with NIS Elements (Nikon).

For analysis of the cell cycle progression images, we used the following morphological features to define temporal landmarks in cell cycle progression: the spindle pole body duplication was identified as the first frame when two SPBs could be spatially resolved; anaphase onset was defined as the first frame when the spindle began to elongate; cytokinesis was defined as the first frame when independent movement of the spindle pole bodies was apparent (in which they moved with respect to each other, indicating complete spindle disassembly).

To assess cortical microtubule contacts and number of productive events, full Z-stacks (19 planes with 0.2 µm spacing) of GFP-labeled microtubules acquired using confocal fluorescence microscopy were converted to XY and XZ maximum intensity projections, which were both used to manually count the number of microtubule–cortex encounters, and the number of productive events (defined as those in which the spindle is observed moving coincident with a microtubule–cortex encounter). All dynein-mediated spindle movements were identified as such by the apparent directed migration of the spindle following and coincident with an astral microtubule associating with the cell cortex.

We purified She1–HaloTag as previously described (Ecklund et al., 2017). Briefly, Escherichia coli BL21 (Rosetta DE3 pLysS) cells transformed with pProEX-HTb-TEV:SHE1-HALO (or pProEX-HTb-TEV:SHE1-NLS^SV40-HALO) were grown at 37°C in LB supplemented with 1% glucose, 100 µg/ml carbenicillin and 34 µg/ml chloramphenicol to OD₆₀₀ 0.4–0.6, shifted to 16°C for 2 h, then induced with 0.1 mM IPTG for 14–16 h at 16°C. The cells were harvested, washed with cold water and stored at −80°C. Cells were thawed in 0.5 volume of cold 2× lysis buffer [1× buffer is 30 mM HEPES pH 7.2, 50 mM potassium acetate, 2 mM magnesium acetate, 0.2 mM EGTA, 10% glycerol, 1 mM DTT and protease inhibitor tablets (Pierce)] and then lysed by sonication (5×30 s pulses) with 1 min on ice between each pulse. The lysate was clarified at 22,000 g for 20 min, adjusted to 0.005% Triton X-100, then incubated with glutathione–agarose beads for 1 h at 4°C. The resin was then washed three times in wash buffer (30 mM HEPES pH 7.2, 50 mM potassium acetate, 2 mM magnesium acetate, 0.2 mM EGTA, 300 mM KCl, 0.005% Triton X-100, 10% glycerol, 1 mM DTT and protease inhibitor tablets) and twice in TEV digest buffer (10 mM Tris-HCl pH 8.0, 150 mM KCl, 0.005% Triton X-100, 10% glycerol and 1 mM DTT). To fluorescently label She1–Halo, the bead-bound protein was incubated with 6.7 µM HaloTag-TMR ligand (Promega) for 15 min at room temperature. The resin was then washed three more times in TEV digest buffer to remove unbound ligand, then incubated in TEV buffer supplemented with TEV protease for 1 h at 16°C. The resulting eluate was collected using a centrifugal filter unit (0.1 µm, Millipore), aliquoted, drop frozen in liquid nitrogen and stored at −80°C.

Purification of ZZ-TEV-6His-GFP-3HA-GST-dynein₃₃₁-HALO (under the control of the galactose-inducible promoter GAL1p) was performed as previously described (Ecklund et al., 2017; Marzo et al., 2020). Briefly, yeast cultures were grown in YPA supplemented with 2% galactose, harvested, washed with cold water, and then resuspended in a small volume of water. The resuspended cell pellet was drop frozen into liquid nitrogen and then lysed in a coffee grinder (Hamilton Beach). After lysis, 0.25 volume of 4× lysis buffer (1× buffer is 30 mM HEPES pH 7.2, 50 mM potassium acetate, 2 mM magnesium acetate, 0.2 mM EGTA, 1 mM DTT, 0.1 mM Mg-ATP, 0.5 mM Pefabloc SC and 0.7 µg/ml Pepstatin) was added, and the lysate was clarified at 22,000 g for 20 min. The supernatant was then bound to IgG sepharose 6 fast flow resin (Cytiva) for 1 h at 4°C, which was subsequently washed three times in wash buffer (30 mM HEPES pH 7.2, 50 mM potassium acetate, 2 mM magnesium acetate, 0.2 mM EGTA, 300 mM KCl, 0.005% Triton X-100, 10% glycerol, 1 mM DTT, 0.1 mM Mg-ATP, 0.5 mM Pefabloc SC and 0.7 µg/ml Pepstatin), and twice in TEV buffer (50 mM Tris-HCl pH 8.0, 150 mM potassium acetate, 2 mM magnesium acetate, 1 mM EGTA, 0.005% Triton X-100, 10% glycerol, 1 mM DTT, 0.1 mM Mg-ATP, 0.5 mM Pefabloc SC). To fluorescently label 6His-GFP-GST-3HA-dynein₃₃₁-HALO (for single molecule analyses), the bead-bound protein was incubated with either 6.7 µM HaloTag-TMR or HaloTag-PEG-biotin ligand (Promega) for 15 min at room temperature. The resin was then washed four more times in TEV digest buffer, then incubated in TEV buffer supplemented with TEV protease for 1 h. Following TEV digest, the bead solution was transferred to a spin column (Millipore) and centrifuged at 20,000 g for 10 s. The resulting protein solution was aliquoted, flash frozen in liquid nitrogen and then stored at −80°C.

Purification of yeast tubulin was performed essentially as described previously (Johnson et al., 2011) with minor modifications. Yeast cells (JEL1) co-transformed with p426Gal1:Tub1 and p424:Tub2-6His were grown in 50 ml of selective SD complete medium (lacking uracil and tryptophan) supplemented with 2% glucose, and then transferred to 1 l of nonselective YPGL (2% peptone, 1% yeast extract, 3% glycerol, and 2% lactate; note, we grew 16 l of cells for a typical preparation). When the cell density reached an OD₆₀₀ between 5 and 9, 20 g of galactose powder was added per liter of YPGL, and after 5 h, cells were harvested, washed with water and stored at −80°C. Approximately 75 g of cells were thawed and resuspended in 70 ml of lysis buffer (50 mM HEPES pH 7.4, 500 mM NaCl, 10 mM MgSO₄, 30 mM imidazole) supplemented with 50 µM GTP and cOmplete protease inhibitor cocktail (Roche), and lysed by 5–6 passes through a microfluidizer (LM10; Microfluidics) at 23,000 PSI, with 5 min on ice between each pass. After clarification (at 22,000 g for 30 min at 4°C), the supernatant was applied to a 5 ml HisTrap Ni-NTA column (Cytiva) pre-equilibrated with 10 column volumes (CVs) of lysis buffer supplemented with GTP using an AKTA FPLC (Cytiva). After washing the column with 10 CVs of lysis buffer supplemented with 50 µM GTP and 10 CVs of nickel wash buffer (25 mM PIPES pH 6.9, 1 mM MgSO₄ and 30 mM imidazole) supplemented with 50 µM GTP, bound protein was eluted with 6 CVs of elution buffer (25 mM PIPES pH 6.9, 1 mM MgSO₄, 300 mM imidazole) supplemented with 50 µM GTP. Peak fractions (determined by absorbance at 260 nm) were pooled and treated with nuclease (Pierce Universal Nuclease; catalog #88702; 10 µl per 20 ml of eluate) for 15 min at room temperature, and then diluted with MonoQ buffer A (25 mM PIPES pH 6.9, 2 mM MgSO₄, 1 mM EGTA) supplemented with 50 µM GTP such that the final imidazole concentration was 100 mM. The protein was then loaded onto a MonoQ 10/100GL anion exchange column pre-equilibrated with 5 CVs of 90% MonoQ buffer A (see above) and 10% MonoQ buffer B (25 mM PIPES pH 6.9, 2 mM MgSO₄, 1 mM EGTA and 1 M NaCl) supplemented with 50 µM GTP, after which bound protein was eluted with a 10–70% MonoQ buffer B gradient over 50 CVs. Peak tubulin fractions (determined by absorbance at 260 nm and SDS-PAGE) were pooled and concentrated (Amicon Ultra-4 30K; catalog #UFC803024) to 2.8 µM (concentration and aggregation were closely monitored using absorbance at 260 nm and 280 nm), and then dialyzed against tubulin storage buffer (10 mM PIPES pH 6.9, 1 mM MgSO₄, 1 mM EGTA) supplemented with 50 µM GTP (Thermo Slide-A-Lyzer; catalog #66810). Resulting protein was aliquoted (50 µl), snap frozen in liquid nitrogen and stored at −80°C.

Anti-His-coated 0.44 µm microbeads [PSS4; Spherotech; prepared as described previously (Driver et al., 2014)] were incubated with purified 6His-GFP-3HA-GST-dynein₃₃₁-HALO in dynein trapping buffer (30 mM HEPES pH 7.2, 2 mM Mg-acetate and 1 mM EGTA) for 1 h at 4°C. During the incubation, flow chambers (assembled from a glass slide, coverslip and double-sided sticky tape) were prepared by sequential addition and incubation with the following solutions: (1) 1 mg/ml biotinylated BSA (Vector Labs #B-2007), (2) BRB80 (80 mM PIPES pH 6.9, 1 mM MgCl₂, 1 mM EGTA, pH 6.9), (3) 0.33 mg/ml Avidin (Vector Labs #A-3100), (4) BRB80, and (5) GMPCPP-stabilized biotinylated microtubules (diluted in BRB80) (Asbury et al., 2006). After a 10 min incubation, the chamber was washed with BRB80, and the dynein-coated microbeads (diluted in trapping buffer supplemented with 1 mg/ml κ-casein, 8 mg/ml BSA, 1 mM DTT, 0.8 mM ATP, 4 mM MgSO₄, 4.5 mg/ml glucose, 250 µg/ml glucose oxidase and 30 µg/ml catalase) were introduced into the chamber, which was then sealed with nail polish, and immediately used for data collection.

The optical trap was essentially as described previously (Franck et al., 2010), and was operated in stationary mode, without feedback control (i.e. in ‘open loop’ mode). Bead-trap separation was saved at 200 Hz and converted into force by multiplying by the trap stiffness, which ranged between 0.025 and 0.045 pN/nm. Custom analysis software written in Igor Pro (Wavemetrics; available upon request) was used to estimate pre-stall speeds, stall forces, and stall times for individual bead motility events. Briefly, we defined the start of an event as the time at which the bead first moved beyond 3× the root-mean-square baseline noise. The end of an event was clearly identifiable as the time when the bead detached from the microtubule. The onset of stalling was chosen as the time at which the bead velocity, averaged over a sliding 2.5-s window, first fell below 2 nm/s. Pre-stall speed was then defined as the slope of a line fit to all the data between the start of the event and the onset of stalling and stall force was defined as the average force level during the stall event.

We used total internal reflection fluorescence (TIRF) microscopy-based microtubule-binding assays to measure the binding affinity of She1 and She1^NLS for yeast microtubules. To prepare microtubules, 15 µl of tubulin polymerization buffer (500 mM PIPES pH 6.9, 5 mM MgSO₄, 25% glycerol) along with epothilone B (50 µM final) and GTP (2 mM final) were added to one 50 µl aliquot of yeast tubulin (see above), which was then incubated at 30°C overnight. Flow chambers constructed from slides, double-sided sticky tape, and plasma cleaned and silanized coverslips were coated with anti-His antibody (100 μg/ml, sc-8036, Santa Cruz Biotechnology), then blocked with 1% Pluronic F-127, after which microtubules (diluted to 0.75 µM in tubulin polymerization buffer supplemented with 50 µM epothilone B and 2 mM GTP) were added. After unbound microtubules were removed by washing with chamber wash buffer [a 9:1 mixture of the dynein assay buffer (30 mM HEPES pH 7.2, 50 mM potassium acetate, 2 mM magnesium acetate, 0.2 mM EGTA, 10% glycerol, supplemented with 1 mM DTT, 50 µM epothilone B, and 2 mM GTP) and TEV buffer, as described above; this mixture was used to account for storage of purified She1-TMR in TEV buffer], purified She1-TMR (wild-type or She1^NLS, diluted such that the final 9:1 buffer mixture described above was achieved for all concentrations of She1 and She1^NLS) was introduced into the chamber. Images were collected on a Nikon Ti-E microscope (controlled with NIS Elements) equipped with a 1.49 NA 100× TIRF objective, a motorized stage, piezo Z-control (Physik Instrumente), and an iXon X3 DU897 cooled EM-CCD camera (Andor). A 561 nm laser (Coherent) was used along with a multi-pass quad filter cube set (C-TIRF for 405/488/561/638 nm; Chroma) and emission filters mounted in a filter wheel (525/50 nm, 600/50 nm, and 700/75 nm; Chroma) to image She1–TMR. To image non-fluorescent yeast microtubules, we used interference reflection microscopy (IRM), as previously described (Mahamdeh et al., 2018). Excitation light for IRM was provided by a Sola SE light engine (Lumencor). To measure the degree of She1-microtubule binding, background-subtracted fluorescence intensities of She1–TMR were determined. Note that we accounted for potential differences in the extent of HaloTag labeling between She1 and She1^NLS by measuring fluorescence band intensities following imaging of acrylamide gels with each on a Typhoon gel imaging system (FLA 9500). Both absolute concentration and relative degree of HaloTag labeling were taken into account when calculating binding affinities. Binding curves and curve fitting for dissociation constants (where appropriate) were generated using GraphPad Prism.

Spindle tracking was performed on maximum intensity projections (XY) using a custom written MATLAB routine [available upon request; as described previously (Marzo et al., 2019)]. Dynein-mediated spindle movements were manually selected from the tracking data to obtain the various metrics described in Figs 4B–E and 5B–E. To determine the fraction of time the spindle centroid resides within 1 µm of the cell cortex (Figs 4F and 5F), an additional MATLAB routine was generated with which the user manually defines the cell cortex of the mother and bud cell. To determine the fraction of time the spindle resides within mother and daughter cell (Figs 4G and 5G), and the relative distance from the bud neck (Figs 4H and 5H), another MATLAB routine was generated in which the user manually defines the bud neck. For this latter routine, the edges of the cell were defined by cropping individual cell images such that the extreme left and right cell edges coincided with the cropped image.

For western blotting, yeast cultures were grown at 30°C in 3 ml SD medium supplemented with 2% glucose and harvested. Equal numbers of cells were pelleted (1500 g for 2 min) and resuspended in 0.2 ml of 0.1 M NaOH and incubated for 10 min at room temperature as described (Kushnirov, 2000). After centrifugation, the resulting cell pellet was resuspended in sample buffer and heated to 100°C for 3 min. Lysates were separated on an SDS polyacrylamide gel and electroblotted to PVDF in 25 mM Tris and 192 mM glycine supplemented with 0.05% SDS and 10% methanol for 30 min. Rabbit anti-c-Myc polyclonal (cat. no A00172, GenScript) and HRP-conjugated goat anti-rabbit-IgG antibody (Jackson ImmunoResearch Laboratories) were used at 1:1000, and 1:3000, respectively. Total protein (using Stain-Free technology; BioRad) and chemiluminescence signal were both acquired on a BioRad ChemiDoc MP gel documentation system without saturating the pixels of the camera. Band intensities (and background values) were measured using NIH ImageJ.

For all datasets, P-values were calculated from Z scores (when comparing proportions) as previously described (Marzo et al., 2019), or by performing unpaired two-tailed Welch's t-test, or the Mann–Whitney test, the latter two of which were performed using GraphPad Prism. These latter tests were selected as follows: the unpaired two-tailed Welch's t-test was used when the datasets in question were both determined to be normal (by the D'Agostino and Pearson test for normality; P>0.05); in the case where only one (or neither) of the datasets were determined to be normal (P<0.05), the Mann–Whiney test was used.

We are extremely grateful to Luke Rice and members of his laboratory for sharing reagents and expertise pertaining to the expression, purification, and assembly of yeast tubulin. We are also grateful to Jeffrey Moore for sharing the nip100^ΔCAP-gly yeast strain, and members of the Markus and DeLuca laboratories for valuable discussions.

The peer review history is available online at https://journals.biologists.com/jcs/article-lookup/doi/10.1242/jcs.258510.