Function of Dynein in Budding Yeast: Mitotic Spindle Positioning in a Polarized Cell

Abstract

Cytoplasmic dynein is a microtubule motor that powers minus-end-directed motility in a variety of biological settings. The budding yeast, Saccharomyces cerevisiae, has been a useful system for the study of dynein, due to its molecular genetics and cell biology capabilities, coupled with the conservation of dynein-pathway proteins. In this review we discuss how budding yeast use dynein to manipulate the position of the mitotic spindle and the nucleus during cell division, using cytoplasmic microtubules, and we describe our current understanding of the genes required for dynein function.

Dyneins are AAA+ ATPases that use the energy from ATP hydrolysis to drive stepwise motility along microtubule tracks. Many dyneins are involved in axonemal function, found exclusively in organisms with cilia and flagella, but the cytoplasmic dynein 1 family is found in nearly all eukaryotes [Wickstead and Gull, 2007]. Cytoplasmic dynein is utilized for a variety of tasks, including transport of cargo along microtubules, positioning of the microtubule organizing center (MTOC), and organization of microtubule networks with respect to the cell cortex.

The study of dynein in budding yeast began with the identification of the DYN1 (or DHC1) gene, which encodes the cytoplasmic dynein heavy chain [Eshel et al., 1993; Li et al., 1993]. The amino acid sequence of yeast Dyn1 is highly similar to that of the heavy chain of cytoplasmic dynein in metazoans (Fig. 1). The most conserved region is the carboxy-terminal motor domain. This region includes a ring of six AAA+ motifs plus a stalk domain that extends outward from the ring. The tip of this stalk contains the microtubule-binding domain of the motor [Carter et al., 2008]. ATP-binding and hydrolysis by the first and third of the AAA+ motifs is required for dynein function in cells, based on mutation of P loops; similar mutations of the other AAA+ motifs do not exhibit noticeable loss-of-dynein phenotypes [Reck-Peterson and Vale, 2004]. In vitro, the motor domain of Dyn1 alone is sufficient for motility in microtubule-gliding assays, suggesting that the function of the amino-terminal tail domain in cells may be to form complexes, bind cargo, or regulate dynein activity [Reck-Peterson et al., 2006].

Fig 1.

Subunits of the dynein complex in Saccharomyces cerevisiae. Protein domain structure shown for dynein heavy chain, Dyn1, dynein intermediate chain, Pac11, dynein light intermediate chain, Dyn3, and dynein light chain (LC8-type), Dyn2. Dyn1 domain structure determined from [Mocz and Gibbons, 2001; Reck-Peterson et al., 2006; Carter et al., 2008; Markus et al., 2009]. Pac11 domain structure from [Geiser et al., 1997]. Dyn3 domain structure from [Hughes et al., 1995]. Dyn2 domain structure from [Dick et al., 1996]. Sequence conservation is shown for each protein: percent identity and percent positive as determined from BLAST searches and ClustalW2 sequence alignments. Dyn1 sequence compared to DYNC1H1 (Homo sapiens, accession number Q14204): Tail domain (aa1-1390), linker domain (aa1393-1745), AAA1 (aa1758-1977), AAA2 (aa 2053-2216), AAA3 (aa2403-2633), AAA4 (aa2732-2932), Stalk (aa3080-3237), MTBD (aa3115-3206), AAA5 (aa3379-3597), and AAA6 (aa3825-3960). Pac11 sequence compared to Dync1li1 (Rattus norvegicus, accession number Q63100): N-term (aa1-140), Serine rich region (aa141-161), and WD40 Domain (aa384-473). Dyn3 sequence compared to Dync1li2 (Rattus norvegicus, accession number Q62698): Dynein Light Intermediate Chain (DLIC) (aa163-278). Dyn2 sequence compared to LC8 (Drosophila melanogaster, accession number NP_525075): Dynein_Light superfamily (aa4-92).

The only known function of dynein in budding yeast is to position the mitotic spindle during cell division. Whereas many eukaryotes employ dynein and microtubule-based mechanisms for the intracellular trafficking of organelles and other cargoes, budding yeast accomplish these processes primarily using the actin cytoskeleton, owing most likely to the small size of the yeast cell [Fagarasanu and Rachubinski, 2007]. Yeast undergo a closed mitosis, and therefore the nucleus and its mitotic spindle must be positioned across the junction between the mother and bud, termed the bud neck, to provide a set of chromosomes to the daughter cell. Null mutants of DYN1 often fail to move the spindle into the bud neck promptly, and cells proceed into anaphase with spindles and nuclei located entirely within the mother [Eshel et al., 1993; Li et al., 1993]. In most of these cases, the spindle does ultimately move into the neck, after a variable delay with cell-cycle arrest in late anaphase.

The Kar9 pathway can compensate for the loss of dynein, using independent molecular mechanisms to move one end of the mitotic spindle into the nascent daughter cell. The Kar9 pathway guides cytoplasmic (i.e. astral) microtubules from one spindle pole body (SPB) into the bud by linking plus ends to a class-V myosin motor (Myo2) that moves along polarized cables of actin filaments [Miller and Rose, 1998; Hwang et al., 2003]. In addition, spindle elongation during anaphase B may help to push one pole into the bud. These compensatory processes require additional time to position the spindle, and this time is provided by a checkpoint mechanism that monitors the position of the spindle, delaying exit from mitosis until one SPB enters the bud [Adames et al., 2001; Molk et al., 2004]. Together, these mechanisms prevent the generation of bi-nucleate mothers and anucleate daughters.

Many dynein components have been identified in yeast, owing to the viability and distinct phenotypes of dynein mutants. Genome-wide screens of haploid null mutants have identified a set of genes producing phenotypes similar to the loss of dynein, including nuclear segregation defects and synthetic genetic interactions with mutations in the Kar9 pathway [Tong et al., 2004; Lee et al., 2005; Li et al., 2005]. These dynein-pathway genes include homologues of many of the dynein constituents and regulators found throughout eukaryota. Divided into functional categories, these genes encode chains of the dynein motor (HC/Dyn1, IC/Pac11, LIC/Dyn3, LC8/Dyn2); the dynactin complex (p150glued/Nip100, p24/Yll049w, dynamitin/Jnm1, Arp1, Arp11/Arp10); the microtubule-associated proteins CLIP170/Bik1 and kinesin, Kip2; and the LIS1 and NudEL homologs, Pac1 and Ndl1. In addition, these screens have identified one gene without a clear functional homologue in higher eukaryotes — NUM1, which encodes a cortical protein.

How does dynein power movement of the spindle and nucleus?

Timelapse microscopy of GFP-labeled microtubules in living cells has provided key insight into how dynein moves the spindle and nucleus. Movies demonstrate that spindle position within the dividing cell depends on interactions of cytoplasmic microtubules with the cell cortex [Carminati and Stearns, 1997; Shaw et al., 1997; Adames and Cooper, 2000; Yeh et al., 2000]. Dynein mutants are specifically defective for what have been termed microtubule “sliding” events, during which the plus end of the microtubule strikes the cortex and then curls along it, maintaining lateral contact along its side while sliding forward in the direction of the plus end (Fig. 2). This sliding produces movement of the spindle by drawing the minus end of the microtubule, and the attached SPB, toward the site of cortical contact. The SPB is embedded in the nuclear envelope, and its movement pulls on the proximal edge of the nuclear envelope [Byers and Goetsch, 1975];(Moore, in press). In other types of interactions, microtubules can attach to the cortex and then appear to shrink or sweep side-to-side, maintaining their end-on connection but without a lateral interaction. Although these types of behaviors appear not to depend on dynein, cells that lack dynein do exhibit an increased frequency of long microtubules that grow along the cell cortex, rather than making end-on interactions [Carminati and Stearns, 1997; Adames and Cooper, 2000]. The basis for this phenotype is not clear, but may be explained by either a suppression of microtubule dynamics or perhaps a role for dynein in promoting end-on capture at the cortex.

Fig 2.

Spindle movement is powered by microtubule sliding along the cell cortex. Timelapse images of GFP-labeled microtubules. Several cytoplasmic microtubules project outward from each end of the short bipolar spindle. These microtubules explore the cytoplasm by growing and shrinking. As a microtubule end hits the cell cortex, dynein pulls the microtubule past the site of interaction, drawing the spindle forward and through the bud neck. Each image is a composite of 9 planes separated by 0.5μm; stacks were captured at 10-second intervals. Scalebar: 1μm.

Microtubule sliding seems to be generated by forces acting at the microtubule-cortex interface. The best evidence for this conclusion comes from observing the movement of free microtubules, which occasionally detach from the SPB. These free microtubules exhibit dynein-dependent sliding over long distances, occasionally circumnavigating the entire cell cortex, including mother and bud [Adames and Cooper, 2000]. During these events, the plus ends of free microtubules move persistently along the cortex, with fluorescent speckles providing fiducial marks; this result argues against the possibility that sliding is driven by a depolymerization or capture-shrinkage mechanism. Instead, force production can be explained by a model in which dynein is anchored to the cortex, and its stepping behavior moves the microtubule past the fixed motor. This model implies that dynein mediates a dynamic interaction between microtubules and the cell cortex, and therefore requires that the motor associate with elements of both structures.

Visualizing the motor complex in living cells provided surprising evidence about how dynein works. Cells expressing a 3X-GFP fusion of dynein heavy chain from the endogenous locus exhibit punctae of dynein at several locations [Lee et al., 2003; Sheeman et al., 2003]. The most prominent localization of Dyn1-3GFP is to cytoplasmic microtubules, with distinct enrichment at distal plus ends and at the minus ends near the SPB. Some signal is also observed along the length of the cytoplasmic microtubules. Remarkably, dynein remains accumulated at the plus end of a microtubule while the microtubule grows and shrinks, adding and losing tubulin subunits. The molecular basis of this plus-end tracking is discussed in detail below.

In addition to its association with cytoplasmic microtubules, dynein is also present in stationary foci on the cell cortex. Only a portion of these foci co-localize with microtubule plus ends, suggesting that the remainder represent dynein in a stable complex with a cortical binding partner, such as Num1. Dynein has not been detected on nuclear microtubules, consistent with the lack of spindle defects in dynein mutants.

Two important unresolved issues are which of these locations dynein needs to occupy or visit in order to function, and at which location dynein exerts force to move microtubules. With regard to the SPB, in other cell systems, dynein at the centrosome helps to organize and anchor microtubule minus ends, but there is no evidence for that function in budding yeast, based on observations of mutants, and we know from speckle microscopy that the minus ends of microtubules are anchored in the SPB [Maddox et al., 2000]. However, some studies have found that dynein accumulates selectively at the daughter-bound SPB and that this might account for higher daughter-cell specific activity of dynein [Shaw et al., 1997; Segal et al., 2000; Grava et al., 2006].

A number of studies have explored the interdependence and functional relevance of these localizations by monitoring Dyn1-3GFP in the presence of mutations in other components of the dynein pathway [Lee et al., 2003; Sheeman et al., 2003; Carvalho et al., 2004; Lee et al., 2005; Li et al., 2005; Caudron et al., 2008; Moore et al., 2008]. Collectively, these studies suggest that localization of dynein to plus ends and to the cortex is necessary for function. In the model that we and others currently favor, dynein is actively targeted to the dynamic plus ends of cytoplasmic microtubules, and then delivered or “offloaded” from these ends to the cortex when the dynamic microtubule plus end encounters a cortical anchor (Fig. 3). Dynein is then activated to drive microtubule sliding and spindle translocation.

Fig 3.

Model for dynein localization and function. Dynein and its activator, dynactin, are targeted to microtubule plus ends by a complex of kinesin and CLIP-170. Retention of dynein-dynactin at the plus end requires LIS1 and NudEl, and may also involve CLIP-170. By associating processively with the plus end, microtubule growth delivers dynein-dynactin to the cortical receptor, Num1. Interaction with Num1 facilitates the docking of dynein-dynactin to the cell cortex. Activation of motor activity causes dynein to then walk toward the microtubule minus end, which pulls the attached spindle pole and proximal edge of the nuclear envelope toward the anchored motor.

Contributions of motor subunits to dynein function

Although a dimer of heavy chain subunits is sufficient for motility in vitro, the dynein complex in cells consists of several conserved subunits, each of which is necessary for proper activity. Yeast homologues of the intermediate chain (IC/Pac11), light-intermediate chain (LIC/Dyn3), and light chain (LC/Dyn2) were identified by a dynein-like mutant phenotype, sequence similarity, and biochemical interactions [Geiser et al., 1997; Kahana et al., 1998; Lee et al., 2005]. The amino-terminal portion of yeast IC/Pac11 contains similarity to the region of vertebrate IC that is responsible for interaction with the dynactin complex and LCs; while the carboxy-terminal half contains WD repeat motifs, which are found in dynein ICs across species (Fig. 1; [Vaughan and Vallee, 1995; Geiser et al., 1997; Lo et al., 2001]. In contrast, LIC/Dyn3 is less similar to its metazoan orthologues, and is thought to be a distant relative of dynein LICs [Lee et al., 2005]. LC/Dyn2 contains a high level of similarity to the LC8 family of dynein light chains [Dick et al., 1996].

The localization of each chain suggests discrete roles in the dynein mechanism. HC/Dyn1, IC/Pacll, LIC/Dyn3, and LC/Dyn2 all localize to microtubule plus ends, based on double staining with fluorescent tubulin and on movie behavior characteristic of plus ends. Pairwise combinations with individual chain mutants have been tested to determine the hierarchy of complex formation on microtubules. The HC/Dyn1 is necessary for localization of IC, LIC, and LC to plus ends [Lee et al., 2005; Moore et al., 2008]. Plus-end localization of HC/Dyn1 is lost in mutants lacking IC/pac11Δ but not ones lacking LC/dyn2Δ or LC/dyn3Δ. In terms of stationary cortical foci, HC/Dyn1 and IC/Pac11 localize at these locations [Lee et al., 2005]. In contrast, LIC/Dyn3 does not appear to localize to the cortex, but it is required for HC/Dyn1 and IC/Pac11 to localize to the cortex [Lee et al., 2005]. Together, these studies suggest that the IC/Pac11 is an obligate partner of the HC/Dyn1, required for dynein function in cells; while LIC appears to be specifically involved in the transfer of dynein from microtubule ends to cortical binding sites, and released from the complex following this transfer. The role of the LC remains unclear. The LC/dyn2Δ mutant exhibits a partial defect in certain spindle position assays, suggesting it is important but not critical for dynein activity (Stuchell-Brereton, Li and Cooper; unpublished).

Delivering dynein via dynamic microtubule plus ends

The plus ends of cytoplasmic microtubules undergo phases of polymerization and depolymerization, leading to frequent collisions with the cell cortex. Dynein localizes to these dynamic ends, and this may allow the motor to search the cortex for docking sites [Sheeman et al., 2003]. The persistent association of dynein with plus ends is paradoxical, given the minus-end-directed motility of dynein. Dynein also targets plus ends in other cell systems, and in each case the localization is mediated by proteins that selectively associate with plus ends [Vaughan et al., 1999; Dujardin et al., 2003].

The accumulation of dynein at plus ends in yeast is accomplished in two steps. First, dynein is targeted to plus ends by the CLIP-170 homologue Bik1, a microtubule-binding protein that decorates the length of cytoplasmic and nuclear microtubules, with enrichment at plus ends and tip-tracking behavior [Carvalho et al., 2004]. In the absence of Bik1, dynein is lost from plus ends, but a small amount is still present at the SPB-associated minus ends [Sheeman et al., 2003]. The localization of Bik1 and dynein to plus ends is enhanced by a plus-end-directed kinesin, Kip2 [Carvalho et al., 2004; Caudron et al., 2008]. Kip2 may transport Bik1 and dynein because punctae of either protein exhibit plus-end-directed movement along microtubules in live cells, and this requires Kip2 [Carvalho et al., 2004; Caudron et al., 2008].

The primary role of Bik1 in plus-end targeting may be to link dynein to Kip2, which is supported by observations of α-tubulin mutants that lack binding sites for Bik1. In these cells, the localization of Bik1 to microtubule plus ends is greatly reduced, but dynein localization is not strongly affected [Badin-Larcon et al., 2004; Caudron et al., 2008]. Thus, targeting of dynein does not absolutely require the binding of Bik1 to plus ends. Nevertheless, the presence of reduced amounts of Bik1 and dynein at plus ends in kip2 null mutants suggests that Bik1 may directly recruit dynein to plus ends as an alternative, albeit less efficient, mechanism.

In the second step for plus-end targeting, retention of dynein depends on Pac1, the yeast homologue of the human lissencephaly protein, LIS1, and its binding partner NudEL/Ndl1. Pac1 and Ndl1 localize to discrete foci at microtubule plus ends [Lee et al., 2003; Li et al., 2005]. Pac1 and dynein are found at a similar frequency of plus ends, and fluorescence intensity measurements indicate that the amount of Pac1 at these ends is similar to that of dynein [Lee et al., 2003; Li et al., 2005]. Ndl1 is also present at similar amounts, but found at a lower frequency of plus ends [Li et al., 2005]. These results support a model in which Pac1-Ndl1-dynein form a complex at the plus end, but suggest that Ndl1 may have a minor or perhaps transitory role. Indeed, Ndl1 appears to promote the localization of Pac1 to plus ends, and the requirement for Ndl1 in the dynein pathway can be rescued by overexpression of Pac1 [Li et al., 2005].

A recent study demonstrates that the motor domain of the heavy chain is necessary and sufficient for the accumulation of dynein at plus ends [Markus et al., 2009]. A motor domain fragment alone displays increased localization to plus ends, and this requires Bik1 and Pac1. The ATPase activity of the third AAA+ domain of dynein HC is not required. Expression of the motor domain fragment can inhibit the localization of full-length dynein to microtubules, resulting in a dominant-negative inhibition of dynein function. These results suggest that Bik1 and Pac1 interact directly with the motor domain, and that the motor-domain fragment may compete with full-length dynein for a limited number of binding sites on the microtubule. Consistent with this hypothesis, overexpression of Pac1 rescues the localization and function of full-length dynein in the presence of the motor domain fragment.

Although Pac1 and Ndl1 appear to form a complex with dynein at plus ends, neither protein co-localizes with dynein at cortical foci [Lee et al., 2003; Li et al., 2005]. These observations suggest that concentrating dynein at plus ends may be the sole of function of Pac1 and Ndl1 in the dynein pathway. Further investigation will be needed to determine whether Pac1 and Ndl1 are obligatory components of the motor complex during microtubule sliding.

Tethering dynein to the cortex

The localization of dynein to a small number of discrete non-motile foci on the cell cortex suggests that docking sites for the motor may be limited to subdomains of the plasma membrane that contain specific receptors. Several lines of evidence indicate that the cortical protein Num1 facilitates the transfer of dynein from microtubules to the cortex or may even function as a binding partner, anchoring dynein to the cortex. Num1 localizes to non-moving patches on the plasma membrane that are enriched at the distal tip of the bud during late mitosis, and Num1 localization does not require dynein or microtubules [Farkasovsky and Kuntzel, 1995; Heil-Chapdelaine et al., 2000]. Null mutations of num1 cause effects similar to those of dynein heavy chain null mutations, and show an absence of dynein at the cortex, with accumulation of dynein at microtubule plus ends over time [Lee et al., 2003; Moore et al., 2008]. In addition, Num1 protein has been found to form a biochemical complex with IC/Pac11 [Farkasovsky and Kuntzel, 2001]. All these data are consistent with Num1 acting as a required component of the receptor for dynein at the plasma membrane.

Num1, a large protein with predicted mass of 313 kDa, does not have a clear homologue in higher eukaryotes based on sequence similarity. The fission yeast protein, Mcp5, is regarded as a functional homologue of Num1, although sequence similarity between the two is poor [Saito et al., 2006; Yamashita and Yamamoto, 2006]. The two regions that display the highest sequence similarity between Num1 and Mcp5 are located near the amino and carboxy-termini (Fig. 4a). The amino-terminal region, predicted to form a coiled-coil structure, is required for Num1 to function [Farkasovsky and Kuntzel, 1995]; however, the precise function of this region is not known. The carboxy-terminal region contains a pleckstrin homology (PH) domain, the only PH domain in budding yeast that binds specifically and tightly to PI(4,5)P₂ [Yu et al., 2004]. The PH domain is necessary and sufficient for the localization of Num1 to the cortex, and excision of the PH domain yields a complete loss of function for Num1 in the dynein pathway (Fig. 4b; [Farkasovsky and Kuntzel, 1995]. This indicates that in addition to providing a docking site for dynein, Num1 may interface with phospho-lipid signaling pathways at the plasma membrane.

Fig 4.

Protein domains and localization of Num1. A) Domain architecture of Num1. The amino-terminal coiled-coil (CC) was identified using the secondary structure prediction program, Coils. The tandem repeat region consists of twelve repeats of 64 amino acids [Verstrepen et al., 2005]. The function of this region is not known. The PH domain was confirmed by a genome-wide analysis of PH domains in budding yeast [Yu et al., 2004]. B) The PH domain is necessary and sufficient to target Num1 to the plasma membrane. Localization of full-length Num1, a truncation that lacks the PH domain (ΔPH), and the PH domain alone. Each allele is tagged with RFP and expressed from the endogenous chromosomal locus (Moore, unpublished).

Roles of the dynactin complex

Dynactin is a multi-subunit complex that biochemically associates with dynein and is necessary for essentially every known function of the dynein motor across species [Schroer, 2004]. For vertebrate dynactin complex, the biochemistry and structure have been relatively well-defined. The complex consists of a short, actin-like filament of the actin-related protein-1 (Arp1) protein, overlaid by a shoulder-sidearm complex of dynamitin, p24, and p150^glued. The Arp1 filament appears to be capped at the pointed end by Arp11 and a complex of the p25, p27, and p62 subunits. The barbed end of the filament appears to contain a molecule of actin capping protein (CapZ) and a molecule of conventional actin.

For budding yeast, homologues of most dynactin subunits have been identified, but not the p25, p27, and p62 subunits of the pointed-end complex [McMillan and Tatchell, 1994; Muhua et al., 1994; Kahana et al., 1998; Clark and Rose, 2006; Amaro et al., 2008; Moore et al., 2008]. Evidence from filamentous fungi suggests that the pointed-end complex is specifically involved in dynein-dependent vesicle transport; thus, these biochemical components may have been lost in budding yeast, along with vesicle transport function [Lee et al., 2001]. In addition, existing biochemical and genetic evidence in budding yeast indicates that capping protein is not part of the complex.

Yeast dynactin appears to function in the off-loading and activation steps of the dynein mechanism. Dynactin co-localizes with dynein at microtubule ends and in foci on the cell cortex [Moore et al., 2008]. In the absence of dynactin, dynein appears at microtubule ends but not on the cortex [Lee et al., 2003; Moore et al., 2008]. Thus, dynactin is not needed to target or retain dynein at microtubule ends, but it may contribute to the offloading process or to docking at the membrane, analogous to a cargo-binding role.

How dynactin promotes the assembly of dynein foci at the cortex is not clear. Although dynactin has been reported to associate with the plasma membrane, the localization of dynactin to cortical foci is dependent on dynein, which argues that dynactin is necessary but not sufficient for cortical localization[Clark and Rose, 2006; Moore et al., 2008]. Perhaps the dynein-dynactin complex must first assemble at the microtubule end before both complexes bind to their cortical docking site.

In addition to facilitating interactions with cargo, vertebrate dynactin can also enhance the level of persistence for dynein moving along microtubules, based on in vitro motility assays [King and Schroer, 2000]. The p150^glued subunit of dynactin contains a Cytoskeletal-Associated-Protein-Glycine-rich (CAP-Gly) domain, which binds directly to microtubules and also to the microtubule-binding proteins, CLIP-170 and EB1[Waterman-Storer et al., 1997; Honnappa et al., 2006; Weisbrich et al., 2007].

We recently discovered that mutation of this conserved CAP-Gly domain in the yeast p150^glued homologue Nip100 diminishes the initiation rate and persistence of microtubule sliding along the cell cortex (Moore, in press). This effect is most apparent during the first sliding event that pulls the nucleus into the bud neck. In contrast, disruption of the CAP-Gly domain did not noticeably alter dynein function later, when microtubule sliding was not coupled with this initial nuclear movement. The specificity of this phenotype suggests that the CAP-Gly domain of p150^glued contributes to dynein activity specifically when the motor is under greater load. Perhaps the Nip100 CAP-Gly domain binds to tubulin and microtubule-associated proteins, enhancing the avidity of the dynein-dynactin complex for its microtubule substrate and resisting load-dependent detachment. In humans and mice, homologous CAP-Gly mutations are sufficient to cause an ALS-like neurodegeneration syndrome.

How is dynein regulated?

Microtubule sliding moves the spindle and nucleus during mitosis, but is not active during G₁, suggesting that the dynein pathway is coordinated with the cell cycle [Yeh et al., 1995]. The activation of dynein-dependent movements may be triggered by the signaling networks that control the initiation of anaphase. Cells arrested at the G₂/M transition in response to DNA damage contain spindles that are positioned at the bud neck, but seldom move across it. However, arrested cells lacking either the Chk1 or Rad53 checkpoint kinases exhibit frequent spindle movement back-and-forth across the bud neck, despite a persistent delay of anaphase [Dotiwala et al., 2007]. Furthermore, cells disrupted for the Cdc14 early anaphase release (FEAR) pathway, which controls several aspects of spindle function, exhibit a hyper-migration phenotype in which the entire nucleus is pulled across the neck and into the bud [Ross and Cohen-Fix, 2004]. These data suggest that cell cycle signaling regulates the timing and perhaps the polarity of dynein activity. A priori, regulation could be accomplished at any one of several points in the dynein pathway — plus-end targeting of dynein, attachment of dynein to cortical binding sites, assembly of the dynein-dynactin complex, and modulating dynein’s ATPase activity.

Dynein localization does vary through the cell cycle. Dynein is most prominently enriched at microtubule plus ends prior to anaphase [Sheeman et al., 2003]. This is thought to be driven by the Kip2 kinesin, which is most active at this stage [Carvalho et al., 2004]. Foci of dynein are only seen at the bud cortex in large-budded cells, suggesting that offloading may only occur in the bud during anaphase [Lee et al., 2005]. Consistent with this notion, mutations that prevent cortical localization lead to increased amounts of dynein at plus ends in late anaphase [Moore et al., 2008].

The molecular mechanisms by which the cell cycle machinery might control dynein are not well understood. Genetic evidence suggests that dynein can be regulated by the type 1 protein phosphatase Glc7. Overexpression of a Glc7 regulatory subunit, Bud14, causes hypermigration of the spindle into the bud prior to anaphase, and this effect depends on dynein and Num1 [Knaus et al., 2005]. Furthermore, mutations in Bud14 exhibit synthetic growth defects when combined with mutations in the Kar9 pathway, which is characteristic of dynein pathway mutations [Knaus et al., 2005]. These data are consistent with Glc7-Bud14 acting as a positive regulator of dynein activity. Phospho-regulation of dynein has been described in other systems; this mechanism will be important to test in yeast [Dell et al., 2000; Kumar et al., 2000; Whyte et al., 2008].

In addition to temporal regulation, dynein force-production may be spatially coordinated by cell polarity. Glc7 is recruited to the tip of the growing bud by Bud14, suggesting a model for generating bud-directed microtubule sliding by promoting dynein activation at this site [Knaus et al., 2005]. Alternatively, polarized force may be generated through the biased recruitment of dynein to a specific set of microtubules. Early observations of fluorescently labeled dynein noted that in cells with short bipolar spindles, dynein was enriched on the set of cytoplasmic microtubules emanating from the SPB that was proximal to the bud [Shaw et al., 1997; Segal et al., 2000]. It has been hypothesized that this asymmetry might favor microtubule sliding in the bud. A study of dynein asymmetry found that the selective accumulation of dynein at one SPB and set of microtubules correlates with the interaction of those microtubules with the bud neck [Grava et al., 2006]. Moreover, asymmetric enrichment of dynein was disrupted by mutations in the cyclin dependent kinase, Cdc28, and several protein kinases that localize to the bud neck [Grava et al., 2006].

Concluding remarks

The use of dynein to organize the microtubule network with respect to the cell cortex is a prevailing theme across eukaryotes. In metazoans, positioning of the mitotic spindle by dynein is critical for the success of asymmetric cell divisions that underlie cell differentiation [Couwenbergs et al., 2007; Nguyen-Ngoc et al., 2007]. Migrating cells such as neuronal precursors use dynein to move the nucleus and assemble microtubule networks along the cell cortex, which are thought to provide a transport network to the leading edge and resist retrograde actin flow during outgrowth [Grabham et al., 2007; Tsai et al., 2007]. In cytotoxic lymphocytes, dynein pulls the microtubule network to the immune synapse, facilitating the delivery of toxic granules to target cells [Combs et al., 2006; Stinchcombe et al., 2006]. In each of these cases, the mechanisms by which dynein is targeted to specific cortical sites in order to generate polarized force are poorly understood. Thus, the characterization of dynein function in yeast will likely provide insight into similar processes in other organisms.

Yeast has also become a powerful tool for the study of fundamental properties of dynein motility. Dynein heavy chain can be purified to near homogeneity from yeast cell extracts, and the movement of single motors can be analyzed with in vitro assays [Reck-Peterson et al., 2006]. These techniques have been used to address questions of how dynein steps along a microtubule; how the addition of load, induced by optical tweezers, influences stepping behavior; and how point mutations in individual AAA+ domains affect these properties [Reck-Peterson et al., 2006; Gennerich et al., 2007; Cho et al., 2008]. Remarkably, the velocities of single dynein motors in these assays are similar to those of sliding free microtubules in cells [Adames and Cooper, 2000]; Moore, in press). Moreover, we have found that the velocity and persistence of sliding are diminished when coupled with the movement of the nucleus into the bud neck (Moore, in press). This supports the idea that dynein motility is altered in scenarios where the motor is asked to move greater loads, and raises the possibility that dynein may be tuned by accessory factors in order to generate more force. In the future, it will be important to combine these techniques with further mutational analysis of dynein and its regulators to improve our understanding of how dynein moves along microtubules and may be directed to produce precise displacements of cargoes.

Abbreviations used

AAA+

ATPase associated with various cellular activities

ALS

amyotrophic lateral schlerosis

CAP-Gly

cytoskeletal-associated protein glycine-rich

CLIP-170

cytoplasmic linker protein-170

EB1

end-binding protein 1

GFP

green fluorescent protein

HC

heavy chain

IC

intermediate chain

LC

light chain

LIC

light intermediate chain

MTOC

microtubule organizing center

PH

pleckstrin homology

RFP

red fluorescent protein

SPB

spindle pole body