Cytoplasmic dynein in fungi: insights from nuclear migration

Cytoplasmic dynein is a microtubule motor that mediates various biological processes, including nuclear migration and organelle transport, by moving on microtubules while associated with various cellular structures. The association of dynein with cellular structures and the activation of its motility are crucial steps in dynein-dependent processes. However, the mechanisms involved remain largely unknown. In fungi, dynein is required for nuclear migration. In budding yeast, nuclear migration is driven by the interaction of astral microtubules with the cell cortex; the interaction is mediated by dynein that is probably associated with the cortex. Recent studies suggest that budding yeast dynein is first recruited to microtubules, then delivered to the cortex by microtubules and finally activated by association with the cortex. Nuclear migration in many other fungi is probably driven by a similar mechanism. Recruitment of dynein to microtubules and its subsequent activation upon association with cellular structures are perhaps common to many dynein-dependent eukaryotic processes, including organelle transport.

four 50-60 kDa light intermediate chains (DLICs) and several 6-22 kDa light chains (DLCs) (King et al., 1998;Bowman et al., 1999) (reviewed by Holzbaur and Vallee, 1994;). DHC is the motor unit of dynein. It belongs to the AAA superfamily (ATPases associated with cellular activities) and has six tandem AAA modules, the first of which contains an ATPbinding catalytic site (reviewed by King, 2000;Asai and Koonce, 2001). The AAA modules form a ring-like, globular head that lies between two extended structures, the stalk and stem (Fig. 1A). DHC is thought to interact with a microtubule at the tip of the stalk and with other cellular structures through the stem (Fig. 1C). DLICs, DICs and DLCs are associated with the stem and probably are involved in binding to such structures and/or regulate motor activity.
Cytoplasmic dynein requires dynactin for its functions (Gill et al., 1991;Schroer and Sheetz, 1991) (reviewed by Allan, 1996). Dynactin is a protein complex that comprises two distinct structural components: a short, actin-like filament and a projecting sidearm (Fig. 1B) (Schafer et al., 1994;Eckley et al., 1999). The actin-like filament consists of a polymer of the actin-related protein Arp1 and several attached proteins [a capping protein and pointed-end-binding proteins (PEBPs) including p62, p25 and Arp11]. The sidearm consists of a dimer of p150 Glued , which contains distinct binding sites for microtubules, DIC and Arp1 (Karki and Holzbaur, 1995). The sidearm is probably connected to the actin-like filament, together with another p24 protein, through dynamitin (Echeverri et al., 1996). Dynactin is proposed to mediate the interaction of cytoplasmic dynein with cellular structures and/or to regulate dynein motility (Fig. 1C). Evidence for a regulatory role of dynactin in the dynein motility includes the observation that dynactin increases the distance that dynein can move along a microtubule in vitro by decreasing the frequency of its dissociation from the microtubule (King and Schroer, 2000). Dynactin also affects the ATPase activity of dynein by changing its phosphorylation state (Kumar et al., 2000).
CLIP-170 and LIS1 are also required for the dynein functions. CLIP-170 belongs to a class of proteins that preferentially interact with growing microtubule plus ends for binding to the growing plus ends (they are referred to as 'plusend-tracking proteins' or '+TIPs') (Diamantopoulos et al., 1999;Perez et al., 1999) (reviewed by Schuyler and Pellman, 2001) and plays a crucial role in endosome transport (Pierre et al., 1992). It recruits dynactin near or at the plus ends (Valetti et al., 1999;Vaughan et al., 1999;Sheeman et al., 2003) and is required for dynein-dependent nuclear migration in fungi (see below). LIS1 was identified as the product of a gene whose mutation causes human lissencephaly ('smooth brain') (Reiner et al., 1993), and was shown to bind to dynein, dynactin and CLIP-170 (Faulkner et al., 2000;Sasaki et al., 2000;Smith et al., 2000;Coquelle et al., 2002;Tai et al., 2002).
In addition to these cytoplasmic factors, other proteins are required for dynein function. In metazoa, spectrin is associated with membranous organelles and is required for dynein-dependent vesicle transport (reviewed by De Matteis and Morrow, 2000;Muresan et al., 2001). In fungi, Num1 or its related factor ApsA is associated with the cell cortex and required for dynein-dependent nuclear migration (see below). Both spectrin and Num1 contain pleckstrin homology (PH) domains and probably interact with the membrane through these domains (Farkasovsky and Kuntzel, 1995;Muresan et al., 2001). Furthermore, spectrin binds to Arp1, a major subunit of the actin-like filament of dynactin (Holleran et al., 1996), and Num1 binds to DIC (Farkasovsky and Kuntzel, 2001). Such factors probably anchor dynein to specific cellular structures and perhaps regulate its motility.

Cytoplasmic dynein in budding yeast
The functions of cytoplasmic dynein have been most extensively studied in budding yeast. In budding yeast mitosis, the nucleus is positioned at the bud neck during division. This is essential for faithful segregation of the chromosomes to the mother and bud cells, and cytoplasmic dynein is required for migration of the nucleus to the bud neck.
The nucleus migrates to the neck in two steps: it first migrates to a position proximal to the neck and then enters the neck (DeZwaan et al., 1997;Adames and Cooper, 2000). The first step occurs during G1/S phase and the second shortly before the onset of anaphase. Two partially overlapping mechanisms drive this two-step process. Nuclear migration is mediated by astral microtubules, which emanate from the spindle pole body (SPB) . The initial migration to the neck-proximal position is mainly driven by shortening of the astral microtubules, whose ends are captured at the cell cortex. This shortening-dependent migration requires actin and a set of proteins including Kar9 and Bim1, which are (respectively) the yeast orthologs of the adenomatous polyposis coli (APC) tumour suppressor and the microtubule-binding protein EB1 (Palmer et al., 1992;Cottingham and Hoyt, 1997;DeZwaan et al., 1997;Schwartz et al., 1997;Cottingham et al., 1999;Lee et al., 1999;Miller et al., 1999;Adames and Cooper, 2000;Beach et al., 2000;Heil-Chapdelaine et al., 2000b;Korinek et al., 2000;Yeh et al., 2000;Yin et al., 2000;Bienz, 2001). Kar9 and Bim1 interact with each other, and Bim1 localizes to microtubule distal ends, recruiting Kar9 to them. Kar9 also localizes to the cell cortex through actin and so is thought to connect microtubule distal ends with the cortex.
The subsequent nuclear migration into the neck is driven mainly by lateral attachment of astral microtubules to the cortex, and their sliding at the attachment site ( Fig. 2) (Adames and Cooper, 2000;Yeh et al., 2000). This sliding-dependent migration requires cytoplasmic dynein and its interactors (Table 1). Microtubule sliding occurs in both the mother and the bud cells during anaphase, generating opposing pulling forces on the nucleus (Fig. 2). The opposing pulling forces orient the spindle along the cell axis, causing oscillatory movements of the nucleus at the neck (Yeh et al., 1995;Yeh et al., 2000;Adames and Cooper, 2000), and contribute to spindle pole separation (Saunders et al., 1995). The shortening-Journal of Cell Science 116 (22)  (Karki and Holzbaur, 1995;Holleran et al., 1996). For simplicity, only DHC, DIC, p150 Glued , p50 and Arp1 are shown. Dynactin interacts with dynein via association of a middle portion of p150 Glued with DIC. Dynein interacts with a microtubule at the tip of the stalk of DHC, whereas dynactin interacts with the microtubule at an Nterminal globular domain of p150 Glued . Dynein/dynactin interacts with a cellular structure via an Arp1 filament. Arrow indicates the direction of dynein/dynactin movement. The minus (-) and plus (+) ends of the microtubule are indicated. C dependent mechanism and sliding-dependent mechanism are functionally redundant for nuclear migration, and thus neither is essential for cell viability. However, loss of both mechanisms is lethal to the cell owing to a complete block of nuclear migration into the neck (Cottingham and Hoyt, 1997;DeZwaan et al., 1997;Geiser et al., 1997;Schwartz et al., 1997;Kahana et al., 1998;Miller et al., 1999;Cottingham et al., 1999;Lee et al., 1999). Driving mitotic nuclear migration is the only known function of dynein in Saccharomyces cerevisiae.
It is generally thought that cortical dynein mediates the attachment of microtubules to the cortex and their cortical sliding. How dynein is associated with the cortex and how its interactors participate in this process or regulate dynein motility have long been poorly understood. Recently, enhancing the fluorescence signal of a tagged dynein molecule by incorporating three tandem copies of green fluorescent protein (GFP), Lee et al. and Sheeman et al. were able to view the localization of dynein at its native expression level (Table  2) (Lee et al., 2003;Sheeman et al., 2003). The tagged DHC localizes to the SPB and astral microtubules, prominently accumulating at the microtubule ends distal to the SPB. This distal-end accumulation of DHC depends on LIS1 and CLIP-170, both of which are localized to the distal ends (Table 2). Interestingly, the distal-end accumulation is increased by depletion of dynactin (Lee et al., 2003;Sheeman et al., 2003) or a mutation that is thought to make the DHC motor immotile (Sheeman et al., 2003). Furthermore, it is also increased by depletion of the cortical protein Num1, which is thought to anchor dynein to the cortex (see above). These results form the basis of the following model (Fig. 3A). Cytoplasmic dynein is initially inactive and is recruited to microtubules, together with

Mother
Spindle Bud SPB Nucleus AMT

Fig. 2.
A role of dynein in the budding yeast S. cerevisiae. Dynein mediates attachment of astral microtubules (AMT; purple lines) to the bud cortex and sliding of the microtubules on the attachment site (red arrow). Consequently, the nucleus (blue sphere) moves into the bud neck (black arrow). Dynein also drives cortical sliding of microtubules in the mother cell (red arrow), generating an opposing force on the nucleus. As a result, the nucleus remains at the neck with oscillations between the mother and bud cells (black arrows). The opposing forces also orient the mitotic spindle along the cell axis, and contribute to spindle pole separation.
*Genes which are not required for the dynein functions. † A gene whose functions in the dynein pathway are unknown. All other genes are required for the dynein functions (see the following references).
(1) Lee et al., 2003, Sheeman et al., 2003; (2) Yamamoto et al., 1999(8) Miki et al., 2002;(9) Yamashita et al., 1997;(10) Brunner and Nurse, 2000; (11) Xiang et al., 1995b, Xiang et al., 2000(12) Zhang et al., 2002;(13) Zhang et al., 2003;(14) Han et al., 2001;(15) Suelmann et al., 1997;(16) Minke et al., 1999b, Minke et al., 2000 Abbreviations: CMTs, cytoplasmic microtubules; SMTs, spindle microtubules; SPB, spindle pole body. the microtubules (Table 2). One possibility is that these fusion proteins are indeed localized at the microtubule distal ends, but their localization is undetectable because they accumulate at low levels. It will be interesting to re-examine the localization of dynactin using three GFPs, as suggested by Sheeman et al. (Sheeman et al., 2003). Second, how is microtubule polymerization/ depolymerization regulated during nuclear migration? This must be regulated, because microtubules normally repeatedly elongate and shorten but, during cortical microtubule sliding, microtubules are relatively constant in length (Adames and Cooper, 2000;Yeh et al., 2000). Moreover, dynein/dynactin promotes microtubule depolymerization (Carminati and Stearns, 1997;Cottingham and Hoyt, 1997;Adames and Cooper, 2000;Yeh et al., 2000) and a kinesin-related protein, Kip2 (which is also required for dynein-dependent nuclear migration), promotes microtubule polymerization (Cottingham and Hoyt, 1997;. Thus, the regulation of microtubule polymerization/depolymerization is probably performed by the concerted actions of dynein/dynactin and Kip2. Microtubule polymerization and depolymerization occur exclusively at the distal ends of astral microtubules . Dynein/dynactin or Kip2p might regulate microtubule polymerization/ depolymerization by transporting a factor that regulates microtubule polymerization to their distal ends. Finally, how is the dynein function regulated during the cell cycle? Dynein-dependent nuclear migration occurs around anaphase. In addition, accumulation of dynein to the microtubule plus ends is more prominent at anaphase than before anaphase (Sheeman et al., 2003). Therefore, dynein activity or its interaction with microtubules and/or the cell cortex must be regulated in a cell-cycle-dependent manner. Recently, Kip2 was shown to translocate Kar9 along microtubules, causing accumulation of Kar9 at the microtubule distal ends (Maekawa et al., 2003), and this process was shown to be regulated by Cdc28, a budding yeast cyclin-dependent kinase (CDK) (Liakopoulos et al., 2003;Maekawa et al., 2003). In addition, kinesin was shown to be required for accumulation of dynein and dynactin to the microtubule distal ends in filamentous fungi (see below) (Zhang et al., 2003). Given these facts, it is probably reasonable to speculate that accumulation of dynein at microtubule distal ends is also mediated by Kip2 and that Cdc28 regulates this process. In such a model, Cdc28 might phosphorylate dynein or a dynein interactor, as it does Kar9 (Liakopoulos et al., 2003;Maekawa et al., 2003). Indeed, dynein or a dynein interactor is phosphorylated at mitosis, probably by a CDK, in some higher eukaryotes (Niclas et al., 1996;Huang et al., 1999;Dell et al., 2000;Addinall et al., 2001;Yan et al., 2003).

Cytoplasmic dynein in fission yeast
The role of fission yeast dynein differs from that of budding yeast dynein at mitosis. In Schizosaccharomyces pombe, although dynein is dispensable for proper nuclear positioning (Yamamoto et al., 1999), it contributes to bipolar attachment of chromosomes to the spindle, unlike dynein in S. cerevisiae (E. L.. Grishchuk and J. R. McIntosh, personal communication). It also plays crucial roles in nuclear migration at meiosis. During meiotic prophase in S. pombe, the nucleus migrates back and forth between the two cell ends, led by the SPB (Fig. 4A) (Chikashige et al., 1994). At this stage, telomeres are clustered near the SPB, whereas centromeres are located away from it (Fig. 4C). The SPB-led nuclear migration is thought to facilitate homologous chromosome pairing by aligning the chromosomes from the telomeres and promoting contact of homologous loci (Chikashige et al., 1994) (reviewed by . Cytoplasmic dynein must be required for this process, because nuclear migration is abolished or impaired by depletion of a subunit of dynein or dynactin (Table 2). In the cells that exhibit defective nuclear migration, recombination frequency is reduced, which supports a crucial role for nuclear migration in homologous chromosome pairing (Yamamoto et al., 1999;Miki et al., 2002). Consistent with a crucial meiotic role for dynein is the observation that DHC expression is upregulated during meiosis (Miki et al., 2002).
In S. pombe, dynein-dependent nuclear migration -like that in S. cerevisiae -is driven by SPB-originating cytoplasmic The nucleus moves until it reaches the cortical attachment site of the microtubules. This series is repeated during meiotic prophase. (B) Telomere clustering. During mitotic interphase, centromeres (white circles) are located near the SPB, whereas telomeres (red rectangles) are located away from it, probably in association with nuclear membrane. Upon entering meiosis, telomeres move towards the SPB to form a telomere cluster, which remains near the SPB during meiotic prophase, whereas centromeres dissociate from the SPB. Dynein plays a role in telomere clustering. (C) Nuclear fusion. Upon nitrogen starvation, haploid fission yeast cells with opposite mating types fuse to form a zygote containing a diploid nucleus. Dynein and Klp2 drive fusion of two haploid nuclei in the zygote.
microtubules. During mitotic interphase, several cytoplasmic microtubules extend along the long cell axis and are associated with the nucleus (see review by Hagan, 1998). These microtubules appear to generate pushing forces by elongating, driving small rocking movements of the nucleus and probably maintaining nuclear position at the middle of the cell (Tran et al., 2001). By contrast, during meiotic prophase, microtubules are exclusively formed from the SPB (Svoboda et al., 1995;Ding et al., 1998). These microtubules appear to generate a pulling force mainly by sliding on the cortex, driving nuclear migration towards it (Fig. 4A) . At this stage, dynein is localized to microtubules and the SPB ( Table  2) (Yamamoto et al., 1999;Miki et al., 2002), and it also regulates microtubule depolymerization at the distal ends . These characteristics suggest that meiotic nuclear migration in S. pombe is driven by a mechanism similar to that in S. cerevisiae. However, unlike S. cerevisiae dynein, S. pombe dynein does not seem to accumulate preferentially at the microtubule distal ends. Instead, it is localized along the length of microtubules and accumulates at cortical sites that the microtubules contact -the closer the nucleus comes to the cortex, the more dynein accumulates (Yamamoto et al., 1999;. One possibility is that an immotile form of dynein is localized along the length of microtubules and, as the cortical dynein pulls in the microtubules, dynein on the microtubules accumulates at the microtubule-pulling site and also participates in generation of the force that drives nuclear migration. Dynein might also play a direct role in telomere clustering (Fig. 4B). Telomere clustering is slightly impaired by depletion of DHC or DLC and is severely impaired by depletion of both (Yamamoto et al., 1999;Miki et al., 2002) (A.Y., unpublished). This suggests that dynein participates in telomere clustering and that DLC has some dynein-independent functions and cooperates with dynein to mediate telomere clustering. Telomeres might be gathered by cytoplasmic microtubules, as has been proposed for centromere clustering during the forced transition from the meiotic cell cycle to the mitotic cell cycle (Goto et al., 2001), but the precise roles of dynein and DLC in the process remain to be determined.
Another role of dynein occurs during karyogamy (Fig. 4C), because nuclear fusion is delayed by depletion of DHC (Yamamoto et al., 1999;Miki et al., 2002) and is nearly eliminated by the combination of DHC depletion and depletion of the putative minus-end-directed, Kar3-related kinesin motor Klp2 (Troxell et al., 2001). During karyogamy, astral microtubules from the two nuclei interdigitate (Ding et al., 1998). It is likely that dynein and the Kar3-related kinesin drive nuclear fusion by bridging the microtubules and driving their sliding, as has been proposed for S. cerevisiae Kar3 (see review by Marsh and Rose, 1997).

Cytoplasmic dynein in filamentous fungi
Filamentous fungi form multinucleate hyphae. In the growing hyphae, nuclei migrate towards the growing tip and distribute evenly along the cell length (Fig. 5A) (Suelmann et al., 1997;Inoue et al., 1998a;Alberti-Segui et al., 2001); cytoplasmic dynein is required for this migration. In the filamentous fungi Aspergillus nidulans, Neurospora crassa, Ashbya gossypii and Nectria haematococca, mutation or deletion of the genes encoding dynein/dynactin subunits or other dynein interactors causes a nuclear migration defect (Table 1).
Like nuclear migration in S. cerevisiae, migration in A. nidulans and N. haematococca appears to be driven by SPBoriginated astral microtubules (Inoue et al., 1998a;Han et al., 2001). This also seems to be the case for N. crassa, because the nucleus is frequently stretched into a pear-shaped structure and the SPB emanating microtubules is located at the apex of the pear-shaped nucleus (Minke et al., 1999a). The localization of dynein or its interactors in A. nidulans and N. crassa is also similar to that seen in S. cerevisiae (Table 2). In addition, in N. crassa, depletion of dynactin appears to increase the accumulation of dynein at microtubule ends, as seen in S. cerevisiae (Minke et al., 2000). Finally, in A. nidulans, dynein promotes microtubule depolymerization, as seen in S. cerevisiae (Han et al., 2001). These similarities suggest that nuclear migration in filamentous fungi is driven by a mechanism similar to that in S. cerevisiae.
There are several differences between S. cerevisiae and filamentous fungi, however. First, nuclear migration requires several additional factors whose orthologues are found in other eukaryotes but not in S. cerevisiae (

Fig. 5. Roles of dynein in filamentous fungi. (A) Nuclear migration.
During the elongation of the hypha by the apical growth (grey arrow), the nuclei migrate towards the growing tip (black arrows) and distribute relatively evenly along the cell length -this migration depends on dynein. Note that the migration distance of each nuclei differs. (B) Organelle transport. In the hyphae, vesicles move in both retrograde (black arrows) and anterograde (blue arrows) directions. It is proposed that dynein drives the retrograde transport of vesicles (blue circles), whereas kinesin drives the anterograde transport (red circles).
Second, LIS1 has a different role in dynein localization. In A. nidulans, dynein does not require LIS1 for its localization at microtubule distal ends and, interestingly, like dynactin depletion in S. cerevisiae, LIS1 depletion causes increased dynein accumulation at the distal ends (Zhang et al., 2002). Third, distal-end accumulation of dynein and dynein interactors is prominent near the hyphal tip but not in the internal hyphal region (Xiang et al., 1995b;Xiang et al., 2000;Minke et al., 1999b;Minke et al., 2000;Han et al., 2001;Zhang et al., 2002;Zhang et al., 2003;Efimov, 2003). Fourth, dynein is required for the formation of astral microtubules. In N. haematococca, depletion of dynein causes loss of astral microtubules emanating from the SPB (Inoue et al., 1998a;Inoue et al., 1998b). Fifth, there are cytoplasmic microtubules in the hypha that are not connected with the SPB in N. crassa (Minke et al., 1999a). These differences probably reflect an even distribution of nuclei along the length of the hypha. Although several models have been proposed to account for this nuclear distribution in fungi (Plamann et al., 1994;Efimov and Morris, 1998), none of them accounts for all of the observations, and so the mechanistic details remain unclear. The differences might also reflect the participation of dynein in organelle transport (see below). There are also some differences among filamentous fungi. Depletion of dynein causes aggregation of nuclei near the hyphal tip in A. gossypii (Alberti-Segui et al., 2001) but it causes their aggregation in regions away from the tip in other fungi (Plamann et al., 1994;Xiang et al., 1994;Inoue et al., 1998a). Therefore, the mechanisms driving nuclear migration, although similar, might not be identical among filamentous fungi.
Dynein mediates organelle transport, as well as nuclear migration, in the hyphae (Fig. 5B). In N. crassa, vacuoles are relatively evenly distributed along the hyphae. In dynein or dynactin mutant cells, however, they are enriched towards the tip and transport of vesicular organelles is reduced (Seiler et al., 1999;Lee et al., 2001). Defective organelle transport in the dynein mutant cells probably accounts for their aberrant hyphal growth (Plamann et al., 1994) because this requires proper transport of cell-wall components to the hyphal tip. Interestingly, in cells containing a mutant form of a conventional kinesin, Nkin, vesicle transport is also reduced but vacuoles are mostly absent around the tip, unlike those in the dynein mutant cells (Seiler et al., 1999). Furthermore, transport is nearly eliminated in the dynein and kinesin double mutant. It was proposed that dynein mediates retrograde organelle transport, whereas kinesin mediates anterograde transport. However, this model is controversial because both motors are implicated in anterograde transport by the observation that the Spitzenkörper, a refractive spheroid body formed near the hyphal tip by accumulation of secretory vesicles (Grove and Bracker, 1970), is reduced in size in kinesin and dynein mutant cells (Seiler et al., 1999;Riquelme et al., 2000).
The interaction of dynein with organelles is essential for dynein-dependent organelle transport. Dynactin plays a crucial role in this interaction. In N. crassa, the C-terminal region of p150 Glued and PEBPs in the actin-like filament have an inhibitory effect on dynein Lee et al., 2001). Interestingly, of the PEBPs, p25 is specifically required for organelle transport but is dispensable for nuclear migration, which indicates that a subset of dynactin subunits have specific biological roles.
Given dynein's localization to microtubule distal ends, organelle transport might also be driven by a mechanism similar to that proposed for nuclear migration (Fig. 3B). An immotile form of dynein would thus be delivered to the organelle by microtubule growth and anchored to the organelle. Dynein would then become activated and transport the organelle along the microtubule. In such a model, when the organelle reached its destination (e.g. a vacuole), dynein would be released from the organelle through the action of dynactin. This model can account for organelle transport but remains to be tested.

Cytoplasmic dynein in dimorphic fungi
Ustilago maydis is a dimorphic, pathogenic fungus that infects plants (see review by Banuett, 1992). Outside its host, U. maydis exists as haploid yeast-like cells but, after the mating of two sporidia, it forms a filamentous dikaryotic hypha that invades the corn tissue. Dynein is required for nuclear migration in U. maydis, as in other fungi (Straube et al., 2001). In the yeast-like form of U. maydis, the nucleus moves from the mother cell into the bud cell before mitosis and then undergoes mitotic division (Fig. 6A). When dynein function is eliminated, the nucleus fails to move into the bud cell and undergoes mitotic division exclusively in the mother cell. Even after the division, the nucleus generally fails to move into the bud cell, unlike those of dynein-mutant S. cerevisiae cells. Loss of U. maydis dynein is lethal to the cell, which supports an essential role for the motor in nuclear migration. Because microtubules do not appear to be formed from the SPB before nuclear migration, it was proposed that nuclear migration in U. maydis is driven by a mechanism different from that in S. cerevisiae Straube et al., 2003). However, the precise microtubule organization during nuclear migration is unknown, and so it remains to be determined whether nuclear migration in U. maydis is indeed driven by a different mechanism. Although the mechanism is unknown, it probably shares some basic principles with that of S. cerevisiae because U. maydis dynein probably promotes microtubule depolymerization, like S. cerevisiae dynein, as shown by the increased number and length of cytoplasmic microtubules in cells lacking dynein.
Dynein also plays a crucial role in endosome transport ( Fig.  6B) (Wedlich-Söldner et al., 2002b). In U. maydis, dynein and an UNC104/KIF1-like kinesin, Kin3, are two major motors driving endosome transport. When both dynein and Kin3 are depleted, endosome movements are nearly eliminated. Dynein probably has a role antagonistic to that of Kin3. Kin3 appears to drive endosome movements along microtubules towards the plus ends, whereas dynein appears to drive movements towards the minus ends. In addition, depletion of dynein causes a phenotype similar to that caused by overexpression of Kin3: in small budded cells, in which endosomes normally accumulate in the small bud, both dynein depletion and Kin3 overexpression cause the accumulation of endosomes at the distal end of the mother cell, where microtubule plus ends are probably located. Proper endosome transport is thought to be required for proper polar growth of U. maydis cells (Wedlich-Söldner et al., 2000) and, indeed, depletion of dynein or Kin3 causes aberrant cell morphology (Straube et al., 2001). In addition to endosome transport, dynein is also required for endoplasmic reticulum motility (Wedlich-Söldner et al., 2002a).
The most amazing finding from studies of this organism is that the N-and C-terminal portions of DHC can be encoded by two genes, dyn1 and dyn2 (Straube et al., 2001). The dyn1 gene product encompasses the first four AAA modules, whereas the dyn2 gene product encompasses the rest. The gene products form a complex and have been shown to function in the same pathway. This shows that DHC consists of two distinct functional domains. Furthermore, this suggests that the first four AAA modules and the rest have distinct functions. Recent electron microscopic analysis of flagellar DHC supports this idea. This indicates that the first four AAA modules, but not other modules, generate conformational changes in DHC that are crucial for dynein motility (Burgess et al., 2003).

Comparison of dynein functions in fungi and higher eukaryotes
Dyneins in other eukaryotes have compelling similarities to fungal dyneins. In other organisms, dynein drives migration of the nucleus or the mitotic spindle by mediating interactions between centrosome-originating astral microtubules and the cortex (Busson et al., 1998;Skop and White, 1998;Gonczy et al., 1999;Koonce et al., 1999;Faulkner et al., 2000). Dynein also drives organelle transport (see review by Hirokawa, 1998;Karki and Holzbaur, 1999;Allan et al., 2002). Moreover, in mammalian cells, dynein localizes to astral microtubules, accumulating at their distal ends with dynactin, CLIP-170 and LIS1 (Busson et al., 1998;Vaughan et al., 1999;Faulkner et al., 2000;Coquelle et al., 2002). Therefore, dynein-dependent processes in higher eukaryotes probably share their basic mechanism with those of fungi. Given that dynein accumulates at the distal ends, it is possible that dynein associates first with microtubules and then with other cellular structures to execute its tasks, as suggested in fungi.
Although dynein-dependent mechanisms seem to be widely conserved, microtubules are apparently dispensable for the association of dynein with certain cellular structures in higher eukaryotes. Dynein is localized at kinetochores together with dynactin, LIS1 and CLIP-170 in these organisms (Pfarr et al., 1990;Steuer et al., 1990;Echeverri et al., 1996;Dujardin et al., 1998;Faulkner et al., 2000). It moves spindle checkpoint factors from the kinetochore to the poles (King et al., 2000;Howell et al., 2001;Wojcik et al., 2001) and perhaps drives poleward chromosome movement (Rieder et al., 1990;Hyman and Mitchison, 1991;Skibbens et al., 1993) by moving on spindle microtubules. In this case, microtubules are not required for the localization of dynein to the kinetochores (King et al., 2000). It is possible that kinetochore dynein is active before it interacts with microtubules because major dynein interactors localizes with it. This might allow dynein to remove checkpoint factors from the kinetochore and/or drive poleward chromosome movement immediately on interacting with microtubules. Such quick action might be crucial for faithful segregation of chromosomes but not for nuclear/ spindle migration or organelle transport.

Conclusions and perspectives
Studies of dynein functions in fungi have significantly advanced our understanding of the mechanisms of dyneindependent processes and revealed novel biological roles of this motor. A simple model for dynein-dependent nuclear migration in S. cerevisiae also explains dynein-dependent nuclear migration in other fungi and might be the basis for dynein-dependent organelle transport. However, additional evidence is needed before we can state conclusively that dynein-dependent nuclear migration in other fungi is indeed driven by a similar mechanism and that nuclear migration and Journal of Cell Science 116 (22) Fig. 6. Roles of dynein in the dimorphic fungus, U. maydis. (A) Nuclear migration. In G2 phase, cytoplasmic microtubules (MT) emanate from a pair of two spherical structures located in the bud cell, unlike those in S. cerevisiae (Steinberg et al., 2000). The structures consist of tubulin and are thus called paired tubulin structures (PTS). The microtubules are probably oriented in such a way that the plus ends are distal to the PTS. The nucleus migrates into the bud cell before undergoing mitotic division (black arrow). This migration depends on dynein. After the nucleus moves into the bud cell, astral microtubules are formed exclusively from the SPB and the nucleus undergoes division. (B) Endosome transport. Endosome distribution changes during the cell cycle (Wedlich-Söldner et al., 2000;Wedlich-Söldner et al., 2002b). In particular, in a small-bud stage, endosomes (green circles) accumulate in the small bud. During this stage, microtubules emanate from PTS located in the small bud. Because of this microtubule organization, the minus ends are probably located in the small bud, whereas the plus ends are located at the distal end of the mother cell. Dynein probably transports endosomes in the minus-end direction, causing accumulation of endosomes in the bud. By contrast, Kin3 probably transports endosomes in the plus-end direction. organelle transport share such a basis. Biochemical approaches must be used in fungi because they have not been extensively used for studying dynein functions in these organisms. Such approaches will undoubtedly help to reveal the relationships between the activation of dynein and its association with cellular structures, as well as the precise roles of dynein interactors in these processes. Studies of other motor proteins involved in dynein-dependent processes will also be important because most dynein-dependent processes are mediated by the concerted actions of dynein and these motors. Given that the functions of dynein are conserved across the eukaryotes, defining the mechanisms of dynein-dependent processes in fungi will make a significant contribution to our understanding of the molecular functions of this protein complex in other organisms as well.