The role of microtubule-based motor proteins in maintaining the structure and function of the Golgi complex

Biochimica et Biophysica Acta 1404 (1998) 113^126

The role of microtubule-based motor proteins in maintaining the structure and function of the Golgi complex Janis K. Burkhardt * Department of Pathology, The University of Chicago, 5841 S. Maryland Ave. MC1089, Chicago, IL 60637, USA Received 11 November 1997; accepted 5 January 1998

Abstract The intimate association between the Golgi complex and the microtubule cytoskeleton plays an important role in Golgi structure and function. Recent evidence indicates that the dynamic flow of material from the ER to the Golgi is crucial to maintaining the integrity of the Golgi complex and its characteristic location within the cell, and it is now clear that this flow is dependent on the ongoing activity of microtubule motor proteins. This review focuses primarily on recent microinjection and expression studies which have explored the role of individual microtubule motor proteins in controlling Golgi dynamics. The collective evidence shows that one or more isoforms of cytoplasmic dynein, together with its cofactor the dynactin complex, are required to maintain a juxtanuclear Golgi complex in fibroblasts. Although questions remain about how dynein and dynactin are linked to the Golgi, there is evidence that the Golgi-spectrin lattice is involved. Kinesin and kinesin-like proteins appear to play a smaller role in Golgi dynamics, though this may be very cell-type specific. Moreover, new evidence about the role of kinesin family members continues to emerge. Thanks in part to recent advances in our understanding of these molecular motors, our current view of the Golgi complex is of an organelle in flux, undergoing constant renewal. Future research will be aimed at elucidating how and to what extent these motor proteins function as regulators of Golgi function. ß 1998 Elsevier Science B.V. All rights reserved. Keywords: Golgi complex; Microtubule; Dynein; Dynactin; Kinesin ; Motor protein

1. Microtubules and the Golgi complex: intimate associates The association between the Golgi complex and the microtubule cytoskeleton was ¢rst recognized over 30 years ago. In the 1950s, cell biologists using the newly developed technique of electron microscopy quickly discovered that the Golgi complex is a feature of all eukaryotic cells, and that in most animal cells, the Golgi is located near the nucleus.

* Fax: +1 (773) 702-3701; E-mail: jburkhar@£owcity.bsd.uchicago.edu

In the mid-1960s, when glutaraldehyde came into common use as a ¢xative, microtubules were found to radiate out from an `organizing center' in the same region of the cell. This region, which became known as the `cell center' was shown to be devoid of most other organelles, including ribosomes, mitochondria, and endoplasmic reticulum, further highlighting the unique association between the Golgi and the microtubule organizing center (MTOC). Though there are exceptions to this rule, the Golgi in most cell types is closely linked to the MTOC. Indeed, the MTOC and the Golgi co-localize even under conditions where cellular architecture is undergoing major remodeling. This occurs, for example,

0167-4889 / 98 / $19.00 ß 1998 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 7 - 4 8 8 9 ( 9 8 ) 0 0 0 5 2 - 4

BBAMCR 14327 5-8-98 Cyaan Magenta

Geel

Zwart

114

J.K. Burkhardt / Biochimica et Biophysica Acta 1404 (1998) 113^126

when ¢broblasts prepare to migrate into a wound site [1], when T-cells polarize toward an antigen presenting cell [2], or during the fusion of myoblasts to form myotubes [3]. From the very beginning, it was obvious that the interaction between microtubules and the Golgi complex was more than coincidence. Evidence that microtubules play an active role in maintaining Golgi structure came originally from studies using drugs such as colchicine, vinblastine, and nocodazole that favor microtubule disassembly. These drugs fragment the Golgi complex, generating shortened, often dilated stacks scattered throughout the cell. This observation was made initially in the 1960s by Robbins and Gonatas [4], and by the mid-1980s, similar observations had been made in a variety of cell types by several groups, most notably that of Thyberg, Moskalewski and coworkers. Indeed, a 1985 review on the topic lists 32 publications showing disruption of the Golgi complex as a result of microtubule-disruptive drugs [5]. Remarkably, it was discovered that the Golgi does not need to be either intact or juxtanuclear to carry out most of its functions. The short stacks formed in the presence of microtubule-depolymerizing agents appear to function normally in glycosylating transiting proteins and delivering them to the cell surface [6,7]. In this regard, the scattered stacks resemble the normal state of a¡airs in plant cells and fungi, where Golgi function occurs without centralization. It appears that the central Golgi complex in animal cells is not important for protein transport and glycosylation per se, but rather for coordinating directed organelle tra¤c. At least in some cases, delivery of proteins to post-Golgi targets is facilitated and directed by the microtubule sca¡old. For example, although microtubules are not required for delivery of constitutive transport vesicles from the Golgi to the cell surface, transport to lysosomes is substantially delayed by the depolymerization of microtubules [8]. Similarly, microtubules are required to facilitate the rapid movement of large regulated secretory granules, and are largely responsible for delivering TGN-derived vesicles to the appropriate cell surface in polarized epithelial cells (reviewed in [9]). The maintenance of a centralized Golgi complex in animal cells may be required for (or arise as a biproduct of) these types of directional transport.

Despite the numerous systems in which Golgi fragmentation was found to occur as a result of microtubule depolymerization, the mechanism underlying this phenomenon has been puzzling. One popular model was that tubular interconnections between individual stacks of £attened cisternae (dictyosomes) were severed [3]. Another was that lateral connections were simply unfolded and extended, allowing Golgi dispersal without disrupting the continuity of the Golgi complex [10]. Neither of these hypotheses could account for the fact that the scattered Golgi stacks were often shorter than normal, with fewer cisternae, and neither could o¡er any mechanism by which these elements could be dispersed through the cell, given the low di¡usion rates in cytoplasm [11]. To further complicate the issue, a mechanistic investigation by Turner and Tartako¡ showed that Golgi fragmentation and dispersal occur subsequent to microtubule depolymerization, and that these processes are energy and temperature-dependent, and sensitive to conditions that block protein transport [12]. Only in the last few years has a plausible mechanism for Golgi fragmentation been o¡ered. Looking back now, it is easy to see why the problem of Golgi fragmentation was ba¥ing. In the current view, it is virtually impossible to ask how microtubules contribute to Golgi structure without considering how microtubules mediate Golgi dynamics. To understand both we must draw on what we have learned in the intervening decade about microtubule-based motor proteins and their interactions with membrane organelles. 2. A motile Golgi complex traverses the microtubule network Prior to the identi¢cation of microtubule motor proteins it was clear that the MTOC played a role in positioning the Golgi complex, but the basis of the interaction was unclear. Did the Golgi bind directly to microtubules or to other material in the MTOC? Was the interaction a static one, mediated by cytoplasmic linking proteins, or was it somehow more dynamic? Studies aimed at determining the basis of Golgi localization at ¢rst focused on determining with which elements the Golgi associates. The centrosome itself is not required, since in karyoplasts

BBAMCR 14327 5-8-98 Cyaan Magenta

Geel

Zwart

J.K. Burkhardt / Biochimica et Biophysica Acta 1404 (1998) 113^126

where the centrosome is surgically removed, a juxtanuclear MTOC is still formed, and the Golgi complex is localized to this region [13]. Moreover, separation of the MTOC from the centrioles occurs naturally in certain specialized cell types such as myotubes. In these cells, the Golgi is found at the MTOC, where the minus ends of microtubules are located [3]. Finally, in fused cells treated with taxol, microtubule bundles are generated which lack both centrosomes and MTOC, yet Golgi membranes were found at the ends of these bundles [14]. Collectively, these data argued that the Golgi complex localizes to the ends of microtubules. It was an important milestone along the road to our current view of the Golgi when it was discovered that Golgi elements can move through the cell in a microtubule-dependent fashion. This was demonstrated by Kreis and coworkers, analyzing the recovery of Vero cells from microtubule depolymerizing agents. Using the vital dye C6 -NBD-ceramide, they followed the re-formation of an intact Golgi complex in living cells, and showed that scattered Golgi elements move centripetally along linear microtubule tracks [15]. Similar studies in a fused-cell system showed that entire Golgi stacks can move along microtubules to congregate and coalesce at the cell center [14]. This movement was shown to be independent of actin or intermediate ¢laments [14,15], suggesting that is driven by a microtubule-motor protein. In ¢broblasts, like those used in these studies, microtubules radiate out from the cell center with their fast-growing or plus ends at the periphery and their slow-growing, or minus ends at the cell center [16]. Given the steady movement of Golgi elements toward the cell center, the motor responsible would have to be minus-end directed. It is important here to point out that cells di¡er signi¢cantly in their microtubule organization, and that the association of the Golgi complex with microtubules is also variable. Unlike the familiar radial microtubule organization of ¢broblasts, in polarized epithelial cells, microtubules are oriented in parallel bundles along the long axis of the cell, with their minus-ends at the apical side of the cell and their plus ends facing the basal domain [9,17,18]. The Golgi complex typically lies just apical to the nucleus, well separated from the microtubule minus-ends [17,19]. Whether this distribution is mediated by

115

the dominant activity of plus-end motor proteins is unclear, and the question has been di¤cult to study under conditions where the overall architecture of the epithelial cells is maintained. However, it is clear that in these cells, Golgi localization is not determined solely by minus-end directed motor activity. Even within the two extremes de¢ned by the textbook ¢broblast and the textbook epithelial cell, more subtle variations exist. As discussed below, astrocytes have a ¢broblast-like radial microtubule array, but the Golgi is extended well out toward the cell periphery [20]. Similarly, while the Golgi complex is compact and supranuclear in many epithelial cells, in the well-studied epithelial line MDCK, the Golgi complex extends upwards from the nucleus to the apical portion of the cell [18]. Though this review will focus primarily on what has been learned in ¢broblasts, where minus-end motors dominate, it is important to recall that in various cell types the balance of power between opposing motor proteins may be quite di¡erent. 3. Microtubule motor proteins move the ¢eld forward The movement of the Golgi complex toward the minus ends of microtubules in ¢broblasts was compared with the better-studied process of retrograde axonal transport, and it was from studies on axoplasm that microtubule motor proteins were soon identi¢ed. Kinesin was the ¢rst to be discovered, and was shown to drive organelle movement toward microtubule plus ends (anterograde transport in axons and transport toward the cell surface in ¢broblasts) [21^23]. Soon after the discovery of kinesin a minus-end directed motor, cytoplasmic dynein, was identi¢ed in brain and C. elegans extracts [24,25]. Cytoplasmic dynein was found to be present in other cell types as well, and shown to mediate the minusend-directed movement of mixed membrane organelles in vitro [26]. Since the structure and motor function of kinesin and cytoplasmic dynein have been extensively studied and several good reviews are available [27^36], only the basic features of these proteins will be described here. Kinesin is a heterotetramer with two V116 kDa heavy chains and two light chains of approximately 70 kDa (Fig. 1). The kinesin heavy chains

BBAMCR 14327 5-8-98 Cyaan Magenta

Geel

Zwart

116

J.K. Burkhardt / Biochimica et Biophysica Acta 1404 (1998) 113^126

Fig. 1. Structure of conventional kinesin. The two heavy chains (blue) form a coiled-coil dimer, with the globular motor domains at one end of the molecule, and sequences contributing to the fan-shaped tail at the other. The globular head is su¤cient for force production, and contains the regions responsible for microtubule binding (circles) and ATP hydrolysis (triangles). A hinge region partially down the stalk allows the protein to fold back on itself. The two light chains (red) bind to the heavy chain through coiled-coil interactions, and contribute to the fan-shaped tail of the molecule.

(KHC) can be divided into three domains: head, stalk and tail. Each globular head contains a microtubule binding site and an ATPase site, and serves as the motor domain [37]. The stalk domain is an Khelical coil^coil, and the C-terminal tail domain is fan-shaped in electron micrographs [38]. The kinesin light chains (KLCs) associate tightly with KHCs via coil^coil interactions along the stalk and tail domains, and they contribute to the fan shaped tail [38]. Because of their location in the molecule, one hypothetical role for KLCs is in binding to membrane organelles, however KHC can bind membranes on its own [39]. Since there exist multiple KLC isoforms, the light chains are proposed to play a regulatory role, a¡ecting organelle speci¢city, motor activity or both [40^42]. Over the past several years, multiple kinesin-like proteins (KLPs, also called kinesin superfamily members, of KIFs) have been identi¢ed (reviewed in [28,43,44]). These have homology to conventional kinesin in the motor domain, and this motor domain may be located at the N-terminus, as in conventional kinesin, at the C-terminus, or centrally in the molecule, £anked by unique sequences. Directional preferences of the KLPs varies; some family members such as ncd and Kar3 are minus-end directed motors. Many KLPs are involved in aspects of mitosis and meiosis; however, there is evidence that some KLPs are involved in membrane organelle motility, and that these proteins have distinct organelle speci¢cities (reviewed in [44]). As discussed below, there is evidence that conventional kinesin and at least one KLP play a role in Golgi dynamics.

Cytoplasmic dynein is considerably larger and more complex than kinesin. Like its axonemal homolog, cytoplasmic dynein is a multisubunit protein complex, with total mass exceeding 1U106 Da. Subunits include two heavy chains (DHC) of V530 kDa, three intermediate chains (DIC) of approximately 74 kDa, several light-intermediate chains in the 50^60 kDa range, and three light chains of 8^22 kDa. The overall structure of cytoplasmic dynein is similar to that of kinesin, with two globular heads connected by extended stalk domains to a single globular tail. As with kinesin, the heavy chains form the globular dynein heads, which contain the ATP and microtubule binding sites responsible for force production [45,46]. The globular tail domain is thought to bind to membrane organelles. In keeping with this view, the 74 kDa dynein intermediate chains, which lie at the base of the molecule [47], have been shown to interact directly with the p150Glued subunit of the dynein cofactor, dynactin [48,49]. A more detailed description of dynactin and its role as part of a membrane anchor for cytoplasmic dynein is given below, and diagrammed in Fig. 2. Recent research has resulted in the identi¢cation of multiple isoforms of several of the dynein chains. The degree to which the di¡erent isoforms can co-associate to form distinct enzymes is still unknown, but the number of possible dynein motors is quite large. Cytoplasmic dyneins play a role in spindle organization and chromosome organization during mitosis [50^56], and in motility of membrane organelles in interphase cells [26,57^63]. At present, relatively little is known about which dynein family

BBAMCR 14327 5-8-98 Cyaan Magenta

Geel

Zwart

J.K. Burkhardt / Biochimica et Biophysica Acta 1404 (1998) 113^126

117

Fig. 2. Dynein and dynactin binding to membrane organelles. Cytoplasmic dynein (red and orange) has a structure similar to that of kinesin, with globular heads that interact with microtubules and hydrolyze ATP. Cytoplasmic dynein is attached to the dynactin complex via the intermediate chains, which bind to the p150Glued subunit of dynactin. The dynactin complex (shades of blue, green and violet) is found in tight peripheral association with membranes. One proposed mechanism for dynactin-membrane attachment is via binding of the Arp-1 actin-like ¢lament to the Golgi-spectrin lattice. Though the location of the dynamitin subunit of dynactin within the complex has not been determined, it is depicted here as joining the arm formed by p150Glued to the Arp-1 ¢lament, since expression of dynamitin in super-stoichiometric amounts can separate these two portions of the dynactin complex, releasing dynein from its cargo organelle.

members are responsible for speci¢c dynein-dependent processes in the cell. As described below, however, it is clear that at least one form of cytoplasmic dynein is critical for maintaining Golgi structure and function. 4. Conventional kinesin and the Golgi complex The evidence that kinesins interact with the Golgi is mixed. Goud and coworkers have identi¢ed a new KLP that apparently interacts with the Rab6. Rabkinesin has been localized to the Golgi apparatus and its overexpression causes disorganization of the Golgi. These ¢ndings are discussed in more detail in their article in this volume. Apart from this newly identi¢ed KLP, the kinesin usually associated with Golgi dynamics is conventional kinesin. Two studies using subcellular fractionation of Golgi membranes followed by immunoblotting with anti-conventional kinesin antibodies have failed to detect kinesin [19,64], while another showed that kinesin is enriched in Gol-

gi fractions [65]. One EM-immunolocalization study showed that kinesin is present primarily on pre-Golgi intermediates, components of the intermediate compartment that cycles between the ER and the Golgi [66], while another showed that kinesin is preferentially localized to the trans-Golgi network [67]. Functional data on the role of kinesin are also mixed. There is evidence that conventional kinesin is involved in transport from the Golgi complex to the cell surface, at least in regulated secretory cells [68] and polarized epithelial cells [61]. But the role of kinesin in the dynamics of the Golgi itself is less clear. One approach which has been used to test the role of kinesin in vivo is transfection of cells with a dominant negative mutant of kinesin. No effect on the Golgi complex in mouse ¢broblasts was observed at the level of immuno£uorescence microscopy [69]; however, it should be pointed out that the mutant used in this study binds to microtubules in rigor, and might therefore tend to lock any a¡ected organelles in place. Contrasting results were obtained by Feiguin et al. [70], who used an antisense ap-

BBAMCR 14327 5-8-98 Cyaan Magenta

Geel

Zwart

118

J.K. Burkhardt / Biochimica et Biophysica Acta 1404 (1998) 113^126

Fig. 3. The extended morphology of the Golgi complex in cultured astrocytes is collapsed by treatment with kinesin antisense oligonucleotides. One-week-old astrocyte cultures were cultured for 24-48 h in the absence of oligonucleotide (A), in the presence of control sense oligonucleotide (B) or in the presence of antisense oligonucleotide corresponding to nucleotides 311 to +14 of the rat kinesin heavy chain sequence (C^F). Cells were ¢xed and the Golgi labeled with antibodies to mannosidase II. Note the collapse of the Golgi toward the nucleus. U700. Figure is adapted from Feiguin et al. [70].

proach to test the e¡ects of drastically decreasing kinesin levels. In cultured astrocytes, the Golgi complex normally has an extended morphology, with numerous elongated tubules, and the extension of these tubular processes has been shown to be microtubule dependent [20]. This structure is collapsed toward the nucleus in cells expressing the kinesin antisense DNA (Fig. 3) [70]. Thus, at least in these cells, kinesin appears to play a role in determining Golgi morphology and distribution. In cells where the Golgi complex is less extended, the comparatively subtle role of kinesin may be masked by the overriding e¡ects of cytoplasmic dynein unless Golgi dynamics are perturbed. Lippincott-Schwartz et al. [66] tested the e¡ects of microinjecting function-blocking anti-kinesin antibody. Under steady-state conditions, the antibody had no e¡ect on Golgi morphology, nor on protein tra¤c. If, however, the injected cells were treated with Brefeldin A, the characteristic drug-induced redistribution of Golgi proteins to the ER was found to be blocked. Tubule formation, a microtubule-dependent

early event in BFA-induced redistribution of Golgi proteins to the ER, was inhibited by the microinjected antibody, indicating that kinesin is involved in generating this plus-end directed tubule growth. In the context of these studies, it should be noted that expression of kinesin antisense by Feiguin et al. [70] did not interfere with the redistribution of Golgi markers to the ER upon treatment with BFA. Whether this discrepancy is due to di¡erences between cell types or to technical di¡erences is not presently clear. Certainly, kinesin cannot be absolutely required for redistribution of proteins from the Golgi to the ER, since this redistribution occurs in nocodazole treated cells, as discussed below. Though the literature about kinesin on the Golgi complex is somewhat confusing, it seems clear that, at least in some cell types, kinesin does interact with the Golgi complex. In ¢broblasts, kinesin is primarily involved in either recycling of Golgi components to the ER or in transport of Golgi-derived vesicles to the cell surface, and in the extension of tubules that precedes these processes. Thus, in these cells, kinesin is important for maintaining tra¤c through the Golgi, whereas, as described below, cytoplasmic dynein is the dominant determinant of Golgi structure and positioning. In epithelial cells and other specialized cell types, the roles of these proteins may di¡er, and the balance of power between the opposing motors may even be reversed. It is likely that kinesin/KLPs play a coordinated role with dynein family members to determine the characteristic balance of Golgi dynamics for each cell type. 5. One or more cytoplasmic dyneins mediate centripetal movement of Golgi elements Cytoplasmic dynein has the appropriate directional preference to mediate juxtanuclear localization of the Golgi complex in ¢broblasts. Corthesy Theulaz et al. [71] tested the interaction of cytoplasmic dynein with Golgi membranes using an in vitro binding assay. Semi-intact CHO cells prepared by osmotic swelling were incubated with puri¢ed rat liver Golgi membranes in the presence of a cytosolic extract and ATP. The exogenous Golgi stacks were `captured' by the broken cells, and the bound Golgi complexes were located near the centrosomes. Capture was

BBAMCR 14327 5-8-98 Cyaan Magenta

Geel

Zwart

J.K. Burkhardt / Biochimica et Biophysica Acta 1404 (1998) 113^126

found to be sensitive to drugs that inhibit dynein activity, and to depletion of cytoplasmic dynein from the cytosolic extract. While this study fell short of demonstrating that dynein drives active movement of Golgi membranes, it showed that cytoplasmic dynein is indeed important for Golgi-microtubule interactions. Within the last 2 years, antibodies have been identi¢ed which interfere with dynein activity, making it possible to test the role of cytoplasmic dynein in living cells. 70.1, an antibody reactive with the 70 kDa intermediate chain of cytoplasmic dynein (DIC) [72] disrupts dynein-dependent motility in vitro [53]. When we injected this antibody into mammalian ¢broblasts, we found that it disrupted the structure of the Golgi complex, leading to a nocodazole-like phenotype [60] (Fig. 4). Together with the studies on dynactin described below, these ¢ndings demonstrate that cytoplasmic dynein mediates the centripetal movement of Golgi elements in vivo. It is important to point out that the antibody actively induced Golgi fragmentation only a few hours after microinjection. This result can only be explained if ongoing dynein activity is needed to maintain an intact Golgi complex, counterbalancing some opposing activity(ies) which tend toward Golgi dispersal. Moreover, since inhibition of dynein activity gave the same phenotype as depolymerizing microtubules altogether, our results argue that cytoplasmic dynein is the dominant protein involved in maintaining Golgi structure and positioning. While these results do not rule out the accessory involvement of cytoplasmic linking proteins (CLIPs) in binding the Golgi to microtubules in a static fashion, one can argue that if such proteins exist, their activity is dominated by the activity of cytoplasmic dynein. Cytoplasmic dynein is a very large and complex motor protein, and interpretation of its role in organelle motility has been complicated by the discovery of multiple new isoforms of several of its component subunits. At the present time, the reactivities of many commonly used anti-dynein antibodies to different family members have not been determined, nor has the ability of the various di¡erent dynein subunits to associate with one another. These constraints, together with the di¤culties inherent to in vivo functional studies, have led to uncertainty about which dynein family members are involved in speci¢c

119

cellular processes. For example, though we found that the anti-DIC antibody 70.1 can disrupt Golgi morphology, microinjection of an antibody reactive with the most abundant isoform of cytoplasmic dynein heavy chain, DHC-1, reportedly had no e¡ect on the Golgi complex [63]. While this negative result does not disprove that DHC-1 participates in Golgi motility, the same antibody was capable of blocking mitotic spindle formation, another dynein-dependent

Fig. 4. Microinjection of anti-dynein intermediate chain antibody disrupts the Golgi complex. Hela cells stably expressing an epitope-tagged sialyltransferase to mark the Golgi were microinjected with anti-dynein intermediate chain together with FITC-dextran to mark the injected cells (top). After 4 h at 37³C, the cells were ¢xed and labeled with antibody reactive with the sialyltransferase (bottom). Whereas in the neighboring uninjected cell, the Golgi complex is compact and juxtanuclear, in the injected cell, the Golgi is fragmented and dispersed throughout the periphery. Figure is adapted from Burkhardt et al. [60].

BBAMCR 14327 5-8-98 Cyaan Magenta

Geel

Zwart

120

J.K. Burkhardt / Biochimica et Biophysica Acta 1404 (1998) 113^126

process [56]. One possible explanation for this is that another DHC isoform is involved. Vaisberg et al. [63] have identi¢ed two new DHC family members, DHC-2 and DHC-3, distantly related to cytoplasmic (DHC-1) and axonemal dynein heavy chains, respectively. Both are present in mammalian cells at levels lower than DHC-1. An antibody speci¢c for DHC-2 reacted with the Golgi complex by immuno£uorescence microscopy, and when microinjected into cells, this antibody perturbed Golgi structure in a fraction of cells [63]. Despite this evidence that DHC-2 interacts with the Golgi complex; however, it seems unlikely that dyneins containing this heavy chain are solely responsible for Golgi localization. DHC-2 is a relatively rare dynein isoform, and the expression of this RNA in some tissues is nearly undetectable. Moreover, the Golgi compartment labeled by antiDHC-2 appears to be variable; the antigen co-localizes with the medial Golgi marker mannosidase II in some cells, and with the intermediate compartment marker ERGIC-53 in others [63,73]. Thus, it remains likely that the more abundant DHC-1 is involved in Golgi localization, perhaps in concert with DHC-2 in some cell types. Another outstanding question is to what degree the di¡erent dynein chains co-assemble. If, for example, DHC-2 does prove to play a dominant role in Golgi localization, our ¢nding that DIC antibody 70.1 disrupts the Golgi complex may mean that DHC-2 shares a common intermediate chain with DHC-1. Alternatively, 70.1 may prove to crossreact with multiple DIC isoforms. Similar issues are on the horizon regarding the light and light intermediate chains of cytoplasmic dynein, subunits which may play an important role in regulating activity of the motor complex. Clearly, analyzing the roles of the dynein family members in mediating the motility of Golgi membranes versus other organelles will be an important area for investigation during the next few years. 6. Attaching dynein to the Golgi complex: the role of dynactin and the spectrin lattice Although puri¢ed cytoplasmic dynein can mediate the gliding of microtubules along a glass coverslip, movement of membrane organelles requires cytoplasmic co-factors [74,75]. One of these, dynactin, has

Fig. 5. Overexpression of the dynamitin subunit of dynactin fragments the Golgi. Hela cells stably expressing an epitopetagged sialyltransferase were transiently transfected with dynamitin. Twenty hours after the addition of the dynamitin DNA, cells were ¢xed and labeled with anti-dynamitin (A) or antibody reactive with the sialyltransferase (B). In the two dynamitin overexpressing cells, the Golgi is fragmented and dispersed in a fashion resembling the e¡ects of treatment with nocodazole or microinjection with anti-DIC antibody. Figure is adapted from Burkhardt, et al. [60].

been well characterized and shown to interact with dynein both biochemically and genetically (reviewed in [35,76,77]), and evidence from a variety of sources indicates that it plays a role in binding cytoplasmic dynein to cargo organelles. Dynactin is a large protein complex with a short actin-like ¢lament and an arm capable of interacting with cytoplasmic dynein and with microtubules [48,49,78,79]. The dynactin ¢lament is made up of several subunits of the actin-related protein Arp-1, capped at either end by actin capping protein and a unique 62 kDa dynactin subunit. The dynein- and microtubule- binding arm of dynactin is composed of the protein p150Glued , which is homologous to the Drosophila Glued protein [80]. In addition to these subunits, the dynactin com-

BBAMCR 14327 5-8-98 Cyaan Magenta

Geel

Zwart

J.K. Burkhardt / Biochimica et Biophysica Acta 1404 (1998) 113^126

plex contains several subunits of a 50 kDa protein, dynamitin, which plays a structural role within the complex ([50], and see below). Recently, two expression studies have shown that dynactin function, like that of cytoplasmic dynein, is required to maintain Golgi structure and localization. These studies involved overexpression of two di¡erent wild type dynactin subunits in ¢broblasts. Distinct e¡ects on Golgi morphology were obtained, shedding new light on the likely function of the subunits involved. In one of these studies, we overexpressed the dynamitin subunit of dynactin [60]. Previously, Echeverri et al. [50] had shown that overexpression of dynamitin disrupts the dynactin complex, separating the Arp-1 ¢lament from the dynein-binding arm of p150Glued . Since this was shown to result in the release of dynein from kinetochores, we reasoned that it might also disrupt dynein-mediated movement of other cargo organelles, including the Golgi complex. This proved to be the case. Though the overexpressed dynamitin remained largely soluble and little perturbation of cytoskeletal elements was observed, we found that the e¡ects on membrane organelles were profound. The Golgi complex was fragmented and dispersed throughout the cytoplasm in a fashion closely resembling the e¡ects of nocodazole treatment or microinjection with anti-dynein intermediate chain antibody (Fig. 5). Our results show that ongoing dynactin activity, like that of dynein itself, is essential for the maintenance of Golgi structure and position. In a parallel study, Holleran et al. [81] tested the function of Arp-1, the dynactin subunit that comprises the actin-like ¢lament. Because the Arp-1 ¢lament lies opposite to the dynein-binding end of the dynactin complex, and because of the resemblance of Arp-1 ¢laments to the short actin ¢laments that lie at the vertices of the spectrin lattice at the plasma membrane, the Arp-1 ¢lament has been proposed to play a role in membrane anchoring [34,77]. When Arp-1 was overexpressed in ¢broblasts, numerous cytoplasmic Arp-1 ¢laments were formed in the transfected cells, and Golgi membranes were found in association with these ¢laments. This phenotype suggests that rather than releasing the Golgi complex from microtubules, overexpression of Arp-1 provides new Golgi attachment sites within the cell. The results of Holleran et al. are thus consistent with the view that

121

the Arp-1 ¢lament is involved in anchoring the dynactin complex, and presumably also cytoplasmic dynein, to the Golgi complex. As described by Nelson and Beck in this issue, a spectrin lattice has been identi¢ed in association with the Golgi complex [82,83], raising the possibility that dynactin is linked to the Golgi via this lattice. Evidence that LI4* spectrin, the Golgi-associated spectrin form, partially cofractionates with the dynactin complex, and that LI4* spectrin and Arp-1 ¢laments co-localize in Arp-1 overexpressing cells [81] supports the idea that dynactin associates via Arp-1 with the Golgi spectrin matrix. This model for dynein/dynactin attachment is diagrammed in Fig. 2. Working in an epithelial cell system, where cytoplasmic dynein is proposed to mediate movement from the Golgi complex to the cell surface, Fath and co-workers have taken a biochemical approach to test the mechanism by which the Golgi complex binds to cytoplasmic dynein [84]. Their results support some, but not all, of the inferences drawn from the overexpression studies above. They ¢nd that while isolated Golgi stacks retain little bound cytoplasmic dynein, they do retain bound dynactin subunits, and that dynein can be recruited to the membranes upon incubation with cytosol. Conditions such as incubation with high salt, and even detergent, leave behind a matrix containing signi¢cant levels of dynactin subunits, and these extracted Golgi membranes retain the ability to bind to cytoplasmic dynein. In contrast, treatment of the membranes with alkaline pH abolishes their ability to re-bind cytoplasmic dynein upon incubation with cytosol. The alkaline pH removes signi¢cant levels of both p150Glued and Arp-1 from the Golgi membranes and though Arp-1 can re-bind e¤ciently when the stripped stacks are incubated with cytosol, p150Glued cannot. These results support the prevailing model for dynactin function, i.e. that the dynactin complex partitions between the cytoplasm and membranes, and functions as a dynein attachment factor. The existence of cytosolic Arp-1 that is able to bind membranes in the absence of p150Glued is rather surprising, since it is thought that most or all of the cytoplasmic dynactin subunits are present as 20S particles. Nonetheless, at least under the conditions tested, these results indicate that Arp-1 alone is insu¤cient to recruit cytoplasmic dynein; p150Glued ,

BBAMCR 14327 5-8-98 Cyaan Magenta

Geel

Zwart

122

J.K. Burkhardt / Biochimica et Biophysica Acta 1404 (1998) 113^126

the subunit already shown to bind directly to DIC, is also required. Probably these subunits are normally recruited together as the intact 20S dynactin complex. Interestingly, Fath et al. found that Golgi membranes extracted with cold 1% Triton X-100 could still bind to cytoplasmic dynein in the same fashion as before, indicating that the detergent-resistant Golgi matrix retains the dynein `receptor'. Under these conditions, dynein intermediate chain, Arp-1 and p150Glued could all re-bind to the extracted stacks upon incubation with cytosol, however, spectrin and ankyrin were released from the Golgi membranes [84]. This remarkable result suggests that interaction of dynactin/dynein with the Golgi can occur independently of the spectrin/ankyrin matrix. While this new evidence does not rule out a role for the spectrin/ankyrin matrix in linking dynactin and dynein to Golgi membranes, it raises questions about whether this model is at least oversimpli¢ed. Should we expect to ¢nd a spectrin-like lattice on all organelles to which cytoplasmic dynein attaches? If this lattice imparts structural rigidity, as proposed for the plasma membrane lattice, how can this be reconciled with the ¢nding, discussed below, that cytoplasmic dynein is usually found in association with dynamic Golgi vesicles, and not with the stack? If factors in addition to the spectrin lattice are involved, what could they be, and to which subunits of dynactin do they bind? Unraveling the full mechanism by which dynein and dynactin bind to the Golgi will be another important area of investigation in the next few years 7. Motor proteins and Golgi dynamics Much of what we know about how the Golgi interacts with motor proteins comes from arti¢cial situations with drug-treated cells, where intermediates which might otherwise not exist are followed through the cell. For example, small Golgi stacks formed by treatment with nocodazole move to the cell center using cytoplasmic dynein, and tubules induced by treatment with Brefeldin A use kinesin to extend toward the ER. But what are the intermediates that normally interact with motor proteins, and how does their steady state £ux add up to the familiar

juxtanuclear Golgi characteristic of most higher eukaryotic cells? Most likely, it is not the Golgi stack that binds to motor proteins, but the dynamic membranes at the cis and trans faces of the stack. Evidence in support of this view comes from both biochemical and morphological analyses. Based upon Western blotting of Golgi subfractions from epithelial cells, cytoplasmic dynein is primarily present on vesicles enriched for markers of the trans-Golgi network, rather on the stack itself [19]. Moreover, isolated Golgi membranes bind cytoplasmic dynein, but the majority of this bound dynein is found in a vesicular fraction if the Golgi membranes are incubated with cytosol under conditions that permit budding from the TGN [84]. In keeping with this ¢nding, our own immunolocalization studies in macrophages show that dynein and dynactin labeling in the Golgi area is over cis- or trans-tubulovesicular pro¢les rather than the Golgi stack (A. Habermann, T. Schroer, G. Gri¤ths and J. Burkhardt, manuscript in preparation). Similarly, kinesin has been found at steady state to be associated primarily with pre-Golgi structures or the TGN [66,67]. Based on in vitro motility studies of isolated ER and Golgi fractions, it appears that motor proteins concentrate in specialized membrane domains [85], and it has been proposed that these sites may actually play a role in de¢ning dynamic exit compartments for these organelles. In the case of cytoplasmic dynein in ¢broblasts, transport intermediates carrying proteins from the ER are very important cargo organelles. Formation of these intermediates occurs at transitional elements, specialized `ER exit sites'. Studies using green-£uorescent protein to trace the movement of proteins from ER to Golgi have shown that the transport intermediates are relatively large structures or clusters of smaller ones [60,86,87] (see QuickTime movie at http://www.elsevier.com/locate/ multimedia). Their inward movement is dependent on microtubules and on the dynactin complex (and by inference, on cytoplasmic dynein). If microtubules are removed or dynactin function is blocked, Golgi proteins can continue to recycle through the ER to accumulate at ER exit sites. In the former case, this apparently happens without help from kinesin, while in the latter, we have evidence that kinesin may facilitate the process by extending long tubules out-

BBAMCR 14327 5-8-98 Cyaan Magenta

Geel

Zwart

J.K. Burkhardt / Biochimica et Biophysica Acta 1404 (1998) 113^126

ward from the Golgi. With net forward transport blocked, the Golgi proteins accumulate, and short Golgi stacks appear at the ER exit sites. This complex model, ¢rst put forth by Cole et al., is currently the best explanation for the dispersal of the Golgi complex upon microtubule depolymerization ([7], see also the article by Storrie and Yang in this issue). According to this view, `fragmentation' never actually occurs. Instead, the Golgi is systematically disassembled in what Storrie and coworkers have shown is a trans-¢rst fashion [88], and new Golgi stacks are built up at the periphery as a result of an interruption in motor-driven Golgi dynamics. Although the Golgi does not need to be either intact or juxtanuclear to carry out most of its functions, the continued dynein-dependent movement of pre-Golgi intermediates is required to maintain an intact juxtanuclear Golgi complex. Once the intermediates have arrived at the Golgi stack, be it juxtanuclear or dispersed, there is no evidence that motor proteins are involved in movement through the stack, or in maintaining the stacked structure. On the contrary, the formation of short peripheral stacks in the presence of microtubule inhibitors indicates that the ability to form a stacked morphology appears to be an intrinsic `self-assembly' property of ER-derived membranes. Whether these stacks have an asymmetric distribution of cisternal markers has not been tested, but glycosylation of newly synthesized proteins occurs with normal kinetics [6,7], so at least by this criterion the Golgi has been correctly assembled. The formation of Golgi stacks under conditions where microtubule-based transport is blocked resembles the reassembly that can occur in vitro starting from mitotic cell extracts [89]. This in vitro system should provide valuable insights as to how such ordered structures are formed. 8. Looking forward to the next anniversary One hundred years after the discovery of the Golgi complex, we see it as one of the most dynamic of all organelles, whose characteristic position in the cell is maintained by constant motion and membrane £ux. Indeed, we know now that the very structure of the Golgi depends on a dynamic balance between opposing microtubule motor proteins. For the most part,

123

this view stems from data that has come to light within the last decade. Though we have identi¢ed kinesins and dyneins that interact with the Golgi, and begun to assemble a coherent view of how these proteins a¡ect Golgi dynamics at steady state, there is still much to learn. In particular, in the coming years we should expect to learn more about the new kinesin and dynein family members, to better understand how these motors attach to the appropriate transport intermediates, how their activity is regulated, and how they interact with actin-based motors to coordinate organelle motility. One area that is certain to be interesting is the question of how motor activity is coordinated with the cycling of coat proteins to generate appropriate forward and backward membrane £ow. Only when we have a better understanding of these issues will we have a complete picture of the numerous ways that these proteins work to control Golgi structure and function in response to changing cellular demands. Acknowledgements Thanks to Drs. Brian Storrie, Jim McIlvain, and Margaret Morgan for critical reading of the manuscript and for many helpful discussions. Thanks to Dr. Feiguin for contributing Fig. 3. Thanks also to Drs. Karl Fath and David Burgess for communicating their data prior to publication, and for pointing out that epithelial cells refuse to follow the ¢broblast-based dogma.

References [1] A. Kupfer, D. Louvard, S.J. Singer, Polarization of the Golgi apparatus and the microtubule-organizing center in cultured ¢broblasts at the edge of an experimental wound, Proc. Natl. Acad. Sci. USA 79 (1982) 2603^2607. [2] A. Kupfer, G. Dennert, S.J. Singer, Polarization of the Golgi apparatus and the microtubule-organizing center within cloned natural killer cells bound to their targets, Proc. Natl. Acad. Sci. USA 80 (1983) 7224^7228. [3] A.M. Tassin, M. Paintrand, E.G. Berger, M. Bornens, The Golgi apparatus remains associated with microtubule organizing centers during myogenesis, J. Cell Biol. 101 (1985) 630^ 638. [4] E. Robbins, N. Gonatas, Histochemical and structural studies on HeLa cell cultures exposed to spindle inhibitors with

BBAMCR 14327 5-8-98 Cyaan Magenta

Geel

Zwart

124

[5] [6]

[7]

[8]

[9]

[10]

[11]

[12]

[13]

[14]

[15]

[16]

[17]

[18]

[19]

J.K. Burkhardt / Biochimica et Biophysica Acta 1404 (1998) 113^126 special reference to the interphase cell, J. Histochem. Cytochem. 12 (1964) 704^711. J. Thyberg, S. Moskalewski, Microtubules and the organization of the Golgi complex, Exp. Cell Res. 159 (1985) 1^16. A.A. Rogalski, J.E. Bergmann, S.J. Singer, E¡ect of microtubule assembly status on the intracellular processing and surface expression of an integral protein of the plasma membrane, J. Cell Biol. 99 (1984) 1101^1109. N.B. Cole, N. Sciaky, A. Marotta, J. Song, J. LippincottSchwartz, Golgi dispersal during microtubule disruption: regeneration of Golgi stacks at peripheral endoplasmic reticulum exit sites, Mol. Biol. Cell 7 (1996) 631^650. J. Scheel, R. Matteoni, T. Ludwig, B. Ho£ack, T.E. Kreis, Microtubule depolymerization inhibits transport of cathepsin D from the Golgi apparatus to lysosomes, J. Cell Sci. 96 (1990) 711^720. F. Lafont, K. Simons, The role of microtubule-based motors in the exocytic transport of polarized cells, Semin. Cell Dev. Biol. 7 (1996) 343^355. A.A. Rogalski, S.J. Singer, Associations of elements of the Golgi apparatus with microtubules, J. Cell Biol. 99 (1984) 1092^1100. K. Luby-Phelps, P.E. Castle, D.L. Taylor, F. Lanni, Hindered di¡usion of inert tracer particles in the cytoplasm of mouse 3T3 cells, Proc. Natl. Acad. Sci. USA 84 (1987) 4910^ 4913. J.R. Turner, A.M. Tartako¡, The response of the Golgi complex to microtubule alterations: the roles of metabolic energy and membrane tra¤c in Golgi complex organization, J. Cell Biol. 109 (1989) 2081^2088. A. Maniotis, M. Schliwa, Microsurgical removal of centrosomes blocks cell reproduction and centriole generation in BSC-1 cells, Cell 67 (1991) 495^504. W.C. Ho, B. Storrie, R. Pepperkok, W. Ansorge, P. Karecla, T.E. Kreis, Movement of interphase Golgi apparatus in fused mammalian cells and its relationship to cytoskeletal elements and rearrangement of nuclei, Eur. J. Cell Biol. 52 (1990) 315^327. W.C. Ho, V.J. Allan, G. van Meer, E.G. Berger, T.E. Kreis, Reclustering of scattered Golgi elements occurs along microtubules, Eur. J. Cell Biol. 48 (1989) 250^263. L.G. Bergen, R. Kuriyama, G.G. Borisy, Polarity of microtubules nucleated by centrosomes and chromosomes of Chinese hamster ovary cells in vitro, J. Cell Biol. 84 (1980) 151^ 159. C. Achler, D. Filmer, C. Merte, D. Drenckhahn, Role of microtubules in polarized delivery of apical membrane proteins to the brush border of the intestinal epithelium, J. Cell Biol. 109 (1989) 179^189. R. Bacallao, C. Antony, C. Dotti, E. Karsenti, E.H.K. Stelzer, K. Simons, The subcellular organization of Madin^Darby canine kidney cells during the formation of a polarized epithelium, J. Cell Biol. 109 (1989) 2817^2832. K.R. Fath, G.M. Trimbur, D.R. Burgess, Molecular motors are di¡erentially distributed on Golgi membranes from polarized epithelial cells, J. Cell Biol. 126 (1994) 661^675.

[20] M.S. Cooper, A.H. Cornell-Bell, A. Chernjavsky, J.W. Dani, S.J. Smith, Tubulovesicular processes emerge from transGolgi cisternae, extend along microtubules, and interlink adjacent trans-Golgi elements into a reticulum, Cell 61 (1990) 135^145. [21] S.T. Brady, A novel brain ATPase with properties expected for the fast axonal transport motor, Nature 317 (1985) 73^ 75. [22] J.M. Scholey, M.E. Porter, P.M. Grissom, J.R. McIntosh, Identi¢cation of kinesin in sea urchin eggs, and evidence for its localization in the mitotic spindle, Nature 318 (1985) 483^ 486. [23] R.D. Vale, T.S. Reese, M.P. Sheetz, Identi¢cation of a novel force-generating protein, kinesin, involved in microtubulebased motility, Cell 42 (1985) 39^50. [24] R.J. Lye, M.E. Porter, J.M. Scholey, J.R. McIntosh, Identi¢cation of a microtubule-based cytoplasmic motor in the nematode C. elegans, Cell 51 (1987) 309^318. [25] B.M. Paschal, R.B. Vallee, Retrograde transport by the microtubule-associated protein MAP 1C, Nature 330 (1987) 181^183. [26] T.A. Schroer, E.R. Steuer, M.P. Sheetz, Cytoplasmic dynein is a minus end-directed motor for membranous organelles, Cell 56 (1989) 937^946. [27] V. Allan, Membrane tra¤c motors, FEBS Lett. 369 (1995) 101^106. [28] G.S. Bloom, S.A. Endow, Motor proteins 1: kinesins, Protein Pro¢le 2 (1995) 1105^1171. [29] S.T. Brady, A.O. Sperry, Biochemical and functional diversity of microtubule motors in the nervous system, Curr. Opin. Neurobiol. 5 (1995) 551^558. [30] H.V. Goodson, C. Valetti, T.E. Kreis, Motors and membrane tra¤c, Curr. Opin. Cell. Biol. 9 (1997) 18^28. [31] E.L. Holzbaur, R.B. Vallee, DYNEINS: molecular structure and cellular function, Annu. Rev. Cell Biol. 10 (1994) 339^ 372. [32] T.A. Schroer, Structure, function and regulation of cytoplasmic dynein, Curr. Opin. Cell Biol. 6 (1994) 69^73. [33] D.A. Skou¢as, J.M. Scholey, Cytoplasmic microtubulebased motor proteins, Curr. Opin. Cell Biol. 5 (1993) 95^ 104. [34] R. Vallee, K. Vaughan, C. Echeverri, Targeting of cytoplasmic dynein to membranous organelles and kinetochores, Cold Spring Harbor Symp. Quant. Biol. 60 (1995) 803^811. [35] R.B. Vallee, M.P. Sheetz, Targeting of motor proteins, Science 271 (1996) 1539^1544. [36] R.A. Walker, M.P. Sheetz, Cytoplasmic microtubule-associated motors, Annu. Rev. Biochem. 62 (1993) 429^451. [37] L.S. Goldstein, Structural features involved in force generation in the kinesin superfamily, Biophys. J. 68 (1995) 260S265S; discussion 265S^266S. [38] N. Hirokawa, K.K. P¢ster, H. Yorifuji, M.C. Wagner, S.T. Brady, G.S. Bloom, Submolecular domains of bovine brain kinesin identi¢ed by electron microscopy and monoclonal antibody decoration, Cell 56 (1989) 867^878. [39] D.A. Skou¢as, D.G. Cole, K.P. Wedaman, J.M. Scholey,

BBAMCR 14327 5-8-98 Cyaan Magenta

Geel

Zwart

J.K. Burkhardt / Biochimica et Biophysica Acta 1404 (1998) 113^126

[40]

[41]

[42]

[43]

[44]

[45]

[46]

[47]

[48]

[49]

[50]

[51]

[52]

[53]

[54]

The carboxyl-terminal domain of kinesin heavy chain is important for membrane binding, J. Biol. Chem. 269 (1994) 1477^1485. J.L. Cyr, K.K. P¢ster, G.S. Bloom, C.A. Slaughter, S.T. Brady, Molecular genetics of kinesin light chains: generation of isoforms by alternative splicing, Proc. Natl. Acad. Sci. USA 88 (1991) 10114^10118. J.M. McIlvain Jr., J.K. Burkhardt, S. Hamm Alvarez, Y. Argon, M.P. Sheetz, Regulation of kinesin activity by phosphorylation of kinesin-associated proteins, J. Biol. Chem. 269 (1994) 19176^19182. L. Lindesmith, J.M. McIlvain Jr., Y. Argon, M.P. Sheetz, Phosphotransferases associated with the regulation of kinesin motor activity, J. Biol. Chem. 272 (1997) 22929^22933. H.V. Goodson, S.J. Kang, S.A. Endow, Molecular phylogeny of the kinesin family of microtubule motor proteins, J. Cell Sci. 107 (1994) 1875^1884. N. Hirokawa, The molecular mechanism of organelle transport along microtubules: the identi¢cation and characterization of KIFs (kinesin superfamily proteins), Cell Struct. Funct. 21 (1996) 357^367. K.A. Johnson, S.P. Marchese-Ragona, D.B. Clutter, E.L. Holzbaur, T.J. Chilcote, Dynein structure and function, J. Cell. Sci. Suppl. 5 (1986) 189^196. R.D. Vale, Y.Y. Toyoshima, Microtubule translocation properties of intact and proteolytically digested dyneins from Tetrahymena cilia, J. Cell Biol. 108 (1989) 2327^2334. W. Ste¡en, J.L. Hodgkinson, G. Wiche, Immunogold localization of the intermediate chain localization within the complex of cytoplasmic dynein, J. Struct. Biol. 117 (1996) 227^ 235. S. Karki, E.L. Holzbaur, A¤nity chromatography demonstrates a direct binding between cytoplasmic dynein and the dynactin complex, J. Biol. Chem. 270 (1995) 28806^28811. K.T. Vaughan, R.B. Vallee, Cytoplasmic dynein binds dynactin through a direct interaction between the intermediate chains and p150Glued , J. Cell Biol. 131 (1995) 1507^1516. C.J. Echeverri, B.M. Paschal, K.T. Vaughan, R.B. Vallee, Molecular characterization of the 50-kD subunit of dynactin reveals function for the complex in chromosome alignment and spindle organization during mitosis, J. Cell Biol. 132 (1996) 617^633. T. Gaglio, M.A. Dionne, D.A. Compton, Mitotic spindle poles are organized by structural and motor proteins in addition to centrosomes, J. Cell Biol. 138 (1997) 1055^1066. T. Gaglio, A. Saredi, J.B. Bingham, M.J. Hasbani, S.R. Gill, T.A. Schroer, D.A. Compton, Opposing motor activities are required for the organization of the mammalian mitotic spindle pole, J. Cell Biol. 135 (1996) 399^414. R. Heald, R. Tournebize, T. Blank, R. Sandaltzopoulos, P. Becker, A. Hyman, E. Karsenti, Self-organization of microtubules into bipolar spindles around arti¢cial chromosomes in Xenopus egg extracts, Nature 382 (1996) 420^425. A. Merdes, K. Ramyar, J.D. Vechio, D.W. Cleveland, A complex of NuMA and cytoplasmic dynein is essential for mitotic spindle assembly, Cell 87 (1996) 447^458.

125

[55] W.S. Saunders, D. Koshland, D. Eshel, I.R. Gibbons, M.A. Hoyt, Saccharomyces cerevisiae kinesin- and dynein-related proteins required for anaphase chromosome segregation, J. Cell Biol. 128 (1995) 617^624. [56] E. Vaisberg, M.P. Koonce, J.R. McIntosh, Cytoplasmic dynein plays a role in mammalian mitotic spindle formation, J. Cell Biol. 123 (1993) 849^858. [57] V. Allan, Protein phosphatase 1 regulates the cytoplasmic dynein-driven formation of endoplasmic reticulum networks in vitro, J. Cell Biol. 128 (1995) 879^891. [58] F. Aniento, N. Emans, G. Gri¤ths, J. Gruenberg, Cytoplasmic dynein-dependent vesicular transport from early to late endosomes, J. Cell Biol. 123 (1993) 1373^1387. [59] A. Blocker, F.F. Severin, J.K. Burkhardt, J.B. Bingham, H. Yu, J.C. Olivo, T.A. Schroer, A.A. Hyman, G. Gri¤ths, Molecular requirements for bi-directional movement of phagosomes along microtubules, J. Cell Biol. 137 (1997) 113^129. [60] J. Burkhardt, C. Echeverri, T. Nilsson, R. Vallee, Overexpression of the dynamitin (p50) subunit of the dynactin complex disrupts dynein-dependent maintenance of membrane organelle distribution, J. Cell Biol. 139 (1997) 469^484. [61] F. Lafont, J.K. Burkhardt, K. Simons, Involvement of microtubule motors in basolateral and apical transport in kidney cells, Nature 372 (1994) 801^803. [62] S.X. Lin, K.L. Ferro, C.A. Collins, Cytoplasmic dynein undergoes intracellular redistribution concomitant with phosphorylation of the heavy chain in response to serum starvation and okadaic acid, J. Cell Biol. 127 (1994) 1009^ 1019. [63] E.A. Vaisberg, P.M. Grissom, J.R. McIntosh, Mammalian cells express three distinct dynein heavy chains that are localized to di¡erent cytoplasmic organelles, J. Cell Biol. 133 (1996) 831^842. [64] P.L. Leopold, A.W. McDowall, K.K. P¢ster, G.S. Bloom, S.T. Brady, Association of kinesin with characterized membrane-bounded organelles, Cell Motil. Cytoskel. 23 (1992) 19^33. [65] D.L. Marks, J.M. Larkin, M.A. McNiven, Association of kinesin with the Golgi apparatus in rat hepatocytes, J. Cell Sci. 107 (1994) 2417^2426. [66] J. Lippincott-Schwartz, N.B. Cole, A. Marotta, P.A. Conrad, G.S. Bloom, Kinesin is the motor for microtubule-mediated Golgi-to-ER membrane tra¤c, J. Cell Biol. 128 (1995) 293^306. [67] K.J. Johnson, E.S. Hall, K. Boekelheide, Kinesin localizes to the trans-Golgi network regardless of microtubule organization, Eur. J. Cell Biol. 69 (1996) 276^287. [68] J.K. Burkhardt, J.M. McIlvain Jr., M.P. Sheetz, Y. Argon, Lytic granules from cytotoxic T cells exhibit kinesin-dependent motility on microtubules in vitro, J. Cell Sci. 104 (1993) 151^162. [69] T. Nakata, N. Hirokawa, Point mutation of adenosine triphosphate-binding motif generated rigor kinesin that selectively blocks anterograde lysosome membrane transport, J. Cell Biol. 131 (1995) 1039^1053.

BBAMCR 14327 5-8-98 Cyaan Magenta

Geel

Zwart

126

J.K. Burkhardt / Biochimica et Biophysica Acta 1404 (1998) 113^126

[70] F. Feiguin, A. Ferreira, K.S. Kosik, A. Caceres, Kinesinmediated organelle translocation revealed by speci¢c cellular manipulations, J. Cell Biol. 127 (1994) 1021^1039. [71] I. Corthesy Theulaz, A. Pauloin, S.R. Pfe¡er, Cytoplasmic dynein participates in the centrosomal localization of the Golgi complex, J. Cell Biol. 118 (1992) 1333^1345. [72] E.R. Steuer, L. Wordeman, T.A. Schroer, M.P. Sheetz, Localization of cytoplasmic dynein to mitotic spindles and kinetochores, Nature 345 (1990) 266^268. [73] E. Vaisberg, P. Grissom, S. Gill, T. Schroer, J.R. McIntosh, Cytoplasmic dynein 2 and dynactin are part of the ER to Golgi transport pathway, Mol. Biol. Cell. 8 (1997) 407a. [74] S.R. Gill, T.A. Schroer, I. Szilak, E.R. Steuer, M.P. Sheetz, D.W. Cleveland, Dynactin, a conserved, ubiquitously expressed component of an activator of vesicle motility mediated by cytoplasmic dynein, J. Cell Biol. 115 (1991) 1639^ 1650. [75] T.A. Schroer, M.P. Sheetz, Two activators of microtubulebased vesicle transport, J. Cell Biol. 115 (1991) 1309^1318. [76] V. Allan, Motor proteins: a dynamic duo, Curr. Biol. 6 (1996) 630^633. [77] T.A. Schroer, Structure and function of dynactin, Semin. Cell Dev. Biol. 7 (1996) 321^328. [78] D.A. Schafer, S.R. Gill, J.A. Cooper, J.E. Heuser, T.A. Schroer, Ultrastructural analysis of the dynactin complex: an actin-related protein is a component of a ¢lament that resembles F-actin, J. Cell Biol. 126 (1994) 403^412. [79] C.M. Waterman-Storer, S. Karki, E.L. Holzbaur, The p150Glued component of the dynactin complex binds to both microtubules and the actin-related protein centractin (Arp-1), Proc. Natl. Acad. Sci. USA 92 (1995) 1634^1638. [80] C.M. Waterman-Storer, E.L. Holzbaur, The product of the Drosophila gene, Glued, is the functional homologue of the

[81]

[82]

[83]

[84]

[85]

[86]

[87]

[88]

[89]

p150Glued component of the vertebrate dynactin complex, J. Biol. Chem. 271 (1996) 1153^1159. E.A. Holleran, M.K. Tokito, S. Karki, E.L. Holzbaur, Centractin (ARP1) associates with spectrin revealing a potential mechanism to link dynactin to intracellular organelles, J. Cell Biol. 135 (1996) 1815^1829. K.A. Beck, J.A. Buchanan, V. Malhotra, W.J. Nelson, Golgi spectrin: identi¢cation of an erythroid beta-spectrin homolog associated with the Golgi complex, J. Cell Biol. 127 (1994) 707^723. P. Devarajan, P.R. Stabach, A.S. Mann, T. Ardito, M. Kashgarian, J.S. Morrow, Identi¢cation of a small cytoplasmic ankyrin (AnkG119) in the kidney and muscle that binds beta I sigma spectrin and associates with the Golgi apparatus, J. Cell Biol. 133 (1996) 819^830. K.R. Fath, G.M. Trimbur, D.R. Burgess, Molecular motors and a spectrin matrix associate with Golgi membranes in vitro, J. Cell Biol. 139 (1997) 1169^1181. V. Allan, R. Vale, Movement of membrane tubules along microtubules in vitro : evidence for specialised sites of motor attachment, J. Cell Sci. 107 (1994) 1885^1897. J.F. Presley, N.B. Cole, T.A. Schroer, K. Hirschberg, K.J. Zaal, J. Lippincott-Schwartz, ER-to-Golgi transport visualized in living cells, Nature 389 (1997) 81^85. S. Scales, R. Pepperkok, T.E. Kreis, Visualization of ER-toGolgi transport in living cells reveals a sequential mode of action for COPII and COPI, Cell 90 (1997) 1137^1148. W. Yang, B. Storrie, Scattered Golgi elements during microtubule disruption are initially enriched in trans Golgi proteins, Mol. Biol. Cell. 9 (1998) 191^207. C. Rabouille, T. Misteli, R. Watson, G. Warren, Reassembly of Golgi stacks from mitotic Golgi fragments in a cell-free system, J. Cell Biol. 129 (1995) 605^618.

BBAMCR 14327 5-8-98 Cyaan Magenta

Geel

Zwart