Cell Motility and the Cytoskeleton 34:299-312 (1996) Comparison of the lntracellular Distribution of Cytoplasmic Dynein and Kinesin in Cultured Cells: Motor Protein Location Does Not Reliably Predict Function Sharron X.H. Lin, K. Kevin Pfister, and Christine A. Collins Department of Cell and Molecular Biology, Northwestern University Medical School, Chicago, Illinois (S.X.H.L., C.A.C.); Department of Cell Biology, School of Medicine, University of Virginia Health Science Center, Charlottesville, Virginia (K.K.F.) While immunolocalization methods have been used as a reasonable means to judge where a given molecule may be active in the cellular milieu, the correlation between distribution and function for proteins involved in intracellular transport may not be clear cut. To address the question of specificity and reproducibility of immunolocalization of microtubule-based motor proteins, we have co-localized cytoplasmic dynein and kinesin by immunofluorescence microscopy using two specific antibodies for each motor molecule. The results indicate that cytoplasmic dynein and kinesin appear to co-localize on a small subset of vesicles, but largely reside or accumulate on morphologically distinct organelles. In addition, antikinesin antibodies differing in their epitope specificity label different cellular compartments. To address the question of whether the distribution of motor molecules is representative of organelles that are undergoing active transport, we have altered the activity of vesicle trafficking pathways in fibroblasts using several different methods, including cytoplasmic acidification and disruption of cellular compartments with brefeldin A, nocodazole and okadaic acid. Analysis of the distribution of cytoplasmic dynein and kinesin under these conditions indicates that immunolocalization data alone are not reliable indicators of sites of likely function for these microtubule-based motors. 0 1996 wiiey-Liss, 1nc. Key words: organelle transport, microtubule-based motility, okadaic acid, immunolocalization, Golgi apparatus INTRODUCTION Cytoplasmic dynein and kinesin are considered the most likely candidates for mediating retrograde and an- terograde ATP- and microtubule-dependent organelle translocations. In developing a model for how cytoplasmic dynein and kinesin carry out regulated, polarized motility, it is important to determine whether one or both motors interact with an individual organelle at a given time, and how the intracellular distributions of these proteins are established and modulated. Existing immunolocalization data support the hypothesis that the motor molecule resides on the vesicle to be translocated, as much of the cellular content of dynein and kinesin ap0 1996 Wiley-Liss, lnc. pears to be associated with membranous organelles. However, there is also evidence for a soluble pool of cytoplasmic dynein and kinesin. Whether this cytoplasmic pool reflects a discrete population of motor protein with altered activities is not known. However, increases Received January 17, 1996; accepted April 19, 1996. Address reprint requests to Christine A. Collins, Department of Cell and Molecular Biology, Northwestern University Medical School, 303 East Chicago Avenue, Chicago, IL 6061 1-3008. Dr. Lin’s current address is Department of Molecular and Cell Biology University of California, Berkeley Life Science Addition Bldg 57 1 Berkeley, CA 94720 300 Lin et al. in the soluble pool of cytoplasmic dynein following certain treatments suggest that the cell has means of altering the equilibrium between membrane bound and “free” motor protein [Lin et al. , 19941. From the literature, the relationship between intracellular localization of motor proteins and their predicted sites of activity in organelle transport is difficult to evaluate. Cytoplasmic dynein has been shown by immunofluorescence microscopy to be associated with membranous organelles in several cell types [Hirokawa et al., 1990; Koonce and McIntosh, 1990; Pfarr et al., 19901, and in particular to be associated with lysosomes in fibroblasts [Lin and Collins, 1992, 19931. Assays of organelle motility and fusion in permeabilized cells or in vitro reconstitution systems predict dynein involvement in transport between early and late endosomes [Bomsel et al., 1990; Aniento et al., 1993; Oda et al., 19951, apical transport in polarized epithelial cells [Lafont et a1., 19941, and perinuclear Golgi vesicle accumulation [Corthesy-Theulaz et al., 19921. However, immunolocalization studies and biochemical fractionation experiments do not indicate specific enrichment of cytoplasmic dynein on organelles involved in these transport processes. The analysis of kinesin distribution is even more complicated, as several distinctly different localization patterns have been described. Pfister and colleagues, using a library of monoclonal antibodies to the kinesin heavy and light chains, demonstrated that kinesin is associated with membranous organelles [Pfister et al., 1989; Hirokawa et al., 1991; Leopold et al., 19921. Recent work used one of these heavy chain antibodies to localize kinesin to a pre-Golgi vesicle compartment in fibroblasts [Lippincott-Schwartz et al., 19951. Using a monoclonal antibody to sea urchin kinesin, Henson et al. [ 19921 found this motor protein regionally codistributed with an ER marker in sea urchin coelomocytes, and Houliston and Elinson  found an association with cortical vesicles in Xenopus eggs. Using the same antisea urchin kinesin antibody, Golgi staining has been observed in a variety of cultured cells [Murphy et al., 1991; Johnson et al., 1993; Marks et al., 19941. Polyclonal antibodies recognizing conserved peptide sequences in kinesin heavy chain stained Golgi in rat hepatocytes, and kinesin was enriched in Golgi membranes and transcytotic and secretory vesicles from liver cells [Marks et al., 19941. However, other investigators have failed to detect kinesin in Golgi following biochemical fractionation [Leopold et al., 1992; Schmitz et al., 1994; Fath et al. , 19941. Functional assays implicate kinesin in anterograde axonal transport [Brady et al., 1990; Saxton et al., 1991; Ferreira et al., 19921, in microtubule-dependent transport of ER membranes [Dabora and Sheetz, 1988, Vale and Hotani, 1988; Feiguin et al., 19941, secretory vesicles [Urmtia et al., 1991; Burkhardt et al., 19931, and pigment granules [Rodionov et al., 19911, and in extension of tubular lysosomes [Hollenbeck and Swanson, 19901 and Golgi to ER vesicular transport [LippincottSchwartz et al., 19951. Though the localization of kinesin on ER- or Golgi-related structures could be considered consistent with a role for lunesin in secretion and maintenance of ER distribution in the cell, the differences observed in immunof luorescence and biochemical localization among various investigators is puzzling. A possible explanation for disparate localization results is that various intracellular distributions reported for kinesin, and for dynein, reflect differences between cell types, in antibody specificity or affinity, or other factors due to differences in methodology among various investigators. It is also possible that different localizations reflect the recognition of distinct populations, e.g. splicing variants, proteins with altered subunit composition, post-translational modifications of the same protein, or motor proteins in different states due to their association with activating complexes or receptors. To directly compare the intracellular distributions of cytoplasmic dynein and kinesin, we determined the localization of these microtubule-based motors in the same cell type under identical conditions. Four specific antibody probes, two which react with cytoplasmic dynein and two distinct antibodies to kinesin, were used. The results for dynein localization were consistent with those previously described [Lin and Collins, 19921, and the kinesin distribution was found to be dependent on the epitope specificity of the particular antibody used. To assess the relationship between localization and function, we investigated whether the distributions of kinesin and dynein are altered by conditions that disrupt the normal pathways of vesicle trafficking thought to be dependent on microtubule motors. These conditions included using nocodazole, brefeldin A, and okadaic acid treatment, and intracellular acidification. The results from these studies indicate that the intracellular localizations of both kinesin and dynein are modulated by the cell. However, the highest concentration or accumulation of motor protein detected by immunolocalization does not necessarily correlate with regions of presumed motor activity. MATERIALS AND METHODS Cell Culture and lmmunofluorescence Microscopy BHK-21, NIH-3T3, and NRK fibroblasts were each cultured in DME with 10% calf serum and antibiotics (100 i.u./ml penicillin and 100 mg/ml streptomycin). Hippocampal cells were isolated from 1-day-old rat pups and were cultured in DMEM [Goslin and Banker, Dynein and Kinesin Localization 19911. Cells were fixed and permeabilized by immersion of coverslips into methanol at -20°C for 5 min. The fixed cells were incubated with primary antibodies for 1 h in a humidified chamber at 37°C. Following washes in PBS (50 mM NaPO,, pH 7.4, 150 mM NaCl), goat anti-rabbit or anti-mouse secondary antibodies conjugated to rhodamine or fluorescein (Jackson Immunoresearch Laboratories, West Grove, PA) were diluted to 45 pg/ml in PBS and incubated with the coverslips for 15 min as above. The coverslips were washed, mounted, viewed, and photographed as previously described [Lin and Collins. 19931. Antibodies to cytoplasmic dynein were a rabbit polyclonal serum raised against native calf brain dynein that reacts primarily with the 74 kDa intermediate chain and weakly with the dynein heavy chain on Western blots [Lin and Collins, 19931 and a mouse monoclonal antibody that reacts with all isoelectric point variants of the 74 kDa intermediate chain of rat brain dynein [Dillman and Pfister, 19941. Two monoclonal antibodies that recognize different epitopes of the kinesin heavy chain were previously characterized by Pfister et al. . In bovine brain extracts, kinesin monoclonal antibody H 1 recognizes a single polypeptide species of 124 kDa, whereas antibody H2 recognizes, in addition, a second species migrating with slightly slower mobility resolved in some gel electrophoresis systems. Cell Treatment Conditions Microtubule disruption was carried out by treating cells with 10 pM taxol or 10 pM vinblastine for 16 h or incubation with 10 p M nocodazole for 2 h. Effects of these drugs on the microtubule cytoskeleton were verified by double-labeling with tubulin antibody (monoclonal anti-a tubulin, Amersham Corp., Arlington Heights, IL) or single-labeling for tubulin in cells on separate coverslips treated under identical conditions for comparison with kinesin distributions. Cells were acidified by incubation with 70 mM sodium acetate, pH 6.5 for 45 min [Heuser, 1989al as previously described [Lin and Collins, 19921. Brefeldin A was used at 10 pg/ml for 15 min. Okadaic acid (1.5 pM) was added to cells in complete culture medium for 45 min. For ATP depletion experiments. 5 mM NaN, was added to glucose-deficient DME in the presence or absence of dialyzed serum. Materials was Obtained from the Drug Synthesis and Chemistry Branch, Division of Cancer Research, National Cancer Institute, and stored as a 10 mM stock in DMSO at -80°C. Vinblastine, nocodazole, and brefeldin A were stored in DMSO at -40°C (10 mM, 6.6 mM, and 5 mg/ml stock solutions, respectively), and okadaic acid (500 pM in DMSO) was stored at 4°C. 301 RESULTS Dynein and Kinesin Localize to Discrete Organelle Populations Studies using a polyclonal antibody raised against native cytoplasmic dynein have demonstrated that this microtubule motor protein localizes to lysosomes clustered in the perinuclear region of fibroblastic cells [Lin and Collins, 1992, 19931. Using a monoclonal antibody specific for the 74 kDa intermediate chain subunit(s) of cytoplasmic dynein [Dillman and Pfister, 19941, essentially complete overlap with the polyclonal anti-dynein antibody staining pattern was observed in NRK fibroblasts (Fig. 1). As the monoclonal antibody recognizes all isoelectric point variants of the intermediate chain subunit [Dillman and Pfister, 19941, this result suggests that staining pattern observed with either antibody is representative of the total population of dynein molecules. Since the anti-kinesin antibodies HI and H2 appear to recognize different epitopes on kinesin heavy chain [Pfister et al., 19891, we wanted to determine whether different populations of organelles are recognized by the two antibodies within the same cell. In Figure 2 we compared staining of BHK and NRK fibroblasts using antikinesin antibodies HI and H2. Results using antibody H1 (Fig. 2A,C) were similar to the intracellular staining patterns found previously using this antibody in other cell types [Pfister et al., 19891, and consistent with staining of a pre-Golgi compartment [Lippincott-Schwartz et al., 19951. However, antibody H2 appeared to recognize a different kinesin population than that identified using antibody H1 in both cell types, and the staining pattern for antibody H2 itself differed in the two fibroblast cell lines. In Figure 2B, the staining pattern using the kinesin antibody H2 in BHK cells was largely diffuse throughout the cytoplasm. In Figure 2D, the staining pattern using kinesin antibody H2 in NRK fibroblasts was distinctly tubulovesicular and concentrated near the nucleus, similar to Golgi localizations for kinesin reported by other investigators [Murphy et al., 1991; Johnson et al., 1993; Marks et al.. 19941, and coincident with the staining pattern obtained in this cell type using wheat germ agglutinin, a marker used to label Golgi structures (data not shown). We identified kinesin on Western blots of BHK and NRK whole cell extracts, and determined that a protein species of identical molecular weight was recognized by both antibodies H1 and H2 in both cell types (data not shown). Therefore it appears that the two antibodies recognize different pools of kinesin within the same cell. Dynein and Kinesin Reside and/or Accumulate on Distinct VesiclePopu1ations Both dynein and kinesin have been shown to be associated with membranous organelles in these and 302 Lin et al. Fig. 1 . Cytoplasmic dynein localizes to discrete organelles in NRK cells. Normal rat kidney fibroblasts were fixed and double-labeled using a previously characterized polyclonal anti-dynein antibody (A) and a monoclonal antibody specific for the cytoplasmic dynein intermediate chain subunit (B). Bar, 10 p.m. other studies, but co-localization of these motor proteins by light microscopy had not previously been examined. In Figure 3, the staining patterns for cytoplasmic dynein and kinesin in two different cell lines double-labeled using polyclonal anti-dynein and monoclonal anti-kinesin H 1 antibodies are shown. The immunofluorescent staining patterns for the two motor proteins were largely distinct from one another. In all cells tested, it appeared that the kinesin-positive organelles were greater in number and more evenly distributed in the cytoplasm than the dyneinpositive organelles. However, there also appeared to be some degree of overlap in the staining patterns, particularly in the 3T3 cells shown in Figure 3 A,B (arrows). Confocal microscopy was also used to evaluate the extent of overlap in the distributions of dynein and kinesin. In Figure 4, a rat glial cell from a culture of primary hippocampal cells was double-labeled using polyclonal anti-dynein antibody and monoclonal anti-kinesin HI antibody. The resulting staining patterns were similar to those found in fibroblasts. Kinesin structures (green) were more numerous and evenly distributed throughout the cytoplasm than dynein positive vesicles (red). A small degree of overlap in signal, indicated by yellow, was found in the perinuclear region. alter the cellular distribution of dynein-positive organelles [Lin and Collins, 19921. In the experiment shown in Figure 5, the effect of microtubule disruption on the relative distributions of dynein and kinesin was evaluated. The organelle distributions were altered by microtubule disruption, but the apparent association between the organelles and the motor molecule was not affected. Dynein and kinesin associated organelles, labeled using anti-dynein polyclonal and kinesin H 1 monoclonal antibodies, were still largely distinct from one another in distribution and appearance. Surprisingly, the staining pattern found for antibody H2 was more significantly altered following microtubule disruption (Fig. 5E) compared with control cells (Fig. 2D). The localization found suggests that the kinesin recognized by antibody H2 is associated with membranous structures that vesiculate following microtubule depolymerization, such as the Golgi stacks [Rogalski and Singer, 19841. Disruption of Organelle Trafficking Alters the Distribution of Kinesin Both kinesin and dynein remained associated with organelles in the absence of microtubules, when these motor proteins would be expected to be non-functional in terms of directed force production. However, we hypothMotor Proteins Remain Associated With esize that binding interactions between motor molecules Organelles in the Absence of Microtubules and their target organelles might be disrupted upon inhiNocodazole treatment of cells leads to vesiculation bition or activation of vesicle trafficking pathways under and dispersion of the Golgi apparatus as well as redistri- other circumstances. To test this premise we examined bution of other cellular organelles [Rogalski and Singer, several conditions that have been shown to affect micro1984; Terasaki et al., 1986; Matteoni and Kreis, 19871. tubule-dependent vesicle transport for their effect on cyVinblastine and taxol treatment have also been shown to toplasmic dynein and kinesin localization. Dynein and Kinesin Localization 303 Fig. 2. Kinesin antibodies recognize different populations of organelles in fibroblasts. BHK (A,B) and NRK (C,D) cells were fixed and labeled using kinesin monoclonal antibody H1 (A$) or antibody H2 (B,D). Bar, 10 pm. Brefeldin A has been used to disrupt secretion in cells, as it causes mixing of Golgi proteins with those of the endoplasmic reticulum via a microtubule-dependent process [Lippincott-Schwartz et al., 19901. Kinesin has been implicated in cyclic transport between endoplasmic reticulum and the Golgi apparatus, as microinjection of kinesin antibody H1 specifically blocked Golgi to ER vesicle trafficking in NRK cells [Lippincott-Schwartz et al., 19951. We expected that upon disruption of the Golgi apparatus with brefeldin A, kinesin might associate to a greater extent with pre-Golgi structures. In addition, accumulation of Golgi components in the perinuclear region has been shown to rely on cytoplasmic dynein [CorthesyTheulaz et al., 19921. Brefeldin A treatment might identify a Golgi-specific pool of cytoplasmic dynein through loss of staining in the perinuclear region of the cell. We examined the localization of cytoplasmic dynein and ki- nesin in NRK cells following a 15 min incubation with brefeldin A. The cellular distribution and organelle association of cytoplasmic dynein did not appear to be significantly changed from control cells (data not shown). Likewise, the kinesin epitope recognized by antibody H 1 (Fig. 6A) did not appear to change in distribution. However, as shown in Figure 6B, the kinesin antigen recognized by antibody H2 was altered significantly in its staining pattern by brefeldin A. The epitope recognized by this antibody behaved as might be expected for Golgi proteins following drug treatment, as it became largely restricted to a small bright perinuclear area, with some additional diffuse staining throughout the cytoplasm. After wash out of the drug, the tubulovesicular pattern observed with this antibody under control conditions was again found (Fig. 6C). Cytoplasmic acidification has been shown to in- 304 Lin et al. Fig. 3. Kinesin and dynein appear to co-localize on only a subset of membranous organelles in cultured cells. Polyclonal antibody to cytoplasmic dynein and kinesin H1 monoclonal antibody were used to double-label 3T3 (A,B) and NRK (C,D) cells. Dynein (A,C); kinesin (B,D). Arrows indicate examples of organelles staining for both dynein and kinesin. Bar, 20 km. hibit endocytosis [Cosson et al., 1989; Heuser, 1989b], alter the shape and distribution of tubular lysosomes in fibroblasts and macrophages [Heuser, 1989a1, and lead to the redistribution of late endosomes in neurons and MDCK cells [Parton et al., 19911. A similar change in vesicle distribution was blocked in acidified hippocampal neurons by suppression of kinesin activity via antisense oligonucleotide administration [Feiguin et al., 19941, suggesting activation of kinesin motor function may be responsible for the anterograde accumulation of organelles. If this were true, cytoplasmic acidification might be expected to lead to the accumulation of kinesinpositive organelles at the plus ends of microtubules. Conversely, dynein may no longer associate with those organelles transported anterogradely or accumulated in the periphery of the cell upon cellular acidification. However, cytoplasmic dynein remains associated with peripheral lysosomes following cytoplasmic acidification [Lin and Collins, 19921, suggesting that dynein on these organelles is not active in transport, or that kinesin binding and motor function predominates in this situation. To evaluate the distribution of kinesin in acidified cells, we treated NRK fibroblasts with 70 mM Na acetate, pH 6.5. The immunofluorescence pattern obtained using kinesin antibody H1 changed in a striking manner following acidification (Fig. 7). The margins of NRK cells became brightly labeled in a discontinuous pattern. This staining was most obvious in, but not limited to, regions of cell-cell contact (Fig. 7A). Perinuclear vesicular staining for kinesin was observed in these cells in addition to the peripheral pattern. The same result was found for antibody H 1 staining following cytoplasmic acidification of 3T3 and BHK fibroblasts, indicating that this cellular response was not unique to one cell type (data not shown). Following return to culture medium of neutral pH for 1 h, the marginal kinesin staining was lost Dynein and Kinesin Localization 305 Fig. 4. Cytoplasmic dynein and kinesin co-localization in rat glia. A primary culture of rat hippocampal cells was fixed and stained using polyclonal anti-dynein and monoclonal anti-kinesin H1 antibodies. A glial cell is shown with superimposed confocal microscopic images indicating kinesin (green) and cytoplasmic dynein (red). Areas of overlap are yellow. Bar, 20 km. (Fig. 7B). This striking redistribution was unique to the kinesin recognized by antibody H1, as the tubulovesicular pattern found using kinesin antibody H2 was unaffected by cytoplasmic acidification (data not shown). While the marginal staining pattern was not strictly punctate, as might be expected for vesicle localization, the kinesin did accumulate in a peripheral compartment, consistent with enhanced microtubule-dependent plusend directed motility. To determine whether the peripheral accumulation of kinesin antibody H1 immunoreactivity under these conditions was due to increased kinesin motor activity, we determined the microtubule-dependence of the acidification-induced effect. Microtubules were depolymerized using nocodazole prior to cytoplasmic acidification and during treatment and subsequent recovery. Nocodazole treatment did not prevent the appearance of kinesin staining at the periphery of the cells, nor did it prevent the recovery of a control kinesin pattern following return to fresh culture medium (data not shown). This result does not support a relationship between accumulation of kinesin and increased kinesin motor activity. To investigate this question further, we examined the energy dependence of motor protein-organelle association. As previously reported, vesicles positive for cytoplasmic dynein appeared somewhat more brightly stained and concentrated in the perinuclear region fol- lowing ATP depletion than under control conditions, and we have interpreted this as an increase in the membraneassociated state of dephosphorylated dynein [Lin et al., 19941. Rather than prevent the appearance of a marginal staining pattern, ATP depletion alone led to a similar change in distribution of the antibody H1 epitope as that observed in acidified cells (Fig. 8). As this effect was observed after several hours of NaN, treatment, it is possible that ATP depletion itself led to alterations in cellular pH or other metabolic changes that indirectly affected kinesin localization. However, it appears that the development and/or maintenance of this particular kinesin distribution does not require metabolic energy. Kinesin localization using antibody H2 was also examined in ATP-depleted cells, and was found to be altered with treatment (Fig. 8C). The kinesin recognized by this antibody appeared to be concentrated around a single perinuclear point in NaN, treated cells. The Golgi apparatus was not obviously disrupted under these conditions (data not shown). Okadaic Acid Effects on Motor Protein Localization Okadaic acid is a phosphoprotein phosphatase inhibitor [Cohen et al., 19901, and previous work showed that okadaic acid treatment leads to an apparent decrease in the association between cytoplasmic dynein and mem- 306 Lin et al. Fig. 5 . Localization of dynein and kinesin following disruption of the microtubule network. Polyclonal anti-cytoplasmic dynein and antikinesin monoclonal antibodies were used for immunofluorescence microscopy of BHK (A-D) or NRK (E) cells following microtubule disruption with taxol (A,B), vinblastine (C,D), or nocodazole (E). Dynein (A,C); kinesin antibody H1 (B ,D); kinesin antibody H2 (E). Bar, 10 prn. Arrows indicate cell margin. Dynein and Kinesin Localization 307 Fig. 7. Disruption of vesicle trafficking by cytoplasmic acidification leads to altered distribution of the kinesin H1 antigen. NRK cells were treated as described in Materials and Methods to lower the cytoplasmic pH (A). Acidified cells were then returned to normal culture medium for 1 h (B). Results of immunolocalization with kinesin antibody H1 are shown. Bar, 10 pm. Fig. 6. Kinesin localization following brefeldin A treatment. Kinesin antibodies HI (A) and H2 (B,C) were used for immunolocalization in brefeldin A treated NRK cells (A,B). Brefeldin A treated cells were incubated in fresh medium for an additional 30 rnin in the absence of drug (C). Bar, 20 pm. branous organelles, in vivo, correlating with increased phosphorylation of the motor molecule itself [Lin et al., 19941. Differences in the phosphorylation state of dynein bound to anterogradely moving organelles compared to the total dynein population also suggest regulation of binding site affinity or motor activity by phosphorylation ,Dillman and Pfister, 19941. we speculated that phosphorylation events may also affect kinesin association with membranous binding sites, and as one means of testing this we examined the effect of okadaic acid on the distribution of kinesin in NRK cells. The localization of 308 Lin et al. kinesin recognized by both antibodies H1 and H2 was altered by okadaic acid treatment (Fig. 9). The staining pattern for the kinesin epitope recognized by antibody H1 was largely diffuse throughout the cytoplasm, similar to the change in staining pattern found for dynein in okadaic acid treated cells [Lin et al., 19941. The staining pattern for kinesin antibody H2 was also diffuse, with some residual staining near the nucleus (arrows). The Golgi apparatus in okadaic acid treated cells vesiculates [Lin et al., 19941. However, rather than the kinesin antibody H2 pattern resembling disrupted Golgi elements as seen following nocodazole treatment, the epitope recognized by this antibody was not associated with any discrete organelle population in okadaic acid treated cells. DISCUSSION Fig. 8. ATP depletion leads to intracellular redistribution of the kinesin H1 antigen. NRK cells were treated with NaN, in glucose-free medium to deplete cellular ATP. Polyclonal antibody to cytoplasmic dynein (A), kinesin antibody H1 (B) and kinesin antibody H2 (C) were used to Stain Cells following 3 (A,B) Or 4 (c)h of treatment. Bar, 20 pm. This study represents the first attempt to examine the co-localization of the microtubule-based motors, cytoplasmic dynein and kinesin, by immunof luorescence microscopy. Though we find some overlap in localization of these proteins in the congested perinuclear region of the cell, they are present primarily on discrete populations of organelles within the cultured fibroblasts and glial cells used in this study. Analysis of these localization patterns raised several questions, and we chose to concentrate on two of them. First, we wanted to determine whether the disparate results found by other investigators for the cellular distribution of cytoplasmic dynein and kinesin could be explained by differences in technique, cell type, or antibody specificity. Secondly, we wanted to address the question of whether the localization determined was representative of the population of organelles undergoing active transport by their respective motor protein. In addressing the first point, we found that the staining pattern for cytoplasmic dynein was consistent between the two antibodies used, and the labeled structures appear to be primarily lysosomes, as previously described [Lin and Collins, 19921. Whether there is also dynein associated with the Golgi apparatus is more difficult to address due to the concentration of dynein positive structures in the perinuclear region. A Golgi association of both kinesin and dynein might explain the degree of overlap in staining patterns observed for these motor proteins, however. The difference in staining patterns between kinesin antibodies H1 and H2 was quite striking. The diffuse pattern in BHK may represent either a even distribution of the kinesin antibody H2 epitope on many membrane structures or a more ‘‘soluble’’ pool of kinesin. However, in NRK cells the staining pattern for antibody H2 closely resembled that of the Golgi apparatus as also Dynein and Kinesin Localization 309 Fig. 9. Kinesin staining of intracellular organelles is altered following treatment with okadaic acid. Control NRK cells (A,B), or those incubated with 1.5 pM okadaic acid for 45 min (C,D) were processed for immunofluorescence microscopy using monoclonal anti-kinesin HI (A,C) or H2 (B,D). Bar, 10 km. described by other investigators. In squid axoplasm and mammalian neurons the H2 antibody recognizes small vesicles and tubulovesicular structures [Brady et al., 1990; Dahlstrom et al., 1991; Hirokawa et al., 19911, suggesting that there may be cell type specific effects on the distribution or accessibility of the kinesin epitope recognized by this antibody. Cell type specific staining has also been demonstrated for the SUK4 anti-sea urchin kinesin antibody, which recognizes a dispersed vesicle population in coelomocytes (the pre-Golgi compartment in this cell?), but stains the Golgi apparatus in a variety of mammalian cell lines. We show in this study that in normal rat kidney fibroblasts both the pre-Golgi transport vesicles and discrete Golgi-like distributions are observed, using the H1 and H2 antibodies, respectively. It should be stressed that both these antibodies recognize bona fide kinesin heavy chain in both brain tissue and cell extracts. The identity of the epitopes on the kinesin heavy chain recognized by these antibodies must be determined before conclusions can be drawn as to the reasons behind the differences in staining pattern seen within the same cell, as well as between different cell types. However, it is intriguing that the two staining patterns for kinesin we obtained in this study within the same cell type, i.e. pre-Golgi dispersed punctate staining and perinuclear Golgi-like staining, describe the two staining patterns described in the literature for kinesins using different antibodies in different cell types. This result indicates the presence of at least two discrete pools of kinesin within the cell that may have distinct roles to play in vesicle motility. The primarily lysosomal distribution for cytoplas- 310 Lin et al. mic dynein is consistent with accumulation of the motor molecule on this organelle once it has accomplished its minus-end transport activity. Kinesin (immunolocalized using antibody H1) is most concentrated on vesicles that are involved in the cyclical movement of material between Golgi and ER [Lippincott-Schwartz et al., 19951. The Golgi-like localization for the kinesin epitope identified with antibody H2 may be a distinct pool of motor protein destined for other transport processes, such as secretion, and a Golgi accumulation could represent the storage depot for this activity. Whether other forms of bona fide kinesin (such as that required for ER distribution?) are also present but undetected with current antibody probes is an interesting problem. However, a soluble pool of kinesin, or dynein, detected as diffuse cytoplasmic staining, would allow for local recruitment of functional motor proteins when and where needed throughout the cell. Localization of motor proteins in fixed cells and tissues provides only a static picture of their cellular distribution. Results of immunofluorescence microscopic localization of cytoplasmic dynein and kinesin under control conditions addressed above support our contention that motor molecules accumulate on transport-inactive organelles which represent intermediates or endpoints in the translocation pathway. To address the correlation between motor protein localization and translocation activity further, we immunolocalized kinesin and cytoplasmic dynein following a variety of treatment conditions known to affect microtubule dependent vesicle trafficking pathways. Our results clearly indicate that cytoplasmic dynein and kinesin remain organelle-associated in the absence of active transport, such as following microtubule depolymerization, and changes in motor protein distribution observed under certain treatment conditions do not correlate with expected effects on microtubule-dependent vesicle motility. Cytoplasmic acidification inhibits endocytosis , and it has been suggested that cytoplasmic pH may regulate the activity of motor molecules, with acidification inhibiting dynein and activating kinesin [Heuser, 1989b; Parton et al., 19911. Upon cytoplasmic acidification, perinuclear vesicle association and peripheral staining were found using antibody H1 to localize kinesin, whereas the staining patterns for cytoplasmic dynein and kinesin epitope H2 were less drastically affected. Whether the kinesin antibody H1 pattern is due to the accumulation of pre-Golgi vesicles at the plasma membrane or binding of kinesin to a different cellular organelle is not yet known. The lack of a microtubule or ATP requirement for this accumulation suggests that it does not result from enhanced plus-end motor activity. Okadaic acid leads to an increase in the soluble pool of cytoplasmic dynein, resulting in a diffuse immu- nofluorescent staining pattern [Lin et al., 19941. A similar change in staining pattern was found in the present study for kinesin localized using either antibody H1 or H2. What is the significance of the soluble pools of motor protein and the observation that that dynein and kinesin can be “forced” into this compartment under certain conditions? Does the normal regulation of vesicle transport involve recruitment of motors from this soluble pool? Evidence for a direct effect of okadaic acid on dynein phosphorylation via inhibition of protein phosphatase 1 has been obtained [Lin et al., 19941, and protein phosphatase 1 has also been implicated in regulating motor activity of cytoplasmic dynein [Allan, 19951. We do not as yet have direct evidence for changes in phosphorylation of the kinesin molecule under our conditions. However, phosphorylation of kinesin [Sato-Yoshitake et al., 1992; Hollenbeck, 1993; Matthies et al., 1993; Lee and Hollenbeck, 19951 or kinesin associated polypeptides [McIlvain et al., 19941 has been implicated in the regulation of kinesin-organelle association as well as modulation of lunesin motor activity. Under the treatment conditions described in this study, as well as in control cells, the distribution of the most concentrated populations of motor protein are not representative of organelles undergoing active transport. It will be difficult to address the relationship between motor activity and localization further by immunochemical studies, even at the level of electron microscopy, if indeed a single or very few motor molecules are capable of mediating organelle translocation. The same problem arises in studies utilizing biochemical fractionation to localize motor proteins, in that organelle binding status does not provide information concerning past or present motility activity, and the presence of one or both motor proteins in an organelle population does not provide sufficient evidence of transport in one or the other direction. Is there then anything that can be learned from determining the localization of motor proteins? The results from this study indicate that kinesin and dynein are localized to largely distinct organelles within the cell, and that their sites of binding/accumulation are differentially affected by treatments that perturb trafficking through the organelle populations to which they are bound. Phosphorylation may be involved in the modulation of both dynein and kinesin distribution, and changes in the integrity or transport activity of the Golgi apparatus may alter the association of at least some population of kinesin. These results demonstrate that monitoring changes in the intracellular distribution of kinesin and dynein may be used as a means of investigating the nature of regulatory mechanisms controlling microtubule motor association with intracellular compartments. How this binding activity relates to motility will require the further development of in vivo activity assays for both kinesin Dynein and Kinesin Localization and dynein. Such an assessment may also require the development of more specific probes which are able to recognize an activated or otherwise functional state of the motor enzyme. Such probes have been useful in determining the sites of action of phosphorylated myosin I species in Acunthumoeba, for example [Baines et al., 19951, and further analysis of microtubule motor function in intact cells awaits similar technical advances. ACKNOWLEDGMENTS We thank Kristina Ferro for technical assistance, and members of the Pfister and Collins labs and Dr. Sam Green for helpful discussions and critical reading of the manuscript. This work was supported by grants from the American Cancer Society (CAC) and NIH and the Jeffress Memorial Trust (KKP). REFERENCES Allan, V. (1995): Protein phosphatase 1 regulates the cytoplasmic dynein-driven formation of endoplasmic reticulum networks in vitro. J. Cell Biol. 128:879-891. Aniento, F., Emans, N., Griffiths, G . , and Gruenberg, G. (1993): Cytoplasmic dynein-dependent vesicular transport from early to late endosomes. J. Cell Biol. 123:1373-1387. Baines, I.C., Corigliano-Murphy, A,, and Korn, E.D. (1995): Quantification and localization of phosphorylated myosin I isoforms in Acantharnoebu casteltunii. J. Cell Biol. 130:591-603. Bomsel, M., Parton, R., Kuznetsov, S.A., Schroer, T.A., and Gruenberg, J. (1990): Microtubule and motor-dependent fusion in vitro between apical and basolateral endocytic vesicles from MDCK cells. Cell 62:719-731. Brady, S.T., Pfister, K.K., and Bloom, G.S. (1990): A monoclonal antibody against kinesin inhibits both anterograde and retrograde fast axonal transport in squid axoplasm. Proc. Natl. Acad. Sci. U.S.A. 87:1061-1065. Burkhardt, J.K., McIlvain, J.M., Jr., Sheetz, M.P., and Argon, Y. (1993): Lytic granules from cytotoxic T cells exhibit kinesindependent motility on microtubules in vitro. J. Cell Sci. 104: 151-162. Cohen, P., Holmes, C.F.B., and Tsukitani, Y. (1990): Okadaic acid: a new probe for the study of cellular regulation. Trends Biochem. Sci. 15:98-102. Corthesy-Theulaz, I., Pauloin, A., and Pfeffer, S.R. (1992): Cytoplasmic dynein participates in the centrosomal localization of the Golgi complex. J. Cell Biol. 118:1333-1345. Cosson, P., de Curtis, I., Pouyssegur, J., Griffiths, G., and Davoust, J. (1989): Low cytoplasmic pH inhibits endocytosis and transport from the trans-Golgi network to the cell surface. J. Cell Biol. 108:377-387. Dabora, S.L., and Sheetz, M.P. (1988): The microtubule-dependent formation of a tubulovesicular network with characteristics of ER from cultured cell extracts. Cell 54:27-35. Dahlstrom, A.B., Pfkter, K.K., and Brady, S.T. (1991): The axonal transport motor “kinesin” is bound to anterogradely transported organelles: quantitative cytofluorometric studies of fast axonal transport in the rat. Acta Physiol. Scand. 141:469-476. Dillman, J.F., 111, and Pfister, K.K. (1994): Differential phosphory- 31 1 lation in vivo of cytoplasmic dynein associated with anterogradely moving organelles. J. Cell Biol. 127:1671-1681. Fath, K.R., Trimbur, G.M., and Burgess, D.R. (1994): Molecular motors are differentially distributed on Golgi membranes from polarized epithelial cells. J. Cell Biol. 126:661-175. Feiguin, F., Ferreira, A., Kosik, K.S., and Caceres, A. (1994): Kinesin-mediated organelle translocation revealed by specific cellular manipulations. J. Cell Biol. 127:1021-1039. Ferreira, A., Niclas, J., Vale, R.D., Banker, G . , and Kosik, K.S. (1992): Suppression of kinesin expression in cultured hippocampal neurons using antisense oligonucleotides. J. Cell Biol. 117~595-606. Goslin, K., and Banker, G.A. (1991): Rat hippocampal neurons in low-density culture. In Banker, G.A., and Goslin, K. (eds.): “Culturing Nerve Cells.” Cambridge, Mass.: MIT Press, pp. 25 1-282. Henson, J.H., Nesbitt, D., Wright, B.D., and Scholey, J.M. (1992): Immunolocalization of kinesin in sea urchin coelomocytes. J. Cell Sci. 103:309-320. Heuser, J. (1989a): Changes in lysosome shape and distribution correlated with changes in cytoplasmic pH. J. Cell Biol. 108:855864. Heuser, J. (1989b): Effects of cytoplasmic acidification on clathrin lattice morphology. J. Cell Biol. 108:401-411. Hirokawa, N., Sato-Yoshitake, R., Yoshida, T., and Kawashima, T. (1990): Brain dynein (MAP 1C) localizes on both anterogradely and retrogradely transported membranous organelles in vivo. J. Cell Biol. 111:1027-1037. Hirokawa, N., Sato-Yoshitake, R., Kobayashi, N., Pfister, K.K., Bloom, G . S . , and Brady, S.T. (1991): Kinesin associates with anterogradely transported membranous organelles in vivo. J. Cell Biol. 114:295-302. Hollenbeck, P. J. (1993): Phosphorylation of neuronal kinesin heavy and light chains in vivo. J . Neurochem. 60:2265-2275. Hollenbeck, P.J., and Swanson, J.A. (1990): Radial extension of macrophage tubular lysosomes supported by kinesin. Nature 346: 864-866. Houliston, E., and Elinson, R.P. (1991): Evidence for the involvement of microtubules, ER and kinesin in the cortical rotation of fertilized frog eggs. J. Cell Biol. 114:1017-1028. Johnson, K.J., Hall, E.S., and Boekelheide, K. (1993): Kinesin localizes to the trans-Golgi network of epithelial cells. Mol. Biol. Cell. 4:274a. Koonce, M.P., and McIntosh, J.R. (1990): Identification and immunolocalization of cytoplasmic dynein in Dictyosteliurn. Cell Motil. Cytoskel. 1 5 51-62. Lafont, F., Burkhardt, J.K., and Simons, K. (1994): Involvement of microtubule motors in basolateral and apical transport in kidney cells. Nature 372:801-803. Lee, K.-D., and Hollenbeck, P.J. (1995): Phosphorylation of kinesin in vivo correlates with organelle association and neurite outgrowth. J. Biol. Chem. 270:5600-5605. Leopold, P.L., McDowall, A. W . , Pfister, K.K . , Bloom, G.S ., and Brady, S.T. (1992): Association of kinesin with characterized membrane-bounded organelles. Cell Motil. Cytoskel. 23: 1933. Lin, S.X.H., and Collins, C.A. (1992): Immunolocalization of cytoplasmic dynein to lysosomes in cultured cells. J. Cell Sci. 101~125-137. Lin, S.X.H., and Collins, C.A. (1993): Regulation of the intracellular distribution of cytoplasmic dynein by serum factors and calcium. J. Cell Sci. 103579-588. Lin, S.X.H., Ferro, K.L., and Collins, C.A. (1994): Cytoplasmic dynein undergoes intracellular redistribution concomitant with 312 Lin et al. phosphorylation of the heavy chain in response to serum starvation and okadaic acid. J. Cell Biol. 127:1009-1019. Lippincott-Schwartz, J., Donaldson, J.F., Schweizer, A,, Berger, E.G., Hauri, H.P., Yuan, L.C., and Klausner, R.D. (1990): Microtubule-dependent retrograde transport of proteins into the ER in the presence of brefeldin A suggest an ER recycling pathway. Cell 60:821-836. Lippincott-Schwartz, J., Cole, N. Marotta, A., Conrad, P.A., and Bloom, G.S. (1995): Kinesin is the motor for microtubulemediated transport from the Golgi complex to the ER. J. Cell Biol. 128:293-306. Marks, D.L., Larkin, J.M., and McNiven, M.A. (1994):Association of kinesin with the Golgi apparatus in rat hepatocytes. J. Cell Sci. 107:2417-1426. Matteoni, R., and Kreis, T.E. (1987):Translocation and clustering of endosomes and lysosomes depends on microtubules. J. Cell Biol . 105:1253-1 255. Matthies, H.J.G., Miller, R.J., and Palfrey, H.C. (1993): Calmodulin binding to and CAMP-dependent phosphorylation of kinesin light chains modulate kinesin ATPase activity. J. Biol. Chem. 268111 176-1 1187. McIlvain, J.M., Jr., Burkhardt, J.K., Hamm-Alvarez, S . , Argon, Y., and Sheetz, M.P. (1994): Regulation of kinesin activity by phosphorylation of kinesin-associated proteins. J. Biol. Chem. 269:19176-19182. Murphy, D.B., Rho, M., Schmitz, F., and Wallis, K.T. (1991): Kinesin binds specifically to the Golgi apparatus in a variety of cell types. J. Cell Biol. 115:391a. Oda, H., Stockert, R.J., Collins, C.A., Wang, H., Novikoff, P.M., Satir, P., and Wolkoff, A.W. (1995):Interaction of the microtubule cytoskeleton with endocytic vesicles and cytoplasmic dynein in cultured rat hepatocytes. J. Biol. Chem. 270:1524215249. Parton, R.G., Dotti, C.G., Bacallao, R., Kurtz, I., Simons, K . , and Prydz, K. (1991): pH-induced microtubule dependent redistri- bution of late endosomes in neuronal and epithelial cells. J. Cell Biol. 113:261-274. Pfm,C.M., Coue, M., Grissom, P.M., Hays, T.S., Porter, M.E., and McIntosh, J.R. (1990): Cytoplasmic dynein is localized to kinetochores during mitosis. Nature 345:263-265. Pfister, K.K., Wagner, M.C., Stenoien, D.L., Brady, S.T., and Bloom, G.S. (1989): Monoclonal antibodies to kinesin heavy and light chains stain vesicle-like structures, but not microtubules, in cultured cells. J. Cell Biol. 108:1453-1463. Rodionov, V.I., Gyoeva, F.K., and Gelfand, V.I. (1991): Kinesin is responsible for centrifugal movement of pigment granules in melanophores. Roc. Natl. Acad. Sci. U.S.A. 88:4956-4960. Rogalski, A.A., and Singer, S.J. (1984): Associations of elements of the Golgi apparatus with microtubules. J. Cell Biol. 99:10921 loo. Sato-Yoshitake, R., Yorifuji, H., Inagaki, M., and Hirokawa, N. (1992): The phosphorylation of kinesin regulates its binding to synaptic vesicles. J. Biol. Chem. 267:23930-23936. Saxton, W.M., Hicks, J., Goldstein, L.S., and Raff, E.C. (1991): Kinesin heavy chain is essential for viability and neuromuscular functions in Drosophilu, but mutants show no defects in mitosis. Cell 64:1093-1102. Schmitz, F., Wallis, K.T., Rho, M., Drenckhahn, D., and Murphy, D.B. (1994):Intracellular distribution of kinesin in chromaffin cells. Eur. J. Cell Biol. 63:77-83. Terasaki, M., Chen, L.B., and Fujiwara, K. (1986):Microtubules and the endoplasmic reticulum are highly interdependent structures. J. Cell Biol. 103:1557-1568. Urmtia, R., McNiven, M.A., Albanesi, J.P., Murphy, D.B., and Kachar, B. (1991): Purified kinesin promotes vesicle motility and induces active sliding between microtubules in vitro. Proc. Natl. Acad. Sci. U.S.A. 88:6701-6705. Vale, R.D., and Hotani, H. (1988):Formation of membrane networks in vitro by kinesin-driven microtubule movement. J. Cell Biol. 107:2233-2241.