A monoclonal antibody against kinesin inhibits both anterograde and ...

Proc. Nati. Acad. Sci. USA Vol. 87, pp. 1061-1065, February 1990

Neurobiology

A monoclonal antibody against kinesin inhibits both anterograde and retrograde fast axonal transport in squid axoplasm (fast axonal transport/microtubules/cytoplasmic dynein)

SCOrr T. BRADY*t*, K. KEVIN PFISTER*, AND GEORGE S. BLOOM*t *Department of Cell Biology and Neuroscience, University of Texas Southwestern Medical Center, 5323 Harry Hines Boulevard, Dallas, TX 75235-9039; and tMarine Biological Laboratory, Woods Hole, MA 02543 Communicated by Keith R. Porter, June 21, 1989

Since discovery of the mechanochemical ATPase kinesin (1, 2), many different physiological functions have been suggested. While kinesins have been proposed to mediate translocation of membrane-bounded organelles (MBOs), cytoskeletal and chromosome movements, secretion, and reorganization of the endoplasmic reticulum (3), these suggestions have remained at the level of speculation. The ability of kinesin to mediate gliding of microtubules (MTs) across a glass coverslip (2, 4) indicated that kinesin could do work, but the physiological correlate of this phenomenon was uncertain. The most widely accepted role for kinesin has been as the motor for fast anterograde axonal transport. However, despite intensive study in a number of laboratories, the evidence for involvement of kinesin in fast transport or in other forms of in vivo motility has remained indirect and correlative. Direct evidence for involvement of kinesin in cellular motility could take several forms. One obvious approach would be reconstitution of the physiological process with well-characterized constituents. For MBO translocation along MTs, suitable purified constituents would consist of

MTs assembled without MT-associated proteins, a wellcharacterized fraction of MBOs, and highly purified kinesin. Alternatively, direct evidence could be provided by suitable immunochemical probes that inhibit kinesin function. Appropriate antibodies could interfere with kinesin function by inhibiting ATPase activity, binding of kinesin to either MTs or MBOs, or conformational changes associated with translocation. A similar approach utilizing myosin-specific antibodies provided direct evidence for involvement of myosin in cell morphology, locomotion, and cytokinesis, but not chromosome movement (5-7). While kinesin has been purified to >90% homogeneity (8) and shown to be a MT-stimulated ATPase (1, 8, 9), the goal of recombining kinesin with MBOs remains elusive. Recombination studies have proved to be technically difficult, and little has been known about interactions of kinesin with MBOs. Surprisingly, direct evidence that kinesin is associated with MBOs in vivo only recently has been obtained (10). Initially, evidence for kinesin association with MBOs was inferred from observations that the nonhydrolyzable analogue of ATP, 5'-adenylyl imidodiphosphate (p[NH]ppA; also called AMP-PNP), caused both vesicles (11) and kinesin (1, 2) to bind to MTs. Direct evidence for kinesin association with MBOs has now been obtained by an immunological approach using our library of monoclonal antibodies to heavy and light chains of bovine kinesin (10). Unlike previously described antibodies to the kinesin heavy chain (12-16), which provided conflicting images for subcellular localization of kinesin, immunofluorescence studies with our library of five monoclonal antibodies established kinesin association with MBOs. All five antibodies (to heavy and light chains) produce similar, detergent-sensitive, punctate immunofluorescence patterns identified as MBOs in a variety of vertebrate cell types (10). Using the same antibodies, we have shown kinesin to copurify with MBOs (unpublished observations). Isolated axoplasm from squid giant axons has proved to be a powerful model for study of the molecular mechanisms of fast axonal transport (17, 18). The absence of a plasma membrane permits introduction of defined amounts of proteins and antibodies that do not readily enter intact cells or axons, while retaining the native organization and concentration of cytoplasmic components (18-20). One of our antibodies to bovine kinesin heavy chain (H2) crossreacted with squid kinesin heavy chain, raising the possibility that H2 could be used as a probe of both the distribution and physiological function of kinesin in isolated axoplasm. The amount of kinesin in axoplasm was determined from quantitative immunoblots with H2. Immunofluorescent studies of

The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Abbreviations: MBO, membrane-bounded organelle; MT, microtubule; p[NH]ppA (AMP-PNP), 5'-adenylyl imidodiphosphate. 1To whom reprint requests should be addressed.

One of our monoclonal antibodies against the ABSTRACT heavy chain of bovine kinesin (H2) also recognized the heavy chain of squid kinesin. The immunofluorescence pattern of H2 in axoplasm was similar to that seen in mammalian cells with antibodies specific for kinesin light and heavy chains, indicating that squid kinesin is also concentrated on membranebounded organeiles. Although kinesin is assumed to be a motor for translocation of membrane-bounded organelles in fast axonal transport, direct evidence has been lacking. Perfusion of axoplasm with purified H2 at 0.1-0.4 mg/ml resulted in a profound inhibition of both the rates and number of organelles moving in anterograde and retrograde directions in the interior of the axoplasm, and comparable inhibition was noted in bidirectional movement along individual microtubules at the periphery. Maximal inhibition developed over 30-60 min. Perfusion with higher concentrations of H2 (>1 mg of IgG per ml) were less effective, whereas perfusion with 0.04 mg of H2 per ml resulted in minimal inhibition. Movement of membranebounded organelles after perfusion with comparable levels of irrelevant mouse IgG (0.04 to >1 mg/ml) were not distinguishable from perfusion with buffer controls. Inhibition offast axonal transport by an antibody specific for kinesin provides direct evidence that kinesin is involved in the translocation of membrane-bounded organelles in axons. Moreover, the inhibition of bidirectional axonal transport by H2 raises the possibility that kinesin may play some role in both anterograde and retrograde axonal transport.

1061

1062

Neurobiology: Brady et al.

Proc. Natl. Acad. Sci. USA 87 (1990)

squid axoplasm resulted in punctate patterns homologous with those seen in mammalian cells. Perfusion of H2 into axoplasm inhibited the rate and number of particles moving in fast axonal transport. Surprisingly, both anterograde and retrograde transport were inhibited to a similar extent, in contrast with in vitro evidence that kinesin is involved only in anterograde transport (15). Collectively, these observations represent direct evidence that kinesin is involved in the translocation of MBOs in neurons.

I-j 2

P

sulfate/polyacrylamide gel electrophoresis (SDS/PAGE)

was run as described (22). Protein was determined by the Bradford assay (23). Samples for immunofluorescence were fixed in 5% acetic acid/95% ethanol for 10 min at -200C and then rinsed in Tris-buffered saline (20 mM Tris HCl, pH 7.4/150 mM NaCl) before incubation with primary antibody. Rhodamine-labeled goat antimouse IgG (Fisher Scientific) was the second antibody, and p-phenylenediamine was used in the mounting medium to reduce photobleaching (24). Chemicals and reagents were obtained from Sigma unless specified otherwise.

RESULTS A library of mouse monoclonal antibodies raised against bovine brain kinesin (10) were checked for cross-reactivity with kinesin-containing MT pellets prepared by using p[NH]ppA and squid optic lobe extracts. One antibody to bovine kinesin heavy chain, H2, recognized squid kinesin heavy chain in p[NH]ppA pellets (Fig. 1 Left). To determine specificity of H2 in squid, immunoblots of unfractionated axoplasm were screened. The H2 antibody recognized a single polypeptide in squid axoplasm corresponding to squid kinesin heavy chain (Fig. 1 Right). H2 had the widest

I 1

I.

METHODS AND MATERIALS Axoplasm and Video Microscopy. Isolated axoplasm was extruded from giant axons of the squid Loligo pealed (Marine Biological Laboratory, Woods Hole, MA). Preparations for video microscopy were as described (17, 18). Axons used for video microscopy were 400-600 ,Am in diameter, and extrusions typically yielded =5 1.l of axoplasm. Actual volumes of axoplasm in each experiment were calculated from axoplasm sizes, and 4 volumes of antibody were perfused into the chamber. Video microscopy used a Zeiss Axiomat with a x 100, 1.3 n.a. objective and differential interference contrast optics, with the zoom lens set for a screen width of 23.8 ,um. The video processing system was a Photonics Microscopy C-1966 and Chalnicon camera. Velocity was measured with a Photonics Microscopy C2117 videomanipulator by matching the speed of cursors moving across the screen to the particle movements. All procedures were conducted at room temperature unless indicated otherwise. Immunological Studies. A full characterization of the antibodies has been published (10, 21). H2 is an IgG2b monoclonal antibody that recognizes the full range of bovine kinesin heavy chain isoforms. Homogeneous IgG2b was purified from ascites fluid by using (i) protein A-Superose chromatography (Pharmacia) (10) or (it) the MASS protein A-based filter system (Nygene; Yonkers, NY). Before perfusion, purified IgG was dialyzed against buffer X2 [10 mM Hepes, pH 7.2/150 mM potassium aspartate/5 mM MgCl2/1 mM EGTA]. Immunoblots were visualized by peroxidaselabeled goat anti-mouse IgG (Cappel Laboratories) second antibody with 4-chloro-1-naphthol as the chromogen (10). Some immunoblots were scanned with an LKB laser scanning densitometer to quantitate the kinesin. A serial dilution of purified kinesin was run on the same gel. All lanes were coincubated with primary and then secondary antibodies and were batch-developed to visualize kinesin. Sodium dodecyl

AV\

-.m,

.ii :

:.

T ....

FIG. 1. Recognition of squid kinesin heavy chain by an antibody raised against bovine kinesin heavy chain. (Left) Lane P shows polypeptide composition of a 1 mM p[NH]ppA pellet obtained by centrifugation of a squid optic lobe extract and taxol-stabilized MTs in SDS/PAGE. The Coomassie blue pattern indicates the presence of squid kinesin heavy chain (120 kDa). In immunoblots of a similar gel lane (H2), a single band corresponding in molecular mass to squid kinesin heavy chain was recognized in the squid p[NH]ppA pellet by H2, a monoclonal antibody against bovine kinesin heavy chain. The positions of tubulin (T) and a series of molecular mass standards (205, 116, 97, 66, 45, and 29 kDa) are indicated. (Right) To confirm specificity of H2 in squid, axoplasmic protein was immunoblotted. Lane AX shows the Coomassie blue-stained pattern for 16 ,ug of axoplasmic protein, and lane H2 is the immunoblot of a lane containing 80 ,g of axoplasmic protein probed with H2 antibody. A single band corresponding in molecular mass to squid kinesin heavy chain is seen in the immunoblot of axoplasm. Similar immunoblots were scanned by laser densitometry to determine the concentration of kinesin in squid axoplasm.

cross-reactivity with kinesins from different sources and appeared to recognize the full complement of isoforms for bovine kinesin heavy chain (8, 10). The concentration of kinesin in squid axoplasm was determined by laser densitometry of quantitative immunoblots. Defined amounts of axoplasm were probed with H2 on blots that contained dilutions of bovine kinesin covering the linear range for densitometry. Quantitative immunoblots gave a kinesin concentration in squid axoplasm of =0.5 uM. By comparison, tubulin concentration in axoplasm is -4 /.M (25), giving a molar ratio of 1:8 for kinesin to tubulin in squid axoplasm. Immunocytochemical localization of kinesin in squid axoplasm by indirect immunofluorescence with H2 (Fig. 2) gave a pattern similar to that seen in mammalian cells with our antibodies specific for kinesin light and heavy chains. The kinesin-positive structures form a characteristically punctate pattern (Fig. 2A). No specific staining was seen either with control IgG (Fig. 2B) or with H3, an antibody that does not recognize squid kinesin in immunoblots (not shown). Immunoreactive structures often formed linear arrays similar to



1063

FIG. 2. Subcellular location of kinesin in squid axoplasm by indirect immunofluorescence with H2. (A) A punctate pattern of fluorescence characteristically obtained with H2 in isolated axoplasm, comparable in appearance to the pattern seen in mammalian tissue culture cells (10). Punctate structures often appear in linear arrays that are thought to represent MBOs attached to MTs. (B) A similar region of isolated axoplasm, processed and photographed as in A, with purified control IgG as the primary antibody. No specific staining of axoplasm was obtained either with control IgG or with H3, a monoclonal antibody to bovine kinesin that did not cross-react with squid kinesin. was

in tissue culture cells. Double-label studies indiare aligned along MTs (10). We assume, therefore, that kinesin-positive structures in axoplasm are MBOs aligned along MTs. As in vertebrate cells, immunocytochemical studies of squid axoplasm indicated that kinesin is associated with MBOs. The effects of H2 on fast axonal transport were evaluated by perfusion of axoplasm. Concentrations of H2 between 0.1 and 0.6 mg/ml resulted in a profound inhibition in the rates and number of MBOs moving in both anterograde and retrograde directions (see Table 1 and Fig. 3). Inhibition of bidirectional transport was comparable at the axoplasm periphery along individual MTs and in the interior of the axoplasm. In both regions, inhibition by H2 had a characteristic time course (Fig. 3). An initial =25% reduction in velocity was seen in the first 10 min, followed by a second 30to 60-min phase of increasing inhibition. At 60 min and optimal antibody concentrations, velocities were typically reduced by 60-70%o. The extent and timing of inhibition were similar but not identical in the axoplasm interior and on isolated peripheral MTs; reductions in velocity became apparent somewhat later in the interior rather than on the periphery, presumably due to the time required for H2 diffusion into axoplasm (18, 20). Inhibition was maximal by 30-60 min throughout the preparation. Optimal inhibition by H2 occurred at molar ratios of antibody/kinesin of -1-5, assuming 0.5 ,uM kinesin. Perfusion with higher H2 concentrations (7-10 ,uM IgG) were less effective at reducing rates, while perfusion with H2 < 0.3 ,uM was ineffective. No effect on fast axonal transport was seen for similar concentrations of control IgG (Table 1 and Fig. 3). The biochemical basis of H2 inhibition was evaluated by in vitro assay systems for MT-stimulated ATPase of kinesin and nucleotide-sensitive binding of kinesin to MTs. Molar ratios of H2 to kinesin from 1 to 5 did not inhibit the MT-stimulated ATPase (M. Wagner, S.T.B., K.K.P., and G.S.B., unpublished data), although higher molar ratios (60- to 110-fold excess of antibody) partially inhibited ATPase activity. Similarly, kinesin-H2 complexes bound to MTs in the presence of p[NH]ppA, and H2-kinesin complexes could be used to decorate MTs in vitro (data not shown). These results are consistent with video microscopy in which MBOs slowed or stopped but were not displaced from MTs. The stoichiometry of H2 to kinesin required for effective inhibition and the time course of inhibition suggest that H2-crosslinked adjacent

those

seen

cate that linear arrays

kinesins on the MBO surface, analogous to immunoprecipitation, were responsible for fast axonal transport inhibition.

DISCUSSION Discovery of the kinesins, a family of mechanochemical ATPases distinct from myosin and dynein, represented a beginning for understanding molecular mechanisms of fast Table 1. Effects of antibody perfusion on velocity Transport velocity, ,um/sec Velocity RetroExp., on fibrils,* AnteroAntibody gradet gradet no. /Lm/sec (mg/ml) 11 0.74 ± 0.23 0.78 ± 0.16 0.78 ± 0.16 H2 (0.1-0.6) (n =6) (n =6) (n =35) 0.9 1.19 2 1.28 ± 0.04 H2 (>1) (n= 1) (n =1) (n =3) 3 1.56 ± 0.06 1.31 ± 0.05 1.16 ± 0.09 H2 (0.04) (n= 3) (n =3) (n =8) 1.19 1.78 6 1.89 ± 0.01 IgG (0.1-1.0) (n 1) (n =1) (n= 12) 2.21 ± 0.02 1.79 ± 0.10 1.09 ± 0.05 Control axoplasm (n 16) (n= 16) (n =9) (- IgG, + 2 mM ATP) (10-30 min) Velocity was measured by matching the rate of calibrated cursor movements to the velocity of multiple particles for each determination. Measurements of velocity on MTs represent values for bidirectional movement on individual MTs at the axoplasm periphery. Anterograde and retrograde moving particles on MTs are combined, because no reliable markers for MT polarity in the periphery are available. When bidirectional transport was seen on a single MT, velocities in both directions were similarly affected. All rates were obtained 40-60 min after perfusion with the antibody solution. *Measurements of velocities on fibrils include data from individual MBOs moving in either direction on MTs in randomly selected fields of view. Velocities indicated were measured for a number of MBOs moving along a MT by two or more individuals (n = number of microtubules for which MBO movement was determined). tVelocities for anterograde and retrograde transport represent average velocities for the population of vesicles moving in each direction in randomly selected fields of view in the axoplasm interior. Each velocity measurement (n = number of velocity measurements) is an average of three to five independent measurements by two or more individuals. This method permits velocity measurements for large numbers of MBOs.

1064



1.500

rb1.000

0. 500

0.000 0

10

1 20

1.

1

30

40

50

60

Min after Perfusion FIG. 3. Time course for inhibition: inhibition of fast axonal on individual MTs in the axoplasm periphery following perfusion with H2. Results of three experiments (o, *, and A) are shown. Aliquots of H2 ranging in concentration from 0.1 to 0.6 mg/ml in buffer X2 were perfused into axoplasm at a H2/kinesin molar ratio of 1-5. Velocities typically were reduced to 75% of control values within the first 10 min. Over the course of 60 min, velocities continued to decline until they were 30-40% of control values. By contrast, velocities at 60 min after perfusion with control IgG (X) were not significantly different from buffer controls. The time course and stoichiometry of H2 inhibition of fast "xonal transport suggest that H2 crosslinks kinesin molecules on the surface of MBOs. Both anterograde and retrograde transport of organelles were affected similarly (see Table 1). transport

axonal transport. Despite a substantial amount of correlative evidence, defining the role of kinesin in intracellular motility proved elusive. Moreover, the discovery of cytoplasmic dyneins (26-29) complicated interpretation of pharmacological assays and in vitro studies implicating kinesin as the motor in fast axonal transport. To understand the physiological roles of kinesin, direct evidence for kinesin involvement in axonal transport or other physiological prqcesses is needed. Such direct evidence has two components- (i) kinesin must have an appropriate subcellular location and (it) molecular (biochemical, pharmacological, or immunochemical) probes specific for kinesin must inhibit the physiological activity. Antibodies specific for kinesin polypeptides have the potential to satisfy both criteria. Our library of monoclonal antibodies (10) constitutes an important set of tools for exploring the role of kinesin in many forms'of cell motility. The specificity of these antibodies, each recognizing a unique epitope, has been demonstrated on both purified kinesin and cellular extracts. Ultrastructural localization of epitopes for these antibodies on kinesin documented specificity and provided details on the molecular architecture of kinesin (21). Immunocytochemical studies with' the antibodies showed kinesin localized on MBOs in both neuronal and nonneuronal cells in culture (10). Kinesin was notably absent from cytoskeletal structures, suggesting that its primary association is with MBOs. The size, distribution, and number of kinesin-positive structures in squid axoplasm (see Fig. 2) and other cell types are consistent with kinesin being localized on transported MBOs. Thus, kinesin has the appropriate cellular and subcellular location to be a motor for fast axonal transport. Once reactivity of H2 with squid kinesin was demonstrated, theeffects of H2 on fast axonal transport in axoplasm could be evaluated. Perfusion of suitable concentrations of H2 into axoplasm inhibited both anterograde and retrograde axonal transport to a similar extent with a similar time course. Thus, studies with H2 satisfy both criteria for direct evidence that kinesin is involved in translocation of MBOs along MTs.

Comparisons between the axoplasm periphery, where transport of MBOs can be followedvalong individual MTs, and the axoplasm interior, where axoplasmic organization was retained, were particularly instructive. The extent of inhibition was similar in both regions, although initial declines in velocity occurred later in the interior than the periphery. By 30-60 min, transport was inhibited throughout the preparation, so inhibition was not limited by restricted access of antibody to the interior. Furthermore, while inhibition in the interior might have been due to nonspecific antibody crosslinks restricting passage of MBOs through the axoplasmic matrix, reduced velocities on isolated MTs at the periphery effectively rule out this possibility. Invariably, both anterograde and retrograde transport were inhibited to a similar extent on individual MTs at the periphery and in the interior (see below). Inhibition of bidirectional transport does not appear to require axoplasmic components other than kinesin, MTs, and MBOs. Although the percent reduction in velocity after H2 treatment is greater for anterograde than retrograde-moving particles (see Table 1), final velocities at maximum inhibition are indistinguishable. Since anterograde and retrograde velocities on individual MTs in the periphery were comparable, retrograde transport rates appear limited by steric constraints imposed by the axoplasmic matrix on the larger, retrogrademoving MBOs (17, 18). This interpretation is consistent with observations that differences between anterograde and retrograde velocities in the interior became less apparent after inhibition by H2 and that transport rates for both directions were comparable on isolated MTs in both H2 and control IgG treatments. While modest differences in effect cannot be excluded, available facts suggest that H2 actions on anterograde and retrograde transport were comparable. Previous studies with other antibodies to kinesin (12, 14, 15) had not inhibited either gliding of MTs or MBO transport along MTs. However, several monoclonal antibodies to sea urchin kinesin were recently described that inhibit MT gliding (30, 31). However, none of these crossreact with preparations suitable for analyzing MBO translocation along MTs. Antibodies to other motors and structural proteins have produced variable results. Antibodies that interfere with myosin activity inhibit cell locomotion (5, 32) and cytokinesis (6, 7) but not translocation of MBOs in most animal cell types. With one exception, antibodies against tubulin have not interfered with MBO transport. Interestingly, an affinity-purified tubulin antibody that inhibited MBO transport reduced bidirectional transport (19). No reports have appeared to date of antibodies against cytoplasmic dyneins suitable for immunocytochemical or functional studies. Thus, little is known at present about the subcellular distribution of cytoplasmic dyneins, and hypotheses regarding the precise roles they play in motility in vivo remain dependent on correlative in vitro studies. Perhaps the best evidence to date for dynein-based organelle transport is in Reticulomyxa (28, 33), which has a dynein-like ATPase but apparently lacks a kinesin homologue. In this system, the dynein-like ATPase may mediate bidirectional MBO movements (33). Several mechanisms of action might be invoked to explain H2 inhibition of fast axonal transport: (i) H2 could inhibit MT-stimulated ATPase, (it) H2 might interfere with kinesin binding to MTs, or (iii) H2 could sterically interfere with mechanochemical movements by crosslinking adjacent kinesins. Since addition of H2 at antibody/kinesin molar ratios of 1-5 had no detectable inhibition on the MT-stimulated ATPase of bovine kinesin and kinesin/H2 complexes still exhibit nucleotide-sensitive binding to MTs, the first two mechanisms are unlikely. The third suggestion is, however, consistent with available experimental evidence. Both the time course for reducing MBO velocities and the incomplete blockade of transport are compatible with a process analogous to kinesin

Neurobiology: Brady et al. immunoprecipitation. This mechanism is plausible, since kinesin contains two heavy chains (22) and two binding sites for H2. Thus, the effect of H2 on fast axonal transport may result from steric constraints on conformational changes in kinesin associated with mechanochemical activity. Inhibition of bidirectional transport by H2 is not consistent with previous suggestions that kinesin was involved only in anterograde transport (15). Earlier suggestions were based on the fact that MT gliding mediated by kinesin was unidirectional (4, 15). Subsequently, cytoplasmic dynein was found to produce MT gliding in the opposite direction, leading to proposals that cytoplasmic dynein was the motor for retrograde transport (27). A number of biochemical and pharmacological differences have been noted between gliding due to kinesin (4, 15) and that due to cytoplasmic dynein (27), including rates of gliding, sensitivity to N-ethylmaleimide, nucleotide specificity, and vanadate sensitivity. However, studies of fast axonal transport do not exhibit these same differences between anterograde and retrograde movements. Pharmacological studies of fast transport in axoplasm using a variety of agents expected to differentiate between kinesinmediated and dynein-mediated transport, including p[NH]ppA (11), erythro-9-[3-(2-hydroxynonyl)]adenine (18), Nethylmaleimide (34), and vanadate (unpublished observations) failed to differentiate clearly between anterograde and retrograde transport in axoplasm. The effects of these agents on MBO transport in axoplasm differ quantitatively and qualitatively from MT gliding because of either kinesin or dynein. Finally, electron microscopic morphometric analyses of cross bridges between MBOs and MTs indicated that cross bridges on anterograde- and retrograde-moving MBOs were indistinguishable (35). Although the hypothesis that kinesin moves MBOs anterogradely and dynein moves them retrogradely remains attractive, other models are also consistent with available evidence. A number of questions must be answered before a specific model can be considered proven. To determine which motors are involved in anterograde and retrograde axonal transport, the identity of kinesinpositive MBOs and the presence of other putative motors must be established. Preliminary experiments aimed at determining whether kinesin is associated with anterogradeand retrograde-moving MBOs suggest that kinesin may be present on MBOs moving in both directions. Similar studies must establish whether cytoplasmic dynein is also associated with MBOs and determine the identity of these MBOs. Until immunological or pharmacological reagents with suitable properties can be obtained, the question of the roles played by kinesin and by cytoplasmic dynein in fast axonal transport will remain uncertain. In summary, the role of kinesin in the intracellular transport of MBOs has been explored by using H2, an antibody specific for kinesin heavy chain. Inhibition of fast axonal transport by H2, when combined with localization of kinesin on MBOs, provides direct evidence that kinesin is involved in axonal transport of MBOs. The fact that bidirectional transport is inhibited by H2 raises the possibility that kinesin may play a role in both anterograde and retrograde axonal transport. Note. Video sequences illustrating the effects of control IgG and H2 antibody on transport of membrane-bounded organelles are to be included in the second video supplement of "Cell Motility and the Cytoskeleton," scheduled for publication in 1989. This work was supported in part by grants from the National Institutes of Health (NS 23320 to S.T.B. and NS23868 to S.T.B. and

Proc. NatL. Acad. Sci. USA 87 (1990)

1065

G.S.B.), from the Welch Foundation (1-1077 to G.S.B. and S.T.B.), from the National Science Foundation Biological Instrumentation Program (DMB-8701164 to G.S.B. and S.T.B.), and by a National Institutes of Health Postdoctoral Fellowship Award (GM 10143 to K.K.P.). 1. 2. 3. 4.

5. 6.

7. 8.

9. 10. 11.

Brady, S. T. (1985) Nature (London) 317, 73-75. Vale, R., Reese, T. & Sheetz, M. (1985) Cell 42, 39-50. Vale, R. (1987) Annu. Rev. Cell Biol. 3, 347-378. Porter, M. E., Scholey, J. M., Stemple, D. L., Vigers, G. P. A., Vale, R. D., Sheetz, M. P. & McIntosh, J. R. (1987) J. Biol. Chem. 262, 2794-2802. Honer, B., Citi, S., Kendrick-Jones, J. & Jockusch, B. M. (1988) J. Cell Biol. 107, 2181-2189. Kiehart, D. P., Mabuchi, I. & Inoue, S. (1982) J. Cell Biol. 94, 165-178. Mabuchi, I. & Okuno, M. (1977) J. Cell Biol. 74, 251-263. Wagner, M. C., Pfister, K. K., Bloom, G. S. & Brady, S. T. (1989) Cell Motil. Cytoskel. 12, 195-215. Kutznetsov, S. & Gelfand, V. (1986) Proc. Natl. Acad. Sci. USA 83, 8530-8534. Pfister, K. K., Wagner, M. C., Stenoien, D. S., Brady, S. T. & Bloom, G. S. (1989) J. Cell Biol. 108, 1453-1463. Lasek, R. J. & Brady, S. T. (1985) Nature (London) 316,

645-647. 12. Neighbors, B. W., Williams, R. C. & McIntosh, J. R. (1988)J. Cell Biol. 106, 1193-1204. 13. Leslie, R. J., Hird, R. B., Wilson, L., McIntosh, J. R. & Scholey, J. M. (1987) Proc. Natl. Acad. Sci. USA 84, 27712775. 14. Scholey, J. M., Porter, M. E.,.Grissom, P. M. & McIntosh, J. R. (1985) Nature (London) 318, 483-486. 15. Vale, R. D., Schnapp, B. J., Mitchison, T., Steuer, E., Reese, T. S. & Sheetz, M. P. (1985) Cell 43, 623-632. 16. Murofushi, J., Ikai, A., Okuhara, K., Kotani, S., Aizawa, H., Kumakura, K. & Sakai, H. (1988) J. Biol. Chem. 263, 1274412750. 17. Brady, S. T., Lasek, R. J. & Allen, R. D. (1982) Science 218, 1129-1131. 18. Brady, S. T., Lasek, R. J. & Allen, R. D. (1985) Cell Motil. 5, 81-101. 19. Johnston, K. T., Brady, S. T., van der Kooy, D. & Connolly, J. A. (1987) Cell Motil. Cytoskel. 8, 155-164. 20. Brady, S. T., Lasek, R. D., Allen, R. D., Yin, H. & Stossel, T. (1984) Nature (London) 310, 56-58. 21. Hirokawa, N., Pfister, K. K., Yorifuji, H., Wagner, M. C., Brady, S. T. & Bloom, G. S. (1989) Cell 56, 867-878. 22. Bloom, G. S., Wagner, M. C., Pfister, K. K. & Brady, S. T. (1988) Biochemistry 27, 3409-3416. 23. Bradford, M. M. (1976) Anal. Biochem. 72, 248-254. 24. Johnson, G. D. & Nogueira Araujo, G. (1981) J. Immunol. Res. 43, 349-350. 25. Morris, J. R. & Lasek, R. J. (1984) J. Cell Biol. 98, 2064-2076. 26. Paschal, B., Shpetner, H. & Vallee, R. (1987) J. Cell Biol. 105, 1273-1282. 27. Paschal, B. M. & Vallee, R. B. (1987) Nature (London) 330, 181-183. 28. Euteneuer, U., Koonce, M. P., Pfister, K. K. & Schliwa, M. (1988) Nature (London) 332, 176-178. 29. Lye, R. J., Porter, M. E., Scholey, J. M. & McIntosh, J. R. (1987) Cell 51, 309-318. 30. Ingold, A. L., Cohn, S. A. & Scholey, J. A. (1989) J. Cell Biol. 107, 2657-2667. 31. Scholey, J. M., Heuser, J., Yang, J. T. & Goldstein, L. S. B. (1989) Nature (London) 338, 355-357. 32. Sinard, J. H. & Pollard, T. D. (1989) Cell Motil. Cytoskel. 12, 42-52. 33. Euteneuer, U., Johnson, K., Koonce, M., McDonald, K., Tong, J. & Schliwa, M. (1989) in Cell Movement, eds. Warner, F. & McIntosh, J. R. (Liss, New York), Vol. 2, pp. 155-167. 34. Pfister, K. K., Wagner, M. C., Bloom, G. S. & Brady, S. T. (1989) Biochemistry 28, 9006-9012. 35. Miller, R. H. & Lasek, R. J. (1985)J. CellBiol. 101, 2181-2193.