19:  19.3 Kinesin, Dynein, and Intracellular Transport

Fast Axonal Transport Occurs along Microtubules

As we discuss in Chapter 21, nerve impulses are transmitted from a neuron by release of neurotransmitters from the terminal of the axon, the very long process that extends from the cell body. The neuron must constantly supply new materials — proteins and membranes — to the terminal to replenish those lost by exocytosis at the junction (synapse) with another cell. Where do these new materials originate? Ribosomes are present only in the cell body and dendrites of nerve cells, so no protein synthesis can occur in the axons and synaptic terminals. Therefore, proteins and membranes must be synthesized in the cell body and then transported down the axon, which can be up to several meters in length, to the synaptic regions. This process of axonal transport, first described in 1948, is now known to occur on microtubules. As noted earlier, the microtubules in axons are all oriented with their (+) ends toward the terminal, which is critical to axonal transport (see Figure 19-7c).

The rate at which proteins are transported along axons, and their identity, can be determined by a pulse-chase experiment, as outlined in Figure 19-19. Such experiments commonly are conducted on neurons in the mammalian sciatic nerve because their cell bodies are conveniently located in the dorsal root ganglion near the spinal cord and their nerve axons are very long. Studies like these have shown that axonal transport occurs in both directions. Anterograde transport, as depicted in Figure 19-19a, proceeds from the cell body to the synaptic junctions and is associated with axonal growth and the renewal of synaptic vesicles. In the opposite, retrograde, direction other substances move along the axon rapidly toward the cell body. These substances, which consist mainly of “old” membrane from the synaptic terminals, are destined to be degraded in lysosomes in the cell body.

Transported materials are divided into three groups according to the speed of movement. The fastest-moving material, consisting of membrane-bounded vesicles, has a velocity of about 250 mm/day, or about 3 μm/s. The slowest material, comprising mostly polymerized cytoskeletal proteins, moves only a fraction of a millimeter per day. Organelles such as mitochondria move down the axon at an intermediate rate.

Axonal transport does not require an intact cell. In fact many studies on fast axonal transport are conducted with extruded axoplasm. In this type of preparation, the cytosol is squeezed from the axon with a roller onto a glass coverslip. When such an extract is provided with ATP, the movement of vesicles along microtubules can be observed by video microscopy. The rate of vesicle movement (1 – 2 μm/s) in this cell-free system is similar to that of fast axonal transport in intact cells. Movement may occur in both the anterograde and retrograde directions; in some cases, two organelles can be seen to move along the same fiber in opposite directions and to pass each other without colliding (Figure 19-20a). Electron microscopy of the same region of the axoplasm has revealed that the transport fibers are individual microtubules (Figure 19-20b). These in vitro experiments established that fast axonal transport occurs along microtubules and that the movement requires ATP. As we will discuss shortly, these two observations led to the identification of microtubule motor proteins, which generate the movements.

Microtubules Provide Tracks for the Movement of Pigment Granules

Another system for observing rapid transport along microtubules is provided by the specialized pigment cells — called melanophores — that are found in the skin of many amphibians and on the scales of many fish. Nerves and hormones control the color of the skin by triggering the transport of membrane-enclosed pigment granules throughout the cell, to darken the color of the skin, or inward, toward the center of the cell, to lighten the color (Figure 19-21a, b). In this way, an animal can adjust its color.

The role of microtubules in color adjustment is studied by placing pigment cells in culture. If a melanophorecovered fish scale is placed in a culture medium, the pigment granules in the melanophores will be seen to move inward and outward spontaneously. In cases when individual granules could be followed, it has been observed that after movement to the cell periphery, each granule always returns to its prior location in the center of the cell. During this movement, microtubules serve as tracks along which the pigment granules can move in either direction (Figure 19-21c).

Intracellular Membrane Vesicles Travel along Microtubules

Having seen that membrane vesicles in specialized cells are transported between the cell body and the cell periphery, we can next ask whether membrane vesicles are transported along microtubules in every eukaryotic cell. The preliminary answer is a qualified yes: some types of vesicle transport are dependent on microtubules, though microfilaments may also be involved in some cases. The best-studied system is the intracellular movement of Golgi vesicles. In cultured fibroblasts, the Golgi complex is concentrated near the MTOC. During mitosis (or after the depolymerization of microtubules by colcemid), the Golgi complex breaks into small vesicles that are dispersed throughout the cytosol. When the cytosolic microtubules re-form during interphase (or after removal of the colcemid), the Golgi vesicles move along these microtubule tracks toward the MTOC, where they reaggregate to form large membrane complexes.

In addition to the Golgi complex, microtubules are also associated with the endoplasmic reticulum (ER). Fluorescence microscopy, using anti-tubulin antibodies and DiOC₆, a fluorescent dye specific for the ER, reveals an anastomosing network of tubular membranes in the cytosol that colocalizes with microtubules (Figure 19-22). If microtubules are destroyed by drugs such as nocodazole or colcemid, then the ER loses its network-like organization. After the drug is washed from the cell, tubular fingers of ER grow as new microtubules assemble. In cell-free systems, the ER can be reconstituted with microtubules and an ER-rich cell extract. Even under this cell-free regime, ER membranes elongate along a microtubule. This close association between ER and intact microtubules suggests that certain proteins act to bind ER membranes to microtubules.

A key feature of the in vitro assembly of the ER is that the tubular membranes seem to elongate along microtubules. This outgrowth of the ER membrane is blocked by microtubule inhibitors like colchicine or nocodazole, suggesting that intact microtubules are required for membrane elongation. This movement along microtubules is one mechanism by which the MTOC organizes cell organelles such as the ER and Golgi.

Kinesin Is a (+) End–Directed Microtubule Motor Protein

To unravel the mechanism of axonal transport, cell biologists sought to identify the protein or proteins in neuronal cytosolic extracts that can propel synaptic vesicles along microtubules assembled in vitro from purified tubulin subunits and stabilized by the drug taxol. When synaptic vesicles were added with ATP to these microtubules, the vesicles neither bound to the microtubules nor moved along them. However, the addition of squid nerve axoplasm (free of tubulin) caused the vesicles to bind to the microtubules and to move along them, indicating that a soluble protein in the nerve cytosol is required for translocation.

When researchers incubated vesicles, nerve cytosol (squid axoplasm), and microtubules in the presence of AMPPNP,

a nonhydrolyzable analog of ATP, the vesicles bound tightly to the microtubules but did not move. However, the vesicles did move when ATP was added. These results suggested that a motor protein in the cytosol binds to microtubules in the presence of ATP or AMPPNP, but movement requires hydrolysis of the terminal phosphoanhydride bond of ATP. To purify the soluble motor protein, scientists used AMPPNP to promote its tight binding to microtubules, which were used as an affinity matrix. A mixture of microtubules, brain extract, and AMPPNP was incubated; the microtubules with any bound proteins then were collected by centrifugation. Treatment of the microtubule-rich material in the pellet with ATP released one predominant protein back into solution; this protein was named kinesin.

Kinesin isolated from squid axoplasm is a dimer of two heavy chains, each complexed to a light chain, with a total molecular weight of 380,000. The molecule is organized into three domains, a pair of large globular head domains connected by a long central stalk to a pair of small globular tail domains, which contain the light chains (Figure 19-23a). Each domain carries out a particular function: the head domain, which binds microtubules and ATP, is responsible for the motor activity of kinesin, and the tail domain is responsible for binding to membrane vesicles. In light of the transport function of kinesin, a bound membrane vesicle is often referred to as kinesin’s “cargo.”

Recent x-ray crystallographic analysis of the kinesin head domain has revealed a surprising feature of its structure (Figure 19-23b). The three-dimensional structure of the kinesin head is similar to that of Ras, a guanine nucleotide – binding protein (see Figure 3-5) and to that of the myosin motor domains (see Figure 18-24). At the core of the head domain is a Ras-like fold, which binds ATP. Microtubule-binding sites lie on the surface of the domain in similar regions to the actin-binding loops of myosin. Thus, these motor proteins may have evolved from a common ancestor, becoming specialized to move along a microtubule or a microfilament.

Kinesin-dependent movement of vesicles can be tracked by an in vitro motility assay using microtubules nucleated from isolated centrosomes. In this assay, a vesicle or a plastic bead coated with kinesin is added to a glass slide along with a preparation of centrosome-nucleated microtubules. The kinesin-coated beads bind to the microtubules, and in the presence of ATP, kinesin carries the beads along a microtubule in one direction (Figure 19-24a). By determining the polarity of the microtubules, researchers found that the beads always moved from the (−) to the (+) end of a microtubule. Alternatively, kinesin alone, adsorbed onto a glass coverslip, causes microtubules to glide across the surface, with their (−) ends in the lead (Figure 19-24b); this direction of movement results from kinesin’s walking toward the (+) end of the microtubule. Thus kinesin is a (+) end – directed microtubule motor protein. Because this direction corresponds to anterograde transport, kinesin is implicated by these studies as the motor protein responsible for anterograde and other (+) end – directed movements, such as the transport of secretory vesicles to the plasma membrane and the radial movements of ER membranes and pigment granules.

In vitro motility experiments using an optical trap have determined two fundamental characteristics of the kinesin motor, its step size and force. In these experiments, similar to those performed on myosin (see Figure 18-23), a two-headed kinesin molecule was found to move in 8-nm steps and exert a force of 6 pN. The step size matches the distance between successive α- or β-tubulin monomers in a protofilament, suggesting that kinesin binds only to one or the other monomer. In other experiments, researchers have established that kinesin moves along a single protofilament, like walking a tightrope. However this poses a question. If the microtubule binding sites are separated by 5 nm in the kinesin dimer, then the kinesin structure must “give” during an 8-nm step. How this occurs is unknown.

Each Member of the Kinesin Family Transports a Specific Cargo

Like myosin, kinesin also belongs to a family of related motor proteins. (Table 19-2). To date, more than 12 different family members have been identified; all contain the kinesin motor domain, but they differ in their tail domains and several other properties. In most kinesins, the motor domain is at the N-terminus (N-type), but in some, the motor domain is central (M-type) or at the C-terminus (C-type). Both Nand M-type kinesins are (+) end – directed motors, whereas C-type kinesins are (−) end – directed motors. In addition, some kinesins are monomeric (i.e., have a single heavy chain); as noted earlier; however, most are dimeric. These two types of kinesins, differing in quaternary structure, may move along a microtubule by different mechanisms.

Kinesins can also be divided into two broad functional groups — cytosolic and spindle kinesins — based on the nature of the cargo they transport. The functional differences between kinesins may be related to their unique tail domains. Cytosolic kinesins are involved in vesicle and organelle transport; they include the classic axonal kinesin, implicated in transport of lysosomes and other membranous organelles. Some cytosolic kinesins, however, are responsible for transport of one specific cargo. For example, KIF1B transports mitochondria, and its relative KIF1A transports synaptic vesicles to the nerve terminal. Spindle kinesins, in contrast, participate in spindle assembly and chromosome segregation during cell division. (These are also known as kinesinrelated proteins, or KRP motors.) This group comprises numerous proteins, including the kinetochore-associated protein CENP-E; the spindle pole protein BimC; and a (−) end – directed motor protein called ncd. We discuss the spindle kinesins in more detail in the later section on mitosis.

Dynein Is a (−) End – Directed Motor Protein

Some microtubule-directed movements, such as retrograde axonal transport or the transit of endocytotic vesicles of the plasma membrane to lysosomes, are in the direction opposite to most kinesin-dependent movements. A second group of motor proteins, the dyneins, were found to be responsible for movement toward the (−) end of microtubules. Dyneins are exceptionally large, multimeric proteins, with molecular weights exceeding 1,000,000. They are composed of two or three heavy chains (MW 470,000 – 540,000) complexed with a poorly determined number of intermediate and light chains. As summarized in Table 19-2, the dyneins are divided into two functional classes: cytosolic dynein, which is involved in the movement of vesicles and chromosomes, and axonemal dynein, which is responsible for the beating of cilia and flagella (discussed later).

Dynein-Associated MBPs Tether Cargo to Microtubules

Like kinesin, cytosolic dynein is a two-headed molecule, with two identical or nearly identical heavy chains forming the head domains. However, unlike kinesin, dynein cannot mediate transport by itself. Rather, dynein-related motility requires a large complex of microtubule-binding proteins that link vesicles and chromosomes to microtubules but by themselves do not exert force to cause movement. The best-characterized complex is dynactin, a heterocomplex of at least eight subunits, including a 150,000-MW protein called Glued, the actin-capping protein Arp1, and dynamatin (Figure 19-25). In vitro binding experiments show that dynactin enhances dynein-dependent motility, possibly through interaction with microtubules and vesicles. The microtubule binding site lies in the N-terminal region of Glued. This region contains a 57-residue microtubule-binding motif that also is present in a vesicle-binding protein called CLIP-170 and the yeast protein BIK1. Like dynactin, these proteins are implicated in the trafficking of endocytotic vesicles. The light chains of dynein interact with the dynamatin subunits of dynactin. One model proposes that dynactin tethers a vesicle to a microtubule, while dynein generates the force and polarity for movement.

As we discuss later, several lines of evidence suggest that the dynein-dynactin complex and another complex, the nuclear/mitotic apparatus (NuMA) protein, mediate association of microtubules with chromosomes during mitosis. In vitro studies show that truncated NuMA protein binds microtubules if the C-terminal region is retained. As in assembly MAPs, the C-terminal region of NuMA protein is highly acidic, and ionic interactions may mediate binding to microtubules.

Multiple Motor Proteins Are Associated with Membrane Vesicles

The identification of (+) and (−) end – directed microtubule motor proteins (kinesin and cytosolic dynein, respectively) explains not only how movement of vesicles is powered but also how the direction of movement is controlled (Figure 19-26). The direction of vesicle transport also depends on the orientation of microtubules, which is fixed by the MTOC. Some cargoes, such as pigment granules, can alternate their direction of movement along a single microtubule. In this case, both anterograde and retrograde microtubule motor proteins must be associated with a microtubule. At any one time, however, only one motor protein is active or, alternatively, only one motor protein is bound to the vesicle.

Increasing evidence suggests that transport of some vesicles is more complicated than depicted in Figure 19-26. For example, during endocytosis vesicles from the plasma membrane are carried inward, while during secretion vesicles from the ER and Golgi are moved outward (Chapter 17). In both processes, a vesicle must traverse microtubule-poor but microfilament-rich regions in the cell. Several complementary experiments indicate that microtubule and microfilament motor proteins bind to the same membrane vesicles and cooperate in their transport. One piece of evidence was obtained from microscopy of vesicle movements in extruded squid axoplasm. As observed many times before, membrane vesicles traveled along microtubule tracks; however, at the periphery of the extruded axoplasm, movement continued even though no microtubules were present. This region contained microfilaments, and subsequent experiments demonstrated that a vesicle could move on a microtubule or a microfilament. Thus at least two motor proteins, myosin and either kinesin or dynein, must be bound to the same vesicle. The discovery that a given vesicle can travel along both cytoskeletal systems suggests that in a neuron, synaptic vesicles are transported at a fast rate by kinesin in the microtubule-rich axon and then travel through the actin-rich cortex at the nerve terminal on a myosin motor.

SUMMARY

• Two families of motor proteins, kinesin and dynein, transport membrane-bounded vesicles, proteins, and organelles along microtubules.

• Nearly all kinesins move cargo toward the (+) end of microtubules (anterograde transport), whereas dyneins transport cargo toward the (−) end (retrograde transport).

• The dimeric kinesin head domain binds microtubules and ATP, and the tail domain binds vesicles (see Figure 19-23). Although the structure of the kinesin head is similar to that of myosin, it lacks a rigid neck domain. Thus the model of myosin-dependent motility may not apply to kinesin.

• Each type of membrane vesicle is transported by its own kinesin motor protein. The specificity of binding may reside in the tail domain, which is unique to each kinesin.

• Cytosolic dyneins are linked to their cargoes (vesicles and chromosomes) by large complexes of microtubule-binding proteins (MBPs), such as dynactin (see Figure 19-25). Once tethered to a cargo, the dynein transports it to the final destination.

• Axonal transport is a well-studied model system for understanding anterograde transport by kinesin and retrograde transport by cytosolic dynein in a typical eukaryotic cell (see Figure 19-26).

• In microtubule-poor regions of the cell, vesicles probably are transported along microfilaments powered by a myosin motor.