Microtubule Transport in the Axon

Peter W. Baas    Department of Neurobiology and Anatomy, MCP Hahnemann University, Philadelphia, Pennsylvania 19129

I Introduction

Microtubules (MTs) form the infrastructure of eukaryotic cells, acting as both architectural elements and railways for the transport of cytoplasmic constituents. To serve these functions, MTs must be organized into a wide variety of configurations, ranging from the bipolar conformation of the mitotic spindle to the dense paraxial arrays that occupy elongated cellular processes. Thus, an important question in cell biology is how different cell types organize their MTs into these various configurations. Traditionally, cell biologists have focused on attachment to structures such as the centrosome as the means by which MTs are organized within the cytoplasm (Brinkley, 1985). The centrosome nucleates a radial array of MTs with their plus ends outward and their minus ends inward. This mechanism can explain how MTs are organized in many simple interphase cell types. However, it cannot explain how MTs are organized into more sophisticated patterns such as those of the mitotic spindle or the elongated processes extended by cells such as neurons and glia. It appears that a variety of different mechanisms contribute to the elaboration of cellular MT arrays. Specific proteins have been identified which regulate the length, distribution, dynamic properties, and polarity orientation of the MTs as well as the manner by which they associate with one another and other structures in the cytoplasm (McNally, 2000; Heald, 2000; Andersen, 2000; Walczak, 2000; Sharp et al., 2000a).

One of the most compelling ideas of how cells organize their MTs was proposed almost three decades ago on the basis of studies on axonal transport in neurons. These studies showed that proteins are actively transported down axons in discrete rate components, thus suggesting that they are conveyed as macromolecular complexes as opposed to individual protein subunits (Lasek, 1988; Baas and Brown, 1997; Brown, 2000). On this basis, it was suggested that tubulin is actively transported down the axon in the form of MT polymers (rather than as free tubulin), and hence that axons must contain “molecular machinery” capable of accomplishing this. It is not difficult to imagine how such machinery might be specialized to transport a MT specifically with either its plus end or its minus end leading and thereby establish distinct patterns of MT polarity orientation (Baas and Ahmad, 1993). There was little discussion in the early work as to whether MT transport is a neuron-specific mechanism or whether it might function across cell types, but during the past decade there has been an explosion of new information relevant to this issue. It is now clear that cells contain enzymes called molecular motor proteins which generate forces that can transport and configure MTs (Vale, 1999). The best studied example of this is the mitotic spindle, which has been shown to require motor-driven transport of MTs for its formation and its transitions from one stage to another (Sharp et al., 1999, 2000b). The transport of MTs has been directly visualized during mitosis (Sharp et al., 2000c) and also within relatively simple interphase cells (Keating et al., 1997; Yvon and Wadsworth, 2000). Today, it is difficult to imagine MT research without live-cell images of MTs moving across video monitors.

The great irony is that although MT transport was first appreciated in axons and is arguably most important in axons given their great length, the issue of whether MTs move has been most controversial in axons. For reasons that are mostly either historical or technical, the view that MTs move down the axon has been repeatedly challenged throughout the years, with strong voices arguing that axonal MTs do not move. During the 1990s a flurry of papers appeared from opposing camps, some of which trumpeted that axonal MTs definitely do not move (Hirokawa et al., 1997) and others of which trumpeted that they definitely do move (Baas and Brown, 1997). Today, it appears that a corner has been turned, and there are only a few voices still arguing that axonal MTs are stationary. Although the end of any good controversy is a sad moment for its participants, this controversy leaves in its wake an exciting future of observational and mechanistic studies on MT transport in the axon. The goals of this article are to discuss the history of the MT transport controversy, to present the strong evidence that MTs are indeed transported down axons, and to explore new frontiers for understanding the mechanisms that orchestrate and regulate MT transport in the axon.

II General Background

A Early History of the Microtubule Transport Controversy

The theory of MT transport in axons was first proposed on the basis of classic axonal transport studies conducted over two decades ago. In these studies, radiolabeled amino acids were introduced into a cluster of cell bodies in which they were rapidly taken up and utilized for protein synthesis (Lasek, 1988; Baas and Brown, 1997; Brown, 2000). The labeled amino acids were presented for a brief period of time, thus creating a pulse of labeled proteins that could then move down the axon with a leading edge and a trailing edge (Fig. 1). After some time, the axon was cut into pieces that could be analyzed for radiolabeled protein content. Using this technique, it was determined that tubulin, the protein that constitutes MTs, is actively transported down the axon in the phase of axonal transport known as the “slow component” at an average rate of about 1 mm per day. The slow component could be further resolved into two subcomponents known as “slow component a” and “slow component b.” Most of the tubulin was found in a, which is the slower moving of the two. The “pulse” of labeled tubulin moved down the axon relatively coherently at first but clearly spread as it moved, indicating a divergence in the rates of movement within the population. Initial considerations focused on the coherence and proposed that tubulin is transported in the form of a highly cross-linked network of MTs. Subsequent considerations took into account the spread and proposed that tubulin is more likely transported in the form of individual MTs that move at a variety of rates. In this modified view the rate of tubulin movement in the slow component represents an average rate of the movement of large numbers of individual MT polymers. In this view, the fastest moving MTs (those at the leading edge) move many times the average rate, at speeds comparable to the “fast transport” of membranous vesicles along MTs. These earlier studies largely predated the so-called “motor revolution”; hence, the movement of MTs was usually discussed in terms of hypothetical “transport machinery” rather than molecular motor proteins.

The first significant challenge to this polymer-transport model came from studies showing that MT poisons stopped axonal growth in culture when applied to the distal tip of the axon but not to the cell body (Bamburg et al., 1986). These studies were enormously influential, but in fact were based on a misconception of MT transport. The axon is full of MTs that are shorter than the length of the axon; being dynamic structures, these MTs can elongate and shorten. The transport of a MT is the actual movement of the polymer through the cytoplasm and is not contingent on the assembly of new polymer. Therefore, the relative sensitivity of axonal growth to MT poisons applied at the cell body...