 |
Before proceeding further in our discussion of microtubule-containing structures and microtubule-based movements, we will take a close look at the assembly, disassembly, and polarity of microtubules. A microtubule can oscillate between growing and shortening phases. This complex dynamic behavior permits the cell to quickly assemble or disassemble microtubule structures. To appreciate fully the function of the microtubule cytoskeleton, we examine the details of microtubule assembly and disassembly, as well as the role of a group of proteins that are integrally associated with microtubules.
(a) At low temperatures, microtubules depolymerize, releasing αβ-tubulin, which repolymerize at higher temperatures in the presence of GTP. (b) The critical concentration (Cc) is the concentration of dimeric αβ-tubulin in equilibrium with microtubules. At dimer concentrations below the Cc, no polymerization occurs. At dimer concentrations above the Cc, tubulin polymerizes into microtubules.
). When warmed to 37 °C in the presence of GTP,
the tubulin dimers polymerize into microtubules.
(a) Fragments of flagellar microtubules act as nuclei for the in vitro addition of αβ-tubulin. The nucleating flagellar fragment can be distinguished in the electron microscope from the newly formed microtubules (MT), seen radiating from the ends of the flagellar fragment. Note that the microtubules at one end are longer than at the other end. (b) Addition and loss of tubulin subunits both occur primarily at one end of a microtubule, the (+) end. [Part (a) courtesy of G. Borisy.]
). Second, the addition
of fragments of flagellar or other microtubules to a solution of
αβ-tubulin accelerates the initial polymerization rate by
acting as nucleation sites. Third, at αβ-tubulin
concentrations higher than the Cc for
polymerization, dimers add to both ends of a growing microtubule, but the
addition of tubulin subunits occurs preferentially at one end. This difference
between the two ends of a growing microtubule is demonstrated by examining
electron micrographs of microtubules that have assembled from the ends of
nucleating flagellar fragments in vitro (Figure
19-10a). The electron micrographs show a tuft of microtubules
sprouting from both ends of the fragment, but one tuft is much longer than the
other. Using the same terminology as in actin assembly, the preferred assembly
end is designated the (+) end and the end that assembles more slowly is
the (−) end. When the tubulin concentration is diluted below the
Cc, the microtubules disassemble twice as
rapidly at the (+) end as at the (−) end. Thus both assembly
and disassembly occur preferentially at the (+) end (Figure 19-10b). The dynamics of
microfilament and microtubule assembly share these three properties.
Free αβ-tubulin dimers associate longitudinally to form short protofilaments (step 1 ). These probably are unstable and quickly associate laterally into more stable curved sheets (step 2 ). Eventually a sheet wraps around into a microtubule with 13 protofilaments. The microtubule then grows by the addition of subunits to the ends of protofilaments composing the microtubule wall (step 3 ). The free tubulin dimers have GTP bound to the exchangeable nucleotide-binding site on the β-tubulin monomer. After incorporation of a dimeric subunit into a microtubule, the GTP on the β-tubulin (but not on the α-tubulin) is hydrolyzed to GDP. If the rate of polymerization is faster than the rate of GTP hydrolysis, then a cap of GTP-bound subunits is generated at the (+) end, although the bulk of β-tubulin in a microtubule will contain GDP. The rate of polymerization is twice as fast at the (+) end as at the (−) end.
).
Microtubules undergoing assembly or disassembly can be quickly frozen in liquid ethane and examined in the frozen state in a cryoelectron microscope. (a) In assembly conditions, microtubule ends are relatively smooth; occasionally a short protofilament is seen to extend from one end. (b) In disassembly conditions, the protofilaments splay at the microtubule ends, giving the ends a frayed appearance. [Micrographs courtesy of E. Mandelkow and E. M. Mandelkow.]
). The
appearance of microtubules undergoing shortening is quite different, suggesting
that the mechanism of disassembly differs from that of assembly (Figure 19-12b). Under shortening
conditions, the microtubule ends are splayed, as if the lateral interactions
between protofilaments have been broken. Once frayed apart and freed from
lateral stabilizing interactions, the protofilaments might depolymerize by
endwise dissociation of tubulin subunits. The splayed appearance of a shortening
microtubule provided clues about the potential instability of a microtubule.
Individual microtubules can be observed in the light microscope, and their lengths can be plotted during stages of assembly and disassembly. Assembly and disassembly each proceed at uniform rates, but there is a large difference between the rate of assembly and that of disassembly, as seen in the different slopes of the lines. During periods of growth, the microtubule elongates at a rate of 1 μm/min. Notice the abrupt transitions to the shrinkage stage (catastrophe) and to the elongation stage (rescue). The microtubule shortens much more rapidly (7 μm/min) than it elongates. [Adapted from P. M. Bayley, K. K. Sharma, and S. R. Martin, 1994, in Microtubules, Wiley-Liss, p. 119.]
). In all cases, the rate of microtubule growth is much
slower than the rate of shortening. When first discovered, this more complex
behavior of microtubules, termed dynamic instability, was
surprising to researchers because they expected that under any condition all the
microtubules in a solution or the same cytosol would behave identically.
Fluorescent-labeled tubulin was microinjected into cultured human fibroblasts. The cells were chilled to depolymerize preexisting microtubules into tubulin dimers and were then incubated at 37 °C to allow repolymerization, thus incorporating the fluorescent tubulin into all the cell’s microtubules. A region of the cell periphery was viewed in the fluorescence microscope at 0 s, 27 s later, and 3 min 51 s later (left to right panels). During this period several microtubules elongate and shorten. The letters mark the position of ends of three microtubules. [From P. J. Sammak and G. Borisy, 1988, Nature 332:724.]
).
Also, within a few minutes, some microtubules appear alternately to grow and
shrink. Since most microtubules in a cell associate by their (−) ends
to MTOCs, we can conclude that their instability is largely limited to the
(+) ends of microtubules.
Dimeric αβ-tubulin subunits with two bound GTP molecules (blue) add preferentially to the (+) end of a preexisting microtubule. After incorporation of a subunit, the GTP bound to the β-tubulin monomer is hydrolyzed to GDP, although the GTP bound to α-tubulin is not. This hydrolysis is apparently catalyzed by the microtubule itself but may be facilitated by cytosolic proteins. Only microtubules whose (+) end is associated with GTP-tubulin (those with a GTP cap) are stable and can serve as primers for polymerization of additional tubulin. Microtubules with GDP-tubulin (red) at the (+) end (those with a GDP cap) are rapidly depolymerized and may disappear within 1 min. At high concentrations of unpolymerized GTP-tubulin, the rate of addition of tubulin is faster than the rate of hydrolysis of the GTP bound in the microtubule or the rate of dissociation of GTP-tubulin from the end; thus the microtubule grows. At low concentrations of unpolymerized GTP-tubulin, the rate of addition of tubulin is decreased; consequently, the rate of GTP hydrolysis exceeds the rate of addition of tubulin subunits and a GDP cap forms. Because the GDP cap is unstable, the microtubule end peels apart to release tubulin subunits. [See T. Mitchison and M. Kirschner, 1984, Nature 312:237; M. Kirschner and T. Mitchison, 1986, Cell 45:329; and R. A. Walker et al., 1988, J. Cell Biol. 107:1437.]
). At concentrations near the
Cc, however, some microtubules grow, while
others shrink. The second condition affecting microtubule stability is whether
GTP or GDP occupies the exchangeable nucleotide-binding site on
β-tubulin at the (+) end of a microtubule (Figure 19-15). A microtubule becomes
unstable and depolymerizes rapidly if the (+) end becomes capped with
subunits containing GDP – β-tubulin
rather than GTP – β-tubulin. This
situation can arise when a microtubule shrinks rapidly, exposing
GDP – β-tubulin in the walls of the
microtubule, or when a microtubule grows so slowly that hydrolysis of GTP bound
to β-tubulin converts it to GDP before additional subunits can be added
to the (+) end of the microtubule. Before a shortening microtubule
vanishes, it can be “rescued” and start to grow if tubulin
subunits with bound GTP add to the (+) end before the bound GTP
hydrolyzes. Thus the one parameter that determines the stability of a
microtubule is the rate at which GTP-tubulin subunits are added to the
(+) end. Possible factors that switch a microtubule between growth and
shrinkage have been identified. One is a microtubule-severing protein, katanin,
which may generate nuclei at centrosomes. Another factor is Op 18, which
increases the frequency of catastrophe, possibly by binding tubulin dimers.
These and other drugs that interfere with normal assembly and disassembly of microtubules have an antimitotic effect that is particularly devastating to rapidly dividing cells, such as cancer cells.
).
Colchicine and a chemical relative, colcemid, have long been used as a mitotic inhibitor. As noted previously, in cells exposed to high concentrations of colcemid, cytosolic micro-tubules depolymerize, leaving an MTOC (see Figure 19-6). However, when plant or animal cells are exposed to low concentrations of colcemid, the microtubules remain and the cells become “blocked” at metaphase, the mitotic stage at which the duplicated chromosomes are fully condensed. When the treated cells are washed with a colcemid-free solution, colcemid diffuses from the cell and mitosis resumes normally. Thus experimenters commonly use colcemid to accumulate metaphase cells for cytogenetic studies; removal of the colcemid leaves a population of cells whose cell cycle is in synchrony. Such synchronous populations are advantageous for studies of the cell cycle (Chapter 13).
Each tubulin dimer has one high-affinity binding site for colchicine. In fact, colchicine binds tubulin irreversibly. This property was used as an early assay for tubulin during the purification of microtubules. Colchicine-bearing tubulin dimers, at concentrations much less than the concentration of free tubulin subunits, can add to the end of a growing microtubule. However, the presence of one or two colchicinebearing tubulins at the end of a microtubule prevents the subsequent addition or loss of other tubulin subunits. Thus colchicine “poisons” the end of a microtubule and alters the steady-state balance between assembly and disassembly. As a result of this disruption of microtubule dynamics, the mitotic spindle does not form in cells treated with low concentrations of colchicine.
Other drugs bind to different sites on tubulin dimers or to microtubules and therefore affect microtubule stability through different mechanisms. For example, at low concentrations, taxol, its chemical derivative taxotere, and vinblastine bind to and stabilize microtubules by inhibiting microtubule dynamics — the lengthening and shortening of microtubules. High concentrations of vinblastine, however, promote depolymerization of microtubules and the assembly of tubulin dimers into nearly crystalline arrays called vinblastine paracrystals.
Drugs that disturb the assembly and
disassembly of microtubules have been widely used to treat various diseases.
Indeed, more than 2500 years ago, the ancient Egyptians treated heart problems
with colchicine. Nowadays, this drug is used primarily in the treatment of gout
and certain other diseases affecting the joints and skin. Other inhibitors of
microtubule dynamics, including taxol and vinblastine, are effective anticancer
agents, since blockage of spindle formation preferentially inhibits rapidly
dividing cells like cancer cells. For instance, taxol treatment of ovarian
cancer cells, which undergo rapid cell divisions, blocks mitosis but does not
affect other functions carried out by microtubules.
Cells were stained with fluorescent- labeled antibodies against tubulin and MAP4 and then viewed in a fluorescence microscope. The colinear arrangement of MAP4 and microtubules, even at the MTOC (arrow), is suggestive of a binding interaction. [Courtesy of J. Chloe-Bulinski.]
). This finding is strong evidence that MAPs
have microtubule-binding activity.
Insect cells transfected with DNA expressing either long-armed MAP2 protein or short-armed Tau protein grow long axonlike processes. Shown here are electron micrographs of cross sections through the processes induced by expression of MAP2 (left) or Tau (right) in transfected cells. Note that the spacing between microtubules in MAP2-containing cells is larger than in Tau-containing cells. Both cell types contain approximately the same number of microtubules, but the effect of MAP2 is to enlarge the caliber of the axonlike process. [From J. Chen et al., 1992, Nature 360:674.]
).
Type II MAPs include MAP2, MAP4, and Tau. These proteins are characterized by the presence of three or four repeats of an 18-residue sequence in the microtubulebinding domain. MAP2 is found only in dendrites, where it forms fibrous cross-bridges between microtubules and also links microtubules to intermediate filaments. MAP4, the most ubiquitous of all the MAPs, is found in neuronal and non-neuronal cells. As discussed later, MAP4 is thought to regulate microtubule stability during mitosis. Tau, which is much smaller than most other MAPs, is present in both axons and dendrites. This protein exists in four or five forms derived from alternative splicing of a tau mRNA. The ability of Tau to cross-link microtubules into thick bundles may contribute to the stability of axonal microtubules.
Transfection of cultured insect cells with either the tau gene or MAP2 gene induces the growth of microtubule-filled processes. These observations indicate that both Tau and MAP2 accelerate the polymerization of tubulin subunits, as well as contribute to cross-linking of microtubules. The importance of Tau in promoting axon growth was further confirmed by the finding that cultured neurons microinjected with DNA encoding tau antisense RNA are unable to grow an axon. Expression of tau antisense RNA inhibits translation of tau mRNA and thus reduces the intracellular level of Tau (see Figure 11-46). Similar experiments with MAP2 antisense RNA showed that MAP2 is critical to formation of dendrites.
When MAPs coat the outer wall of a microtubule, tubulin subunits are unable to dissociate from the ends of a microtubule. Although the rate of microtubule disassembly is generally dampened by bound MAPs, the assembly of microtubules is affected to varying degrees: some MAPs, like Tau and MAP4, stabilize microtubules, whereas other MAPs do not.
Because of the effect of MAPs on microtubule dynamics, the length of microtubules can be controlled by modulating the binding of MAPs. In most cases, this is accomplished by the reversible phosphorylation of the MAP projection domain. Phosphorylated MAPs are unable to bind to microtubules; thus they promote microtubule disassembly. MAP kinase, a key enzyme for phosphorylating MAPs, is a participant in many signal-transduction pathways, indicating that MAPs are targets of many extracellular signals (Chapter 20). MAPs, especially MAP4, also are phosphorylated by a second kinase, cdc2 kinase, which plays a major role in controlling the activities of various proteins during the cell cycle (Chapter 13).
). Addition and loss of
subunits occur preferentially at one end, the (+) end.Microtubules exhibit two dynamic phenomena that are pronounced at Cc: treadmilling, the addition of subunits at one end and their loss at the other end, and dynamic instability, the oscillation between lengthening and shortening.
).Various drugs, including colchicine and taxol, disrupt microtubule dynamics and have an antimitotic effect. Some of these drugs are useful in the treatment of certain cancers.
Microtubule-associated proteins (MAPs) co-purify with microtubules and have the same cellular localization.
An important class of MAPs, the assembly MAPs, prevent cytosolic microtubules from depolymerizing, organize them into bundles, and cross-link them to membranes and intermediate filaments.
