Reconstitution of Physiological Microtubule Dynamics Using Purified Components

See allHide authors and affiliations

Science  09 Nov 2001:
Vol. 294, Issue 5545, pp. 1340-1343
DOI: 10.1126/science.1064629


Microtubules are dynamically unstable polymers that interconvert stochastically between polymerization and depolymerization. Compared with microtubules assembled from purified tubulin, microtubules in a physiological environment polymerize faster and transit more frequently between polymerization and depolymerization. These dynamic properties are essential for the functions of the microtubule cytoskeleton during diverse cellular processes. Here, we have reconstituted the essential features of physiological microtubule dynamics by mixing three purified components: tubulin; a microtubule-stabilizing protein, XMAP215; and a microtubule-destabilizing kinesin, XKCM1. This represents an essential first step in the reconstitution of complex microtubule dynamics–dependent processes, such as chromosome segregation, from purified components.

Microtubules polymerize and depolymerize by the addition and loss of αβ-tubulin dimer subunits from their ends (1). Polymerizing and depolymerizing microtubules coexist and infrequently interconvert between these two states, a behavior known as dynamic instability (2). The transition of a polymerizing microtubule to a depolymerizing state is referred to as a catastrophe, and the converse transition is referred to as a rescue (3). Microtubules exhibit dynamic instability when assembled from purified tubulin (3, 4) and in a physiological cytoplasmic environment (5–10), but there are notable differences between the two. In a physiological environment, microtubules polymerize about fourfold faster than a similar concentration of purified tubulin. At the polymerization rates observed in physiological conditions, purified tubulin has a near-zero rate of catastrophe. In contrast, microtubules in cells and in cytoplasmic extracts have a high catastrophe rate despite their high polymerization rate.

Clues to the dynamic behavior of microtubules in physiological conditions have come from the identification of proteins that modulate microtubule dynamics. Microtubule-associated proteins (MAPs) increase the polymerization rate of microtubules, whereas destabilizing proteins increase the rate of catastrophe. In Xenopus egg extracts, the dominant stabilizing MAP appears to be XMAP215, a member of an evolutionary conserved protein family (11, 12). Depleting XMAP215 prevents microtubule growth, whereas depletion of other MAPs has so far had little effect (12, 13). The dominant catastrophe factor in Xenopus extracts is XKCM1 (14, 15), a member of the KinI subfamily of kinesins (16). Depletion of XKCM1 markedly stabilizes microtubules, whereas depletion of other catastrophe factors have to date had more modest effects (14,17).

Depletion experiments suggest that these two factors oppose each other to determine the stability of the microtubule lattice inXenopus extracts. Depletion of XMAP215 destabilizes microtubules, and subsequent inhibition of XKCM1 causes the microtubules in XMAP215-depleted extracts to become stable again (12). These results suggest that the coordinate action of only these two proteins on tubulin may explain why microtubules in cells can both polymerize rapidly and exhibit high catastrophe rates. We tested this hypothesis by combining these two factors with purified tubulin in vitro and examining the behavior of microtubules. We first produced full-length recombinant His6-tagged XMAP215 and XKCM1 (18). As expected, XKCM1 reduced microtubule length whereas XMAP215 increased microtubule length (Fig. 1) (15,18–20). However, when added together, XMAP215 opposed the ability of XKCM1 to decrease the average length of microtubules (Fig. 1) (18). Therefore, we conclude that the relative activities of XMAP215 and XKCM1 can determine the steady-state length of microtubules assembled from purified tubulin in the absence of other factors.

Figure 1

XMAP215 opposes XKCM1 in the regulation of microtubule length in vitro. (A andB) XKCM1 (62.5 nM, +XKCM1) or control buffer (−XKCM1) was mixed with 15 μM tubulin and centrosomes in the absence (A) or presence (B) of 125 nM XMAP215. Microtubule asters are shown after 5 min of incubation at 37°C (20). Bars, 25 μm. (C) Quantification of microtubule length in (A) (−XMAP; open columns) and (B) (+XMAP; solid columns). Arrowheads indicate the average length of microtubules (−XMAP, open arrowheads; + XMAP, solid arrowheads).

In Xenopus extracts, XMAP215 suppresses the ability of XKCM1 to induce microtubule catastrophe (12). To test if this suppression can be observed with purified proteins, we set up a real-time assay in which the behavior of individual microtubules was monitored by video-enhanced differential interference contrast (VE-DIC) microscopy (21). Briefly, purified centrosomes were adsorbed to the surface of a perfusion chamber, and tubulin was perfused into the chamber and allowed to polymerize (18). Different combinations of factors with tubulin were then perfused into the chamber, and the number of microtubules transiting from polymerization to depolymerization after perfusion was quantified (18). After perfusion of 150 nM XKCM1 together with tubulin, 60 to 70% of the microtubules transited to depolymerization (Fig. 2). Adding XMAP215 together with 150 nM XKCM1 and tubulin suppressed the number of microtubules transiting to depolymerization in an XMAP215 concentration–dependent manner (Fig. 2). This result with three purified components leads us to conclude that the suppression of XKCM1-induced catastrophes by XMAP215 in cytoplasmic extracts is a direct consequence of the action of these two proteins at microtubule ends.

Figure 2

XMAP215 inhibits catastrophes induced by XKCM1. (A) Images of microtubules before and 4 min after perfusion (21). Microtubules were polymerized from centrosomes with 33 μM tubulin (left). Perfusion chambers were perfused with tubulin (33 μM) and control buffer (i), 0.15 μM XKCM1 (ii), or 0.15 μM XKCM1 + 1.2 μM XMAP215 (iii). Bar, 10 μm. (B) The percentage of shrinking microtubules after perfusion of XKCM1 (0.15 μM) + XMAP215 (0, 0.6, or 1.2 μM) was plotted versus the concentration of XMAP215.

The above results suggested that potentially a steady-state mixture of tubulin, XMAP215, and XKCM1 could reconstitute the two characteristic physiological features of dynamic instability: rapid polymerization and high catastrophe rates (Table 1). To explore this idea, we developed conditions in which the dynamic behavior of these microtubules could be directly observed by DIC microscopy (22). The concentrations of these proteins inXenopus extracts have been estimated to be 25 μM (tubulin) (23), 60 nM (XKCM1) (14), and 0.6 μM (XMAP215) (12). Using a mixture of 25 μM tubulin, 0.2 μM XKCM1, and 0.8 μM XMAP215, we were able to reconstitute physiological parameters of dynamic instability (18). A typical aster growing under these conditions is shown in Fig. 3. Quantification of microtubule dynamics revealed that this simple three-component mixture recapitulated the essential features of physiological microtubule dynamic instability (Table 1).

Figure 3

Reconstitution of physiological microtubule dynamics with XMAP215 and XKCM1. (A) Dynamic instability behavior of microtubules in the presence of 25 μM tubulin, 0.8 μM XMAP215, and 0.2 μM XKCM1 at 30°C (22). VE-DIC images of microtubules are shown at various time points (t 1 tot 4). g, growing microtubule; p, pausing microtubule; s, shrinking microtubule. The colors of arrows and arrowheads in (A) correspond to the traces of microtubules in (B). (B) Life-history traces of the microtubules. (C) Fate of a single microtubule. Scale bars, 10 μm.

Table 1

Comparison of the parameters of microtubule dynamics in vitro and in vivo. Dashes indicate that no microtubules were formed. NA, not applicable.

View this table:

Why does the three-component mixture reconstitute physiological dynamic instability? XMAP215 alone accounts for the observed fast polymerization rate, whereas XKCM1 alone prevents any assembly of microtubules, presumably as a consequence of inducing a catastrophe rate that is too high (Table 1) (18). In the three-component mixture, XMAP215 must partially suppress the catastrophe-promoting activity of XKCM1 in order to generate the combination of fast polymerization and high catastrophe rates. This partial suppression of XKCM1-induced catastrophes by XMAP215 (Fig. 2) without influencing the polymerization rate (Table 1) is central to understanding why the three-component mixture reconstitutes the combination of fast polymerization and high catastrophe rates.

Taken together with the results in Xenopus extracts (12, 14), we believe that the essential features of dynamic instability in Xenopus egg extracts are derived from the sole action of these two factors on tubulin. These factors are conserved from yeast through mammals (24–31), and at least in Saccharomyces cerevisiae, orthologous factors oppose each other in the control of microtubule length in vivo (31). Thus, this three-component system may represent a conserved module that generates the characteristic behavior of physiological microtubules. The simple mixture we describe here will serve as a starting point for analyzing the effect of other microtubule regulators, such as XMAP230 and Op18 (13, 17), in a reconstruction-type approach. Furthermore, it represents an important step in the pursuit of the eventual reconstitution of complex dynamic microtubule assemblies, such as the mitotic spindle, from purified components.

  • * To whom correspondence should be addressed. E-mail: kinoshita{at}, hyman{at}

  • Present address: Equipe ATIPE, UMR 6026, Université de Rennes 1, Campus de Beaulieu, Bt 13, 35042 Rennes Cedex, France.


View Abstract

Navigate This Article