Project Topics
Get Complete Project Material File(s) Now! »
The microtubule associated proteins (MAPs)
To be able to catch chromosome, for example, microtubule dynamics must be somehow controlled, which is achieved by microtubule associated proteins (MAPs). MAPs are proteins that can bind in a nucleotide insensitive manner to the microtubule lattice.
MAPs can influence microtubule dynamics by stabilizing or destabilizing microtubules. MAPs that stabilize microtubules are often seen at microtubule plus-ends, such as EB1 or XMAP215. Although it is still not understood how plus-end tip proteins stabilize microtubules, two mechanisms of microtubule destabilization have been reported for MAPs. The first one is to sequester free tubulin subunits, such as Op18 or stathmin, while the second is to sever microtubule lattice, such as katanin (McNally and Vale, 1993). Katanin is an heterodimeric enzyme, which promotes microtubule disassembly by generating internal breaks within the microtubule lattice. This severing activity is cell cycle dependent and particularly high in mitosis (Vale, 1991) (McNally and Thomas, 1998). Moreover, MAPs can also act as nucleating factors, for example TPX2.
MAPs are often found complexed with each other and their activity is also regulated, usually through phosphorylation.
Motor properties of cytoplasmic dynein
In contrast to kinesins, the detailed mechanism of dynein stepping is unknown, probably due to its high molecular weight in the mega dalton range, which makes its study difficult. Although dynein alone is sufficient to drive microtubule gliding in vitro, addition of yet another multi-protein complex, dynactin, considerably enhances dynein motor properties.
Dynactin can interact in vivo and in vitro with dynein intermediate chain but can also directly bind microtubules via its p150glued subunit. This additional microtubule binding site of dynactin could explain the increase in dynein processivity in vitro (King and Schroer, 2000).
In addition, dynein has been proposed to function like a ÔgearÕ as its step size decreases while its produced force increases under increasing load (Mallik et al., 2004).
Function and regulation of cytoplasmic dynein
Unlike kinesins, which have expanded and specialized in evolution, often resulting in Ôone kinesin=one functionÕ, dynein completes its various functions by interacting with different accessory proteins.
First, the composition of dynein itself can be subjected to changes. Heterogeneity in dynein light intermediate chain number could lead to differential regulation of its motor domain. In this way, it has been reported that dynein intermediate chain could act as a negative regulator of the motor domain (Mallik and Gross, 2004).
As already mentioned, dynein can interact with dynactin, probably in a phosphorylation dependent manner. Formation of the dynein/dynactin complex is essential for dynein targeting and function. For example, dynein localization at the kinetochore is achieved through dynactin binding to ZW10, a kinetochore component. This ternary complex may be involved in a tension-sensitive checkpoint mechanism, which would delay anaphase if chromosome bipolar attachment is defective (Starr et al., 1998). MAP-2 has also been reported to stimulate dynein detachment from microtubules, probably by interfering with dynein microtubule binding sites (Schroer, 1994).
Motor properties of kinesins
The mechanism of movement of KRPs along microtubules varies between different KRPs and it seems that even the same KRP can switch occasionally between different modes of movement. Nevertheless, 3 categories of movement can be described for KRPs: processive movement, non-processive movement and diffusion.
A processive motor makes several steps before detaching from the microtubule. It is now well accepted that the hand-over-hand model, in which the release of the rearward head is tightly coupled with the binding of the forward head, can describe KRP processive movement. Succinctly, the forward head releases its adenosine diphosphate (ADP) upon microtubule binding. Subsequent ATP binding promotes docking of the flexible neck linker, ÒthrowingÓ the rearward head in the front. The new forward head binds the microtubule. ATP hydrolysis occurs on the new rearward head followed by phosphate release and detachment. Thus, ADP release is catalyzed by KRP binding to the microtubule, whereas unbinding of KRPs from the microtubule requires ATP hydrolysis (Howard, 2001). The kinetic properties of the conventional kinesin, which belongs to the kinesin-1 subfamily, have been extensively studied: it moves along microtubules with a speed of 20m/min, hydrolyses one ATP per step and makes on average 100 steps of 8nm, corresponding to a tubulin dimer, per run. Optical-trapping studies have also revealed that conventional kinesin can sustain hindering loads up to 6pN, with its velocity slowing down until the motor finally stalls (Valentine and Gilbert, 2007). Most of the KRPs studied so far move processively along microtubules (Fig. 6).
Motor properties of Eg5 in vitro
Thanks to the development of single molecule assays, insight into Eg5 motor properties was gained and in particular the role of its unique neck linker configuration for its processivity. Eg5 processivity has remained for a long time under debate, whereas Eg5 plus-end directed microtubule gliding, with a velocity of 2m/min, was clearly demonstrated (Sawin et al., 1992a). A study by Valentine et al (Valentine et al., 2006), using optical-trapping shed light on Eg5 processivity.
In this study, Valentine et al demonstrated that individual purified Eg5 dimers (hsEg5-513-5His) step processively along microtubules. Eg5 dimers make on average 8 steps per run (compared to 100 for conventional kinesin), with 1 step per tubulin dimer and one ATP hydrolysed per step (like conventional kinesin). Moreover, they showed that Eg5 velocity was less force-sensitive when compared to conventional kinesin. While kinesin velocity slows and stalls at high hindering loads, Eg5 dissociates from microtubules after a slight slowdown. Furthermore, they interpreted this low processivity as being a consequence of Eg5 rigid neck linker. In order that the Eg5 forward head binds ATP, Eg5 neck linker must undergo a conformational change from perpendicular to parallel with respect to the long axis of the motor core. Once this isomerisation has occurred and ATP has bound, the rearward head can move forward, initiating the processive run. This rearrangement of Eg5 neck linker was also described by Rosenfeld et al (Rosenfeld et al., 2005), and is believed to be slow: 0.5 – 1 s-1 (Valentine and Gilbert, 2007). This also suggests that unlike kinesin, for which stepping is limited by phosphate release, ATP hydrolysis is rate limiting for Eg5 stepping. Recombinant Eg5 tetramer has also been reported to be processive by Kwok et al, who also described an additional diffusive component to Eg5 motility (Kwok et al., 2006). Eg5 diffusive behavior is ATP independent, in contrast to its directional stepping, and is favored when Eg5 is inhibited by monastrol (Crevel et al., 2004). Korneev et al. (Korneev et al., 2007) also confirmed that Eg5 tetramer is processive with an 8nm step size and an average of 10 steps per run, similar to Eg5 dimer. However, they reported that individual Eg5 tetramers are released at lower force compared to individual Eg5 dimers (respectively 2pN in low salt buffer and 7pN in high salt buffer), suggesting that Eg5 load-dependent detachment could be regulated by the Eg5 tail. These final evidences that Eg5 tetramer is processive corroborate and furthermore give a possible explanation to previous experiments such as those of Kapitein et al in which they observed that Eg5 tetramers crosslink, align and drive microtubules sliding relative to each other (Kapitein et al., 2005). This would have been difficult to explain for a non-processive motor as its probability of simultaneously binding two microtubules would be extremely low. One should also emphasize that this study was the first to directly show that Eg5 could crosslink parallel and slide antiparallel microtubules. These studies have shed light on Eg5 in vitro motor properties, providing basis to understand Eg5 in vivo motor properties and in particular in relation to its function.
Eg5 function in bipolar spindle formation and maintenance
The monopolar phenotype observed when the Eg5 inhibitor, monastrol, is added to BS-C-1 cells was further characterized by Kapoor et al (Kapoor et al., 2000). They concluded that monastrol inhibits centrosome separation, hence leading to monopolar spindles. This is consistent with previous data involving Eg5 in centrosome separation and consequently bipolar spindle formation, although these observations are still under debate in Drosophila melanogaster (Sharp et al., 1999b).
Interestingly, the role of Eg5 in bipolar spindle assembly seems to be evolutionarily conserved since it has been described in A. nidulans (Enos and Morris, 1990) in fission yeast (Hagan and Yanagida, 1990), in Saccharomyces cerevisiae 11 (Hoyt et al., 1992), in Drosophila melanogaster (Heck et al., 1993), in Xenopus (Sawin and Mitchison, 1995), in Homo Sapiens (Blangy et al., 1995). In addition to bipolar spindle assembly, Eg5 has also been implicated in bipolar spindle maintenance, probably in a Òpush and pullÓ mechanism, as addition of anti-Eg5 antibodies or monastrol disrupt preformed spindle in Xenopus egg extract (Sawin et al., 1992b) (Kapoor et al., 2000). Both bipolar spindle formation and maintenance intimately rely on Eg5 capacity to link interpolar microtubules together in vivo as described by Sharp et al (Sharp et al., 1999a), and remarkably reflect Eg5 in vitro motor properties (Kapitein et al., 2005).
Eg5 function in microtubule poleward flux
It is no longer controversial that Eg5 also plays a role in microtubule poleward flux. Poleward flux is composed of 3 coordinated activities: microtubule minus-end depolymerization, microtubule translocation and microtubule plus-end polymerization (Kwok and Kapoor, 2007) (Rogers et al., 2005).
Spindle elongation rate and Eg5 microtubule sliding rate have been correlated (Shirasu-Hiza et al., 2004), suggesting that Eg5 is responsible for the Ômicrotubule translocation activityÕ of poleward flux. Eg5 involvement in microtubule poleward flux has also been reported in Xenopus egg extract (Miyamoto et al., 2004). Indeed, addition of Eg5 inhibitors inhibited spindle microtubule poleward flux in a dose-responsive manner, suggesting that flux is driven by an ensemble of non-processive Eg5. In addition, microtubules move poleward with a speed of 2m/min in Xenopus egg extract 2,22m/min in Drosophila S2 (Rogers et al., 2004), which is found to match Eg5 microtubule gliding speed in vitro. However, Eg5 is probably not the only flux driver as Eg5 independent poleward flux as been observed for kinetochore microtubules, which interestingly are organized in parallel arrays, in PtK1 cells (Cameron et al., 2006).
Since both spindle bipolarity and microtubule poleward flux clearly involve Eg5 antiparallel microtubule cross-linking and sliding activities, it is not surprising that spindle bipolarization itself relies on microtubule poleward flux, at least in Xenopus egg extract, as suggested by Mitchison et al (Mitchison et al., 2004).
Eg5 function in a potential spindle matrix
Owing to Eg5 intriguing localization to spindle poles, it was proposed earlier that Eg5 could be part of a Òspindle matrixÓ (Sawin et al., 1992a). Thus Eg5 localization to minus-ends of microtubules would not depend on its motor activity as such. However, the existence of the matrix itself is still hypothetical as its composition, regulation and role are unknown. The current hypothesis proposes that such a matrix would help to stabilize, organize microtubules and serve as a stationary substrate against which motors could slide microtubules (Scholey et al., 2001).
To date, many different proteins (e.g.: NuMa, skeletor) have been proposed to form or be part of the matrix. The best candidate so far is lamin B as it associates with the mitotic spindle, is nocodazole insensitive and retains a certain number of proteins involved in spindle formation such as Eg5 (Tsai et al., 2006). Moreover, another study also implicated Eg5 as being part of a spindle matrix (Kapoor and Mitchison, 2001).
As described above, the evidence for the matrix existence and thus, Eg5 being part of it, are rather poor. As a consequence, the actual role of Eg5 in the matrix remains unclear, in particular if the matrix is a non-microtubule based structure, raising the question of how the microtubule-dependant molecular motor Eg5 can bind to it.
Table of contents :
1 Introduction
1.1 Cell cycle and cell division
1.1.1 The cell cycle
1.1.2 Cell division
1.2 The mitotic bipolar spindle
1.2.1 Microtubules
1.2.2 The microtubule associated proteins (MAPs)
1.2.3 The molecular motors
1.2.3.1 The dynein superfamily
1.2.3.1.1 Classification and structure of dyneins
1.2.3.1.2 Motor properties of cytoplasmic dynein
1.2.3.1.3 Function and regulation of cytoplasmic dynein
1.2.3.2 The kinesin superfamily
1.2.3.2.1 Classification and structure of kinesins
1.2.3.2.2 Motor properties of kinesins
1.2.3.2.3 Function and regulation of kinesins
1.2.4 Spindle architecture
1.3 The Xenopus kinesin-5: XlEg5
1.3.1 Eg5 homologs
1.3.2 Structure of Eg5
1.3.3 Motor properties of Eg5 in vitro
1.3.4 Function of Eg5 in vivo
1.3.4.1 Eg5 function in bipolar spindle formation and maintenance
1.3.4.2 Eg5 function in microtubule poleward flux
1.3.4.3 Eg5 function in a potential spindle matrix
1.3.5 Regulation of Eg5
1.3.5.1 Cdk1 phosphorylation of Eg5
1.3.5.2 Eg2 phosphorylation of Eg5
1.3.5.3 RanGTP regulation of Eg5
2 Motivation
3 Results
3.1 Role of Eg5 phosphorylation in bipolar spindle formation in Xenopus egg extract
3.1.1 Eg2 phosphorylates full length Eg5 at serine 543
3.1.1.1 Eg2 phosphorylates truncated GST-Eg5 at serine 543 in vitro and in mitotic egg extract
3.1.1.2 Eg2 phosphorylates full length Eg5 on Serine 543 in vitro
3.1.2 Cdk1 is the major kinase that phosphorylates Eg5 in Xenopus egg extract .
3.1.3 Eg5 S543A’ and Eg5 T937A are functional motors in vitro
3.1.4 Cdk1 phosphorylation, but not Eg2 phosphorylation is required for spindle assembly in Xenopus egg extract
3.1.5 Eg5 T937A localization to spindle microtubule is disrupted in Xenopus egg extract
3.2 Role of Eg5 intrinsic motor properties in spindle formation in Xenopus egg extract.
3.2.1 Eg5 chimeras
3.2.1.1 Kid-Eg5, Dkhc-Eg5 support microtubule gliding
3.2.1.2 The specific properties of Eg5 motor domain are required for spindle formation in Xenopus egg extract
3.2.2 Eg5 motility in mitotic Xenopus egg extract, development of an assay
3.2.2.1 Microtubule gliding assay in mitotic Xenopus egg extract
3.2.2.2 Eg5 motility in mitotic Xenopus egg extract
4 Discussion
4.1 Eg5 phosphorylation by Eg2 is not required for bipolar spindle formation in Xenopus egg extract.
4.2 Cdk1 phosphorylation regulates Eg5 efficient binding to spindle microtubule
4.3 Eg5 intrinsic motor properties are required for spindle assembly in Xenopus egg extract
4.4 Eg5 motility in Xenopus egg extract
5 Material and methods
5.1 Cloning
5.1.1 Full length Eg5 constructs
5.1.2 Chimeric Eg5 constructs
5.2 Proteins expression and purification
5.2.1 Expression and purification of Eg5 constructs
5.2.2 Expression and purification of Eg2
5.2.3 Expression and purification of p50
5.2.4 Antibody purification
5.2.4.1 Eg5 antibody
5.2.4.2 XKCM1 antibody
5.2.4.3 Katanin antibody
5.3 Phosphorylation experiments
5.4 Xenopus egg extract experiment
5.4.1 Preparation of Xenopus egg extract
5.4.2 Depletion-add back experiments
5.5 Microtubule gliding assay
5.5.1 In vitro microtubule gliding assay
5.5.2 In extract microtubule gliding assay
6 References
7 Appendix
7.1 Résumé
7.2 Abstract
7.3 Abbreviations
7.4 Publications
