This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.
In contrast to those of metaphase chromosomes, the shape, length, and architecture of human interphase chromosomes are not well understood. This is mainly due to technical problems in the visualization of interphase chromosomes in total and of their substructures. We analyzed the structure of chromosomes in interphase nuclei through use of high-resolution multicolor banding (MCB), which paints the total shape of chromosomes and creates a DNA-mediated, chromosome-region-specific, pseudocolored banding pattern at high resolution. A microdissection-derived human chromosome 5-specific MCB probe mixture was hybridized to human lymphocyte interphase nuclei harvested for routine chromosome analysis, as well as to interphase nuclei from HeLa cells arrested at different phases of the cell cycle. The length of the axis of interphase chromosome 5 was determined, and the shape and MCB pattern were compared with those of metaphase chromosomes. We show that, in lymphocytes, the length of the axis of interphase chromosome 5 is comparable to that of a metaphase chromosome at 600-band resolution. Consequently, the concept of chromosome condensation during mitosis has to be reassessed. In addition, chromosome 5 in interphase is not as straight as metaphase chromosomes, being bent and/or folded. The shape and banding pattern of interphase chromosome 5 of lymphocytes and HeLa cells are similar to those of the corresponding metaphase chromosomes at all stages of the cell cycle. The MCB pattern also allows the detection and characterization of chromosome aberrations. This may be of fundamental importance in establishing chromosome analyses in nondividing cells.
Interphase chromosomes analyzed with currently available routine cytogenetic techniques do not exhibit any recognizable structures such as bands, centromeres, telomeres, or specific shapes. It has therefore been assumed that chromosomes in interphase are relatively decondensed (Comings 1968). Until now, the concept of condensation and decondensation of chromosomes during mitosis was well established and has profoundly influenced our understanding of the structure and function of chromosomes during mitosis. This concept implies that chromosomes in interphase are very long and condense after S phase. Clustering of chromatin loops results in condensation, giving very long prophase chromosomes, which show thousands of bands (Yunis 1981). These further condense to prometaphase, metaphase, and anaphase chromosomes. Decondensation takes place as cells return, via telophase, to interphase.
Surprisingly, there have been no scientific investigations directly confirming this dynamic concept-for example, by comparing the total length of chromosomes in interphase and metaphase. There is, however, indirect evidence, such as countless daily observations that harvesting chromosomes shortly after S phase results in elongated prophase and prometaphase chromosomes. Also, pictures of cells in G2 phase showing interlaced threads of elongated chromosomes that are sometimes much longer than the diameter of interphase cells are well known. Furthermore, the extent of compaction of chromosomes in interphase cells, estimated in both yeast (Guacci et al. 1994) and human lymphoma cell lines (Lawrence et al. 1988) by measuring the distances between fluorescently labeled DNA probes, was found to be significantly different when compared with mitotic chromosomes, with a two- to tenfold compaction of mitotic chromosomes in comparison with interphase chromosomes. In addition, the phenomenon of premature chromosome condensation (Johnson and Rao 1970) and results obtained from high-resolution chromosome preparations (Yunis 1976) provide further support for our understanding of chromosomes in interphase nuclei. It was observed that human prophase chromosomes can be very long, showing as many as 3,000 bands per haploid set (Yunis 1981). In addition, numerous investigations dealing with proteins such as topoisomerase II, HMC, SMC, H1, and H3, which have been confirmed to be involved in-or responsible for-chromosome condensation, have been published (for a review, see Koshland and Strunnikov ). None of these data, however, can fully explain the speed with which chromosomes change their structure during mitosis or the fact that homologous metaphase chromosomes sometimes show extreme differences in size.
Experience with artificial stretching of human chromosomes (Claussen et al. 1994; Hliscs et al. 1997a) indicates that a significant chromosome-elongation process takes place during routine chromosome preparation. This leads to doubts as to whether chromosomes in interphase are indeed very long. Furthermore, microirradiation experiments (Cremer et al. 1982) and molecular cytogenetic investigations with whole-chromosome paints (Cremer et al. 1988; Lichter et al. 1988; Pinkel et al. 1988) or region-specific microdissection probes (Cremer et al. 1993; for review, see Chevret et al. 2000; Cremer and Cremer 2001) have confirmed a territorial organization of chromosomes in interphase nuclei that is more likely to be equivalent to short chromosomes in interphase.
To completely paint chromosomes in interphase, to measure their length, and to analyze their DNA-based banding structure, the microdissection-based high-resolution multicolor banding (MCB) technique was used. This is a FISH approach first described by Chudoba et al. (1999). The results show that chromosomes in interphase are not decondensed but are as short as metaphase chromosomes and show almost the same MCB pattern. This pattern can be used for the identification of small structural aberrations in chromosomes.
Length of Chromosomes in Interphase Nuclei
Chromosomes in interphase are thought to (1) be much longer than chromosomes in prophase, (2) further condense to metaphase and anaphase chromosomes, and (3) become decondensed and longer after telophase. However, our results show that chromosomes in interphase are very similar in length to metaphase chromosomes. Therefore, doubts arise about the concept of chromosome condensation in general. We propose that all the convincing experiments published to date that deal with H3- and SMC-phosphorylation with respect to chromosome condensation (Gurley et al. 1978; Hirano 1999; Strunnikov and Jessberger 1999; Wei et al. 1999) may explain phenomena related to the formation and/or compaction of chromosome loops that influence the width of chromosomes in two dimensions (Weise et al. 2002) and probably their volume in three dimensions. The length of the chromosome axis, the third component of chromosome condensation (Koshland and Strunnikov 1996), is clearly not influenced as much, although this has not been investigated directly by FISH. Technical difficulties preventing such an investigation have been solved here by the use of MCB.
The distances between fluorescently labeled DNA probes on human lymphoma cells indicate as much as a tenfold compaction of mitotic chromosomes, when compared with interphase chromosomes (Lawrence et al. 1988). In addition, Yunis (1976, 1981) showed that prophase chromosomes are very long, in comparison with metaphase chromosomes. The discrepancy between these observations and our findings may be due to prophase chromosomes being artificially elongated during drying of fixative on the slide during chromosome preparation (Hliscs et al. 1997b). It also indicates that cell cycle-specific differences in the preparation-induced chromosome-elongation process may occur. In comparison with metaphase and interphase chromosomes, prophase chromosomes appear to be more sensitive to preparation-induced chromosome elongation. This elongation is a prerequisite for obtaining well-spread metaphases useful for chromosome analysis (Hliscs et al. 1997b). Chromosomes, routinely harvested for chromosome analysis shortly after completion of S phase, probably contain chromosome region-specific proteins that can easily be stretched during drying of the fixed cells on the slide, resulting in prophase chromosomes. Chromosomes harvested late after completing their S phase, however, are more resistant to preparation-induced chromosome stretching, leading to metaphase chromosomes that are relatively short. Chromosome-stretching experiments indicate that G-light bands are the stretchable units and that a hierarchy exists in their elasticity (Hliscs et al. 1997a; Küchler et al. 2001). Consequently, the stable G-banding patterns occur, which provide the basis for most chromosome analyses.
The standard deviations of the length of flattened interphase chromosomes obtained after routine chromosome preparation and of intact three-dimensional chromosomes after preparation at reduced humidity are small. This indicates a stable length of interphase chromosomes. The mean length of flattened interphase chromosomes 5 obtained after routine chromosome preparation is 12 Î¼m, about twice the length of those obtained after drying the fixative at a humidity of â‰¤7%. Nearly identical results were found in a similar comparison of the length of metaphase chromosomes (Hliscs et al. 1997b). Metaphase chromosomes of the C group obtained after short-term evaporation of the fixative at 80Â°C were 3 Î¼m in length, and routinely harvested and dried metaphase chromosomes were 5.6 Î¼m.
Chromosome-stretching analysis (Hliscs et al. 1997a; Küchler et al. 2001) revealed that G-light bands represent chromosome regions that can be stretched. Chromosome stretching takes place during drying of the fixed suspension on the slide (Hliscs et al. 1997b) and, as discussed above, this is a prerequisite to obtaining metaphase spreads of sufficient quality for chromosome analyses. Without chromosome preparation-induced stretching, chromosomes are short and probably do not show any bands. The banding pattern may be regarded as a preparation-induced artifact, although this does not sufficiently reflect the fact that chromosome preparation leads to characteristic and highly reproducible morphological changes of chromosomes on which chromosome analyses are based. The biology behind the formation of these "reproducible artifacts" may be related to a chromosome region-specific fixed hierarchy in the potential of proteins of the chromatin to be stretched, but this has not been investigated. The methanol, acetic acid, and water of the fixative may interact with proteins of the chromatin, especially in G-light bands, thus resulting in their elongation. Following this interpretation, one may assume that routine chromosome preparation using fixative and drying of the fixed suspension on the slide induces a dramatic genomewide elongation of all chromosome regions in which housekeeping genes are located, leading to a banded metaphase spread. A similar mechanism may operate in living cells, restricted to single chromosome regions, to open the chromatin (G-light chromosome regions) for the transcription machinery, probably in a function-specific manner.
The Banded Structure of Interphase Chromosomes
Ideas about the chromosomes in interphase nuclei have been influenced by the previous inability to visualize analyzable structures. We show for the first time, to our knowledge, that interphase chromosomes are structured, with a highly reproducible DNA-specific banding pattern. This will profoundly enhance our understanding of the architecture of the interphase nucleus. The existence of such a banding pattern can be regarded as a logical consequence of the observation by Dietzel et al. (1998) that the chromosome arm domains in interphase nuclei show separate signals.
Some interphase chromosomes 5 show an incomplete MCB pattern. This partial loss of chromosome 5-specific MCB signals may be due to technical problems, because FISH techniques are predominantly adapted to metaphase chromosomes and not to interphase nuclei. This is supported by the observation that an MCB probe mixture of reduced hybridization quality resulted in increased partial loss of MCB signals. The margins of interphase chromosome 5 in S phase appear to be less sharply marked, which may also be a preparation-induced artifact. We decided not to investigate this in more detail.
Interphase nuclei in living cells are three-dimensional. Routine chromosome preparation, however, leads to flattened interphase nuclei, and significant alterations of the structure of chromosomes in interphase nuclei of cells flattened during preparation are expected. However, the similarity between the DNA-based banding pattern of flattened interphase chromosomes and that of metaphase chromosomes suggests that, in living cells, both three-dimensional intact interphase chromosomes and metaphase chromosomes may be more similar than expected. The preparation technique we used to gain information about three-dimensional intact interphase chromosomes led to half-spherical interphases, which are not optimal. In an attempt to preserve the three-dimensional chromatin structure of the interphase nuclei, we fixed cells with buffered formaldehyde; however, subsequent MCB did not give sufficient hybridization signals.
The DNA-based banding pattern of interphase chromosomes is very similar to that of metaphase chromosomes. This holds true not only for the G2 phase but also for the G1 and S phases. Consequently, the terms "chromosome territory" and "chromosome domain" will no longer be needed and can be replaced by the term "interphase chromosome." Chromosomes are chromosomes throughout the whole cell cycle, and ideas and concepts are needed in dealing with structural changes and reorganization of chromosomes during the cell cycle, with regard to functional aspects and tissue specificity. Several models for the organization and architecture of chromosomes in interphase nuclei have been discussed (for review, see Chevret et al.  and Cremer and Cremer ) that indicate that the basic information about interphase chromosomes available at present is not sufficient to define the correct model. Technical advances are needed, especially in FISH technology, to enable the dynamic changes in interphase nuclei to be visualized.
The position of genes inside chromosomes in interphase depends on the transcriptional activity of the gene (Dietzel et al. 1999). Active genes are located at, or close to, the surface of interphase chromosomes. Inactive genes are more central, suggesting that dynamic position changes probably interfere with functional aspects. The chromosome preparation-induced elongation mechanism described by Hliscs et al. (1997b), which probably takes place exclusively in G-light bands (Hliscs et al. 1997a) where the housekeeping genes are placed (Holmquist 1992), may have an equivalent in the interphase. The mechanisms to open chromatin structures for transcription could relocate structures from an internal to a more external position, depending upon whether the genes are active or inactive. For the human X chromosomes in interphase cells, Eils et al. (1996) showed that the active X chromosome has the larger and more irregular surface. An extreme variation of this is the observation that transcriptionally upregulated genes in the major histocompatibility complex on the short arm of chromosome 6 were found to be on an external loop, apparently outside the chromosome 6 territory (Volpi et al. Volpi et al., 2000 EV Volpi, E Chevret, T Jones, R Vatcheva, J Williamson, S Beck, RD Campell, M Goldsworthy, SH Powis, J Ragoussis, J Trowsdale and D Sheer, Large-scale chromatin organization of the major histocompatibility complex and other regions of human chromosome 6 and its response to interferon in interphase nuclei, J Cell Sci 113 (2000), pp. 1565-1576. View Record in Scopus | Cited By in Scopus (234)2000). Also, tissue-specific aspects of the same mechanism of transcriptional activation have been shown for the epidermal differentiation complex at 1q21 in keratinocytes, where the genes are active, and in lymphoblast interphase nuclei, where they are silenced (Williams et al. 2002). We therefore assume a hierarchy in the distance from actively transcribed chromosomal DNA to their interphase chromosome territory, which corresponds to the hierarchy of the splitting of bands into their sub-bands for pro-, prometa-, and metaphase chromosomes.
Chromosome 5 aberrations in HeLa interphase cells, such as i(p) or deletions of the p and q arms, are clearly visible (fig. 3). In clinical cases (fig. 4), chromosome 5 aberrations can also be detected. Therefore, we speculate that, in the future, the form of all human chromosomes, together with their substructures, will be visualized in three dimensions and will be used to detect chromosome aberrations in interphase nuclei. This opens new possibilities for cytogenetic investigations.(Lemke et al. 2002)
During the cell cycle, two critical steps involving chromatin are the duplication of the genome, occurring in S-phase, and the segregation of the two sets of replicated chromosomes during anaphase. These two events are highly regulated in order to ensure that the genetic information is integrally and faithfully distributed into each daughter cell. In mitosis, the chromatin is packaged into compact entities, the mitotic chromosomes. This condensation process greatly reduces the volume of the replicated genomes to interphase chromatin, contributes to the resolution of the replicated sister chromatids and finally facilitates the subsequent separation of sister chromatids in anaphase.
In the past two decades, genetic and biochemical approaches have greatly contributed to the knowledge of the molecular actors and the underlying mechanisms supporting chromosome condensation and segregation. Although separated in time, these events are highly co-ordinated with each other and with the replication process. A dysfunction in one or more of these events may result in chromosome breakage or aneuploidy, which could potentially contribute to the development of tumour.
The first part will present an overview of the replication process, highlighting molecular aspects involved in coupling replication with chromatin dynamics in mitosis. The second part will present the current understanding of chromosome condensation and segregation during mitosis in higher eukaryotes. Finally, concluding remarks will underline the links that exist between replication and mitosis.
2. DNA replication and sister chromatid cohesion
2.1. DNA replication
Once a cell begins to proliferate, the first crucial event is the duplication of its genetic material.
In eukaryotes, this semi-conservative replication, catalysed by the holoenzyme point in the DNA polymerase, fires on chromatin at numerous sites called origins of replication. In budding yeast, origins of replication are short, consensus sequences; on the contrary, replication can start at any DNA sequence during early development of some metazoans (Xenopus laevis, Drosophila melanogaster). In mammalian cells, the situation is somewhere between these two extremes. Eukaryote origins of replication direct the formation of protein complexes that lead to the assembly of replicative complex. Briefly, the six subunit origin recognition complex (ORC) binds DNA sequences thus defined as origins of replication. This complex binds DNA throughout the cell cycle in yeast. In metazoans, however, it is still not clear whether the ORC remains bound to chromosomes in mitosis. After mitosis, two additional proteins, Cdc6p and Cdt1p, and a multisubunit complex, the MCM2-7 complex associate with the ORC, together forming the pre-replication complex (pre-RC). The pre-RC is matured into pre-initiation complex (pre-IC) by the recruitment of additional factors, including Cdc45 and Sld3. DNA polymerases are then recruited to origins of replication, and synthesise new DNA strands. The initiation of S-phase is triggered by different CDK-Cyclin combinations depending on the species. CDKs are Cyclin Dependent Kinases that regulate numerous cell events. The first identified CDKs regulate progression through the cell cycle. In addition to initiating replication, specific CDKs inhibit re-initiation by preventing new pre-RC assembly, so that replication occurs once and only once per cycle. Replication elongation depends on additional factors, among which the DNA helicase PCNA (Proliferating Cell Natural Antigen). During DNA synthesis, checkpoints ensure that the replication machinery faithfully copies the genome (for review on replication process, see [Bell and Dutta, 2002]Â ; for detailed description of the order of events during the chromosome replication cycle, see [Diffley and Labib, 2002]). Replication is coupled to several other processes, such as sister chromatid cohesion, post-replicative repair and imprinting (epigenetic inheritance).
From one origin of replication emanate two replication forks that progress in opposite directions along the DNA fibre (bi-directional replication). Because of the double helical structure of DNA, progression of replication forks generates strains and supercoiling that must be dissipated by topoisomerase activitiesÂ : actually, the region of DNA in front of a running replication fork becomes overwound or positively supercoiled. Some of the overwinding of the helix is transmitted to the region behind the replication fork causing the intertwining of the two replicated regions of DNA fibres. The main part of the resulting topological links are resolved during replication by topoisomerases I and II (for details see [Lucas et al., 2001]). Nonetheless, some links between newly synthesised sister chromatids will persist until the metaphase (see below).
2.2. Sister chromatid cohesion
Once DNA replication is achieved, each chromosome is composed of two catenated sister chromatids that have to be held together until they segregate, this can occur long after duplication is completed. The maintenance of this tightly-held-chromatid state, known as sister chromatid cohesion is a prerequisite for the accurate distribution of the genetic information into the two daughter cells. Sister chromatid cohesion is established during replication and provides a mechanism by which paired sisters are recognised as such by the spindle apparatus during mitosisÂ : sister chromatids associate with spindle microtubules via kinetochores. These are huge proteinaceous complexes that assemble on centromeres, which are the main point where sister chromatids are held together. Centromeres contain over 3Â million DNA base pairs in humans, organised in blocks of tandemly repeated sequence. In mitosis, sister kinetochores associate with microtubules emanating from opposite poles of the mitotic spindle, generating poleward-pulling forces that tend to separate sister chromatids. Sister chromatid cohesion opposes these pulling forces by maintaining the two sister chromatids together, thereby generating tension that facilitates the correct biorientation of sister chromatids on the mitotic spindle and subsequently, their proper segregation. In addition to topological links between sister chromatids, sister chromatid cohesion is ensured by a multisubunit complex called cohesin. Cohesin contains a core complex of a heterodimer composed of two SMC proteins (for Structural Maintenance of Chromosome), namely Smc1 and Smc3. SMC family proteins are characterised by an amino-terminal nucleotide binding domain (walker A motif), two central interacting coiled-coils separated by a hinge region and a carboxy-terminal domain (walker B motif), that, when associated with the walker A motif, constitutes an ATPase domain. Smc1 and Smc3 interact through their hinge regions. This dimer binds via its globular ATPase domains a third protein, Scc1 that in turn binds a fourth cohesin subunit, Scc3.
The cohesin complex associates with chromosomes from telophase to the onset of the subsequent anaphase. The association of the cohesin complex with chromatin depends on a distinct additional complex containing Scc2 and Scc4 proteins. Recent electron microscopy observations [Anderson et al., 2002 and Haering et al., 2002] together with biochemical experiments [Haering et al., 2002 and Gruber et al., 2003] have provided new insights into cohesin structure leading to the "ring model", depicted in Figure 1 . Each Smc protein constitutes one half of the cohesin ring by their homotypically interacting coiled coil regions. Interacting by their hinge regions, the two Smc proteins are also linked by the Scc1 that interacts with Smc1 and Smc3 via its C-terminal and N-terminal regions, respectively. The cohesin ring binds either unreplicated chromatids from telophase to S-phase or duplicated chromatids from DNA replication to the onset of anaphase, when proteolysis of Scc1 opens the ring, allowing sister chromatids to segregate.
Full-size image (28K)
Fig.Â 1. SMC protein structure and the ring model of cohesin.A) SMC proteins are composed of an amino-terminal nucleotide binding region (green) called the Walker A motif, two central interacting coiled-coils (blue) separated by a flexible hinge region (orange), and finally, a carboxy-terminal region, the Walker B motif (red). SMC proteins fold via the interaction of their two coiled-coils. The association of the Walker A and B motifs constitutes an ATPase domain. In the cohesin complex, Smc1 (dark colours) and Smc3 (light colours) interact through their hinge region. B) The ring model of cohesin structureÂ : the heterodimer Smc1/Smc3 constitutes a ring that encircles sister chromatids. This ring is closed by Scc1 (yellow) that binds the ATPase domains of both Smc1 and Smc3, respectively, through its C-terminal and N-terminal domains. In addition, Scc1 binds Scc3 (purple).
View Within Article
Before entry into mitosis, human cells possess in their 20Â Î¼m-diameter nucleus 46Â chromosomes, each made of two replicated sister chromatids, topologically linked and held together by cohesin rings, that have to be separated and equally distributed into each daughter cells. This separation requires that chromatin is condensed into mitotic chromosomes and spatially organised relative to the mitotic spindle.
3. Mitotic chromosome condensation and segregation
3.1. Condensin complex, topoisomerase II and mitotic chromosome formation
As mitosis begins, chromatin starts to condense into chromosomes. Biochemical experiments in Xenopus egg extracts have provided major insights into the molecular actors of chromosome condensation by identification and characterisation of a multisubunit complex associated with mitotic chromosomes [Hirano and Mitchison, 1994]. This complex, named condensin, is composed of a heterodimer of SMC proteins, but different from those in cohesin, namely Smc2/XCAP-E (for Xenopus Chromosome Associated Protein) and Smc4/XCAP-C. This dimer, considered as the core complex, interacts with three additional subunits, XCAP-D2, XCAP-G and XCAP-H. Some condensin subunits have also been characterised by genetic and biochemical approaches in independent experiments [Saka et al., 1994, Strunnikov et al., 1995, Cubizolles et al., 1998 and Sutani et al., 1999]. When mitotic extracts are immunodepleted of condensin, exogenous chromatin fails to condense into mitotic chromosomes. The condensation is restored by adding back condensin, indicating that this complex plays an important role in the condensation process. Condensin complex has also been shown capable of introducing positive supercoils into plasmids in the presence of topoisomerase I, in vitro. This indicates that the condensin complex is able to maintain constrained positive supercoils in plasmid DNA. This activity is thought to be involved in the mechanism of DNA compaction, although the way this occurs still remains speculative.
Condensation of chromatin into mitotic chromosomes is not the mere compaction of a linear chromatin fibre, because it has to deal with the topological links that remain between sister chromatids once replication is complete. The decatenation of sister chromatids is ensured by the topoisomerase II activity, which as a consequence is required for resolution and segregation of replicated genome in anaphase. As topoisomerase II does not discriminate between catenation and decatenation, its activity must be directed towards sister chromatid resolution. Interestingly, topoisomerase II has been shown to interact with Barren, the Drosophila melanogaster homologue of the XCAP-H condensin subunit, providing a physical link between compaction and resolution [Bhat et al., 1996]. This could explain the phenotypes of condensin subunit disruption. Indeed, they mainly show resolution defects in vivo, rather than dramatic condensation defects, at least in higher eukaryotes. Indeed, even if results from different organisms indicate a clear role of condensin in chromosome condensation, it is also clear that condensin cannot account for the entire chromatin compaction process. Rather, condensin appears to be a key player in the resolution of sister chromatids, possibly by regulating topoisomerase II activity. However, recent experiments in cycling extracts report that topoisomerase II activity is required prior to mitosis and independently of condensin for chromatin to condense properly during mitosis [Cuvier and Hirano, 2003].
In addition to these catalytic roles of compaction and resolution, the condensin complex and topoisomerase II also play a structural role in maintaining the mitotic chromosome architecture. In mitosis, condensin and topoisomerase II display a similar chromosomal localisation, staining the axis of sister chromatids (see Figure 2, A ). Topoisomerase II is present on chromosomes throughout the cell cycle, whereas immunofluorescence experiments indicate that the condensin complex is loaded onto chromosomes during late prophase, a stage where chromosome condensation is already largely completed (Figure 2, b). This indicates that condensin is not the only chromatin condensing complex. The chromosomal localisation of condensin is not yet elucidated, but it has been shown that XCAP-D2 is stoichiometricaly required to localise XCAP-H, but not XCAP-E [Watrin et al., 2003] in Xenopus egg extract. Moreover, AKAP-95, a nuclear-matrix protein, is required to address condensin onto chromosomes in mitotic human cell extract, via a direct interaction with hCAP-H, the human homologue of XCAP-H. Interestingly, an interaction has also been described between XCAP-H and hCAP-D2 (the human homologue of XCAP-D2) in a two-hybrid assay [Eide et al., 2002]. These results suggest that both CAP-D2 and AKAP-95 are involved in chromosomal localisation of condensin via their interaction with CAP-H. Finally, the chromosomal localisation of condensin depends on the presence of aurora B kinase in both Drosophila melanogaster (Giet et al., 1998) and Caenorhabditis elegans [Kaitna et al., 2002](see below). Dissociation of condensin complexes from chromatin occurs at the end of anaphase, by an unknown mechanism, at about a time when cohesin complexes are loaded onto chromatin (for comprehensive review on condensin and chromosome condensation, see [Swedlow and Hirano, 2003]).
Full-size image (65K)
Fig.Â 2. Dynamic localisation of topoisomerase II and condensin during mitosis. Pictures show HeLa cells fixed in methanol and stained with antibodies against topoisomerase II and hCAP-D2.A) Topoisomerase II (red) and condensin localise on the chromosomal axis in metaphase as revealed by immunofluorescence. B) Topoisomerase associates with chromosomes during mitosis, whereas condensin binds to chromosomes from the end of prophase to late anaphase.
View Within Article
3.2. Passenger proteins and chromosome bi-orientation
Passenger proteins are characterised by their behaviour during mitosisÂ : localised onto chromosomes upon entry into mitosis, passenger proteins are enriched in centromeric regions from the end of prophase to metaphase. At the onset of anaphase, they relocalise to the central region of the microtubule spindle (see Figure 3 ). Up to now, four proteins share this particular behaviourÂ : the mitotic kinase aurora B, the inner centromere protein INCENP, the IAP-related (inhibitor of apoptosis) protein survivin and finally, a fourth newly described passenger protein, CSC-1. These proteins seem to be mainly involved in coordinating the behaviour of chromosomes, mitotic spindle and cleavage furrow during mitosis. In particular, the first three proteins are involved in phosphorylation of histone H3, bipolar attachment of spindle microtubules to kinetochores, chromosome condensation and segregation.
Full-size image (22K)
Fig.Â 3. Dynamic behaviour of passenger proteins in mitosis.The pictures show HeLa cells fixed in methanol and stained with antibodies against Aurora B. When mitosis starts, Aurora B localises on chromosomes, mainly at centromeres. At the onset of anaphase, Aurora B dissociates from chromosomes and relocalises to the middle part of the mitotic spindle, and finally binds to the midbody in telophase.
View Within Article
Aurora B is involved in both chromosomal localisation of condensin and histone H3 phosphorylation, two concomitant events that occur during prophase [Giet and Glover, 2001, Kaitna et al., 2002, Losada et al., 2002 and MacCallum et al., 2002]. However, although an attractive hypothesis is that this chromatin modification triggers condensin recruitment, there is no evidence for a causal link between these two events.
To allow sister chromatid segregation at anaphase, spindle microtubules have to bind sister chromatid kinetochores. Among the different ways microtubules can attach to kinetochores (merotelic, syntelic and amphitelic attachment see Figure 4 ), only amphitelic attachment will allow migration of sister chromatids to opposite poles. Microtubules act as "blind kinetochore-catchers". Control mechanisms ensure the correct (amphitelic) attachment of microtubules to kinetochores. One of the major role of aurora B-INCENP-survivin complex is to favour the correct bipolar attachment of spindle microtubules to kinetochores by destabilising incorrect attachments rather than directly inducing amphitelic attachment (see Figure 4Â ; for review see [Tanaka, 2002]. It must be noted that misattached or unattached chromosomes trigger the activation of the mitotic spindle checkpoint that prevent premature separation of sister chromatids. This checkpoint involves regulatory proteins that bind Cdc20, the activator of the anaphase-promoting complex or cyclosome ubiquitin ligase (APC/c). This blocks ubiquitination and thereby degradation of both securin and cyclin B (see belowÂ ; for review concerning mitotic spindle checkpoint, see [Musacchio and Hardwick, 2002]).
Full-size image (9K)
Fig.Â 4. Aurora B favours correct attachments of microtubules to kinetochores by destabilising incorrect ones.During prometaphase and metaphase, microtubules emanating from opposite poles of the mitotic spindle bind kinetochores in a random manner. As only amphitelic attachments are suitable for correctly executed anaphase, incorrect attachments are converted via the Aurora B signalling pathway to monotelic attachments, until they develop amphitelic ones. Until every attachment is correctly aligned, the spindle checkpoint is activated, preventing anaphase and precocious chromosome separation.
View Within Article
3.3. Cohesin release and chromosome segregation
To allow sister chromatids to separate at the onset of anaphase, topological links that persist after replication are resolved by topoisomerase II activity (see above). In addition, cohesin that encircles replicated sister chromatids after replication has to be removed from chromosomes. In vertebrates, two distinct pathways trigger cohesin release during mitosis [Waizenegger et al., 2000 and Sumara et al., 2002].
During prophase, the bulk of cohesin present on chromosome arms is removed by its phosphorylation by Polo-like kinase, probably reducing the affinity of cohesin for chromatin. A small amount of cohesin (about 5%) remains on chromosomes after prophase, mainly on pericentromeric regions. It is not known if the remaining pool of cohesin is insensitive to Polo-like kinase phosphorylation due to differences in composition or to localisation in regions inaccessible to Polo-kinase. This remaining pool of centromeric cohesin participates to the bi-orientation of chromosomes on the mitotic spindle by holding sister chromatids together.
Once all chromosomes are correctly bi-attached to spindle microtubules, the spindle checkpoint is inactivated, thereby allowing anaphase to proceed. The onset of anaphase is determined by the degradation of cyclin B by the 26Â S proteasome (for review see [Peters, 2002]), via its ubiquitination by the APC/c ubiquitin ligase. Dissociation of the remaining cohesin is triggered by the cleavage of cohesin's Scc1 subunit by separase, a cysteine protease. Separase is kept inactive till anaphase by binding to securin, an inhibitory chaperone that is degraded by the APC pathway at the onset of anaphase. Once all cohesin has been removed, poleward-pulling forces exerted by spindle microtubule allow the two sets of chromosomes to move to opposite poles of the cell (for outstanding reviews on segregation see [Nasmyth, 2002] and [Petronczki et al., 2003]).
4. Concluding remarks
Chromosome condensation and segregation are fundamental events that have to proceed correctly in order to ensure that the two sets of replicated chromosomes are faithfully transmitted to each daughter cells.
As mitotic chromosome condensation is linked to DNA replication (replicated chromatin is the template for condensation), it is not surprising that perturbing the latter could have an effect on the former. Indeed, in addition to replication defects, mutations in the origin of replication complex (ORC), as well as in PCNA and MCM4 cofactors, lead to mitotic chromosomes that exhibit abnormal condensation [Pflumm and Botchan, 2001], highlighting the importance of replication on subsequent chromosome condensation, even if it cannot be ruled out that replication defects would de facto lead to condensation defects.
Interestingly, several components of mitotic chromosome condensation seem to be somehow involved in the replication processÂ : AKAP-95, the nuclear matrix protein involved in the chromosomal localisation of condensin, has been shown to interact with MCM2 and its disruption leads to replication initiation defects (Eide et al., 2003). Also, condensin subunit mutants exhibit replication defectsÂ : in fission yeast, genetic studies have shown that mutants of Cnd2, the homologue of XCAP-H, cannot activate the replication checkpoint [Aono et al., 2002]. Other genetic studies in budding yeast have revealed a possible involvement of condensin in replication elongation [Lavoie et al., 2000].
In addition to this crosstalk between condensation and replication machineries, several reports in the literature indicates that condensin may function in additional interphase processes, such as interphase chromatin organisation [Freeman et al., 2000 and Uzbekov et al., 2003] and interestingly, regulation of transcription [Lieb et al., 1998, Lupo et al., 2001 and Bhalla et al., 2002]. It appears then that future studies on chromatin dynamics in mitosis should take into account other cellular processes such as gene expression or replication (for review, see [Hagstrom and Meyer, 2003]. This promises to be an exciting field of research, and will provide an integrated view of chromatin rearrangements within global cellular context in tight relation with cell cycle.(Watrin and Legagneux 2003)
Three checkpoints controlling mitosis
As cells undergo mitosis, their replicated genetic material must be distributed evenly between two daughter cells. In prophase, a bipolar spindle is formed between two microtubule-organizing centers (Fig. 1). Condensed sister chromatids are attached to the spindle via their kinetochores in prometaphase and metaphase, and are pulled to opposite poles as they undergo anaphase. As the missegregation of sister chromatids leads to aneuploidy, this process must be tightly controlled by checkpoints that monitor the completion of critical steps in the pathway. In human cells, prometaphase is delayed when centrosome separation (a prerequisite for the formation of a bipolar spindle) does not take place . The metaphase-to-anaphase (M-A) checkpoint inhibits the separation of sister chromatids until all of the kinetochores are attached to microtubules [2 and 3]. Finally, exit from mitosis is controlled by a Bub2-dependent checkpoint pathway that inhibits the mitotic exit network (MEN) [2, 4 and 5] until the completion of chromosome separation.
Full-size image (15K)
Fig. 1. Schematic illustration of a metaphase arrest after activation of the M-A checkpoint pathway. One chromosome is not attached to the microtubules. The remaining chromosomes are already attached to the bipolar spindle at their kinetochores and are aligned at the metaphase plate.
View Within Article
In this review, we focus on recent advances that explain how the M-A checkpoint arrests cells in mitosis when there are failures in the attachment of the mitotic spindle to the chromosomes. Furthermore, the consequences of loss of mitotic-checkpoint control in mouse and human cells in relation to the development and progression of cancer are discussed.
1.2. Cell-cycle regulation in mitosis
To understand how mitotic checkpoints arrest the cell cycle, a brief review of the molecular mechanisms regulating mitosis is necessary. Key steps in the progression through mitosis are controlled by the destruction of mitotic inhibitory proteins, which occurs when the anaphase-promoting complex/cyclosome (APC/C) ubiquitinates them and targets them for degradation by the 26S proteasome (for reviews, see [6, 7 and 8]). The M-A transition takes place after the degradation of an 'anaphase inhibitor' called Pds1 in Saccharomyces cerevisiae and Securin in mammalian cells (Fig. 2). Loss of Pds1, in turn, leads to the liberation of Separin/Esp1, which causes loss of sister-chromatid cohesion. The APC/C must be in a complex with Cdc20 (Cdc20APC/C) in order to ubiquitinate Pds1. Ubiquitination and degradation of cyclin B is required for the exit from mitosis, and depends on Cdc20 and Cdh1/Hct1 (Cdh1/Hct1APC/C) as co-activators of the APC/C in budding yeast .
Full-size image (27K)
Fig. 2. Cell-cycle control in mitosis. The APC/C drives cells through mitosis by ubiquitinating and thereby targeting key regulatory proteins for degradation by the 26S proteasome (for review, see [6, 7 and 8]). Recent advances both in S. cerevisiae and mammalian cells suggest the following order of events: the APC/C needs to associate with either Cdc20 or Cdh1/Hct1 to become active towards early or late mitotic substrates, respectively. Cdc20 is phosphorylated but phosphorylation does not seem to be important for regulation of its activity. Phosphorylation by Cdk1-associated kinase activity retains Cdh1/Hct1 in an inactive state. Cdh1/Hct1 is activated through dephosphorylation by the phosphatase Cdc14, which is a component of the MEN in S. cerevisiae. Early substrates of Cdc20APC/C in S. cerevisiae are the Securin Pds1 and the S-phase cyclin Clb5. Degradation of Pds1 leads to the liberation of the protease Esp1, which in turn allows separation of sister chromatids by cleaving the cohesion protein Scc1. Overexpression of a non-destructible form of Pds1 blocks sister-chromatid separation. Contradictory to the overexpression result, a deletion of pds1, although behaving normally at low temperature, also delays sister-chromatid separation at elevated temperature. In fission yeast, a deletion in the Securin cut2 as well as expression of a non-degradable form prevents sister-chromatid separation. Therefore, Securins seem not only to inhibit sister-chromatid separation but also to promote it by triggering full activation of the protease Esp1. In accordance with this view, overexpression of Securin, the vertebrate homolog of Pds1, has been shown to transform NIH 3T3 cells [45 and 47]. Mitotic cyclins, whose associated kinase activity is high as cells enter mitosis, need to be degraded to exit mitosis. It has been shown that Clb5 in yeast and cyclin A in Drosophila  are degraded at the M-A transition. Cdh1/Hct1APC/C was thought to confer substrate specificity for late substrates such as Clb2 in S. cerevisiae but it has been shown recently that Cdc20APC/C targets a pool of Clb2 for degradation at the M-A transition . In agreement with results obtained in yeast, degradation of cyclin B in mammalian cells also starts taking place at the M-A transition and has been shown to affect the pools of cyclin B localized to the chromosomes and the spindles . In S. cerevisiae, the decrease of overall Cdk1-associated kinase activity is a prerequisite for activation of Cdh1/Hct1APC/C, which then targets the remaining pool of B-type cyclins for degradation [9 and 50]. This figure illustrates where mitotic checkpoints inhibit cell-cycle progression: the M-A checkpoint inhibits Cdc20APC/C (see text) and the Bub2-dependent checkpoint inhibits the MEN.(Wassmann and Benezra 2001)
Lemke, J., et al. 2002. The DNA-Based Structure of Human Chromosome 5 in Interphase. The American Journal of Human Genetics. 71(5): pp.1051-1059.
Wassmann, K. and Benezra, R. 2001. Mitotic checkpoints: from yeast to cancer. Current Opinion in Genetics & Development. 11(1): pp.83-90.
Watrin, E. and Legagneux, V. 2003. Introduction to chromosome dynamics in mitosis. Biology of the Cell. 95(8): pp.507-513.