Mutagenesis, Vol. 16, No. 2, 109-114,
March 2001
© 2001 UK Environmental Mutagen Society/Oxford University Press
Sequence of centromere separation: effect of 5-azacytidine-induced epigenetic alteration
Departamento Biologia Animal y Genetica, Facultad de Medicina, Universidad del Pais Vasco/Euskal Herriko Unibertsitatea, Lejona Vizcaya, Spain 48940 and 1 Department of Biology/314, University of Nevada, Reno, NV 89557-0015, USA
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and methodsResultsDiscussionReferences |
|---|
The factors which control the sequential separation of the various chromosomes in a genome at the meta-anaphase junction are not well understood. In genomes in which separation is correlated with the quantity of pericentric heterochromatin one factor appears to be the epigenetic nature, namely condensation, of pericentric heterochromatin. When we induced decondensation of pericentric heterochromatin in mouse cells with 10–6, 4x 10–6 and 6x10–6 M 5-azacytidine (5-AC) for 8 h, it resulted in alteration of the sequence of centromere separation. The centromeres which lacked pericentric heterochromatin appeared not to have been affected because there could not be an epigenetic alteration induced by 5-AC. The major effect was on chromosomes with the largest quantity of pericentric heterochromatin. These chromosomes separated at significantly higher frequency than in the untreated population. We also treated human cells, in which separation does not depend upon the quantity of heterochromatin, with 2x10–5 and 6x10–6 M 5-AC for 5 and 8 h. Compared with the control, 5-AC treatment resulted in an increased frequency of separated centromeres of acrocentric chromosomes in relation to those of non-acrocentric chromosomes. In the control the acrocentric chromosomes are the last to separate; in the treated population there was almost random separation of the two types of chromosomes. This epigenetic alteration might be another factor which results in genesis of aneuploidy.
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
At the meta-anaphase junction the centromeres of various chromosomes in a given genome separate into two daughter units in a genetically controlled, non-random manner (Vig et al., 1993
). In the human genome, for example, the chromosomes split at the centromeres in sequence such that chromosomes 18, 17, 2 and 8 separate earliest. These are followed by others, for example chromosomes 1, 9 and 16. The last chromosomes to split at the centromere are the acrocentrics, namely 13, 14, 15, 21 and 22. In mouse and wood lemming the sequence of separation closely follows the quantity of pericentric heterochromatin (Vig, 1982); the greater the quantity of heterochromatin, the later the separation. We argued that the delay in separation associated with quantity of heterochromatin might reflect the sequence in which this region of the chromosome completes replication. This prompted us to look into the timing of replication of the pericentric region. Interestingly, we found that neither in human nor in mouse does there exist a correlation between the timing when a centromere completes replication of its DNA and the order in which the centromeres separate (Garcia-Orad et al., 2000). Apparently, it is not the timing of replication of DNA but some other factor, like the protein composition of the centric region, which might control the timing of separation of a given centromere. The pericentric heterochromatin and the centromere are associated with a host of proteins, e.g. CENPs (centromere proteins), dyneins, dynactins, condensins, cohesins and topoisomerases (for a review see for example Yanagida, 1995). It is not yet clear as to which protein performs which specific function. However, it is known that proteins like heterochromatin protein 1 (HP1) are associated with condensation of the pericentric region. Hamkalo's laboratory (see Hamkalo et al., 1993) carried out competition experiments using Hoechst 33258, a derivative of bis-benzimidazole, and showed that this chemical, as well as some others, can compete with HMG-1 protein. These chemicals, when added during late S phase, inhibit proper reconstitution of the region of the chromosome which carries A:T-rich repetitive DNA, like heterochromatin in mouse. These regions become highly decondensed and fiber-like in appearance.
If the timing of DNA replication is not a factor governing centromere separation, then this process could be governed by epigenetic factors, e.g. association/disassociation of proteins, modification of proteins, such as phosphorylation/dephosphorylation, or even physical changes in the compaction of the centric/pericentric region. We have started looking into the role of compaction of the pericentric region on centromere separation using Hoechst 33258 (Vig and Willcourt, 1998). In these studies we reported that decondensation and elongation of the heterochromatic region generated by Hoechst 33258 results in aberration of the sequence of centromere separation in mouse. The Hoechst 33258 molecules intercalate into the major groove of the A:T-rich heterochromatin but not into the euchromatic region of mouse, which is not as A:T rich. The details of how Hoechst 33258 affects only the heterochromatin and why it causes alterations in the sequence of centromere separation are not well understood.
The alteration in the sequence of centromere separation might result from a direct effect of Hoechst 33258 on the centromere or might result from the epigenetic phenomenon of decondensation of the pericentric heterochromatin. Distinction between these two alternatives is possible by studying the effect of other molecules which can also cause decondensation similar to that induced by Hoechst 33258. One such molecule is 5-azacytidine (5-AC). This chemical, unlike Hoechst 33258, does not bind to the DNA but is incorporated into the replicating DNA, replacing cytidine. In addition, it affects DNA methylation (Haaf, 1995). Even though its mechanism of action is different from that of Hoechst 33258, the result in terms of epigenetic phenomena appear similar. In this communication we report that decondensation of the pericentric heterochromatin of mouse induced by 5-AC results in alterations of the sequence of centromere separation.
|
Materials and methods |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
A-9 and L-929 mouse cells as well as MDA 435 human cells were employed in these studies. The A9 cells (ATCC no. CRL-5319) originated from mouse connective tissue and exhibit a fibroblast-like morphology; the L929 cells (ATCC no. CCL-1) were originally derived from strain L of mouse and are composed of subcutaneous connective tissue; the MDA 435 cells were derived from ductal carcinoma (ATCC no. HTB 129) of the human breast. All cell lines were grown in Iscove's modified Dulbecco's medium (Life Technologies) with 12% newborn calf serum, 100 IU/ml penicillin, 100 µg/ml streptomycin and 292 mg/l Na(HCO)2 supplemented with additional L-glutamine (0.5%). Cells grown without CO2 were treated with 5-AC (Sigma Chemicals) at final concentrations ranging from 10–3 to 10–6 M for periods of between 6 and 24 h. Concentrations >10–4 M resulted in a drastic reduction in mitotic index and, hence, did not yield any usable data, even though decondensation of heterochromatin was observed in some cases. For the mouse cells the best results were obtained with 10–6 and 6x10–6 M 5-AC treatments given for 8 h to A9 cells and 4x10–6 and 6x10–6 M treatment for 8 h given to L929 cells. For human cells, however, concentrations of 2x10–5 and 6x10–6 M applied for 5 and 8 h, respectively, yielded the best results. Since the purpose was to study decondensed heterochromatin, the cells were not treated with a spindle inhibitor, which generally results in recondensation of the extended regions. In order to obtain comparable results, the controls were also not treated with a spindle poison. [It has, however, previously been reported (Figueroa and Vig, 1983
) that application of colcemid does not result in an alteration in the sequence of centromere separation compared with untreated cells.] The cells were fixed in 3:1 methanol:acetic acid mixture. The cells were C-banded by 0.2 N HCl, saturated Ba(OH)2 treatment followed by reconstitution in 2x SSC. Since the karyotype of cells in long-term cultures varies dramatically, we analyzed several hundred chromosomes in order to obtain a valid sample size. The time period for resolution of a centromere into two units (centromere separation or splitting) between the first and the last chromosome to split is very short. Therefore, it was normally difficult to find cells in which only some chromosomes had separated at the centromeres while others were still held as one unit. In these studies the cells used in the analysis had at least two chromosomes separated or at least two chromosomes still unseparated. The same criterion was applied to the 5-AC-treated cells in order to be included in the analysis. The problem with the 5-AC-treated cells was compounded because the chemical appears to drastically delay the separation of centromeres. We tried to resolve the situation by providing a recovery period after 5-AC treatment, but this resulted in recondensation of the pericentric region.
The various chromosomes in the mouse genome, as mentioned earlier, separate in a hierarchical manner depending upon the quantity of pericentric heterochromatin. For mouse cells, therefore, the goal was to study the sequence of centromere separation in relation to the quantity of heterochromatin in the pericentric region. In a previous publication (Vig and Willcourt, 1998) we divided the genome of L929 cells into five categories based on quantity of pericentric heterochromatin. However, we found it difficult to distinguish between Class 1 and 2 chromosomes in A9 cells. Therefore, the C-banded chromosomes in both cell lines were classified into four categories: Class 0 chromosomes were those without any detectable heterochromatin; Class 1 with the smallest quantity; Class 2 with larger blocks of heterochromatin; Class 3 with the largest blocks in the karyotype (Figure 1). In 5-AC-treated cells the quantity of heterochromatin was estimated by the degree of decondensation of this region. C-banding of decondensed heterochromatin is neither effective nor does it generate any additional parameter of distinction between the various categories. These criteria are, unquestionably, subjective and some overlap may be unavoidable. However, it does not appear to cause distortion of the general picture.
|
In no two experiments was the relative number of separated centromeres the same or even similar. This is because in various populations the cells can be, and usually are, at different stages of meta-anaphase separation. This precludes the application of a simple
2 test using the raw data. Therefore, it was necessary to compare the frequency of separated chromosomes in the various treatments in the context of relative separation in the control (see Results).
In human cells the acrocentric chromosomes as a group separate later than the non-acrocentric chromosomes. This pattern of separation is independent of the quantity of pericentric heterochromatin. In these cells, therefore, chromosomes were classified into acrocentric and non-acrocentric and the number of separated versus unseparated chromosomes in each class was recorded. Since it was not the purpose of this study to correlate the degree of extension with the timing of separation in MDA 435 cells, the human cells were not scored for chromosomes with decondensed heterochromatin. Comparisons between the treated and untreated populations were made using the 2 test.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
Centromere separation for this study is defined as the resolution of a centromere into two microscopically distinct units. In untreated populations of L929 cells a sample of 1000 chromosomes was analyzed. The relative frequency of separated/total chromosomes in the various categories was as follows: 60/67 (89.6%) in Class 0, 275/417 (65.9%) in Class 1, 148/315 (47%) in Class 2 and 49/201 (24.4%) in Class 3; the overall frequency of separated chromosomes in all classes combined was 53.2%. The data indicate that the various classes of chromosomes do not separate with equal frequency; the larger the quantity of pericentric heterochromatin, the lower the frequency (and, hence, the later the separation). There is clearly a significant effect of the quantity of pericentric heterochromatin on centromere separation.
The A9 cell line exhibited a similar pattern of separation. The overall frequency of separation was 42.6% (244/573 chromosomes). However, 100% (14/14) of chromosomes in Class 0 showed separation whereas the respective figures for Classes 1–3 were 58.8% (164/279 chromosomes), 26.1% (57/218 chromosomes) and 14.5% (9/62 chromosomes). The two sets of control data indicate that the separation of centromeres is delayed in proportion to the increasing quantity of pericentric heterochromatin. This pattern of separation is in agreement with the literature data (see for example Vig and Willcourt, 1998). In summary, both L929 and A9 cells show a direct relationship between the quantity of pericentric heterochromatin and the delay in timing of centromere separation.
The cells treated with 5-AC exhibit decondensation of heterochromatin (Figure 2). In this study only undercondensed chromosomes were used for analysis. This approach also provides support for a possible correlation between under-condensation and centromere separation in the most decondensed chromosomes (Classes 2 and 3) and those with no or only a little deondensation (Class 0 or 1). In experiments in which decondensation was achieved the centromeres were mostly held as one unit; there appeared to be a drastic delay in centromere separation caused by the treatment. It was, therefore, rare to find cells with only part of the genome having separated at the centromeres. In the cells in which centromere separation was available for study the pattern of separation was not in conformity with what was seen in the control. In contrast to the control, several chromosomes belonging to Class 3 had separated at their centromeres while those in Classes 1 and 2 were still held as one unsplit unit at their centromeres (Figure 3). The data on the pattern of centromere separation in 5-AC-treated cells from four individual experiments, two each for A9 and L929 cell lines, are reproduced in Table I. The overall frequencies of separated centromeres in A9 cells for Classes 0–3, respectively, were 76.6, 53.6, 55.2 and 54.2%, compared with 56.8% separation for all classes combined. The results of 5-AC treatment of L929 cells were similar. In the two experiments a total of 68.1% (832/1222) of chromosomes had separated. This percentage of separated chromosomes is somewhat higher than that reported above for A9, however, the pattern of separation was similar. The separated chromosomes constituted 95.4% of the Class 0 population, 71.1% of Class 1, 63.5% of Class 2 and 56.3% of Class 3. Even though there are some minor differences in the frequencies of separated chromosomes in the two treated populations from A9 and L929 cells, the overall impression is that there is a far higher frequency of separated chromosomes in Classes 2 and 3 amongst the 5-AC-treated populations compared with these classes in the control. If the data from both cell lines are pooled, a similar picture emerges, i.e. a relatively greater frequency of Class 3 and Class 2 chromosomes had separated when compared with the data in the control population. For instance, Class 3 exhibits 55.6% (203/464) chromosomes having separated and the figure for Class 2 is 60.2% (426/708). When these data are compared with the control data, based on the relative separation of Class 0 or Class 1 chromosomes in the control populations, the differences between the control and treated populations turn out to be quite significant. The data in Table I show that decondensation of heterochromatin causes an alteration in the sequence of centromere separation.
|
|
) has separated before one with a smaller quantity (small
|
If an alteration in the sequence of centromere separation results from epigenetic changes generated by 5-AC treatment, then the least affected chromosomes should belong to Class 0. This would mean that this group should show the highest frequency of separated chromosomes, as in the untreated population. This is what is observed. Pooling the data for all controls the frequency of separated centromeres in Class 0 is 91.3%, compared with a frequency of 87.4% for this class for all treated populations. These differences are insignificant (0.8 < P > 0.5,
2 = 0.08).
If epigenetic change is the reason for the alteration in the sequence of centromere separation, then this effect should be most pronounced on centromeres associated with pericentric heterochromatin, i.e. a major shift in separation sequence should be observed for Classes 2 and 3. Superficially it clearly emerges that the frequency of separation of these classes of chromosomes is higher than that in the control. However, a comparison of the absolute frequencies of separation, as discussed for the raw data above, can be misleading since the samples available may be at different stages of the meta-anaphase transition. Hence, statistical comparisons cannot be made using the 2 test on this raw data. We therefore analyzed the data considering the relative frequency of separated chromosomes in various classes in the control and applying these relationships to the treated populations. Since 5-AC does not have an effect on the sequence of separation of centromeres which are not associated with heterochromatin (Class 0), it is logical to compare the relative separation of chromosomes with a smaller quantity of heterochromatin (Class 1) with those with a larger quantity (Classes 2 and 3). The expected frequency of separated chromosomes in the treated populations for Classes 2 and 3 are given in Table I
. The analysis of the data to arrive at the expected frequency is explained by the following example. For the L929 control population the separated chromosomes in Class 1 constitute a fraction of 0.66 (275/417 chromosomes). For Class 3 chromosomes this fraction is 0.24 (49/201). The comparable figures for the cumulative data from the treated populations are 0.71 (303/426) and 0.56 (138/245), respectively. Should the relationship between the separation coefficient for the control be applicable to the treated population, there should have been 63/245 separated chromosomes in Class 3. In other words, if we consider the somewhat higher percentage of separated chromosomes in Class 1 in the treated population than in the control (0.71 versus 0.66) and apply this relationship for separated chromosomes in Class 3 to the treated population, the latter should have 63.4 (0.24/0.66x0.71x245) separated chromosomes in the total population of 245 chromosomes analyzed. However, the actual number of separated chromosomes is 138/245, a figure about twice what is expected. Clearly, the differences are significant (P < 0.005). Based on these parameters, the expected numbers for various other treatments are provided in Table I. At a glance the data show that the larger the extent of heterochromatin, the higher the degree of aberration (Class 3 versus Class 2).
It is possible that the errors in centromere separation are caused by a long delay in the separation of decondensed centromeres, as might be evident from the study of mitotic index. When the mitotic index of the control populations was compared with those of the 5-AC-treated cells it appeared that 5-AC treatment induced a prolonged metaphase. For example, the mitotic index in A9 controls ranged between 1.1 (44/4000 cells) and 1.13% (113/10 000 cells). It was more variable for the 5-AC-treated cells, ranging between 1.5% (36/2400 cells) for 6x10–6 M and 4.8% (202/4200 cells) for 10–6 M treatment. However, a prolonged metaphase period, as also observed with application of colcemid (Figueroa and Vig, 1983), does not per se influence the sequence of centromere separation (see Discussion).
Treatment of MDA435 cells also provided evidence of 5-AC affecting the sequence of centromere separation. The control population of cells in the MDA 435 line was analyzed for separated and unseparated centromeres in the acrocentric as well as the non-acrocentric classes (Table II). In untreated MDA 435 cells the acrocentrics were amongst the last to separate (Figure 4). In this sample, when 35.9% acrocentrics showed separation the corresponding figure for the non-acrocentric chromosomes was 72.6%. In other words, separated acrocentrics comprised only 5.3% of all chromosomes while separated non-acrocentrics constituted 62%. Based on an equal probability of separation for the two types of chromosomes, 148 acrocentric chromosomes should have separated; the observed number was a little less than half. A 2 test of the data shows that the differences are highly significant (P < 0.001).
|
|
In the experiment in which cells were treated with 6x10–6 M 5-AC for 8 h a total of 627 chromosomes were analyzed. Among these, 57.8% acrocentric chromosomes and 71.1% non-acrocentrics exhibited separation. Even though the frequency of separation of non-acrocentric chromosomes in this treatment is about the same as in the control (71.1 versus 72.6%), the differences between the two classes in the control compared with the respective classes in the treated populations are significant. This is because relatively more chromosomes in the acrocentric category have separated in the treated population than in the control (57.8 versus 35.9%). Interestingly, however, the variation observed in the frequency of separation of acrocentrics and non-acrocentrics in the treated cell population is not significant (
2, 0.2 < P > 0.5)
MDA 435 cells treated with 2x10–5 M 5-AC for 8 h yielded 906 analyzable chromosomes (Table II). In this sample, 50% of acrocentric and 57.1% of non-acrocentric chromosomes had undergone centromere separation. These data also show that in these cells the acrocentric and non-acrocentric chromosomes do not differ from each other in the frequency of separation (2, 0.2 < P > 0.5) or that the two groups of chromosomes exhibit random separation.
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
The foregoing data show that a 5-AC-induced epigenetic change (i.e. decondensation of pericentric heterochromatin) in mouse affects the sequence of centromere separation, especially in the chromosomes which harbor a large quantity of pericentric heterochromatin. In human cells, in which the quantity of pericentric heterochromatin is not a defining factor for the timing/sequence of centromere separation, the data show that 5-AC alters the sequence of separation by changing the pattern of separation of acrocentrics from late to early separation in comparison with non-acrocentric chromosomes. Apparently, the details of the effect of 5-AC on the mechanisms of centromere separation in the two species cannot be compared in terms of epigenetic changes; in human it can be done only by comparing chromosomes with differential quantities of pericentric heterochromatin with chromosomes in which the quantity of heterochromatin is uniform, i.e. only with 1, 9 or 16. Nonetheless, acrocentrics in human cells, as a class, have highly repetitive DNA and all have moderate quantities of pericentric heterochromatin. This might be a factor in the alteration of the pattern of centromere separation in MDA cells. Statistical treatment of the data in Table II
shows that 5-AC treatment tends to bring the pattern of centromere separation close to random. The data from mouse as well as human cells support the fact that, by some mechanism, 5-AC alters the sequence of centromere separation.
The two chromatids of a chromosome separate from each other in late prophase. However, the two daughter centromeres remain attached until the meta-anaphase junction. These constraints ensure the high fidelity of chromosome transmission to daughter cells. The biochemical and biophysical control of these events is poorly understood, even though cell cycle regulator molecules, like the cell division kinases, have been implicated in the process. A discussion of many proteins and their effects on chromatid segregation have been the subject of recent reviews (see for example Yanagida, 1995; Hirono, 1998). It has been shown that pericentric heterochromatin plays a crucial role in centromere separation and function in mammals. Even in lower eukaryotes disturbances in the heterochromatin-equivalent segments of the pericentric chromatin disrupt proper chromatid segregation: one example is mutations derepressing silent centromeric domains in fission yeast (Allshire et al., 1995). At least in mouse cells one epigenetic change, namely decondensation of pericentric heterochromatin induced by Hoechst 33258, causes a disturbance in the sequence of centromere separation (Vig and Willcourt, 1998). The present report shows that decondensation induced by 5-AC also results in such aberrations. These two studies support the conclusion that it is the alteration in the physical state of pericentric heterochromatin which causes changes in the sequence of centromere separation and not other possible effects of treatment with 5-AC or Hoechst 33258. The two drugs interact with the chromatin material differently; 5-AC demethylates cytosine residues in the DNA and is incorporated into DNA replacing cytidine, whereas Hoechst 33258 intercalates in the large groove of the DNA. It was also casually noted that 5-AC-treated cells in which the heterochromatin was only minimally decondensed showed no or very little effect on the sequence of centromere separation in mouse.
We have previously shown that the sequence of centromere separation in mouse depends upon the quantity of pericentric heterochromatin: centromeres surrounded by larger quantities are delayed in separation in proportion to the quantity of heterochromatin. We thought that the timing of centromere separation might reflect the timing of replication of the pericentric/centric region; the larger the block of heterochromatin, the later the completion of replication and, hence, the later the separation. It turns out, however, that the timing of replication does not necessarily correlate with the sequence of separation. In mouse, for example, large blocks of heterochromatin do not necessarily exhibit the latest vestiges of replication, as detected by 5-bromodeoxyurine labeling (Garcia-Orad et al., 2000). In human the last separating centromeres belong to groups D and G, yet these are not the last to replicate at the centromere region. If completion of DNA synthesis does not control the timing of centromere separation, then it might be a function of one or more of the host of proteins associated with the centromere region. Whereas there is no sure answer to this question, the presence of phopho epitopes at metaphase and their absence at anaphase (Gorbsky and Ricketts, 1993) brings us a step closer to understanding the dilemma. To the best of our knowledge none of the known changes caused by Hoechst 33258 or 5-AC explain the shift in the pattern of centromere separation. For 5-AC the idea of demethylation of heterochromatic DNA is unappealing since, following common wisdom, such an effect would have to be mediated through transcriptional activity in the centromere region and there is no conclusive proof for that.
The interesting factor emerging in the control of centromere separation is the physical state of the pericentric heterochromatin. In cells treated with Hoechst 33258 or 5-AC decondensation of the pericentric region is the only common factor which appears to exert control over the timing of separation. This epigenetic change might also function through mediation of proteins specific to the centromeric/pericentric region. Changes in HMG-1 binding in the presence of a ligand like Hoechst 33258 are known to occur and might be the result of (or result in) decondensation of the A:T-rich heterochromatin (Hamkalo et al., 1993). These observations point to the significance of physical changes in the various phases of the cell cycle and support the notion that DNA base composition is not necessarily the only, or always the primary, factor in the ultimate outcome of the cell cycle (see Vig, 1998). This fact gains support from the observation that 5-AC-treated cells with pronounced extensions representing large C-bands showed a more aberrant pattern of centromere separation than those with small extensions.
From a practical point of view, such studies are of significance in that phenomena like premature centromere separation have been linked to non-disjunction and, hence, the genesis of aneuploidy. This suggestion was made in the early 1980s from the study of prematurely separating human X chromosomes (Fitzgerald, 1987) and out-of-phase separation of centromeres (Vig et al., 1993). More recently, Angell (1997) has shown a correlation between aberrant centromere separation and non-disjunction in human oocytes. In similar studies, Mailhes et al. (1998) have shown a relationship between premature separation of centromeres and aneuploidy in a population of mouse oocytes. It was in consideration of such correlations that the present study, analyzing the relationship between centromere separation and 5-AC-induced decondensation of heterochromatin, was undertaken.
|
Acknowledgments |
|---|
The authors are grateful to Mr William H.Hallett for help.
|
Notes |
|---|
* To whom correspondence should be addressed. Tel: +1 775 784 6544; Fax: +1 775 784 1302; Email: vig{at}med.unr.edu
|
References |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
-
Allshire,R.C., Nimmo,E.R., Ekwall,K., Javarzat,J.-P. and Cranston,G. (1995) Mutations derepressing silent centromeric domains in fission yeast disrupt chromosome segregation. Genes Dev., 15, 218–233.
Angell,R. (1997) First meiotic division non-disjunction in human oocytes. Am. J. Hum. Genet., 61, 23–32.[Web of Science][Medline]
Figueroa,M. and Vig,B.K. (1983) Sequence of centromere separation: a lack of colcemid effect on the Chinese hamster genome. Cytogenet. Cell Genet., 36, 627–632.[Medline]
Fitzgerald,P.H. (1987) Premature centromere division and aneuploidy. In Vig,B.K. and Sandberg,A.A. (eds), Aneuploidy, Part A. Liss, New York, NY, pp. 249–264.
Garcia-Orad,A., Vig,B.K. and AuCoin,D. (2000) Separation vs. replication of inactive and active centromeres in neoplastic cells. Cancer Genet. Cytogenet., 120, 18–24.[Medline]
Gorbsky,G.J. and Ricketts,W.A. (1996) Differential expression of a phosphoepitope at the kinetochores of moving chromosomes. J. Cell Biol., 122, 1311–1321.
Haaf,T. (1995) The effects of 5-azacytidine and 5-azadeoxycytidine on chromosomes structure and function: implications for methylation-associated cellular processes. Pharmacol. Ther., 65, 19–46.[Web of Science][Medline]
Hamkalo,B,A., Lundgren,K., Radic,M.Z. and Saghbini,M. (1993) Molecular features of heterochromatin condensation. In Vig,B.K. (ed.), Chromosome Segregation and Aneuploidy, NATO ASI Series H: Cell Biology. Springer-Verlag, Vol. 72, pp. 151–164.
Hirono,T. (1998) SMC protein complexes and higher order chromosome dynamics. Curr. Opin. Cell, Biol., 10, 317–322.[Web of Science][Medline]
Mailhes,J.B., Young,D. and London,S.N. (1998) Postovulatory oocyte ageing and aneuploidy. Biol. Reprod., 58, 1206–1210.
Vig,B.K. (1982) Sequence of centromere separation: role of centromeric heterochromatin. Genetics, 102, 795–806.
Vig,B.K. (1998) Centromere: a candidate for face lift. Environ. Mol. Mutagen., 32, 197–199.[Medline]
Vig,B.K. and Willcourt,M. (1998) Decondensation of pericentric hetero- chromatin alters the sequence of centromere separation in mouse cells. Chromosoma, 107, 417–423.[Web of Science][Medline]
Vig,B.K., Paweletz,N. and Schroeter,D. (1993) Review: The centromere–kinetochore complex in cancer. Cancer J., 6, 243–252.
Yanagida,M. (1995) Frontier questions about sister chromatid separation in anaphase. Bioessays, 17, 519–526.[Web of Science][Medline]
Received on May 5, 2000; accepted on October 20, 2000.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  A. Zijno, C. Andreoli, P. Leopardi, F. Marcon, S. Rossi, S. Caiola, A. Verdina, R. Galati, A. Cafolla, and R. Crebelli Folate status, metabolic genotype, and biomarkers of genotoxicity in healthy subjects Carcinogenesis, June 1, 2003; 24(6): 1097 - 1103. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||