Effect of 5-azacytidine and trichostatin A on somatic centromere association in wheat

Transcription

1 399 Effect of 5-azacytidine and trichostatin A on somatic centromere association in wheat Maria Vorontsova, Peter Shaw, Steve Reader, and Graham Moore Abstract: Both homologous and non-homologous chromosomes in wheat associate via their centromeric heterochromatin in the developing xylem vessel cells of the root. The antimetabolite 5-azacytidine (which reduces DNA methylation) decreases the overall level of centromere association. Treatment with 5-azacytidine caused a more marked reduction in the level of homologous chromosome association observed in a wheat line carrying a pair of marked chromosomes. On the other hand, treatment of wheat seedlings with trichostatin A (which increases histone acetylation) raises the overall level of centromere association. The Ph1 locus controls the specificity of both somatic and meiotic pairing of homologous centromeres in wheat. The level of non-homologously associated centromeres is, however, reduced in the presence of Ph1 compared with its absence, even after treatment with either drug. Thus these two drugs, which have been shown to affect chromatin structure, do affect chromosome association, but Ph1 must act at least in part by a different mechanism. Key words: pairing, roots, cereals, Ph1, polyploids. Résumé : Chez le blé, tant des chromosomes homologues qu homéologues se joignent via leur hétérochromatine centromérique dans les cellules conductrices du xylème en développement chez la racine. La 5-azacytidine (laquelle réduit la méthylation de l ADN) diminue de telles associations d une manière globale. Chez une lignée de blé portant une paire de chromosomes marqués, un traitement à la 5-azacytidine a entraîné une réduction plus marquée de l association entre chromosomes homologues. Par contre, le traitement de jeunes plantules de blé avec la trichostatine A (laquelle augmente l acétylation des histones) a induit un accroissement généralisé de l association des centromères. Le locus Ph1 contrôle la spécificité de l appariement des centromères homologues tant mitotique que méiotique chez le blé. Le niveau d association entre centromères non-homologues était cependant réduit en présence de Ph1 comparé à son absence, même après traitement avec l un ou l autre composé. Ces deux composés, qui sont connus pour affecter la structure de la chromatine, ont un effet sur l association des chromosomes, mais le locus Ph1 doit agir, en partie à tout le moins, via un autre mécanisme. Mots clés : appariement, racine, céréale, Ph1, polyploïdes. [Traduit par la Rédaction] Vorontsova et al. 403 Introduction Allopolyploids like hexaploid wheat (Triticum aestivum) not only possess pairs of identical chromosomes (homologues), but also related chromosomes (homoeologues). Wheat homoeologues possess a similar gene order, but differ from each other in their repetitive content (Gale et al. 1998). A single locus, Ph1, in both hexaploid and tetraploid (Triticum durum) wheat, has a major controlling effect on the pairing of the chromosomes, such that pairing is restricted to homologues at meiosis (Riley and Chapman 1958). Thus, hexaploid wheat pairs its chromosomes like a Received 11 March Accepted 17 November Published on the NRC Research Press Web site at on 26 March Corresponding Editor: J.P. Gustafson. M. Vorontsova, P. Shaw, 1 S. Reader, and G. Moore. 1,2 John Innes Centre, Colney, Norwich. U.K. 1 This work was carried out jointly in the laboratories of these authors. 2 Corresponding author ( graham.moore@bbsrc.ac.uk). diploid species at meiosis, which is necessary for maintaining a stable polyploid genome. We have recently shown that the chromosomes of hexaploid wheat, tetraploid wheat, and their polyploid relatives associate via their centromeric heterochromatin before meiotic prophase I during meiosis in the floral tissues and also in developing xylem vessel cells in the roots (Aragon-Alcaide et al. 1997; Martinez-Perez et al. 1999; Martinez-Perez et al. 2001). It has been shown that between 70% and 80% of premeiotic associations in hexaploid wheat are not initially homologous interactions, being either non-homologous or homoeologous (Aragon-Alcaide et al. 1997; Maestra et al. 2002). However, at the onset of meiosis and just before the telomere bouquet formation, the level of association of homologues increases to more than 90% (Aragon-Alcaide et al. 1997). This is coincident with the chromosomes associating into seven groups via their centromeres (Martinez-Perez et al. 2003). However, at this stage in the absence of Ph1, the level of homologue association is less than in its presence, being 50% or under (Aragon-Alcaide et al. 1997; Maestra et al. 2002). This implies that in the absence of Ph1, homologues are not found in the same chromosome group in many meiocytes and may Genome 47: (2004) doi: /G03-138

2 400 Genome Vol. 47, 2004 therefore not be processed in the same manner during meiosis. Moreover, Ph1 increases the ability to resolve incorrect associations during meiosis (Martinez-Perez et al. 2001). The ability to resolve such associations may again be a consequence of how and what chromosomes are processed in each of the seven groups. Ph1 does increase the specificity of interactions made between chromosomes (Martinez-Perez et al. 2001). Thus, this may be the key effect of Ph1 in that it causes a subtle change in how the chromosomes associate via their centromeres premeiotically, which in turn influences how the chromosomes are processed during meiosis. Ph1 affects both the specificity of somatic association in the roots and premeiotic association in the anthers (Aragon-Alcaide et al. 1997; Martinez-Perez et al. 2001). The somatic association of the centromeres in the xylem vessel cells of the roots mimics that occurring between centromeres during premeiotic development in the anthers (Martinez-Perez et al. 2001). Heterochromatin has been previously suggested to play an important role in chromosome pairing (reviewed Renauld and Gasser 1997; McKee 1998). At least three mechanisms have been proposed to explain such pairing of heterochromatin at the molecular level (Karpen et al. 1996). First, it may occur by means of DNA homology searching (similar to recombinational pairing). Second, there may be special pairing proteins that interact with identical sequences on both homologues. Third, higher-order three-dimensional chromatin structure may constitute a unique landscape through which the homologues fit together. The observed association of the centromeres in specific root cells of wheat provides an opportunity to study the pairing process in the absence of other processes that occur during meiosis (such as recombinational pairing). Moreover, root tissue is easier to perform drug treatments on than anther tissue. DNA methylation and histone acetylation affect both chromatin organisation and gene expression (Wolffe and Matzke 1999). Genes that are active are associated with undermethylated chromatin regions. The nucleosomes in these regions contain core histones that exhibit increased acetylation (Wolffe and Matzke 1999; Wolffe and Guschin 2000). Two drugs, 5-azacytidine (5-AC) and trichostatin A (TSA), can be used, respectively, to reduce DNA methylation and increase histone acetylation of chromatin (Kokalj-Vokac et al. 1993; Loidl 1994; Santos et al. 2002). Both drug treatments have been previously shown to cause significant changes in chromatin organization in wheat (Santos et al. 2002). The present study shows that both drugs affect centromere association in wheat, but that Ph1 can act independently of this. Materials and methods Plant material The following plant lines were used in this study: Triticum aestivum Chinese spring ; a T. aestivum Chinese Spring mutant (ph1b; Sears 1977) lacking the Ph1 locus; a T. aestivum line carrying a rye chromosome arm (1RS) translocation attached to the wheat 1BL chromosome arm; two T. aestivum Chinese Spring Aegilops variabilis F 1 hybrids, one with and one without the Ph1 locus; and Triticum monococcum, a diploid wheat relative. Seeds were Table 1. Effect of drug treatment on centromere association. germinated for 4 days at 24 C on filter paper soaked in water alone or water containing 80 µm 5-AC (Sigma, St. Louis, Mo.) or water containing 15 µm TSA (Sigma) (diluted just before use from a 10-mM stock solution in dimethylsulfoxide). The 5-AC solution was freshly dissolved in water and changed daily. A previous study had shown that these concentrations were sufficient to induce hypomethylation and hyperacetylation in wheat seedlings (Santos et al. 2002). Fluorescence in situ hybridization The centromere (CCS1) and telomere (TTTAGGG repeat) probes used, the tissue sectioning and specimen preparation, in situ hybridization and probe preparation, and the labelling of the two homologous rye arms have all been described previously (Martinez-Perez et al. 1999). Fluorescence microscopy and image processing Confocal optical section stacks were collected using a Leica TCS SP or SP2 confocal microscope as described by Martinez-Perez et al. (1999). Confocal images were processed by Adobe Photoshop and the public domain program NIH Image (W. Rasband). The models were created using Object-image 1.62n7 (a modified version of NIH Image; N.O. Visscher) and Rotater 3.5. Results T. monococcum 11.9 (1.6) 11.9 (1.2) 11.9 (1.7) (14 centromeres) T. aestivum Ph (4.8) 20.4 (3.7) 26.9 (4.5) T. aestivum Ph (5.0) 21.4 (4.4) 27.3 (4.4) (42 centromeres) ANOVA P=0.25 P=0.49 P=0.73 Note: ANOVA test for the null hypothesis that the association is the same with and without TSA treatment in the presence of Ph1 (P = 0.004) and its absence (P = 0.008) at 5% confidence can be rejected. The ANOVA test for the null hypothesis that association is the same with and without 5-AC in the presence of Ph1 (P = 0.005) treatment can also be rejected, but accepted in the absence of Ph1 (P = 0.094) at 5% confidence. The ANOVA test for the null hypothesis that association is the same with each treatment in the presence and absence of Ph1 is accepted. The standard deviation is given in parentheses. Examples of the types of images that were counted are shown in Fig. 1. We made an initial survey of what factors might affect the association of centromeres in developing xylem vessel cells of the root. We assessed a range of temperatures for seedling growth and a number of different genotypes. Temperatures from 5 to 37 C did not affect association in any of the following genotypes: euploid wheat ( Chinese spring ), nulli5dt5a, or nulli5dt5b (data not shown). However, germination of the seedlings in the presence of 5-AC or TSA did affect the level of centromere association, with 5-AC having a greater effect than TSA. Treatment with 5-AC The 42 centromeres present in hexaploid wheat associated in the xylem vessel cells, displaying approximately half this number of sites. The level of centromere association in xy-

3 Vorontsova et al. 401 Fig. 1. Centromere behaviour in the xylem vessel cells. Confocal sections of xylem vessel cells showing centromeres (green) in the presence and absence of Ph1, trichostatin A (TSA) and 5-azacytidine (5-AC), showing that the sites are clearly countable. Hexaploid wheat possesses 42 centromeres, the images show centromere association in treated and untreated conditions. The centromere counts for the cells presented are as follows; water Ph1+ (23), water Ph1- (26), TSA Ph1+ (23), TSA Ph1- (20), 5-AC Ph1+ (28) and 5-AC Ph1- (30). Scale bar 10 µm. Fig. 2. Homologue behaviour during root development. A hexaploid wheat root germinated in water and in the presence of 5-AC and sectioned to show the vessel cell nuclei (white arrows), with homologues (green) associated via their centromeres (red arrows) and separated after 5-AC treatment. Scale bar 10 µm. lem vessel cells was significantly lower after treatment with 5-AC (Table 1; Fig. 1). This lower level of association could represent a higher level of homologous, homoeologous, or non-homologous association. The effect of 5-AC on homologous chromosome association was assessed using a wheat line carrying a pair of 1RS 1BL wheat rye translocated chromosomes in which the rye arms can be visualized after hybridization with total rye genomic DNA. The presence of a sub-telomeric heterochromatin knob on the rye (1RS) arm allowed the easy iden- tification of the distal telomere from the centromere region. The association of homologous centromere regions was seen as a V-shaped configuration of the rye arms with the centromeres at the apex and the two heterochromatin knobs at the other end (Fig. 2). 5-AC treatment results in a reduction the level of association of these labelled chromosome arms at their centromeres (Table 2). Any effect of the drugs on non-homologous and (or) homoeologous centromere association was assessed by studying wheat alien hybrid lines. The hybrid line

4 402 Genome Vol. 47, 2004 (T. aestivum Ae. variabilis) contains a haploid set of 21 wheat chromosomes and a haploid set of 14 Ae. variabilis chromosomes; a total of 35 chromosomes, of which none are homologous. Thus the level of centromere association in the xylem vessel cells of this hybrid provides an indication of the level of non-homologous centromere association. Table 3 shows that there are 24 centromere sites visible, indicating significant association of the 35 centromeres. Centromeres will associate even in the absence of their homologous partners. Treatment with 5-AC significantly decreased the level of non-homologous centromere association in this situation. Thus 5-AC treatment reduced the overall level of centromere association, both homologous and non-homologous. Treatment with TSA In contrast to the reduction in centromere association induced in xylem vessel cells with 5-AC treatment, TSA treatment increased centromere association in wheat (Table 1). Moreover, labelling of the two translocated chromosomes showed that there was less homologous centromere association in these roots after TSA (Table 2; Fig. 2). However, TSA did not significantly increase the level of non-homologous centromere association in the xylem vessel cells of the T. aestivum Ae. variabilis hybrids (Table 3). Previous studies have shown that centromere association in the floral tissues and in the xylem vessel cells of the roots occurs only in hexaploid and tetraploid wheat and in polyploid relatives. The diploid relatives of wheat and its wild diploid relatives do not exhibit centromere association. Since TSA increased centromere association in hexaploid wheat under conditions in which association is induced, we determined whether or not TSA treatment could actually induce centromere association in the diploid relative T. monococcum under conditions where no centromere association had been observed. Table 1 shows that the centromeres in the xylem vessel cells of this species remained unpaired even after TSA. Effect of Ph1 The presence or absence of the Ph1 locus did not affect the overall level of centromere association in the wheat root xylem vessel cells after treatment with either TSA or 5-AC (Table 1; Fig. 1). However, previous studies have shown that Ph1 reduces the overall level of non-homologous association (Martinez-Perez et al. 2001). Therefore, we investigated the effect of Ph1 on the level of non-homologous centromere association after treatment with the two drugs. Table 3 shows about 19 sites in xylem vessel cells of untreated roots of the hybrid in the absence of Ph1, and 22 sites after 5-AC treatment. In the presence of Ph1, there are 24 sites in untreated seedlings and 28 in 5-AC treated seedlings. Although 5-AC reduces the overall centromere association in both the presence and absence of Ph1, Ph1 further reduced the non-homologous centromere association (Table 3). TSA treatment did not significantly increase the level of non-homologous centromere association in both the presence and absence of Ph1. Ph1 is still effective in reducing the level of non-homologous centromere association at 7% confidence. Table 2. Effect of drug treatment on homologous centromere pairing. Thus Ph1 was effective irrespective of whether centromere association was increased or reduced by the drug treatments. Discussion % paired chromosomes 29 (4.4) 20.6 (4.1) 16.6 (3.4) χ 2 P=0.02 P<0.001 Note: A total of 960 chromosome pairs were scored in 91 columns of xylem vessel cells. A χ 2 test for the null hypothesis that the pairing levels are the same with and without 5-AC treatment and with and without TSA treatment can be rejected at 5% confidence. Standard deviation is given in parentheses. Table 3. Effect of drug treatment on the number of centromere sites in the presence and absence of Ph1. T. aestivum Ae. variabilis Ph1+ (35 centromeres) ANOVA test compared with water T. aestivum Ae. variabilis Ph1 (35 centromeres) ANOVA test compared with water ANOVA test Ph1+ compared with Ph (2.9) 21.4 (5.4) P= (3.4) 16.7 (2.1) P= (4.9) P= (2.1) P=0.005 P=0.003 P=0.074 P=0.001 Note: The ANOVA test for the null hypothesis that TSA does not have an effect on homoeologous and (or) non-homologous association compared with the control is P = in the presence of Ph1 andp = in its absence at 5% confidence, thus the hypothesis can be accepted. The ANOVA test for the null hypothesis that 5-AC does not have an effect on homoeologous and (or) non-homologous association compared with the control is P = 0.04 in the presence of Ph1 and P = in its absence at 5% confidence, thus the hypothesis can be rejected. The ANOVA test for the null hypothesis that there is no difference at 5% confidence between association in the presence or absence of Ph1 can be rejected for treatments of water and 5-AC at 5% confidence and for TSA at 7% confidence. The standard deviation is given in parentheses. At least three models have been proposed to explain the molecular basis of chromosome pairing, including DNA homology searching, specific pairing proteins that bind to homologous sequences, or proteins that provide a landscape for fitting homologues together. TSA and 5-AC have been shown to cause changes in histone acetylation, DNA methylation, large-scale chromatin structure, and gene expression. The present study indicates that these two drugs also affect chromosome pairing with 5-AC producing a greater effect than TSA. It suggests that chromatin does indeed have a role in pairing of the centromere regions in the very least. Most centromeres associate as pairs in the xylem vessels of roots. Despite this level of association, it could be shown that 5-AC treatment significantly reduces overall

5 Vorontsova et al. 403 centromere association both of homologues and non-homologues, suggesting that it has a general effect on all centromeres. On the other hand, TSA treatment significantly increases overall level of association of wheat centromeres. TSA also significantly reduced the pairing between the two homologously marked centromeres, but did not significantly increase the level of non-homologous centromere associations between wheat and Aegilops variabilis centromeres. There is a substantial reduction of homologous centromere association after treatment with 5-AC and TSA, which suggests that the specificity of pairing at centromeres is affected by chromatin structure, DNA methylation, or gene expression. Since centromeres can associate irrespective of the presence of homologous centromeres, the centromeres could still potentially associate even if homologous centromere associations are disrupted by loss of centromere specificity owing to drug treatment. Thus the more pronounced effect of the drug treatments on specificity rather than overall centromere association is likely to reflect its sensitivity to disruption. A previous study indicated that multiple homologous transgene loci located at a number of sites along a chromosome arm could associate in the interphase nuclei of wheat. Treatment with either 5-AC or TSA reduced the association (Santos et al. 2002). The effects of both drugs on the association of the marked centromeres had a similar effect in reducing the homologous centromere association. Ph1 has been shown to have an effect on chromosome association by increasing the specificity of interactions (Martinez-Perez et al. 2001). The present study does indicate that Ph1 is still effective after the drug treatments, suggesting its activity is indeed independent of modulation of chromatin structure or gene expression. One possibility is that Ph1 is coating the heterochromatin and thereby changing the specificity of interactions made between these regions (Moore 2002). Acknowledgements This work was funded by a Biotechnology and Biological Research Council and Dupont Pioneer Case studentship. The authors would like to thank Dr. Roger Stern from the Statistics Unit, University of Reading, for help and advice on the analysis of the data. References Aragon-Alcaide, L., Reader, S., Beven, A., Shaw, P., Miller, T., and Moore, G Association of homologous chromosomes during floral development. Curr. Biol. 7: Gale, M.D., Atkinson, M.D., Chinoy, C.N., Harcourt, R.L., Jia, J., et al Genetic maps of the hexaploid wheat. In Proceedings of the International Wheat Symposium. 8th ed. Edited by Z.S. Li and Z.Y. Xin. China Agric Scientech Press, Beijing, China. pp Karpen, G.H., Le, M.H., and Le, H Centric heterochromatin and efficiency of achiasmate disjunction in Drosophila female meiosis. Science (Washington, D.C.), 273: Kokalj-Vokac, N., Almeida, A., Viegas-Pequignot, E., Jeanpierre, M., Malfoy, B., and Dutrillaux, B Specific induction of uncoiling and recombination by azacytidine in classical satellite-containing constitutive heterochromatin. Cytogenet. Cell Genet. 63: Loidl, P Histone acetylation: facts and questions. Chromosoma, 103: Maestra, B., de Jong, J.H., Shepherd, K., and Naranjo, T Chromosome arrangement and behaviour of the rye homologous telosomes at the onset of meiosis in disomic wheat 5RL addition lines with and without the Ph1 locus. Chromosome Res. 10: Martinez-Perez, E., Shaw, P., Reader, S., Aragon-Alcaide, L., Miller, T., and Moore, G Homologous chromosome pairing in wheat. J. Cell Sci. 112: Martinez-Perez, E., Shaw, P., and Moore, G The Ph1 locus is needed to ensure specific somatic and meiotic centromere association. Nature (London), 411: Martinez-Perez, E., Shaw, P., Aragon-Alcaide, L., and Moore, G Chromosomes form into seven groups in hexaploid wheat and tetraploid wheat as a prelude to meiosis. Plant J. 36: McKee, B.D Pairing sites and the role of chromosome pairing in meiosis and spermatogenesis in male Drosophila. Curr. Top. Dev. Biol. 37: Moore, G Meiosis in allopolypoids the importance of Teflon chromosomes. Trends Genet. 18: Renauld, H., and Gasser, S.M Heterochromatin: meiotic matchmaker. Trends Cell Biol. 7: Riley, R., and Chapman, V Genetic control of the cytological diploid pairing behaviour of hexaploid wheat. Nature (London), 182: Santos, A.P., Abranches, R., Stoger, E., Viegas, W., and Shaw, P.J The architecture of interphase chromosomes and gene positioning are altered by changes in DNA methylation and histone acetylation. J. Cell Sci. 115: Sears, E.R An induced mutant with homoeologous pairing in common wheat. Can. J. Genet. Cytol. 19: Wolffe, A.P., and Guschin, D Chromatin structural features and targets that regulate transcription. J. Struct. Biol. 129: Wolffe, A.P., and Matzke, M.A Epigenetics: regulation through repression. Science (Washington, D.C.), 286: