Higher eukaryotic centromeres contain thousands of satellite repeats organized into tandem arrays. As species diverge, new satellite variants are homogenized within and between chromosomes, yet the processes by which particular sequences are dispersed are poorly understood. Here, we isolated and analyzed centromere satellites in plants separated from Arabidopsis thaliana by 5–20 million years, uncovering more rapid satellite divergence compared to primate -satellite repeats. We also found that satellites derived from the same genomic locus were more similar to each other than satellites derived from disparate genomic regions, indicating that new sequence alterations were homogenized more efficiently at a local, rather than global, level. Nonetheless, the presence of higher-order satellite arrays, similar to those identified in human centromeres, indicated limits to local homogenization and suggested that sequence polymorphisms may play important functional roles. In two species, we defined more extensive polymorphisms, identifying physically separated and highly distinct satellite types. Taken together, these data show that there is a balance between plant satellite homogenization and the persistence of satellite variants. This balance could ultimately generate sufficient sequence divergence to cause mating incompatibilities between plant species, while maintaining adequate conservation within a species for centromere activity.

---

LARGE tandem arrays of satellite repeats, comprising up to hundreds of kilobases on a single chromosome, are prominent features of eukaryotic centromeres (HENIKOFF et al. 2001). As species diverge, some satellite variants disappear while novel repeats are generated to form a distinct set, or library, of satellite sequences (MESTROVIC et al. 1998; NIJMAN and LENSTRA 2001; PONS et al. 2004). The predominant mechanisms by which individual satellites are amplified and homogenized are not understood. Particular sets of satellite variants could be dispersed through the genome in a homogenization process known as molecular drive (DOVER 1982), or alternatively the maintenance of specific repeats could reflect natural selection that ensures proper centromere function.

Higher eukaryotic centromeres mediate kinetochore formation, promote sister chromatid cohesion, and suppress reciprocal crossing over during meiosis. Centromere satellites were once viewed as "junk" or "parasitic" DNA (ONO 1972; ORGEL and CRICK 1980), and their absence from human neocentromeres called their importance into question (BARRY et al. 1999). More recent studies, however, have compiled strong evidence that satellites contribute to centromere functions (WILLARD 1998, 2001; YANG et al. 2000; SCHUELER et al. 2001; GRIMES et al. 2002; NAGAKI et al. 2003). For example, when tandem arrays of the 171-bp human -satellite repeat are placed on a linear or circular artificial chromosome vector, they promote efficient inheritance (WILLARD 1998, 2001; YANG et al. 2000; SCHUELER et al. 2001; GRIMES et al. 2002); similar results have been obtained using satellites and cell lines from mice (TELENIUS et al. 1999; CO et al. 2000). In addition, chromatin immunoprecipitation experiments in Arabidopsis thaliana have shown that an essential centromere-binding protein, HTR12, associates primarily with the 178-bp satellite repeat that occurs exclusively in the genetically defined centromere region (COPENHAVER et al. 1999; ARABIDOPSIS GENOME INITIATIVE 2000; NAGAKI et al. 2003). HTR12 is homologous to centromere protein A (CENP-A) or Cid, an essential histone H3-like protein that marks the site of centromere function in species ranging from yeast to human (HENIKOFF et al. 2001).

The requirement for partitioning a complete chromosome complement during each mitotic and meiotic cell division places a strong selective pressure on the interactions between satellites and centromere-binding proteins (CLARKE 1998). As satellites evolve, centromere proteins must maintain the ability to bind to satellite sequences, thus creating the need for rapid protein evolution to maintain DNA-protein binding specificity (HENIKOFF et al. 2001). Evidence of positive selection has been detected in centromere-binding proteins from organisms such as Drosophila, grasses, and the plant family Brassicaceae (HENIKOFF et al. 2001; MALIK and HENIKOFF 2001; TALBERT et al. 2002, 2004; COOPER and HENIKOFF 2004). In plants, HTR12 and CENP-C, both of which are essential for centromere function, show evidence of positive selection in regions predicted to be DNA-binding and centromere-targeting domains (TALBERT et al. 2002, 2004; COOPER and HENIKOFF 2004). Understanding the relationships between new satellite variants and evolving centromere-binding proteins will require a facile system that can monitor the effects of sequence variation on chromosome transmission. A. thaliana provides an attractive genetic model for such studies—it can be transformed easily with heterologous satellite arrays and genes encoding centromere-binding proteins, and it offers convenient and quantitative assays for chromosome segregation in both mitosis and meiosis (COPENHAVER 2003).

While it is clear that new satellite variants can become homogenized across a genome, the recombination mechanisms that mediate satellite dynamics remain unclear. In the -satellite DNA of humans, an abundance of recombination breakpoints in the centromere over evolutionary time have been pinpointed to a 20- to 25-bp segment of the -satellite repeat (WARBURTON et al. 1993). These breakpoints may mediate homogenization of satellites over multiple generations of cell divisions by serving as a site that initiates gene conversion, unequal crossovers, or replicative transposition events (DOVER 1982, 2000, 2002; OHTA and DOVER 1983; CHARLESWORTH et al. 1994). Reciprocal meiotic crossovers probably do not play a role, as they are typically absent from higher eukaryotic centromeres (MAHTANI and WILLARD 1998; COPENHAVER et al. 1999). Indeed, analysis of tetrads from 1000 meioses in A. thaliana detected no exchanges of markers flanking the centromere satellites (COPENHAVER et al. 1999).

Our previous studies of satellite variation in 41 populations of A. thaliana showed that the centromere satellites evolve more rapidly than other genomic sequences, including those within the centromeric and pericentromeric regions (HALL et al. 2003). Here, we expanded our analysis to the satellites in closely related Brassicaceae species, a family that has expanded over 40–50 million years (MY) to include >3350 members (AL-SHEHBAZ 1984; KOCH et al. 2001). While other studies of satellites have often relied on sequences amplified by polymerase chain reaction (PCR), we instead sequenced individual bacterial artificial chromosome (BAC) clones derived from libraries corresponding to A. arenosa (5 MY divergence), Olimarabidopsis pumila and Capsella rubella (10–14 MY), and Sisymbrium irio (16–21 MY) (KOCH et al. 2001). This approach allowed an assessment of local variation and, consequently, detection of local homogenization events, as well as a comparison of satellite homogeneity across the genome and between species. Our data indicate that local rates of satellite homogenization exceed global rates and that satellites in these plant families diverge more rapidly than their counterparts in primate genomes.

Satellite identification and sequencing:

Genomic DNA was isolated using a modified CTAB protocol (CSAIKL et al. 1998) and digested with a panel of restriction enzymes to produce a regular pattern of satellite monomers and multimers corresponding to restriction sites within the satellite array. Satellite dimer bands were cloned, sequenced, and used to probe the corresponding BAC libraries from A. arenosa (Aa), O. pumila (Op), C. rubella (Cr), and S. irio (Si) (Amplicon, Pullman, WA). For each species, three BACs with strong hybridization signals were chosen for further analysis, and in S. irio three additional BACs with fainter hybridization signals were also selected. BAC DNA was isolated using a Large-Construct kit (QIAGEN, Valencia, CA) and sheared randomly into 2- to 5-kb fragments (Hydroshear, Genemachines). Recovered DNA was end repaired, cloned into pBluescript to produce a shotgun library, and sequenced to yield 1x coverage (University of Chicago CRC DNA Sequencing Center, Chicago). Cloning efficiencies were initially low for three S. irio BACs; this was resolved by producing shotgun libraries with the cloning vector pSMART (Lucigen). GenBank accession numbers of the resulting sequences are AY640635, AY640924, AY642784, AY642785, AY642786, AY642787, AY642788, AY642789, AY642790, AY642791, AY642792, AY642793, AY642794, AY642795, AY642796, AY642797, AY642798, AY642799, AY642800, AY642801, AY642802, AY642803, AY642804, AY642805, AY642806, AY656017, AY656018, AY656019, AY656020, AY656021, AY656022, AY656023, AY656024, AY656025, AY656026, AY656027, AY656028, AY656029, and AY792367, AY792498.

Analysis of satellite sequences:

Satellite sequences were separated into monomers at a conserved restriction site (Aa, HindIII; Op, HindIII; Cr, DraI; Si, HindIII and BamHI) and aligned with SeqmanII (DNAStar, Madison, WI). Consensus bases were defined as nucleotides that occur three times more frequently than all other nucleotides at that site; sites that did not meet these criteria were considered polymorphic and were noted using the standard IUPAC symbols (HALL et al. 2003). Differences between BAC clone consensus sequences were considered significant if the observed nucleotide frequencies were different from those expected in a ²-test. To measure the conservation of the consensus sequence, the average frequency of the most common base was determined by calculating the consensus base frequency at each site and averaging the values across the entire monomer. Neighbor-joining phylogenetic trees were made with ClustalX and bootstrapped for 100 trials (JEANNMOUGIN et al. 1998). Conserved and variable regions were determined by examining nucleotide occurrence values over a sliding window of 15 bp. Windows were considered significantly conserved or variable by using a z-score analysis (z = [x – µ]/, where x is each window data point, µ is the average of all windows, and is the standard deviation); statistical cutoffs were adjusted so that 20% of the windows were considered significant. If two highly significant windows overlapped, the windows were combined.

Southern blot analysis:

Length variants in O. pumila satellites were detected using a modified tetramethyl ammonium chloride (TMAC) hybridization procedure (SCHENA and DAVIS 1992). Oligos (PUM, TAGAACTTCCAAAMYAACGG; PUM, GAGTGTAGAAAAYGGTTCTA) were end labeled with [-³²P]ATP. To confirm probe specificity, nylon filters were spotted with a dilution series of BAC DNA ranging from 100 to 0.1 ng. BAC DNA was denatured 10 min in 0.1% NaOH before spotting, and filters were UV irradiated to crosslink DNA and incubated 16 hr at 37° in prehybridization buffer (0.5 M sodium phosphate pH 7.2, 7% SDS, 1% BSA, 1 mM EDTA, and 10 µg/ml salmon sperm DNA). After adding the probes, the hybridization temperature was elevated to 42° for 24 hr. Filters were washed three times in 20° 6x SSC/0.05% sodium pyrophosphate and three times at 20° and twice at 54° in TMAC buffer (3 M tetramethylammonium chloride, 50 mM Tris HCl pH 8.0, 0.2 mM EDTA); washes were 15 min each. Filters were exposed on phosphorimager screens for 30 min and visualized using ImageQuant 5.2 (Molecular Dynamics, Sunnyvale, CA). The same protocol was used to screen O. pumila BAC library filters.

FISH analysis:

To test whether the sequenced satellites localized to centromere regions, pachytene chromosomes from O. pumila, C. rubella, and S. irio were probed with species-specific satellite probes labeled with biotin using a nick translation kit (Roche, Indianapolis), and slides were prepared using immature flower buds as described (SCHWARZACHER and HESLOP-HARRISON 1994). Depending on the species, enzyme digestion (5% cellulase/0.5% pectinase) required 60–90 min. The tissue was macerated on the slide, and the cells were spread under a coverslip by gently tapping with a sharpened dowel. After freezing in liquid nitrogen and removing the coverslip, two drops of cold 3:1 EtOH:glacial acetic acid were added to the slide and dried. The slides were incubated at 37° in 2x SSC for 2 min, 2x SSC with 100 µg/ml RNase for 60 min, 10 mM HCl for 5 min, and 10 mM HCl with 5 µg/ml pepsin for 7 min. At room temperature, the slides were washed in dH₂O for 2 min; treated with 2x SSC for 2 min, 4% paraformaldehyde for 5 min, and 2x SSC for 1 min; dehydrated in 70, 80, 95, and 100% EtOH for 2 min each; and dried. Hybridization, wash conditions, and antibody detection were based on WALLING et al. (2005); slides were denatured in 70% deionized formamide/2x SSC at 86° for 90 sec and then rinsed in ice-cold 70, 90, and 100% EtOH for 3 min each. Denatured probe mixture (50% deionized formamide, 2x SSC, 10% dextran sulfate, 2 µg salmon sperm DNA, and 0.1 µg of labeled probe) was added and incubated under a sealed coverslip for 24 hr in a humid 37° chamber. Slides were rinsed in 2x SSC and 50% formamide/2x SSC solutions at 42° for 5 min each, followed by rinses in 2x SSC, 2x SSC/0.1% Tween-20, and 1x PBS at room temperature for 5 min each. Slides were prepared for antibody staining in 4% BSA/2x SSC for 30 min at 37° and rinsed in room temperature 2x SSC; the antibody mixture was added [5 µg/ml FITC-labeled streptavidin Fab fragments (Roche) in 4% BSA/2x SSC], and the slides were incubated in a dark, 37° humid chamber for 45 min and washed at room temperature once in TNT buffer (100 mM Tris, 150 mM NaCl, 0.05% Tween-20) and three times in 1x PBS for 5 min each. Slides were mounted in Vectashield (Vector Labs, Burlingame, CA) and 1 ng/µl DAPI and photographed using an Axioplan microscope (Carl Zeiss, Thornwood, NY) equipped with a Photometrics FxHQ CCD camera.

To determine the genomic location of satellite variants in O. pumila and S. irio, interphase cells from immature flower buds were prepared for FISH. Slide preparation and prehybridization washes were as described above. The probes (BAC DNA for Si; PUM and PUM probes for Op) were fluorescently labeled using a nick translation kit (Molecular Probes, Eugene, OR) and added to the slides in a 30-µl hybridization mixture (50% formamide, 2x SSC, 10% dextran sulfate, 1 µg salmon sperm DNA, and 2.5 µl of each probe mixture). The target and probe were denatured on a heating block at 86° for 90 sec, 76° for 30 sec, 66° for 30 sec, 50° for 30 sec, and 30° for 30 sec and then hybridized overnight in a 37° humid chamber. For S. irio, slides were washed at 42° for 1 min in 2x SSC, 5 min in 0.1x SSC/20% formamide, and 5 min in 2x SSC three times and then at room temperature for 10 min in 1x TBST/1 g/ml DAPI and 5 min in 1x TBS. After O. pumila hybridization, slides were washed in TMAC solution. All slides were mounted in anti-fade solution (Molecular Probes, P-7481) and photographed as described above.

Chromatin immunoprecipitation assays:

Chromatin immunoprecipitation (ChIP) assays were performed with nuclei isolated from 1 g of plant leaves ground to a fine powder and mixed with 10 ml of ice-cold nuclei isolation buffer (10 mM Tris-HCl pH 9.5, 10 mM EDTA, 100 mM KCl, 0.5 M sucrose, 4 mM spermidine, 1.0 mM spermine, 0.1% mercaptoethanol, 0.05% Triton X-100). Ice-cold homogenate was filtered sequentially through 250-, 90-, 50-, and 20-pm nylon mesh, pelleted at 4° at 4000 rpm for 10 min, and resuspended in isolation buffer (without Triton X-100) to a final volume of 0.5 ml; formaldehyde was added to a final concentration of 1%. The nuclei were incubated at 37° for 10 min, washed twice using ice-cold PBS containing protease inhibitors (Sigma, St. Louis), and pelleted at 4° and 2000 rpm for 4 min. Nuclei were resuspended in 200 µl of SDS lysis buffer (Upstate Biotechnology, Lake Placid, NY), incubated on ice for 10 min, sonicated to shear DNA, and centrifuged at 4° for 10 min at 13,000 rpm. The supernatant was transferred to a new tube and diluted 10-fold in ChIP dilution buffer (Upstate Biotechnology) and protease inhibitors. An aliquot was saved as an "input DNA" control. To reduce background, the suspension was incubated with 75 µl of a salmon sperm DNA/protein-A Agarose-50% slurry (Upstate Biotechnology) or salmon sperm DNA/donkey anti-chicken IgY agarose (for CENP-C antibody; Gallus Immunotech) at 4° with agitation for 30 min; agarose was removed by brief centrifugation, and the supernatant was collected. Ten microliters of each immunoprecipitating antibody [anti-trimethyl-histone H3 Lys4 (Upstate Biotechnology), anti-trimethyl-histone H3 Lys9 (Upstate Biotechnology), anti-HTR12 (kindly provided by S. Henikoff), and anti-CENP-C (made using A. thaliana CENP-C synthesized peptides CDVQLNPIPNKRERR and CKQTKGKSNEREEKK)] was added to the supernatant and the solution was incubated overnight at 4° with rotation. To collect the antibody-bound chromatin, 60 µl of salmon sperm DNA/protein-A agarose slurry or salmon sperm DNA/donkey anti-chicken IgY agarose was added to the supernatant and incubated at 4° for 1 hr and then centrifuged at 4° for 1 min at 1000 rpm. The pellet containing the antibody-bound chromatin was washed for 5 min with 1 ml of low-salt immune complex wash buffer (Upstate Biotechnology), high-salt immune complex wash buffer (Upstate Biotechnology), and LiCl immune complex wash buffer (Upstate Biotechnology) and twice with TE buffer. Chromatin was eluted from the antibody complexes by adding 250 µl elution buffer (1% SDS, 0.1 M NaHCO₃), vortexing briefly, and incubating at room temperature for 15 min with shaking. The agarose/antibodies were removed by centrifugation, and the supernatant was transferred to a new tube. The elution step was repeated once and the supernatants were combined. DNA-protein crosslinks were reversed by adding 20 µl 5 M NaCl and incubating at 65° for 4 hr, and DNA was recovered following a standard phenol/chloroform extraction. PCR of isolated DNA was performed with species-specific satellite primers and a noncentromeric control. Primer sequences for satellites were: A. arenosa, forward (F) 5' AGCTTCTTCTTGCTTCTCA, reverse (R) 5' GCTGAGGTTTTGTGATTGG; O. pumila, F 5' TTCCGAAAGCTAAAGTGGTGTG, R 5' ACCGGTTTCATCCCAAGTTC; C. rubella, F 5' GGATAAGAACCCAAAGCCATTG, R 5' GGTTTTGGGTTCTTATCCCTTTTG; and S. irio, F 5' TTGCATGGTTGAGGACTGG, R 5' AAACGGGAATCTGGGACTG. The noncentromeric control primers amplify the rDNA spacer: F 5' GCGGCTAGGAACTGGAACGAGACG, R 5' GCGCGAGCATGGAGCCTACGAA.

Identification of satellites from Brassicaceae species:

A. thaliana has only one major centromere satellite repeat (MARTINEZ-ZAPATER et al. 1986; ARABIDOPSIS GENOME INITIATIVE 2000). This tandemly arrayed 178-bp sequence comprises 2–3% of the genome (ARABIDOPSIS GENOME INITIATIVE 2000) and can be visualized easily with agarose gel electrophoresis after digesting genomic DNA with HindIII, a restriction enzyme that cuts 0–1 times in the satellite and consequently yields a ladder of bands representing satellite monomers and multimers. We used a similar strategy to identify satellites in relatives of A. thaliana; digesting genomic DNA with HindIII (Aa, Op, and Si), BamHI (Si), or DraI (Cr) yielded ladders of satellites similar in size to those found in A. thaliana. Several published lines of evidence indicate that these sequences indeed represent centromere satellites: (i) Like A. thaliana satellites, they are highly abundant in the genome, hybridizing to 1.3% (Cr) to 5.5% (Si) of a BAC library (LUO et al. 2004); (ii) they were extensively methylated, a feature that is restricted to centromeric heterochromatin in A. thaliana (LUO et al. 2004); and (iii) in previously published FISH experiments, they hybridized to a unique site on each chromosome [KAMM et al. 1995 and HESLOP-HARRISON et al. 2003 (Aa); LUO et al. 2004 (Cr); HESLOP-HARRISON et al. 2003 (Op)].

To verify the centromeric origin of the satellites, we performed FISH on pachytene chromosomes using a species-specific satellite probe for O. pumila, C. rubella, and S. irio (Figure 1). A. arenosa was not tested, as its satellites share 73% identity to the centromere satellites of A. thaliana (see below). Each satellite probe hybridized to a DAPI-bright region of heterochromatin, consistent with localization to the centromeric regions. Similar patterns were observed on numerous slides containing pachytene spreads. For both O. pumila and C. rubella, every DAPI-bright region in the genome overlapped with the satellite probes (Figure 1, C and F). In S. irio, some DAPI-bright regions did not hybridize to the satellite probe (Figure 1I), reflecting the highly variable nature of the satellites in this species (see below).

View larger version (78K): In this window In a new window Download PPT slide   FIGURE 1.— Colocalization of satellite probes with DAPI-stained heterochromatin. Meiotic pachytene bivalents are shown from O. pumila (A–C), C. rubella (D–F), and S. irio (G–I). DAPI-stained chromosomes (A, D, and G), FITC-labeled (green) species-specific satellite probes (B, E, and H), and merged DAPI-stained and satellite probed images (C, F, and I) are shown.

 
To assess whether the satellites we identified are associated with the biochemical markers characteristic of heterochromatic or centromere regions, we performed ChIP assays in triplicate (Figure 2). Antibodies to the tri-methyl-histone H3 K9, tri-methyl-histone H3 K4, and the essential centromere proteins HTR12 and CENP-C were used to isolate heterochromatin, euchromatin, and centromere DNA, respectively. For all species, the satellites we isolated were enriched in the DNA precipitated with tri-methyl-histone H3 K9 and CENP-C antibodies, indicating that these satellites are located in heterochromatin and at the active centromere (Figure 2). As expected, the HTR12 antibody did not immunoprecipitate satellites from these species above background levels because it was directed against the rapidly evolving N-terminal tail of the A. thaliana protein (TALBERT et al. 2002). The satellites were also not immunoprecipitated above background levels by the euchromatic marker, tri-methyl-histone H3 K4 antibody (Figure 2). As a noncentromeric control, we also examined the immunoprecipitation profile of the rDNA spacer region. This DNA was not present above background levels when precipitated with HTR12 and CENP-C antibodies, but was associated with the euchromatic marker in A. arenosa and both euchromatic and heterochromatic markers to varying degrees in the other species. Thus, the abundance, methylation content, association with condensed chromosomal regions, heterochromatic histones, and CENP-C strongly suggest that the satellite repeats we isolated reside in centromeric arrays. Although the FISH results indicate localization of satellites to the centromere, the possibility remains that some satellites are also located elsewhere in the genome; further analysis is required to see if these species are analogous to A. thaliana, where the 178-bp satellites occur exclusively in the genetically defined centromeres (COPENHAVER et al. 1999; ARABIDOPSIS GENOME INITIATIVE 2000; NAGAKI et al. 2003).

View larger version (73K): In this window In a new window Download PPT slide   FIGURE 2.— Chromatin immunoprecipitation (ChIP) experiments show satellites associated with centromere-binding protein and modified histones. Immunoprecipitation of chromatin from A. arenosa, C. rubella, O. pumila, and S. irio was performed using antibodies to HTR12, CENP-C, Histone H3 Lys4, and Histone H3 Lys9. Primers for species-specific satellite repeats and rDNA were used to amplify DNA from immunoprecipitations using each antibody, immunoprecipitation using no antibody, and input DNA. For S. irio, the rDNA control shows clear bands in one HTR12 and one histone H3 Lys9 immunoprecipation reactions; the remaining reactions show an absence of a band or nonspecific DNA smear. Three independent PCR reactions were performed for each experiment.

 

Analysis of Brassicaceae satellite sequences:

To characterize the satellite sequence content of the four Brassicaceae species, we isolated BAC DNA identified by library hybridization, confirmed an abundant satellite content by restriction digestion analysis, and performed DNA sequencing (1x coverage) of three clones each from Aa, Cr, and Op and six clones from Si. Aligning the satellite monomers yielded a consensus sequence for each BAC clone and each species (Figure 3, Table 1). Our previously derived satellite consensus from 41 A. thaliana populations showed that each position within satellite monomers randomly amplified with PCR had an average nucleotide conservation of 90.3%, that pairs of monomers shared an average identity of 83.4%, and that satellite length was highly conserved, averaging 178 ± 0.01 bp (HALL et al. 2003). Thirteen of the 15 BAC clones sequenced here had an average satellite nucleotide conservation that was higher, ranging from 90.8 to 96.9%. The two remaining BACs (Si 51B6 and Si52G1) were more polymorphic (Figure 3) and had a nucleotide conservation of 85.7 and 84.2%, respectively (Table 1). In addition, the pairwise identity of satellite monomers in each of the 15 BACs was elevated relative to that of randomly amplified A. thaliana satellites, measuring 85.6 to 95.0%. These results suggest a greater level of conservation over localized genomic regions than across the species as a whole; indeed, when the BAC data sets were combined to generate a species consensus, nucleotide conservation dropped in each species, ranging from 91.6 to 95.7%, and satellite identity similarly fell, ranging between 82.7 and 90.9%.

View larger version (52K): In this window In a new window Download PPT slide   FIGURE 3.— Consensus sequences and alignments of Brassicaceae species. (A) Centromere satellite consensus sequences of BAC clones from each individual Brassicaceae species. The consensus sequence for the species is shown on top. Statistically conserved (C) and variable (V) regions are indicated by open boxes (HALL et al. 2003); consensus nucleotides that differ significantly from the species consensus are shaded gray. Aa, A. arenosa; Cr, C. rubella; Op, O. pumila; Si, S. irio. (B) Interspecies satellite alignments, showing the sites of conservation (black) and variability (gray) as indicated in A.

 

View this table: In this window In a new window   TABLE 1 Satellite properties and conservation in Brassicaceae species

 
Satellite monomer lengths were notably consistent among the BACs from A. arenosa and C. rubella (averaging 178.7 ± 1.7 and 168.4 ± 1.0 bp, respectively), but were more variable in the other species. In O. pumila, two different, yet highly conserved monomer lengths were identified: BACs 33O5 and 48F7 encoded a 178- to 179-bp monomer, whereas BAC 48K10 contained a 169-bp monomer that represented a 9- to 10-bp deletion of the longer form (Figure 3). These dimorphic forms were noted previously (HESLOP-HARRISON et al. 2003), but their abundance and chromosomal locations were not investigated. Because most genomes contain satellites that differ by nucleotide substitutions rather than by polymorphic deletions or insertions (BRUTLAG 1980; HENIKOFF et al. 2001; HALL et al. 2003), we examined the genomic distribution of this size polymorphism in more detail (see below). Additionally, four of the S. irio BAC clones showed unusually large variations in monomer length. In BAC Si51J1 this variation resulted from a 47-bp insertion into a single monomer, and in the other Si BACs (51B6, 51C15, and 52G1) the length variation resulted from numerous single-base insertions and deletions. Nonetheless, the consensus sequences for all of the Si BACs were highly similar in length (218 bp in 50E21 and 219 bp in the remaining BACs; Figure 3 and Table 1). Most species studied to date have a monomer consensus that is more consistent in length than our observations in S. irio (HENIKOFF et al. 2001; UGARKOVIC and PLOHL 2002). This raises the possibility that some genomes undergo less efficient homogenization, a phenomenon that could be enhanced in S. irio due to the relatively larger size of its genome (Table 1).

Local homogenization of satellites:

Our data indicate that satellite monomers in each species have a higher degree of similarity when derived from an individual BAC clone than when compared across multiple BACs (Table 1). To examine this difference in more detail, we derived phylogenetic trees for the satellites from each species (Figure 4). For A. arenosa, C. rubella, and O. pumila, satellite monomers from each BAC grouped into their own distinct clades (Figure 4, A–C); only four satellite monomers from these three species (two from Cr and two from Op) did not follow this pattern. The nucleotide changes that account for these distinct groupings are evident in the consensus sequences (Figure 3), with individual BACs containing unique sequence differences (14 in Aa and 4 in Op, as well as a 10-bp deletion) and significant polymorphisms (10 in Aa, 17 in Cr, and 15 in Op).

View larger version (20K): In this window In a new window Download PPT slide   FIGURE 4.— Phylogenetic trees comparing centromere satellites. Satellites from each BAC clone are represented by a different color as follows: (A) A. arenosa 8M8 (red), 8C6 (green), and 1E13 (blue); (B) O. pumila 48F7 (red), 33O5 (green), and 48K10 (blue); (C) C. rubella 5M1 (red), 7A5 (green), and 31F1 (blue); and (D) S. irio, type I, 51J1 (light blue), 49C3 (green), 51C15 (orange), and 50E21 (purple) and type II, 52G1 (red) and 51B6 (blue). Arrows highlight examples of satellites that reside next to each other in the genome, yet are located in separate clades. The numbers indicate the order of monomers in the clone.

 
The phylogenetic tree of S. irio showed a different pattern: the three BACs we initially selected (49C3, 51C15, and 51J1) had only two significant differences between their consensus sequences and, consequently, were not sorted into distinct clades. We considered two explanations for this observation: Either we inadvertently selected three BACs that were derived from the same genomic region, or S. irio is more efficient at homogenizing satellite monomers across its genome. The former possibility seemed more likely for the following reasons. First, the high degree of monomer length polymorphism (see above) suggests less efficient satellite homogenization in this species, and second, the probe used to screen the S. irio library hybridized more strongly to the BACs we selected for sequencing than to other S. irio satellite-containing BACs. FISH results supported this interpretation (see below): The probe hybridized to every centromere, but more strongly to a subset, suggesting the presence of a homologous, yet divergent satellite sequence. Accordingly, we sequenced three additional Si BACs (50E21, 51B6, and 52G1) that hybridized more weakly to the probe. Whereas BAC 50E21 had a consensus sequence that was highly similar to that observed before (type I satellite, Figure 4D), BACs 51B6 and 52G1 contained a distinct repeat (type II). Placing these sequences into the phylogenetic tree sorted monomers from 50E21, 51B6, and 52G1 into separate clades that were distinct from the clade interspersed with monomers from 49C3, 51C15, and 51J1 (Figure 4D). Consequently, the overall satellite consensus for S. irio, when compared to the other species, was derived from a more diverse set of satellite monomers. Individual S. irio BACs contained 136 conserved polymorphisms and 78 sequence differences from this consensus (Figure 3). Taken together, our data show that phylogenetic groupings in all four species provide clear evidence for sequence homogenization events over a small genomic region (Figure 4). In the cases of O. pumila and S. irio, these local homogenization events maintain highly distinct satellite classes confined to a subset of the satellite-containing BACs.

Although our results strongly point to efficient local sequence homogenization, the satellite repeat arrays of A. arenosa, C. rubella, O. pumila, and S. irio nevertheless maintain significant variation. While the satellites of a particular BAC are organized into the same large clade in the phylogenetic trees, monomers that are adjacent to each other often reside in 2–4 different subclades (Figure 4, arrows). Such an arrangement is indicative of a higher-order repeat array structure. Adjacent satellite monomers fell into the same subclade only in rare cases (12 instances in the 330 clones analyzed). In humans, higher-order arrays range from 4 to 35 monomers (CHOO et al. 1991); similarly, our data indicate a higher-order repeat structure in these species that is constructed from the 2–11 subclades present in each BAC (Figure 4). The challenges of assessing array structure from sequencing reads that typically span only 4 monomers make it difficult to determine the organization of these higher-order arrays. Analysis of restriction site periodicities, or other approaches that can measure physical continuity, will be necessary to determine the exact composition of higher-order repeats in these species.

Genomic distribution of satellite variants:

Although chromosome-specific -satellite variants have been found previously in humans and other primates, it is not clear how such variants are homogenized on a genome-wide level (WILLARD 1985; WAYE and WILLARD 1987; WILLARD and WAYE 1987a; ALEXANDROV et al. 1988; BALDINI et al. 1991; CHOO et al. 1991; WARBURTON et al. 1996; HAAF and WILLARD 1997). The detection of two distinct satellite lengths in O. pumila (179 and 169 bp) and two satellite types in S. irio afforded us an opportunity to examine the distribution of dimorphic satellite forms and discern whether they are located on different chromosomes, present in distinct regions of the same centromeres, or, alternatively, uniformly mixed throughout all centromeric regions. For O. pumila, we designed 20-mer oligonucleotide probes that distinguish between the satellite forms and verified their specificity on a dot blot of the sequenced O. pumila BACs (Figure 5, A and B). To assess the genomic abundance of each satellite variant, we used these oligonucleotides to probe the O. pumila BAC library. This BAC library contains inserts averaging 115 kb and sufficient clones for 5x coverage of the O. pumila genome. Of the 9216 clones on the filter, 199 hybridized to the full-length probe and 296 clones to the probe corresponding to the deletion (2.1 and 3.2% of the library, respectively). Remarkably, only one clone hybridized to both probes, ruling out the possibility that the two variants are highly interspersed and suggesting that this BAC clone represents either a rare junction between these satellite variants or, alternatively, a chimeric clone produced during library construction.

View larger version (37K): In this window In a new window Download PPT slide   FIGURE 5.— Genomic distribution of O. pumila satellite variants. (A) Probe sequences (PUM and PUM) used for BAC library hybridization and FISH experiments are boxed; only the relevant sequence of the centromere satellite is shown. (B) Dot blots of a dilution series of BAC DNA confirm the probe specificity; PUM hybridizes only to the 169-bp length variant; PUM hybridizes to the 179-bp satellite and to a lesser extent to the 178-bp variant. (C–N) FISH analysis of chromosomes from O. pumila flower bud cells, probed with PUM and PUM. (C, G, and K) DAPI staining. (D and H) Chromosomes probed with full-length copies of the O. pumila satellite that labels both sequence variants. (E and L) Chromosomes probed with PUM. (I and M) Chromosomes probed with PUM. (F, J, and N) Merged images showing that PUM and PUM are present on only a subset of centromeres. Arrowheads, centromeres containing only the 179-bp (PUM) variant. Bars in F, J, and N, 10 µm.

 
To observe the distribution of the satellite variants on individual O. pumila chromosomes, we performed FISH, using the same 20-mer oligonucleotide probes and highly stringent hybridization conditions (Figure 5, C–N). Both the probe corresponding to the 169-bp satellite variant (Figure 5, C–F) and the probe recognizing the longer variant (Figure 5, G–J) failed to hybridize with the same intensity to every centromere. When chromosomes were labeled with both probes (Figure 5, K–N), merged fluorescent images showed that most centromeres contain both variants in significant amounts, but two centromere foci were predominantly labeled with the probe corresponding to the longer variant (Figure 5N, arrowheads). For the other centromeres, the relative abundance of each probe signal varied (Figure 5N, yellow indicates approximately equal quantities of both probes; orange represents a higher abundance of the deletion variant). Of the 30 chromosomes, 20 were enriched in the 169-bp variant and 8 contained more of the longer variant, consistent with the number of BACs that hybridized to each probe. These results are consistent with previous observations showing differential FISH signal strength among chromosomes when full-length centromere monomers were used as probes (HESLOP-HARRISON et al. 2003). Because large quantities of both the 169- and 178/179-bp variants rarely reside in the same BAC clone, these results suggest that centromeres are composed of arrays of each satellite, with stretches of nonsatellite sequence residing between the size variants that are larger than the size of a typical BAC insert. O. pumila is a tetraploid; however, whether this species is an autotetraploid or allotetraploid is unclear (I. A. AL-SHEHBAZ, personal communication). The presence of significant amounts of two-centromere satellite variants suggests it arose by hybridization of two plants that carried different satellites, followed by dispersal of the satellite variants to nonhomologous chromosomes via gene conversion or replicative transposition mechanisms (see DISCUSSION). Alternatively, an insertion/deletion event could have occurred spontaneously, and the O. pumila genome may be in the process of homogenizing its genome to ultimately eliminate one of these satellite variants.

We also performed FISH using the sequenced S. irio BACs as probes to investigate the genomic relationship between type I and II satellites. Type I BACs 49C3, 51C15, and 51J1 hybridized strongly to only two chromosomes, with weaker hybridization to other centromeres (Figure 6, B, F, and J), indicating they were likely derived from a single centromere. DNA fingerprinting indicated the three type I BACs are nonoverlapping (data not shown), and analysis of the unique nonsatellite DNA sequences from each BAC identified only one shared 149-bp sequence between Si 51C15 and Si 51J1. Thus, the high degree of satellite similarity, coupled with the apparent lack of overlap between unique sequences on these BACs, suggests that the satellite population of this centromere is homogenized over adjacent regions. Although additional mapping will be required to dissect the organization of this centromere, it may differ from human -satellite arrays, where the satellites within a centromere can diverge as much as 30% (WARBURTON and WILLARD 1990; SCHUELER et al. 2001). BAC 50E21 contained predominately type I satellite, but its monomers were sorted to a distinct clade (Figure 4) and it hybridized to six centromeres, two of which overlap with the strong signal from 49C3, 51C15, and 51J1 probes (Figure 6, A–D). Type II BACs 52G1 and 51B6 each hybridized to chromosomes different from those identified by type I probes (Figure 6, E–L). These results indicate that the S. irio genome contains chromosome-specific satellite variants and that homogenization occurs at a higher rate on a given chromosome than between arrays on nonhomologous chromosomes.

View larger version (20K): In this window In a new window Download PPT slide   FIGURE 6.— Genomic distribution of satellite variants in S. irio. FISH analysis of chromosomes from S. irio flower bud cells is shown, using satellites from BACs as probes. (A, E, and I) DAPI staining. (B, F, and J) Chromosomes probed with type I satellite from 49C3, 51C15, and 51J1. (C, G, and K) Chromosomes probed with type I satellites (C) 50E21 and type II satellites (G) 51B6 and (K) 52G1. (D, H, and L) Merged images showing that type I and type II satellites are located on different chromosomal subsets. Bars in D, H, and L, 10 µm.

 

Satellite evolution between Brassicaceae species:

The rapid divergence of satellite sequences often makes relationships between species undetectable. Here, we targeted closely related plants in an effort to better understand the mechanisms behind satellite divergence during speciation. We performed pairwise alignments between each of the satellite consensus sequences of A. thaliana and the Brassicaceae species in this study (Figure 3) and determined percentage of identity (Figure 3B; Table 2). To align A. thaliana and A. arenosa, three gaps were introduced, producing a sequence identity of 73.3% (Figure 3B). A. thaliana and O. pumila have a significant alignment, spanning only 87 bp with a sequence identity of 65.5%; O. pumila and A. arenosa similarly share 65.8% sequence identity, yet have a longer significant alignment of 154 bp (Figure 3B). The remainder of the comparisons between other species yielded no significant overlap, indicating that these satellite sequences have diverged considerably (see below, Table 2). These species-specific differences point to efficient homogenization mechanisms that cause the sequences within a species to share a higher sequence identity than they do with their orthologous sequences in near relatives, an observation known as concerted evolution (DOVER 1982; ELDER and TURNER 1995). Even A. thaliana and A. arenosa, which diverged within the past 5 MY, have highly distinct centromere satellite repeats, indicating that species divergence has been followed by rapid satellite divergence and efficient homogenization.

View this table: In this window In a new window   TABLE 2 Satellite identity in closely related species

 
Interestingly, the highly conserved regions identified in the satellite consensus from each species were not conserved across A. thaliana, A. arenosa, and O. pumila. In fact, the highly variable V1 and V3 regions from A. thaliana and A. arenosa, respectively, overlap with the conserved O. pumila C1 and C2 domains. Initially we postulated that these conserved regions could correspond to specific protein-binding sites (HALL et al. 2003); however, the lack of conservation of these regions between species suggests either protein-binding adaptation on a very rapid timescale or an alternative function of the conserved regions. For instance, conserved regions could provide the necessary sites of homology for gene conversion or other types of exchange events within the satellite population of the genome.

Compared to published sequence data from primates, the plant satellites examined here are evolving at a faster rate. For example, the 5-MY separation between A. thaliana and A. arenosa is similar to the separation between human and chimpanzee, yet the plant satellites share 73.3% identity (Table 2), while the primate satellites are 84–91% identical (BALDINI et al. 1991; WARBURTON et al. 1996; HAAF and WILLARD 1997; GLAZKO and NEI 2003). Likewise, A. thaliana and O. pumila have been separated as long as human and orangutan, yet show much less satellite identity (Table 2, 65.5% vs. 92%, respectively) (GLAZKO and NEI 2003). Even New World monkeys and humans, separated by 32–36 MY, share more satellite similarity than plant genera that diverged over the last 10 MY. Clearly, plant satellites, while potentially modulated by the same genetic mechanisms of recombination that are active in other species, are evolving at a much faster rate. This difference in the rate of satellite divergence could also reflect the differences in life cycles, the vastly different number of cell divisions required to produce gametes, a difference in ploidy (A. arenosa is likely an autotetraploid and O. pumila and S. irio are tetraploids of unknown origin), or a differential impact of mutagens on DNA repair processes.

To elucidate the genetic mechanisms that mediate the divergence of centromere satellites, we sequenced satellite-containing BACs from four closely related plant species: A. arenosa, O. pumila, C. rubella, and S. irio. Our results show that satellite homogenization occurs more rapidly at a local level compared to homogenization across the genome. Even so, regional satellite variants were maintained, forming higher-order repeat arrays in each species and significant regional variation in the O. pumila and S. irio genomes. In O. pumila, two length variants of a highly similar satellite sequence resided in arrays at most centromeres, but were generally not contained on the same BAC and thus were separated by DNA segments >100 kb in length. In S. irio, divergent satellite variants reside at many centromeres, but are preferentially localized to a subset of chromosomes. As described below, these observations require models that can explain the dispersion of satellite arrays. In addition, our results indicate that satellites in this plant family evolve at rates that are more rapid than those observed in primate species. Rapid divergence may take place soon after speciation, followed by homogenization that generates uniform satellite content across the genome.

Genetic mechanisms with the potential for driving centromere evolution include unequal crossover, gene conversion, and satellite transposition; each of these could occur intrachromosomally, between sisters of the same chromosome, between homologous chromosomes, or between nonhomologous chromosomes (CHARLESWORTH et al. 1994). These exchanges may preferentially initiate within conserved domains; for example, the 20- to 21-bp region that mediates recombination breakpoints in human satellites directly overlaps a conserved region we previously described (WARBURTON et al. 1993; HALL et al. 2003). In humans, unequal crossovers between misaligned arrays of tandem repeats on sister chromatids and homologous chromosomes create higher-order repeat units and variations in array length (WILLARD and WAYE 1987b; WEVRICK and WILLARD 1989) that are predicted by an unequal crossover model (SMITH 1976). This model also predicts that satellite monomers located at the edges of arrays are more divergent than the array core, due to the lower frequency of being included in an unequal crossover event (SMITH 1976). Indeed, examination of human -satellite arrays on chromosome X strongly supports the predominant role of unequal crossover in centromere satellite evolution (SCHUELER et al. 2001). The observations of varying satellite array lengths (HOSOUCHI et al. 2002) and higher-order satellite repeats in plants also support a model in which unequal crossovers have a prominent role in centromere satellite evolution (SMITH 1976). However, such exchanges would likely involve sister chromatids, given the extremely low rate of recombination between homologs during meiosis in A. thaliana (COPENHAVER et al. 1999).

Several theoretical simulations, however, argue that unequal crossover alone cannot account for satellite evolution; gene conversion, satellite transposition, or amplification mechanisms would also be needed to expand satellite arrays removed by unequal crossovers (STEPHAN 1986; WALSH 1987). While selection for a minimal array size could counteract satellite loss due to unequal crossovers (YANG et al. 2000; GRIMES et al. 2002), some studies of human satellite variants provide evidence that additional recombination processes, such as sequence conversion, also play a role in centromere satellite evolution (WARBURTON et al. 1993; SCHINDELHAUER and SCHWARZ 2002). Our results are also consistent with a model in which sequence conversion events preferentially take place within a chromosome (intrachromosomal), between sister chromatids, or between homologous chromosomes. Our analysis of individual BAC clones showed that satellites are homogenized locally, which is similar to the local homogeneity found in human -satellite arrays on chromosomes 17 and X (WARBURTON and WILLARD 1990). Thus, conversion events likely occur most often intrachromosomally or between sisters and less frequently between homologs and nonhomologs. An alternative mechanism, satellite transposition, has been advocated by some studies as another important feature in the evolution of higher-order satellite repeats in humans (ALEXANDROV et al. 1988; ALKAN et al. 2002). Extrachromosomal circles of centromere satellite DNA, formed by intrachromosomal recombination, have been isolated in human cell lines and Drosophila and could potentially mediate satellite transposition (JONES and POTTER 1985; OKUMURA et al. 1987; COHEN et al. 2003). While these circles could serve to eliminate diverged, nonfunctional satellites from the genome (JONES and POTTER 1985), they also could reinsert into regions of homology to spread satellite variants to nonhomologous chromosomes (WALSH 1987).

Our data indicate that Brassicaceae species have a robust process for genome-wide homogenization of satellite monomers. Nonetheless, because we observed select variants located in chromosome-specific arrays or on distinct DNA segments, the rate of exchange throughout the genome must be less frequent than the homogenization that is restricted to localized regions. In O. pumila, satellite length variants were dispersed on many chromosomes (Figure 5N) but were rarely interspersed; only one BAC clone contained both sequence variants. In S. irio, three BACs derived from the same chromosome contained nearly identical type I satellites, suggesting broad homogenization of these variants across the centromere, but predominant localization on a single chromosome (Figure 6). In their study of gene family dispersal, OHTA and DOVER (1983) concluded that the relative frequency of gene conversion events between nonhomologous chromosomes decreases when overall conversion rates are low or when the number of chromosomes containing the sequences is high. The species we analyzed here vary approximately twofold in genome size and threefold in chromosome number (Table 1), which creates the potential for differential genome homogenization rates (OHTA and DOVER 1983). Indeed, S. irio has the largest genome and the highest degree of satellite variation, while C. rubella has the smallest genome and satellites that are more homologous (Table 1). Consistent with this, human centromeres have numerous centromere-specific satellite families and a genome size 20 times larger distributed on more chromosomes, which raises the possibility that satellite dispersal to nonhomologous chromosomes may be highly dependent on a genome's physical size. The plant genomes examined here are smaller than the primate genomes (Table 2), resulting in fewer required recombination events within and between chromosomes in plants for centromere satellite homogenization across the genome. Our results support this hypothesis; although primate satellites undergo efficient local homogenization to generate chromosome-specific variants, their more modest divergence between species suggests a slower pace of genome-wide homogenization compared to the plants examined here. In addition, this rapid turnover could be due to the unique genome biology of plants, where hybridization, ploidy changes, and chromosome rearrangements are common (WENDEL 2000; CHARLESWORTH and WRIGHT 2001).

The requirement for a single functional centromere on each chromosome could create a strong meiotic selection for centromere-binding proteins that interact favorably with satellite repeats (HENIKOFF et al. 2001). Indeed, the rapid diversification of plant centromere satellite repeats is accompanied by adaptive evolution of centromere-binding proteins (TALBERT et al. 2002, 2004; COOPER and HENIKOFF 2004). CENP-A, a histone H3-like protein first identified in human cells (PALMER et al. 1991), is evolving adaptively in the N-terminal tail of the plant homolog, HTR12, between A. thaliana and A. arenosa (TALBERT et al. 2002). Recent analysis of HTR12 evolution has been expanded to other plants in the Brassicaceae, including O. pumila, localizing adaptive evolution domains in the N-terminal tail and the Loop 1 region, both of which are important for DNA-protein interactions (COOPER and HENIKOFF 2004). Adaptive evolution of Loop 1 was not detected between A. thaliana and A. arenosa, where a separation of only 5 MY resulted in satellites that retain 73.3% identity. Over broad evolutionary distance, the rapid molecular coevolution of centromere satellites and centromere-binding proteins could conceivably create mating barriers between species (DOVER 2000). Nonetheless, synthetic and natural allotetraploids of A. thaliana and its closest relatives do not have reduced fitness, perhaps because recruitment of HTR12 variants from both species into histone tetramers enables interactions with both satellite types (TALBERT et al. 2002). The centromere-binding protein CENP-C has been shown to bind to human satellites in vivo and is essential for proper chromosome segregation (POLITI et al. 2002; TRAZZI et al. 2002). Although a specific binding site in plants has not been identified, CENP-C is evolving adaptively in A. thaliana and grasses, undergoing adaptive evolution in the region homologous to the human DNA-binding domain (TALBERT et al. 2004).

Interactions with centromere-binding protein domains could be expected to provide a selective pressure that limits the range of changes that are tolerated in satellite repeats. Such conservation could be difficult to detect, as binding sites for centromere protein B (CENP-B) are known to be highly polymorphic (MASUMOTO et al. 1989; MURO et al. 1992; IKENO et al. 1994). Although we did not find satellite regions conserved across the Brassicaceae that could potentially act as binding sites for centromere proteins, we did find regions of conservation in individual species. These short conserved domains could serve to enhance binding of centromere proteins or, alternatively, to mediate homogenization of arrays through molecular drive processes. The ongoing characterization of rapidly evolving centromere satellites and proteins in relatives of A. thaliana will provide an opportunity to resolve this question directly; by assessing the binding efficiency of variant satellite monomers to divergent centromere-binding proteins, it will become possible to test the predictions of molecular drive hypotheses or to find evidence for natural selection.

We are grateful to G. Copenhaver, K. von Besser, and R. Palanivelu for critical reading of the article and to G. Kettler for computer programming support. We thank S. Henikoff for providing the antibody to HTR12 and J. Walling, S. Jackson, K. Reddy, and J. Mach for excellent advice on FISH methods. This research was funded by the Howard Hughes Medical Institute, the Packard Foundation, and the Atlantic Philanthropies.

Sequence data from this article have been deposited with the EMBL/GenBank libraries under accession nos. AY640635, AY640924, AY642784, AY642785, AY642786, AY642787, AY642788, AY642789, AY642790, AY642791, AY642792, AY642793, AY642794, AY642795, AY642796, AY642797, AY642798, AY642799, AY642800, AY642801, AY642802, AY642803, AY642804, AY642805, AY642806, AY656017, AY656018, AY656019, AY656020, AY656021, AY656022, AY656023, AY656024, AY656025, AY656026, AY656027, AY656028, AY656029, and AY792367, AY792498.

AL-SHEHBAZ, I. A., 1984 The tribes of Cruciferae (Brassicaceae) in the southeastern United States. J. Arnold Arbor. Harv. Univ. 65: 343–373.

ALEXANDROV, I. A., S. P. MITKEVICH and Y. B. YUROV, 1988 The phylogeny of human chromosome specific alpha satellites. Chromosoma 96: 443–453.[CrossRef][Medline]

ALKAN, C., J. A. BAILEY, E. E. EICHLER, S. C. SAHINALP and E. TUZUN, 2002 An algorithmic analysis of the role of unequal crossover in alpha-satellite DNA evolution. Genome Inf. Ser. Workshop Genome Inf. 13: 93–102.

ALVES, G., H. N. SEUáNEZ and T. FANNING, 1994 Alpha satellite DNA in neotropical primates (Platyrrhini). Chromosoma 103: 262–267.[Medline]

ARABIDOPSIS GENOME INITIATIVE, 2000 Analysis of the genome sequence of the flowering plant Arabidopsis thaliana. Nature 408: 796–815.[CrossRef][Medline]

BALDINI, A., D. A. MILLER, O. J. MILLER, O. A. RYDER and A. R. MITCHELL, 1991 A chimpanzee- derived chromosome-specific alpha satellite DNA sequence conserved between chimpanzee and human. Chromosoma 100: 156–161.[CrossRef][Medline]

BARRY, A. E., E. V. HOWMAN, M. R. CANCILLA, R. SAFFERY and K. H. CHOO, 1999 Sequence analysis of an 80 kb human neocentromere. Hum. Mol. Genet. 8: 217–227.[Abstract/Free Full Text]

BRUTLAG, D. L., 1980 Molecular arrangement and evolution of heterochromatic DNA. Annu. Rev. Genet. 14: 121–144.[CrossRef][Medline]

CHARLESWORTH, B., P. SNIEGOWSKI and W. STEPHAN, 1994 The evolutionary dynamics of repetitive DNA in eukaryotes. Nature 371: 215–220.[CrossRef][Medline]

CHARLESWORTH, D., and S. I. WRIGHT, 2001 Breeding systems and genome evolution. Curr. Opin. Genet. Dev. 11: 685–690.[CrossRef][Medline]

CHOO, K. H., B. VISSEL, A. NAGY, E. EARLE and P. KALITSIS, 1991 A survey of the genomic distribution of satellite DNA on all the human chromosomes, and derivation of a new consensus sequence. Nucleic Acids Res. 19: 1179–1182.[Free Full Text]

CLARKE, L., 1998 Centromeres: proteins, protein complexes, and repeated domains at centromeres of simple eukaryotes. Curr. Opin. Genet. Dev. 8: 212–218.[CrossRef][Medline]

CO, D. O., A. H. BOROWSKI, J. D. LEUNG, J. VAN DER KAA, S. HENGST et al., 2000 Generation of transgenic mice and germline transmission of a mammalian artificial chromosome introduced into embryos by pronuclear microinjection. Chromosome Res. 8: 183–191.[CrossRef][Medline]

COHEN, S., K. YACOBI and D. SEGAL, 2003 Extrachromosomal circular DNA of tandemly repeated genomic sequences in Drosophila. Genome Res. 13: 1133–1145.[Abstract/Free Full Text]

COOPER, J. L., and S. HENIKOFF, 2004 Adaptive evolution of the histone fold domain in centromeric histones. Mol. Biol. Evol. 21: 1712–1718.[Abstract/Free Full Text]

COPENHAVER, G. P., 2003 Using Arabidopsis to understand centromere function: progress and prospects. Chromosome Res. 11: 255–262.[CrossRef][Medline]

COPENHAVER, G. P., K. NICKEL, T. KUROMORI, M. I. BENITO, S. KAUL et al., 1999 Genetic definition and sequence analysis of Arabidopsis centromeres. Science 286: 2468–2474.[Abstract/Free Full Text]

CSAIKL, U. M., H. BASTIAN, R. BRETTSCHNEIDER, S. GAUCH, A. MEIR et al., 1998 Comparative analysis of different DNA extraction protocols: a fast, universal maxi-preparation of high quality plant DNA for genetic evaluation and phylogenetic studies. Plant Mol. Biol. Rep. 16: 69–86.

DOVER, G., 1982 Molecular drive: a cohesive mode of species evolution. Nature 299: 111–117.[CrossRef][Medline]

DOVER, G., 2000 How genomic and developmental dynamics affect evolutionary processes. BioEssays 22: 1153–1159.[CrossRef][Medline]

DOVER, G., 2002 Molecular drive. Trends Genet. 18: 587–589.[CrossRef][Medline]

ELDER, JR., J. F., and B. J. TURNER, 1995 Concerted evolution of repetitive DNA sequences in eukaryotes. Q. Rev. Biol. 70: 297–320.[CrossRef][Medline]

FANNING, T. G., H. N. SEUáNEZ and L. FORMAN, 1993 Satellite DNA sequences in the New World primate Cebus apella (Platyrrhini, Primates). Chromosoma 102: 306–311.[CrossRef][Medline]

GLAZKO, G. V., and M. NEI, 2003 Estimation of divergence times for major lineages of primate species. Mol. Biol. Evol. 20: 424–434.[Abstract/Free Full Text]

GRIMES, B. R., A. A. RHOADES and H. F. WILLARD, 2002 -Satellite DNA and vector composition influence rates of human artificial chromosome formation. Mol. Ther. 5: 798–805.[CrossRef][Medline]

HAAF, T., and H. F. WILLARD, 1997 Chromosome-specific-satellite DNA from the centromere of chimpanzee chromosome 4. Chromosoma 106: 226–232.[CrossRef][Medline]

HAAF, T., and H. F. WILLARD, 1998 Orangutan –satellite monomers are closely related to the human consensus sequence. Mamm. Genome 9: 440–447.[CrossRef][Medline]

HALL, S. E., G. KETTLER and D. PREUSS, 2003 Centromere satellites from Arabidopsis populations: maintenance of conserved and variable regions. Genome Res. 13: 195–205.[Abstract/Free Full Text]

HENIKOFF, S., K. AHMAD and H. S. MALIK, 2001 The centromere paradox: stable inheritance with rapidly evolving DNA. Science 293: 1098–1102.[Abstract/Free Full Text]

HESLOP-HARRISON, J. S., A. BRANDES and T. SCHWARZACHER, 2003 Tandemly repeated DNA sequences and centromeric chromosomal regions of Arabidopsis species. Chromosome Res. 11: 241–253.[CrossRef][Medline]

HOSOUCHI, T., N. KUMEKAWA, H. TSURUOKA and H. KOTANI, 2002 Physical map-based sizes of the centromeric regions of Arabidopsis thaliana chromosomes 1, 2, and 3. DNA Res. 9: 117–121.[Abstract]

IKENO, M., H. MASUMOTO and T. OKAZAKI, 1994 Distribution of CENP-B boxes reflected in CREST centromere antigenic sites on long-range –satellite DNA arrays of human chromosome 21. Hum. Mol. Genet. 3: 1245–1257.[Abstract/Free Full Text]

JEANNMOUGIN, F., J. D. THOMPSON, M. GOUY, D. G. HIGGINS and T. J. GIBSON, 1998 Multiple sequence alignment with Clustal X. Trends Biochem. Sci. 23: 403–405.[CrossRef][Medline]

JOHNSTON, J. S., A. E. PEPPER, A. E. HALL, Z. J. CHEN, G. HODNETT et al., 2005 Evolution of genome size in Brassicaceae. Ann. Bot. 95: 229–235.[Abstract/Free Full Text]

JONES, R. S., and S. S. POTTER, 1985 Characterization of cloned human alphoid satellite with an unusual monomeric construction: evidence for enrichment in HeLa small polydisperse circular DNA. Nucleic Acids Res. 13: 1027–1042.[Abstract/Free Full Text]

KAMM, A., I. GALASSO, T. SCHMIDT and J. S. HESLOP-HARRISON, 1995 Analysis of a repetitive DNA family from Arabidopsis arenosa and relationships between Arabidopsis species. Plant Mol. Biol. 27: 853–862.[CrossRef][Medline]

KOCH, M., B. HAUBOLD and T. MITCHELL-OLDS, 2001 Molecular systematics of the Brassicaceae: evidence from coding plastidic MatK and nuclear Chs sequences. Am. J. Bot. 88: 534–544.[Abstract/Free Full Text]

LUO, S., A. E. HALL, S. E. HALL and D. PREUSS, 2004 Whole-genome fractionation rapidly purifies DNA from centromeric regions. Nat. Methods 1: 67–71.[CrossRef][Medline]

MAHTANI, M. M., and H. F. WILLARD, 1998 Physical and genetic mapping of the human X chromosome centromere: repression of recombination. Genome Res. 8: 100–110.[Abstract/Free Full Text]

MALIK, H. S., and S. HENIKOFF, 2001 Adaptive evolution of Cid, a centromere-specific histone in Drosophila. Genetics 157: 1293–1298.[Abstract/Free Full Text]

MARTINEZ-ZAPATER, J. M., M. A. ESTELLE and C. R. SOMERVILLE, 1986 A highly repeated DNA sequence in Arabidopsis thaliana. Mol. Gen. Genet. 204: 417–423.[CrossRef]

MASUMOTO, H., H. MASUKATA, Y. MURO, N. NOZAKI and T. OKAZAKI, 1989 A human centromere antigen (CENP-B) interacts with a short specific sequence in alphoid DNA, a human centromeric satellite. J. Cell Biol. 109: 1963–1973.[Abstract/Free Full Text]

MESTROVIC, N., M. PLOHL, B. MRAVINAC and D. UGARKOVIC, 1998 Evolution of satellite DNAs from the genus Palorus—experimental evidence for the "library" hypothesis. Mol. Biol. Evol. 15: 1062–1068.[Abstract]

MURO, Y., H. MASUMOTO, K. YODA, N. NOZAKI, M. OHASHI et al., 1992 Centromere protein B assembles human centromeric –satellite DNA at the 17-bp sequence, CENP-B box. J. Cell Biol. 116: 585–596.[Abstract/Free Full Text]

NAGAKI, K., P. B. TALBERT, C. X. ZHONG, R. K. DAWE, S. HENIKOFF et al., 2003 Chromatin immunoprecipitation reveals that the 180-bp satellite repeat is the key functional DNA element of Arabidopsis thaliana centromeres. Genetics 163: 1221–1225.[Abstract/Free Full Text]

NIJMAN, I. J., and J. A. LENSTRA, 2001 Mutation and recombination in cattle satellite DNA: a feedback model for the evolution of satellite DNA repeats. J. Mol. Evol. 52: 361–371.[Medline]

OHTA, T., and G. A. DOVER, 1983 Population genetics of multigene families that are dispersed into two or more chromosomes. Proc. Natl. Acad. Sci. USA 80: 4079–4083.[Abstract/Free Full Text]

OKUMURA, K., R. KIYAMA and M. OISHI, 1987 Sequence analyses of extrachromosomal Sau3A and related family DNA: analysis of recombination in the excision event. Nucleic Acids Res. 15: 7477–7489.[Abstract/Free Full Text]

ONO, S., 1972 So much "junk" in our genome. Brookhaven Symp. Biol. 23: 366–370.[Medline]

ORGEL, L. E., and F. H. CRICK, 1980 Selfish DNA: the ultimate parasite. Nature 284: 604–607.[CrossRef][Medline]

PALMER, D. K., K. O'DAY, H. L. THRONG, H. CHARBONNEAU and R. L. MARGOLIS, 1991 Purification of the centromere-specific protein CENP-A and demonstration that it is a distinctive histone. Proc. Natl. Acad. Sci. USA 88: 3734–3748.[Abstract/Free Full Text]

POLITI, V., G. PERINI, S. TRAZZI, A. PLISS, I. RASKA et al., 2002 CENP-C binds the alpha-satellite DNA in vivo at specific centromere domains. J. Cell Sci. 115: 2317–2327.[Abstract/Free Full Text]

PONS, J., B. BRUVO, E. PETITPIERRE, M. PLOHL, D. UGARKOVIC et al., 2004 Complex structural features of satellite DNA sequences in the genus Pimelia (Coleoptera: Tenebrionidae): random differential amplification from a common "satellite DNA library". Heredity 92: 418–427.[CrossRef][Medline]

SCHENA, M., and R. W. DAVIS, 1992 HD-Zip proteins: members of an Arabidopsis homeodomain protein superfamily. Proc. Natl. Acad. Sci. USA 89: 3894–3898.[Abstract/Free Full Text]

SCHINDELHAUER, D., and T. SCHWARZ, 2002 Evidence for a fast, intrachromosomal conversion mechanism from mapping of nucleotide variants within a homogeneous –satellite DNA array. Genome Res. 12: 1815–1826.[Abstract/Free Full Text]

SCHUELER, M. G., A. W. HIGGINS, M. K. RUDD, K. GUSTASHAW and H. F. WILLARD, 2001 Genomic and genetic definition of a functional human centromere. Science 294: 109–115.[Abstract/Free Full Text]

SCHWARZACHER, T., and J. S. HESLOP-HARRISON, 1994 Direct fluorochrome-labeled DNA probes for direct fluorescent in situ hybridization to chromosomes. Methods Mol. Biol. 28: 167–176.[Medline]

SMITH, G. P., 1976 Evolution of repeated DNA sequences by unequal crossover. Science 191: 528–535.[Abstract/Free Full Text]

STEPHAN, W., 1986 Recombination and the evolution of satellite DNA. Genet. Res. 47: 167–174.[Medline]

TALBERT, P. B., R. MASUELLI, A. P. TYAGI, L. COMAI and S. HENIKOFF, 2002 Centromeric localization and adaptive evolution of an Arabidopsis histone H3 variant. Plant Cell 14: 1053–1066.[Abstract/Free Full Text]

TALBERT, P. B., T. D. BRYSON and S. HENIKOFF, 2004 Adaptive evolution of centromere proteins in plants and animals. J. Biol. 3: 18.[CrossRef][Medline]

TELENIUS, H., A. SZELES, J. KERESO, E. CSONKA, T. PRAZNOVSZKY et al., 1999 Stability of a functional murine satellite DNA-based artificial chromosome across mammalian species. Chromosome Res. 7: 3–7.[CrossRef][Medline]

TRAZZI, S., R. BERNARDONI, D. DIOLAITI, V. POLITI, W. C. EARNSHAW et al., 2002 In vivo functional dissection of human inner kinetochore protein CENP-C. J. Struct. Biol. 140: 39–48.[CrossRef][Medline]

UGARKOVIC, D., and M. PLOHL, 2002 Variation in satellite DNA profiles—causes and effects. EMBO J. 21: 5955–5959.[CrossRef][Medline]

WALLING, J. G., J. C. PIRES and S. A. JACKSON, 2005 Preparation of samples for comparative studies of plant chromosomes using in situ hybridization methods. Methods Enzymol. 395: 442–459.

WALSH, J. B., 1987 Persistence of tandem arrays: implications for satellite and simple-sequence DNAs. Genetics 115: 553–567.[Abstract/Free Full Text]

WARBURTON, P. E., and H. F. WILLARD, 1990 Genomic analysis of sequence variation in tandemly repeated DNA: evidence for localized homogeneous sequence domains within arrays of –satellite DNA. J. Mol. Biol. 216: 3–16.[CrossRef][Medline]

WARBURTON, P. E., J. S. WAYE and H. F. WILLARD, 1993 Nonrandom localization of recombination events in human alpha satellite repeat unit variants: implications for higher-order structural characteristics within centromeric heterochromatin. Mol. Cell. Biol. 13: 6520–6529.[Abstract/Free Full Text]

WARBURTON, P. E., T. HAAF, J. GODSEN, D. LAWSON and H. F. WILLARD, 1996 Characterization of a chromosome-specific chimpanzee alpha satellite subset: evolutionary relationship to subsets on human chromosomes. Genomics 33: 220–228.[CrossRef][Medline]

WAYE, J. S., and H. F. WILLARD, 1987 Nucleotide sequence heterogeneity of alpha satellite repetitive DNA: a survey of alphoid sequences from different human chromosomes. Nucleic Acids Res. 15: 7549–7569.[Abstract/Free Full Text]

WENDEL, J. F., 2000 Genome evolution in polyploids. Plant Mol. Biol. 42: 225–249.[CrossRef][Medline]

WEVRICK, R., and H. F. WILLARD, 1989 Long-range organization of tandem arrays of alpha satellite DNA at the centromeres of human chromosomes: high-frequency array-length polymorphism and meiotic stability. Proc. Natl. Acad. Sci. USA 86: 9394–9398.[Abstract/Free Full Text]

WILLARD, H. F., 1985 Chromosome-specific organization of human alpha satellite DNA. Am. J. Hum. Genet. 37: 524–532.[Medline]

WILLARD, H. F., 1998 Human artificial chromosomes coming into focus. Nat. Biotech. 16: 415–416.[CrossRef][Medline]

WILLARD, H. F., 2001 Genomics and gene therapy: artificial chromosomes coming to life. Science 290: 1308.

WILLARD, H. F., and J. S. WAYE, 1987a Chromosome-specific subsets of human alpha satellite DNA: analysis of sequence divergence within and between chromosomal subsets and evidence for an ancestral pentameric repeat. J. Mol. Evol. 25: 207–214.[CrossRef][Medline]

WILLARD, H. F., and J. S. WAYE, 1987b Hierarchical order in chromosome-specific human alpha satellite DNA. Trends Genet. 3: 192–198.[CrossRef]

YANG, J. W., C. PENDON, J. YANG, N. HAYWOOD, A. CHAND et al., 2000 Human mini- chromosomes with minimal centromeres. Hum. Mol. Genet. 9: 1891–1902.[Abstract/Free Full Text]

Communicating editor: S. HENIKOFF

This article has been cited by other articles:

				N. Mestrovic, P. Castagnone-Sereno, and M. Plohl Interplay of Selective Pressure and Stochastic Events Directs Evolution of the MEL172 Satellite DNA Library in Root-Knot Nematodes Mol. Biol. Evol., December 1, 2006; 23(12): 2316 - 2325. [Abstract] [Full Text] [PDF]

				H.-R. Lee, P. Neumann, J. Macas, and J. Jiang Transcription and Evolutionary Dynamics of the Centromeric Satellite Repeat CentO in Rice Mol. Biol. Evol., December 1, 2006; 23(12): 2505 - 2520. [Abstract] [Full Text] [PDF]

				A. E. Hall, G. C. Kettler, and D. Preuss Dynamic evolution at pericentromeres Genome Res., March 1, 2006; 16(3): 355 - 364. [Abstract] [Full Text] [PDF]

International Access Link

Online ISSN: 1943-2361  Print ISSN: 0016-6731
Copyright 2009 by the Genetics Society of America
phone: 412-268-1812   fax: 412-268-1813   email: genetics-gsa{at}andrew.cmu.edu