The centromeres of higher plants and animals share many common features, though current models fail to account for all aspects of centromere composition and function. This dilemma is likely to be resolved in the next few years in Arabidopsis where robust assays for centromere function are available and the sequence of the entire genome will be determined.
Addresses
University of Chicago, 1103 E. 57th Street, Chicago, IL 60637, USA *e-mail: gpcopenh@midway.uchicago.edu
Current Opinion in Plant Biology1999, 2:104–108 http://biomednet.com/elecref/1369526600200104 © Elsevier Science Ltd ISSN 1369-5266
Introduction
Most eukaryotic chromosomes contain a specialized domain — the centromere — that mediates critical mitotic and meiotic functions, including kinetochore nucleation, spindle attachment, and sister chromatid cohesion. These processes ensure faithful chromosome partitioning during cell division. Intense scrutiny of the DNA and proteins that comprise the centromere revealed its fundamental roles in chromosome mechanics and, in some cases, enabled the creation of artificial chromosomes. The cen-tromeric DNA of several lower eukaryotes has been completely defined; much less is known about the cen-tromeric DNA of higher eukaryotes, including plants. In this review we define parallels with lower eukaryotic sys-tems, survey current models for higher eukaryotic centromere function, suggest guidelines for assessing models, and summarize the short-term prospects for inves-tigating plant centromeres.
Centromeres from lower eukaryotes
The centromeres from many fungal organisms were mapped by tetrad analysis, a powerful genetic tool used to map sites of genetic exchange relative to the centromere (Figure 1) [1]. Tetrad analysis is possible when meiotical-ly-related daughter cells can be recovered and analyzed; yeasts such as Saccharomyces cerevisiae and Schizosaccharomyces pombe package their meiotic products into asci that can be readily separated and characterized. Using tetrad analysis and chromosome walking, Clarke and Carbon [2] identified the centromere of S. cerevisiae chro-mosome III, mapping the chromosomal regions that attach to the spindle and migrate to opposite poles in meiosis I, thus providing a functional definition of the centromere. Chromosome walking delineated a small region of ~20 kb, and subsequently, artificial chromosomes were used to show that only 125 bp were needed for centromere func-tion [3]. A highly conserved copy of this sequence resides at the centromere of every S. cerevisiae chromosome. This minimal region meets two stringent criteria: first, artificial
chromosomes containing the 125 bp region segregate effi-ciently through mitosis, with loss rates approaching those of normal chromosomes (>99% transmission in the absence of selection); and second, these constructs partition effi-ciently through meiosis, as assayed by tetrad analysis [4]. Similar genetic analyses led to the elucidation of the S. pombecentromere [5]. Chromosome walking on all three chromosomes pinpointed regions ranging from 40 to 100 kb. Unlike S. cerevisiaecentromeres, each S. pombe cen-tromere contains a divergent core region surrounded by repetitive elements. S. pombe artificial chromosomes define a minimal centromere region that confers high-fidelity transmission: this ~7 kb fragment consists of a portion of the chromosome II centromere core and a single flanking repeat element.
Centromeres from higher eukaryotes
Tetrad analysis has not been used to map the centromeres of most higher eukaryotes because the products of individ-ual meioses are relatively difficult to obtain and analyze. This limitation has hampered progress toward genetically defining these centromeres in their native contexts. In an alternative strategy, the DNA sequences that comprise higher eukaryotic centromeres are typically localized genetically by analyzing the assortment of chromosome fragments and rearrangements. These techniques have limited resolution — breakpoints close to the centromeres cannot always be obtained, and the activity of centromeres can be altered when they are removed from their natural context. Nonetheless, such methods have localized cen-tromere functions to regions ranging from a few hundred kilobases (Drosophila and maize) to several megabases (tomato, rice, and humans) [6••,7–11]. The relatively large chromosome size typical of these organisms facilitates cytological analysis of centromeres, and the genetically defined regions typically correlate with the highly con-densed chromatin known as the centromeric constriction [12,13•]. These studies have been the subject of several recent reviews [14–17]. The emerging trend is that the large regions that encompass higher eukaryotic cen-tromeres resemble the cencen-tromeres defined in S. pombe. All higher organisms examined thus far, including plants, contain distinctive and numerous repetitive elements at each centromere; other than these repeats, a conserved sequence that resides at the centromere of each chromo-some has not been found [15,18,19•].
Genomic sequencing is underway in several multicellular organisms, promising to yield a dramatically improved understanding of centromeric DNA in the coming years. The first complete sequence of a higher eukaryote, that of the nematode C. elegans, unfortunately does not clarify the understanding of centromeres: C. elegans has holocentric chromosomes that attach to spindle fibers along their
Centromeres in the genomic era: unraveling paradoxes
entire length [20]. In contrast, the complete sequence of Arabidopsis thaliana, an organism with typical centromeres, will be obtained within the next two years [21]. Arabidopsis is particularly suited for elucidating cen-tromere functions — among all multicellular model organisms, it is uniquely amenable to tetrad analysis. The Arabidopsis quartet mutation causes the four products of male meiosis to remain attached, yielding tetrads of pollen grains [22]. Recently, tetrad analysis was used to geneti-cally map all five centromeres in Arabidopsis [23••]; high resolution genetic mapping will allow a precise definition of the DNA sequences that confer centromere function. Tetrad analysis in combination with sequence information will clarify whether centromeres consist entirely of repeats or if a unique centromeric element resides within the repetitive region.
Centromere paradoxes
Neither the DNA sequences that comprise higher eukary-otic centromeres, nor the structure of their centromeric chromatin are fully understood; consequently, only conjec-tural models for centromere function are available. Appropriate explanations must account for several puz-zling observations, including the diversity of centromere DNA sequences, the evolutionary conservation of some centromere binding proteins, and the accumulation of repetitive DNA in the centromeric regions.
In many higher eukaryotes, arrays of repetitive DNA measuring approximately 170 bp have been identified at the centromere, yet the sequence of these repeats is typ-ically species specific. The centromeres from the budding yeasts Kluyveromyces lactis, Kluyvermyces marxi-anus, Candida glabrata, and S. cerevisiaeare more similar in size and sequence composition, yet none function effi-ciently when introduced into S. cerevisiae or S. pombe artificial chromosomes [1]. By contrast, the proteins that bind to centromeres display remarkable conservation between species. For example, homologs of the human centromere binding proteins CENP-A, CENP-B, and CENP-C have been identified in S. cerevisiae or S. pombe. CENP-A encodes a modified histone H3 protein, and CENP-B, which binds to repetitive arrays, shares sequence similarity with transposase [1]. The diverse nature of centromeric DNA sequences in higher eukary-otes together with the observed homology among centromere binding proteins support the idea that the conservation of centromere functions derives from a sec-ondary or tertiary structure, rather than a particular DNA sequence. The lack of a complete DNA sequence from any higher eukaryotic centromere, however, leaves open the possibility that a conserved sequence nucleates cen-tromere formation in multicellular organisms.
The ubiquity of repetitive elements at the centromeres of higher eukaryotes led to the suggestion that repeats are required for centromere function. These repetitive arrays have been postulated to form essential higher order
struc-tures, bind key centromeric proteins, or serve as targets for critical DNA modification. Support for the requirement for repetitive centromeric DNA has come from the construc-tion of human artificial chromosomes [24••,25•]. Constructs containing long arrays of the human cen-tromere associated repeat known as α satellite DNA together with telomeres acquire modest mitotic stability in cell lines. However, these constructs often acquire other sequences in vivo, raising the possibility that α satellite DNA alone is insufficient to confer centromere function. Importantly, the role of repetitive DNA has been called into question by the recent characterization of highly sta-ble human neocentromeres — centromeric constrictions that assemble at an atypical site on the chromosome. Many of these neocentromeres contain neither α satellite DNA nor display associated CENP-B protein [26••]. These recent findings raise the possibility that centromeric repeats are not required for centromere functions, but rather that the intrinsic nature of the centromeric regions, particularly the relatively low level of meiotic recombina-tion, may lead to the accumulation of repetitive DNA.
Models for centromere function in higher
eukaryotes
In multicellular organisms, the similarities between cen-tromere functions and the paradoxes still remaining have lead to the generation of several competing models that aim to explain centromere activity (Figure 2). These mod-els can be grouped into two categories: firstly, sequence-based models that claim higher centromeres, like those from S. cerevisiae, depend on a specific sequence, but that this sequence may be dispersed or appear in
Figure 1
A tetrad of meiotically related cells is formed following meiosis I (MI) and meiosis II (MII). Genetic analysis of marker assortment in tetrads can define sites where paternal (white chromosomes) and maternal (hatched chromosomes) DNA always migrate to opposite poles. These regions constitute a genetically defined centromere.
M I MII
multiple copies along the centromeric region, and second-ly, sequence-independent models that invoke epigenetic modification of centromeric DNA, higher order DNA structures, or an entirely passive role for sequences that comprise the centromere.
Although the paradigms established in lower eukaryotes are appealing, they have yet to be demonstrated in more complex systems. Currently, the most complete under-standing of higher centromeres comes from extensive characterization of Drosophila mini-chromosomes [6••]. These studies argue that Drosophila centromeres may be composed entirely of repetitive DNA. Final proof of this possibility will be achieved after complete DNA analysis rules out the presence of small islands of unique sequence. In addition, because these minichromosomes have under-gone significant rearrangements, they may reflect neocentromeric rather than native centromeric activity. Nonetheless, this body of work demonstrates strong dis-tinctions between higher and lower eukaryotic centromeres context.
If higher eukaryotes do rely on particular DNA sequences for centromere function, then those sequences may span a relatively large portion of the chromosome. The constrict-ed regions of higher chromosomes clearly encompass megabases of DNA, with multiple spindle fiber attach-ments during cell division. In some cases, for example the maize B chromosomes, it is possible to split a centromere,
yielding two chromosome fragments that segregate nor-mally [10]. These observations suggest that critical DNA sequences may be spread throughout the centromeric region, with the space between such elements filled by apparently nonfunctional DNA.
Because a unique centromere sequence has not been demonstrated in higher eukaryotes, several recent models focus on non-sequence based mechanisms. Such models are appealing and could explain why centromeres moved from their native host into another eukaryotic species are stable, yet centromeres transferred first through E. coli can rarely function in a heterologous species. Two strong examples of epigenetic regulation have been demonstrat-ed in Drosophila and S. pombe. Fragments of Drosophila chromosomes that are normally telomeric can acquire mod-est segregation stability after residing near a centromere [27•]. Control fragments that have never resided near a centromere do not share this property. These observations led to the proposal that centromere activity can spread to adjacent DNA through an unknown epigenetic mecha-nism. Although these results are intriguing, the relatively low transmission rates (14–24%) indicate that complete stability is not acquired by these fragments. Epigenetic control of centromere function in S. pombehas been pro-posed following the quantitative analysis of artificial chromosome transmission [28]. Curiously, a construct bear-ing a partial centromere was transmitted with a high fidelity in some colonies and a low fidelity in others. This difference in transmission frequency was not due to muta-tion, and rapid switching between states, through an unknown mechanism, was observed. These findings are quite unusual for yeast artificial chromosomes and reveal a potentially exciting level of epigenetic control; however, only partial centromeres display this activity, and inclusion of the complete centromeric sequence always resulted in highly stable transmission (5×10–4). Thus, for both Drosophila and S. pombe, the proposed epigenetic regula-tion may be most critical in cases where only a partial centromere is present.
Recently, Csink and Henikoff [29] proposed an alternative model in which centromeric DNA excludes a specific activity, such as the DNA replication machinery, and thus creates an unusual region that de facto becomes a cen-tromere. In this model, the sequence of the region is not important as long as it does not participate in the critical activity — in this example, replication early in the cell cycle. This model requires no sequence specificity, could account for the abundance of repetitive centromeric DNA, and can be experimentally tested by manipulating reagents that are currently available. There is a vast amount of experimental evidence, however, that indicates large chromosome fragments that lack a centromere from the wheat, maize and human genomes are not segregated efficiently, despite the expectation that they would contain sizable regions of late-replicating or repetitive DNA. Consequently, this model must be modified to account for
Figure 2
Models for centromere function in higher eukaryotes: (a)a unique sequence present at a single location on each chromosome; (b) dispersed centromere interrupted by nonfunctional DNA; (c)fully functional elements represented multiple times across the centromere region; (d)higher order DNA structures, DNA modification, or association with specific chromatin elements establishes a specialized domain subject to epigenetic regulation; and (e)DNA in the
centromere region passively excludes an activity, thereby establishing a specialized domain recognized by the cell as a centromere.
X
a
b
c
d
e
the observation that only certain repetitive regions possess the necessary qualities to create a centromere, whereas other neocentromeric regions that lack repetitive arrays can function as efficient centromeres.
Conclusions
Many aspects of higher eukaryotic centromere function promise to be elucidated in the coming years. It will be interesting to explain features uncommon in lower eukaryotes, including the presence of large arrays of repetitive DNA, the dispersal of centromere functions along a large chromosomal region, and the potential for epigenetic regulation. Attempts to meet these chal-lenges require rigorous criteria, like the elegant assays performed in lower eukaryotes, for assessing centromere function. In particular, a thorough understanding of cen-tromeres requires quantitative genetic tests that measure chromosome transmission per cell division, a complete understanding of DNA sequences surrounding the cen-tromere in its native context, and ideally, the development of artificial chromosomes that can reconsti-tute centromere activity.
A quantitative distinction between complete centromere activity and partial centromere function must be estab-lished. A fully functional centromere is expected to segregate, in the absence of selection, with fidelity approaching a native chromosome. Moreover, it should also display stability through both the first and second divi-sions of meiosis, reflecting proper sister chromatid cohesion and release. Cell autonomous color markers make it possible to achieve these goals, even in a multicellular organism. Like the colony color assays developed by Heiter and colleagues [30], cell autonomous assays that rely on green fluorescent protein or other vital dyes will enable chromosome transmission to be studied in plants and animals. With such assays, chromosome loss rates per cell division can be determined; these calculations will depend on monitoring cell lineage either in intact tissues with defined patterns of division or in cells cultured on a solid substrate.
The vast genomic sequencing effort will have a profound impact on the understanding of centromere functions in plants and animals. First, the availability of a complete genome sequence, along with robust genetic assays for centromere function, will allow an assessment of the role of repetitive, low copy or unique centromeric sequences. As functional genomic assays are developed, such as DNA microarrays, it will become feasible to identify deletions and insertions that may alter centromere function, as well as to characterize the expression of all of the genes in the centromeric region. The rapid advances in genomic tech-nology, the availability of a powerful genetic assay for centromere function in Arabidopsis, and the potential for developing artificial chromosomes in higher eukaryotes promise to yield a much improved understanding of high-er centromhigh-eres in the next decade.
Acknowledgements
We thank Jake Mayfield for critical reading of the manuscript.
References and recommended reading
Papers of particular interest, published within the annual period of review, have been highlighted as:
• of special interest ••of outstanding interest
1. Clarke L: Centromeres: proteins, protein complexes, and repeated domains at centromeres of simple eukaryotes.Curr Opin Genet Devel1998, 8:212-218.
2. Clarke L, Carbon J: Isolation of a yeast centromere and construction of functional small circular chromosomes.Nature
1980, 287:504-509.
3. Cottarel G, Shero JH, Hieter P, Hegemann JH: A 125 base-pair CEN6 DNA fragment is sufficient for complete meiotic and mitotic centromere functions in Saccharomyces cerevisiae.Mol Cell Biol
1989, 9:3342-3349.
4. Carbon J, Clarke L:Centromere structure and function in budding and fission yeasts.New Biol1990, 2:10-19.
5. Baum M, Ngan VK, Clarke L: The centromeric K-type repeat and the central core are together sufficient to establish a functional
Schizosaccharomyces pombecentromere.Mol Biol Cell1994, 5:747-761.
6. Sun X, Wahlstrom J, Karpen GH: Molecular structure of a functional
•• Drosophilacentromere.Cell1997, 91:1007-1019
This paper provides the first detailed description of a complex eukaryotic centromere. From deletion derivatives of the X chromosome, a 420 kb region was inferred to be capable of supporting centromere function, although with less efficiency than native centromeres. Importantly, after careful DNA hybridization analyses, this region was found to be composed primarily of repetitive DNA and retroelements. Because these sequences are neither centromere specific nor found an all chromosomes, a non-sequence specif-ic basis for centromere function was suggested.
7. Farr CJ, Bayne RA, Kipling D, Mills W, Critcher F, Cooken HJ: Generation of a human X-derived minichromosome using telomere-associated chromosome fragmentation.EMBO J1995, 14:5444-5454.
8. Heller R, Brown KE, Burgtorf C, Brown WR: Mini-chromosomes derived from the human Y chromosome by telomere directed chromosome breakage.Proc Natl Acad Sci USA1996, 93:7125-7130.
9. Harushima Y, Yano M, Shomura A, Sato M, Shimano T, Kuboki Y, Yamamoto T, Lin SY, Anonio BA, Parco A et al.: High-density rice genetic linkage map with 2275 markers using a single F2 population. Genetics 1998, 148:479-494.
10. Kaszas E, Birchler J: Misdivision analysis of centromere structure in maize.EMBO J1996, 15:5246-5255.
11. Frary A, Presting GG, Tanksley SD: Molecular mapping of the centromeres of tomato chromosomes 7 and 9.Mol Genet1996, 250:295-304.
12. Jiang J, Nasuda S, Dong F, Scherrer CW, Woo SS, Wing RA, Gill BS, Ward DC: A conserved repetitive DNA element located in the centromeres of cereal chromosomes.Proc Natl Acad Sci USA
1996, 93:14210-14213.
13. Fransz P, Armstrong S, Alonso-Blanco C, Fischer TC, Torres-Ruiz RA,
• Jones G: Cytogenetics for the model system Arabidopsis thaliana.
Plant J1998, 136:867-876.
This study highlights the power of cytogenetics in Arabidopsis, showing that FISH analysis can be used to map and determine the size of DNA elements, including those that reside in centromeric heterochromatin.
14. Birchler JA: Do these sequences make CENs yet?Genome Res
1997, 7:1035-1037.
15. Richards EJ: Plant centromeres: structure and control.Curr Opin Plant Biol1998, 1:130-135.
16. Willard HF: Centromeres: the missing link in the development of human artificial chromosomes.Curr Opin Genet Devel1998, 8:219-225.
18. Dong F, Miller JT, Jackson SA, Wang GL, Ronald PC, Jiang J: Rice (Oryza sativa) centromeric regions consist of complex DNA.Proc Natl Acad Sci USA1998, 95:8135-8140.
19. Round EK, Flowers SK, Richards EJ: Arabidopsis thaliana
• centromere regions: genetic map positions and repetitive DNA structure.Genome Res1997, 7:1045-1053.
An important paper which used pulsed field gel technology to genetically map large arrays of the 180 bp centromere associated sequence pAL1 in
Arabidopsis. This paper demonstrates that the 180 bp repeats positions correlate with the position of the functional centromeres defined in [23••].
20. The C. elegansSequencing Consortium: Genome sequence of the nematode C. elegans: a platform for investigating biology.Science
1998, 282:2012-2018.
21. Meinke DW, Cherry MJ, Dean C, Rounsley SD, Koorneef M:
Arabidopsis thaliana: a model plant for genome analysis.Science
1998, 282:662-682.
22. Preuss D, Rhee SY, Davis RW: Tetrad analysis possible in
Arabidopsiswith mutations of the QUARTET(QRT) genes.
Science1994, 264:1458-1460.
23. Copenhaver GP, Browne WE, Preuss D: Assaying genome-wide
•• recombination and centromere functions with Arabidopsistetrads.
Proc Natl Acad Sci USA1998, 95:247-252.
This work demonstrates the first use of tetrad analysis in Arabidopsisusing the quartetmutation. Tetrad analysis was used to functionally define the regions containing all five Arabidopsis centromeres and determine their location on the genetic map.
24. Ikeno M, Grimes T, Nakano M, Saitoh K, Hoshino H, McGill NI,
•• Cooke H, Masumoto H: Construction of YAC-based mammalian artificial chromosomes.Nat Biotech1998, 16:431-439.
This work represents the state of the art in the construction of mammalian artificial chromosomes (MACs). Recombination-defective strains of yeast were used to construct MACs containing telomeres, selectable markers and
~100 kb of αsatellite DNA. These constructs displayed modest mitotic sta-bility in the absence of selection.
25. Harrington JJ, Van Bokkelen G, Mays RW, Gustashaw K, Willard HF:
• Formation of de novocentromeres and construction of first-generation human artificial microchromosomes.Nat Genet1997, 15:345-355.
This ground breaking work illustrated the feasibility of reconstituting cen-tromere function from cloned fragments of DNA including large arrays of α
satellite DNA.
26. du Sart D, Cancilla MR, Earle E, Mao J, Saffery R, Tainton KM,
•• Kalitsis P, Martyn J, Barry AE, Choo KHA: A functional neocentromere formed through activation of a latent human centromere and consisting of non-alpha satellite DNA.Nat Genet1997, 16:144-153. This careful analysis reveals that the neocentromeric DNA of a naturally occurring human minichromosome does not contain detectable amounts of
αsatellite sequences despite the fact that it binds the centromere associat-ed proteins CENP-A and CENP-C. These findings challenge the idea that repetitive DNA is necessary for centromere function.
27. Williams BC, Murphy TD, Goldberg ML, Karpen GH: Neocentromere
• activity of structurally acentric mini-chromosomes in Drosophila.
Nat Genet1998, 18:30-37.
This paper provides compelling evidence for an epigenetic model of cen-tromere function. DNA not normally associated with the cencen-tromere was shown to acquire centromeric activity in higher eukaryotes after being near a native centromere.
28. Steiner NC, Clarke L: A novel epigenetic effect can alter centromere function in fission yeast.Cell1994, 79:865-874.
29. Csink AK, Henikoff S: Something from nothing: the evolution and utility of satellite repeats.Trends Genet1998, 14:200-204.
30. Heiter P, Mann C, Snyder M, Davis RW: Mitotic stability of yeast chromosomes: a colony color assay that measures
![Figure 2yielding two chromosome fragments that segregate nor-mally [10]. These observations suggest that critical DNA](https://thumb-ap.123doks.com/thumbv2/123dok/1038539.929777/3.612.56.299.89.294/figure-yielding-chromosome-fragments-segregate-observations-suggest-critical.webp)
