Centromeres: Controllers of Inheritance
Summary: Simon Chan studied the role of centromeres in chromosome segregation and develops new plant-breeding tools by manipulating the properties of genetic inheritance.
It is with great sadness that we report the death of our colleague, Dr. Simon W.-L. Chan, an associate professor of plant biology at the University of California, Davis, and an HHMI-Gordon and Betty Moore Foundation Investigator. Simon, 38, passed away on Wednesday, August 22, 2012.
More information, including an obituary from the university, can be found through the two links below:
My lab has two major goals: we study basic mechanisms of genetic inheritance, and we manipulate these mechanisms to create novel plant-breeding methods. Centromeres control chromosome segregation during cell division, because they are the site where spindle microtubules attach to chromosomes via the kinetochore protein complex.
Despite their essential function, centromere DNAs are hypervariable in structure and in sequence. Most plant and animal centromeres contain megabases of tandem repeats, but these sequences can be dispensable for kinetochore function. Because centromeres may assemble on different DNA sequences, the most important determinant of centromere location is CENH3, a histone H3 variant that replaces conventional H3 in centromeric nucleosomes. CENH3 recruits other kinetochore proteins and is therefore essential for centromere function. Like centromere DNA, CENH3 evolves rapidly, particularly in the amino-terminal “tail” domain that protrudes from the globular nucleosome core. We use Arabidopsis thaliana as an experimental model, in four areas of research, to understand the function of large plant and animal centromeres.
Comparative Genomics of Centromere DNA
Many of our assumptions about centromere DNA evolution are based on a few well-studied organisms, raising the following questions: (1) How prevalent are centromere tandem repeats in eukaryotes? (2) What are the general properties of centromere tandem repeats? (3) How do centromere tandem repeats evolve?
We use comparative genomics to study how centromere DNA evolution has been constrained by functional demands. In collaboration with Ian Korf (University of California, Davis), we developed bioinformatic methods that detect high-copy tandem repeats—candidate centromere DNAs—in small amounts of shotgun genomic sequence. Candidate centromere DNA repeats from more than 200 eukaryote genomes lack conserved properties: they show a wide range of repeat lengths, GC content, and abundance. These findings suggest that the presence of tandem repeats, rather than any specific sequence feature, may provide some functionally important property for centromere DNA. In the future, we will use high-throughput sequencing to study the relationship between centromere DNA sequences, chromosome structure, and organismal life history.
Meiosis-Specific Centromere Reorganization
Meiosis halves the ploidy of the parental diploid cell to produce gametes. This requires specialized centromere behavior during meiosis I; sister chromatids remain together (in contrast to mitosis), and homologous chromosomes are segregated. We replaced the fast-evolving tail domain of CENH3 with the tail of regular H3 in a GFP (green fluorescent protein)-tagged variant. This GFP-tailswap variant of CENH3 loads normally during mitosis and supports accurate chromosome inheritance but is completely removed from meiotic centromeres, causing random chromosome segregation and sterility. Our results indicate that centromeres have a meiosis-specific assembly pathway that may be important for specifying meiotic centromere function.
Our analysis of CENH3 variants that cause meiosis-specific defects is complemented by recent imaging experiments. The tandem repeat nature of centromere DNA has hampered quantitative analysis of kinetochore structure: we do not know what subset of centromere DNA is bound by CENH3, nor the stoichiometric relationship between CENH3 and other kinetochore proteins. To address these questions, we developed a method in which GFP fluorescence intensity is used to count the absolute number of proteins in individual plant kinetochores (Saccharomyces cerevisiae strains with GFP-tagged proteins serve as controls). The ratio of CENH3 molecules to the NDC80 complex (a kinetochore component essential for microtubule binding) is much higher in Arabidopsis than in budding yeast, suggesting that plant kinetochores have a distinct organization. In meiosis, transient removal of the NDC80 complex and the elongated shape of kinetochores further confirm that centromeres have a distinct assembly pathway.
Parental Centromere Differences Cause Chromosome Missegregation in Zygotes, Creating Haploid Offspring
A model proposed by Steven Henikoff (HHMI, Fred Hutchinson Cancer Research Center) and Harmit Malik (also at the Hutchinson Center) states that centromere differences between two parents may reduce fertility or vigor in an interspecies hybrid, causing speciation. We have discovered a mechanism by which centromere differences can affect fitness: competition between parental chromosomes during mitosis in the fertilized zygote. We crossed Arabidopsis plants expressing altered CENH3 proteins to wild type. After fertilization, chromosomes with altered centromeres can be completely missegregated and lost. In GFP-tailswap × wild-type crosses, up to 50 percent of progeny are haploids with only chromosomes from their wild-type parent (we also saw many aneuploids caused by chromosome segregation errors). Such progeny are not seen when either parent self-fertilizes, indicating that chromosome segregation errors are caused by competition between centromeres containing different CENH3 variants. Genome elimination has been observed in interspecies crosses from diverse plant taxa, and even in salmonid fish. Many of these classical genetic observations may be explained by centromere differences.
Centromere competition is a subtle assay for mitotic defects, because some of the CENH3 alterations that cause drastic chromosome missegregation in a cross have a wild-type phenotype and are completely fertile. We now aim to look for chromosome missegregation in crosses that feature naturally occurring differences in CENH3 and in other kinetochore proteins. Surprisingly, Arabidopsis thaliana cenh3 mutants cannot be rescued by the closely related Brassica rapa CENH3 (tagged with GFP), despite the fact that the alien protein localizes to kinetochores. We are continuing this analysis with CENH3 proteins from many Brassicaceae species. These experiments may further validate the idea that centromere differences can be a speciation mechanism.
Manipulating Chromosome Inheritance to Create New Plant-Breeding Technologies
Haploid plants that can be converted back to diploids are extremely useful in plant breeding. Such “doubled haploids” are the fastest way to generate fully homozygous true-breeding lines from heterozygous F1s, a process that normally takes 8–10 generations of inbreeding. Current protocols for producing haploid plants involve anther or ovule culture, or wide crosses followed by embryo rescue. These methods are genotype specific, inefficient, and difficult to apply to a wide variety of crops.
Our method of engineering centromeres to induce genome elimination is a novel approach for producing haploid plants. Haploids are produced through seed in Arabidopsis (removing the need for tissue culture) and are easily converted into fertile diploids. This method is immediately useful for studying natural variation in Arabidopsis. QTL (quantitative trait loci) mapping populations can now be easily constructed by any laboratory, encouraging investigators to study the most interesting accessions rather than working with preexisting recombinant inbred populations.
Because CENH3 is found in all eukaryotes, our haploid production method should theoretically work in any plant species. We are testing our method in tomato (which completely lacks a haploid production method) and in Brassica species related to cabbage and canola. These plants can be crossed with many wild relatives, allowing us to test the effect of parental centromere DNA differences on haploid production. We have also initiated a haploid production project in cassava and banana/plantain, two crops that lack breeding tools yet are crucial for food security in sub-Saharan Africa.
Clonal reproduction through seeds (apomixis) could revolutionize agriculture, by allowing farmers to propagate vigorous hybrid genotypes. In collaboration with Raphael Mercier (French National Institute for Agricultural Research/INRA) and Imran Siddiqi (Centre for Cell and Molecular Biology, India), we crossed mutants that produce clonal diploid gametes to a plant with altered CENH3 proteins. Loss of chromosomes from the cenh3 mutant yielded up to 35 percent of progeny that were clones of their parent, showing that genome elimination can produce seeds from clonal gametes. This proof-of-principle experiment still required a cross between two mutants. We are working to create a clonal reproduction system in a self-fertilizing plant. In principle, apomixis can now be engineered by manipulating just four conserved genes.
Grants from the National Science Foundation and the March of Dimes provided partial support for these projects.