Adaptive evolution of centromere proteins in plants and animals

Journal of Biology BioMed Central Open Access Research article Adaptive evolution of centromere proteins in plants and animals Paul B Talbert, Terri D Bryson and Steven Henikoff Address: Howard Hughes Medical Institute, Fred Hutchinson Cancer Research Center, 1100 Fairview Avenue N, Seattle, WA 98109-1024, USA. Correspondence: Steven Henikoff. E-mail: Published: 31 August 2004 Received: 25 May 2004 Revised: 20 July 2004 Accepted: 22 July 2004 Journal of Biology 2004, 3:18 The electronic version of this article is the complete one and can be found online at http://jbiol.com/content/3/4/18 © 2004 Talbert et al., licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Abstract Background: Centromeres represent the last frontiers of plant and animal genomics. Although they perform a conserved function in chromosome segregation, centromeres are typically composed of repetitive satellite sequences that are rapidly evolving. The nucleosomes of centromeres are characterized by a special H3-like histone (CenH3), which evolves rapidly and adaptively in Drosophila and Arabidopsis. Most plant, animal and fungal centromeres also bind a large protein, centromere protein C (CENP-C), that is characterized by a single 24 amino-acid motif (CENPC motif). Results: Whereas we find no evidence that mammalian CenH3 (CENP-A) has been evolving adaptively, mammalian CENP-C proteins contain adaptively evolving regions that overlap with regions of DNA-binding activity. In plants we find that CENP-C proteins have complex duplicated regions, with conserved amino and carboxyl termini that are dissimilar in sequence to their counterparts in animals and fungi. Comparisons of Cenpc genes from Arabidopsis species and from grasses revealed multiple regions that are under positive selection, including duplicated exons in some grasses. In contrast to plants and animals, yeast CENP-C (Mif2p) is under negative selection. Conclusions: CENP-Cs in all plant and animal lineages examined have regions that are rapidly and adaptively evolving. To explain these remarkable evolutionary features for a single-copy gene that is needed at every mitosis, we propose that CENP-Cs, like some CenH3s, suppress meiotic drive of centromeres during female meiosis. This process can account for the rapid evolution and the complexity of centromeric DNA in plants and animals as compared to fungi. Background Centromeres are the chromosomal loci where kinetochores assemble to serve as attachment sites for the spindle microtubules that direct chromosome segregation during mitosis and meiosis. Despite this essential conserved function in all eukaryotes, centromere structure is highly variable, ranging from the simple short centromeres of budding yeast, which have a consensus sequence of approximately 125 base Journal of Biology 2004, 3:18 18.2 Journal of Biology 2004, Volume 3, Article 18 Talbert et al. pairs (bp) on each chromosome, to holokinetic centromeres that span the entire length of a chromosome [1]. In plants and animals, centromeres are large and complex, typically comprising megabase-sized arrays of tandemly repeated satellite sequences that are rapidly evolving [2] and may differ significantly between closely related species [3-5]. The failure of conventional cloning and sequencing assembly tools to adequately characterize rapidly evolving satellite sequences at centromeres has made them the last regions of most eukaryotic genomes to be well understood [1]. Although there is no discernable conservation of centromeric DNA sequences in disparate eukaryotes, considerable progress has been made in identifying common proteins that form the kinetochore [6]. A universal protein component of centromeric chromatin found in all eukaryotes that have been examined is a centromere-specific variant of histone H3 (CenH3), which replaces canonical H3 in centromeric nucleosomes [7,8]. CenH3s are essential kinetochore components yet, like centromeric DNA, they are rapidly evolving [1]. In both Drosophila [9] and Arabidopsis [10], this rapid evolution of CenH3s is associated with positive selection (adaptive evolution), and involves regions of CenH3 that are predicted to contact the centromeric DNA [9,11,12]. The finding of positive selection in a protein that is required at every cell division is remarkable. Ancient proteins with conserved function are expected to be under negative selection because they typically have achieved an optimal sequence, so new mutations tend to produce deleterious variants that are quickly eliminated from populations. The canonical histones are extreme examples of this type of protein. In contrast, recurrent positive selection generally occurs as a consequence of genetic conflict, for example in the ‘arms race’ between pathogen surface antigens and the immune-cell proteins that recognize them. In this case, a mutation in a surface antigen that allows the pathogen to escape detection and proliferate will trigger selection for a new immune receptor to fight the mutated pathogen, which can then mutate again, and so on. The evidence for positive selection of CenH3 proteins specifically in the regions that contact DNA thus suggests a conflict between centromeric DNA and a histone component of the nucleosome that packages it. Is it commonplace for eukaryotes to have such a conflict at their centromeres? Is the conflict unique to centromere-specific histones, or are other proteins that bind centromeres also involved in this conflict? Is conflict responsible for centromere complexity? To answer these questions, we investigated the evolution of a second common DNA-binding kinetochore protein. Of the handful of essential kinetochore proteins that are widely distributed among eukaryotes, only one class other http://jbiol.com/content/3/4/18 than CenH3 has been shown to bind centromeric DNA: centromere protein C (CENP-C), a conserved component of the inner kinetochore in vertebrates [13-16]. Human CENP-C binds DNA non-specifically in vitro [17-19] and binds centromeric alpha satellite DNA in vivo [20,21]. Vertebrate CENP-C and the yeast centromere protein Mif2p [22,23] share a 24 amino-acid motif (CENPC motif) that has also been found in kinetochore proteins in nematodes [24] and plants [25]. As expected for kinetochore proteins, disruption or inactivation of genes encoding proteins containing a CENPC motif (CENP-Cs) results in the failure of proper chromosome segregation [16,23,24,26-28]. Other than the defining CENPC motif, these proteins are dissimilar in sequence across disparate phyla. Such a small stretch of sequence conservation, accounting for less than 5% of the length of these 549-943 amino-acid proteins, is unexpected considering that CENP-Cs are encoded b (...truncated)