Genetics, epigenetics and back again: Lessons learned from neocentromeres
Section snippets
Genetic elements instructing centromere formation
Initial studies were performed in the budding yeast Saccharomyces cerevisiae, where the site of microtubule attachment became known as a ‘‘point centromere’’. Pioneering studies showed a mere 125 bp DNA sequence containing three conserved DNA elements (CDE I, II and III) to be key for active centromere formation [[1], [2], [3]] (Fig. 1A). The finding that the centromeric DNA sequence placed onto a plasmid carrying a yeast replication origin is all that is required for the plasmid to behave as a
CENP-A, an epigenetic centromere mark
Contemporaneous with the discovery of neocentromeres, it became apparent that centromeres feature an unusual chromatin structure. One of the initial set of centromere proteins discovered along with CENP-B and CENP-C was CENP-A, a histone-like component tightly associated with chromatin at centromeres, suggesting that it may function as a centromere-specific core histone [13,[20], [21], [22]], a notion confirmed by its cloning that revealed it to be a novel histone H3 variant [23].
CENP-A-based
Neocentromeres, origins of a centromere paradigm shift
The first clear indication for the epigenetic regulation of the centromere came from the discovery of human neocentromeres, the first of which was described in 1993. This neocentromere was derived from a rearrangement of chromosome 10, resulting in a centromere-containing ring chromosome and an acentric linear chromosome lacking any centromeric DNA. The latter acquired centromere proteins at a novel location, constituting a functional centromere that rescued mitotic maintenance of the
Artificial systems for neocentromere generation
In the pathogenic yeast Candida albicans, regional centromeres can expand from 4 to 15 kb and CENP-A is assembled into unique sequences that cover around 3 kb on each chromosome which are surrounded by direct or inverted repeats lacking classical pericentric heterochromatin [84]. By taking advantage of the higher rates of homologous recombination in Candida, a selectable marker was used to replace the centromere of chromosome 5. Surviving cells maintained the chromosome by neocentromere
Common elements driving the formation of neocentromeres
The accumulated collection of different neocentromeres, as well as the artificial generation of neocentromeres in different species allows us to define shared features that are required for centromere specification. All neocentromeres described up to date are universally marked by CENP-A [17,[91], [92], [93]] (Fig. 3A). While perhaps not surprising, it underscores the critical role of the centromere-specific histone variant in building the structural core of the kinetochore [94]. In addition to
Centromere size
Mapping neocentromere positions along the genome has allowed for a comparison of centromere domains and specific DNA sequences supporting centromere formation. One relevant parameter is centromere size. In S.pombe, the CENP-ACnp1 domain expanded to around 20 kb in the artificially generated neocentromeres, a size comparable to the endogenous centromeres (15 kb) [87]. A similar picture emerges from chicken DT40 cells whose endogenous non-repetitive centromeres expand to ~35 kb while
General features of centromeric chromatin
At the primary DNA sequence level, the most conserved feature across neocentromeres is the tendency to form preferentially on AT rich sequences [81,90], although the preference is slight and by no means sufficient to explain centromere formation. Moreover, some human neocentromeres are enriched in repetitive LINE elements and a potential role of the L1 retrotransposons in the regulation of neocentromere activity has been suggested [102]. At the chromatin level, more similarities arise. The
The role of higher order chromatin features in centromere function
From a more recent perspective, analysis performed in Candida, revealed that neocentromeres generated at different loci along the chromosome resulted in kinetochores that cluster with active endogenous centromeres within the nucleus [86]. This result suggested that the higher order chromatin organization may be another important feature in defining sites that can support centromere formation. Indeed, 4C analysis performed in chicken DT40 cells revealed that following neocentromere formation,
Centromere drifting of evolutionary new centromeres
As we have seen above, there is no universal agreement about what local features are responsible for neocentromere specification, and it is quite likely that random chance plays a major role. It is generally clear that chromatin-based epigenetic identity plays a dominant role. If indeed centromere inheritance is largely uncoupled form the underlying DNA sequence, this would make a clear prediction: Pure chromatin-based replication of centromere identity would be subject to stochastic
From genetic centromere specification to epigenetics and back again
The local drift of CENP-A domains is expected from a self-templating epigenetically defined mechanism for centromere specification. In fact, due to the existence of ENCs in repetitive DNA tracks, it has been hypothesized that a possible role for the repetitive centromeric DNA is to generate a safe space for CENP-A drifting to occur, avoiding migration towards genomic regions containing essential genes [135] (Fig. 4). Thus, the presence of alpha-satellite DNA may promote centromere maintenance
Future perspectives
Through the study of neocentromeres we have discovered several salient features of centromeric chromatin structure that specifies centromere positions. While there is no broad consensus on exactly what are the local DNA sequence and chromatin features favourable for centromeres to form, many advances have been made regarding the elements required for centromere specification.
Artificial systems for neocentromere formation, developed in different model organisms have been particularly powerful
Acknowledgements
Salary and research support to MMP and LETJ is provided by an ERC-consolidator grant ERC-2013-CoG-615638 and a Senior Research Fellowship in Basic Biomedical Science.
References (147)
Centromeres: proteins, protein complexes, and repeated domains at centromeres of simple eukaryotes
Curr. Opin. Genet. Dev.
(1998)Centromeres of mammalian chromosomes
Trends Genet.
(1990)- et al.
Transmission of a fully functional human neocentromere through three generations
Am. J. Hum. Genet.
(1999) - et al.
The case for epigenetic effects on centromere identity and function
Trends Genet.
(1997) - et al.
Cse4p is a component of the core centromere of Saccharomyces cerevisiae
Cell
(1998) - et al.
Characterization of chicken CENP-A and comparative sequence analysis of vertebrate centromere-specific histone H3-like proteins
Gene
(2003) - et al.
Discovery of unconventional kinetochores in kinetoplastids
Cell
(2014) - et al.
The CENP-L-N complex forms a critical node in an integrated meshwork of interactions at the centromere-kinetochore interface
Mol. Cell.
(2015) - et al.
CENP-C is a structural platform for kinetochore assembly
Curr. Biol.
(2011) - et al.
CCAN makes multiple contacts with centromeric DNA to provide distinct pathways to the outer kinetochore
Cell
(2008)