Clusters of alpha satellite on human chromosome 21 are dispersed far onto the short arm and lack ancient layers

Abstract

Human alpha satellite (AS) sequence domains that currently function as centromeres are typically flanked by layers of evolutionarily older AS that presumably represent the remnants of earlier primate centromeres. Studies on several human chromosomes reveal that these older AS arrays are arranged in an age gradient, with the oldest arrays furthest from the functional centromere and arrays progressively closer to the centromere being progressively younger. The organization of AS on human chromosome 21 (HC21) has not been well-characterized. We have used newly-available HC21 sequence data and an HC21p YAC map to determine the size, organization, and location of the AS arrays, and compared them to AS arrays found on other chromosomes. We find that the majority of the HC21 AS sequences are present on the p-arm of the chromosome and are organized into at least five distinct isolated clusters which are distributed over a larger distance from the functional centromere than that typically seen for AS on other chromosomes. Using both phylogenetic and L1 element age estimations, we found that all of the HC21 AS clusters outside the functional centromere are of a similar relatively recent evolutionary origin. HC21 contains none of the ancient AS layers associated with early primate evolution which is present on other chromosomes, possibly due to the fact that the p-arm of HC21 and the other acrocentric chromosomes underwent substantial reorganization about 20 million years ago. AbstractClick here to download Abstract ABSTRACT.docx List of AbbreviationsAS -alpha satellite HCX -human chromosome X HC21 -human chromosome 21HC21p -human chromosome 21 short arm HOR -higher order repeat SD -segmental duplications

A Game of Thrones at Human Centromeres I. Multifarious structure necessitates a new molecular/evolutionary model

“…Some HORs are shared among nonhomologous autosomes, e.g., chromosome groups (1, 5, and 19), (13 and 21) and (14 and 22) and each share a unique HOR in common. At least half of the chromosomes have more than one HOR present simultaneously (UCSC genome browser, GRCh38; Ziccardi et al 2016). On some chromosomes, more than one HOR is active (Maloney et al 2012;McNulty and Sullivan 2018), although current evidence indicates that only one HOR is active at one time on a single chromosome (e.g., Aldrup-MacDonaldet al 2016).…”

Classification and monomer-by-monomer annotation of suprachromosomal family 1 alpha satellite higher-order repeats in hg38 human genome assembly

“…For instance, in chromosome 8 which has SF2 live centromere, there are chunks of S1CMH1d in both p-and q-arms (chr8:43,936,718-43,966,062 and chr8:45,945,726-45,971,761, respectively) which seem to be segment duplications (SDs) of one another (parts are 96.5% identical). These SDs however are not not reflected in the UCSC SD track (see discussion of AS SDs in hg38 assembly in [1] and [21]). Whether these SDs indicate the former presence of SF1 centromere in chromosome 8, or they just came from another chromosome is not clear.…”

A Game of Thrones at Human Centromeres II. A new molecular/evolutionary model

“…To spread population-wide, assuming no natural selection or transmission advantage (as might occur from centromere drive; see below), the stochastic and slow process of random genetic drift would be required -unless some other mechanism spreads the longer HOR horizontally to other chromosomal lineages. The concerted evolution that has homogenized the active HORs on chromosomes 1, 5, and 19 (and also a different HOR on chromosomes 13 and 21 and another on chromosomes 14 and 22; Ziccardi et al 2016) demonstrates that some mode of horizontal transfer must be occurring at a nontrivial rate between different chromosomes sharing the same HOR. For the case of horizontal transfer between homologs, work by Roizes 2006 and Pironon et al 2010 provides strong empirical evidence for substantial horizontal transfer of 'diagnostic variant nucleotides' (DVNs) between HOR arrays found in different lineages -despite strong evidence against meiotic and/or mitotic crossovers between homologs.…”

“…About half of the 23 humans chromosomes have more than one centromeric HOR array (UCSC genome browser [GRCh38]; Ziccardi et al 2016), although current evidence indicates that only one…”

“…This arrangement predicts that repeat elements that are far from the active HOR array should have lost their HOR structure due to the long time period that was available for the accumulation of selectively favored point mutations, small inversions and indels. This structure is observed in the distal pericentric regions surrounding active HOR arrays (Ross et al 2005;Nusbaum et al 2006;Shepelev et al 2009;Ziccardi et al 2016): the distal flanks of the pericentric heterochromatin are made up of unorganized monomers that lack HOR structure and contain frequent small inversions and indels.…”