Figure 5

Hypothesis on satellite DNA evolution, based on the fact that all kinds of elements can be found clustered or non-clustered, at cytological level and show many common satellitome properties.

The birth of a satDNA implies a de novo duplication of a genomic sequence of two or more bp. This can occurs, for instance, by means of replication slippage in the case of short satDNAs or rolling circle replication in the case of long ones. This gives rise to a short array (<1.5 kb, i.e. the sensitivity FISH threshold) at a single genomic location (small dot in (a)). This short array is then disseminated throughout the genome by unknown mechanisms, although transposable elements or rolling circle replication and reinsertion elsewhere might be good candidates (b). All satDNAs remain at this stage in prokaryote species, where genomic constraints and natural selection (represented by double vertical bars) pose rigid limits to satDNA accumulation and some of them remain this way in eukaryotes appearing as non-clustered satDNAs (b). In eukaryotes, however, any of the short arrays can undergo local amplification surpassing the 1.5 kb thus becoming into a clustered satDNA and being visible by FISH (c). The fact that all clustered satDNAs found in the L. migratoria genome were found in many different contigs of the assembled genome provides strong support to the hypothesis that dissemination precedes clustering. Local amplification implies rapid increase in array size and could take place, for instance, by unequal crossing over. Based on our simulation of random genomes of the L. migratoria size, satDNA arrays of 15 bp or less can appear by chance at many genomic locations (>15) (Supplementary Table S9). For this reason, all microsatellites and the shortest minisatellites can start their life-cycle at stage (b). Of course, further research is necessary to unveil the precise mechanisms involved in reaching each stage, as those included here are only suggestions based on current literature.