Satellitome comparison of two oedipodine grasshoppers highlights the contingent nature of satellite DNA evolution

Abstract

Background: The full catalog of satellite DNA (satDNA) within a same genome constitutes the satellitome. The Library Hypothesis predicts that satDNA in relative species reflects that in their common ancestor, but the evolutionary mechanisms and pathways of satDNA evolution have never been analyzed for full satellitomes. We compare here the satellitomes of two Oedipodine grasshoppers (Locusta migratoria and Oedaleus decorus) which shared their most recent common ancestor about 22.8 Ma ago.

Results: We found that about one third of their satDNA families (near 60 in every species) showed sequence homology and were grouped into 12 orthologous superfamilies. The turnover rate of consensus sequences was extremely variable among the 20 orthologous family pairs analyzed in both species. The satDNAs shared by both species showed poor association with sequence signatures and motives frequently argued as functional, except for short inverted repeats allowing short dyad symmetries and non-B DNA conformations. Orthologous satDNAs frequently showed different FISH patterns at both intra- and interspecific levels. We defined indices of homogenization and degeneration and quantified the level of incomplete library sorting between species.

Conclusions: Our analyses revealed that satDNA degenerates through point mutation and homogenizes through partial turnovers caused by massive tandem duplications (the so-called satDNA amplification). Remarkably, satDNA amplification increases homogenization, at intragenomic level, and diversification between species, thus constituting the basis for concerted evolution. We suggest a model of satDNA evolution by means of recursive cycles of amplification and degeneration, leading to mostly contingent evolutionary pathways where concerted evolution emerges promptly after lineages split.

Keywords: Cytogenomics; Library Hypothesis; Satellite DNA; Satellitome Evolution.

Fig. 1

FISH analysis of a pair of orthologous families, belonging to OSF12, in O.decorus (a) and L. migratoria (b). a OdeSat59-185 showed no FISH bands on this meiotic metaphase I cell, thus showing the NS pattern. b LmiSat01A-193 showed conspicuous pericentromeric FISH bands on most chromosomes of this embryo mitotic metaphase cell, thus showing the B pattern

Fig. 2

FISH analysis of three pairs of orthologous families in O.decorus and L. migratoria: showing the B pattern in both species for OSF1 (a and b), the NS and B patterns, respectively, for OSF3 (c and d), and the B and NS patterns, respectively, for OSF7 (e and f) (see also Table 2 for satDNA classification into OSFs). a Presence of pericentromeric FISH bands for OdeSat01-287 on all chromosomes of this meiotic metaphase II cell of O. decorus. b Note the presence of its orthologous family (LmiSat09-181) on a single chromosome pair of this embryo mitotic metaphase cell of L. migratoria. c Absence of FISH bands for OdeSat17-176 in a meiotic metaphase I cell of O. decorus. d Presence of its orthologous LmiSat02-176 on pericentromeric regions of several chromosome pairs and on whole B chromosome length (B) of this embryo mitotic metaphase cell of L. migratoria. e Presence of a pericentromeric FISH band on a single chromosome of the haploid set shown in this meiotic metaphase II cell of O. decorus. f Absence of FISH bands for LmiSat24-266 on the haploid chromosome set shown in this embryo mitotic metaphase cell of L. migratoria

Fig. 3

Definition of satDNA parameters in respect to abundance and divergence. The distribution of the abundances of groups of sequences differing by 1% divergence constitutes a repeat landscape (RL). It may be seen as a curve (left) or an histogram (right). In addition of variation in kurtosis, represented by several curves on the left, three properties of satDNA can be defined on RLs: DIVPEAK is the divergence class showing the highest abundance (3% in the histogram); PEAK-SIZE is the sum of the abundances of the five classes included around DIVPEAK, thus constituting the sum of all sequences differing by less than 5%, thus coinciding with our definition of satDNA subfamily; RPS is the relative peak size and represents the fraction of abundance which is included in the 5% amplification peak.

Fig. 4

Repeat landscape (RL) and minimum spanning tree (MST) of two orthologous superfamilies of satellite DNA in O. decorus and L. migratoria (OSF02 and OSF12). a OSF02 showed the highest consensus turnover rate (CTR = 2.86) found among the 20 values estimated between orthologous pairs of families in both species. Note that OSF02 showed large amplification peaks in both species (green curve in O. decorus and red curve in L. migratoria) and that the MST showed complete separation of OdeSat02 and LmiSat03 sequences. b OSF12 showed the lowest CTR estimate (0.26 between OdeSat59 and LmiSat01) and the MST (on the right) reveals that the consensus DNA sequences of these two satDNA families showed only two differences. Also note in the RL (on the left) that the OdeSat59 curve is very close to zero, as this is the satDNA family in O. decorus showing the lowest abundance, indicating that OSF12 is represented in this species as relict remains which, by chance, almost coincide in consensus sequence with the most abundant subfamily in L. migratoria (LmiSat01A), thus evidencing extreme incomplete lineage sorting (see other cases in Additional file 2: Fig. S1)

Fig. 5

Gardner-Altman plots comparing RPS, kurtosis and DIVPEAK between the L migratoria satDNA families being shared or non-shared with O. decorus. Note that shared satDNAs showed higher homogenization (higher RPS and kurtosis) and lower degeneration (5% effect size for mean difference in DIVPEAK) than non-shared ones, suggesting most recent amplification of the shared ones

Fig. 6

A model of satDNA evolution. We consider that evolutionary events are rather different at intra- and intergenomic levels. At intragenomic level, tandem duplication gives birth to a new tandem repeat and its reiteration yields many copies of identical noncoding sequences (satDNA amplification). The newly amplified satDNA displays RLs sharply leptokurtic (a). As time goes by, point mutation increases divergence among the amplified sequences and the curve progressively is flattened (b–e) and DIVPEAK (i.e., the divergence value showing the higher abundance) increases (i.e., the peak moves to the right in the a–e graphs). At any moment of this first amplification-degeneration cycle, another sequence undergoes amplification and begins a new cycle. This sets the satDNA family farther from degeneration and extinction because its average divergence decreases and now predominates a newly amplified subfamily with leptokurtic RL (we represent here three successive cycles of amplification; note that the differences in size among cycles are to facilitate drawing and have nothing to do with amplification level). In parallel, an intragenomic spread of the satDNA can occur at higher or lower extent (brown stars). A conceivable exit of these cycles is satDNA degeneration, when homology with the original sequence is lost. At intergenomic level, individual reproduction will mark the destiny of the different satDNA sequences in populations. When reproduction is differential, albeit random (drift) or non-random (selection), some sequences may become prevalent above others. At this respect, the mutational-hazard hypothesis is applicable to explain the limits to purifying selection in some species showing extremely high abundance of satDNA. Finally, we cannot rule out that, in some case, transmission drive could help satDNA to prosper and, even that positive selection may recruit satDNA for important functions, such as telomeric or centromeric functions