Centromeres in eukaryotes are composed of highly repetitive DNAs, which evolve rapidly and are thought to achieve a favorable structure in mature centromeres. However, how the centromeric repeat evolves into an adaptive structure is largely unknown. We characterized the centromeric sequences of Gossypium anomalum through chromatin immunoprecipitation against CENH3 antibodies. We revealed that the G. anomalum centromeres contained only retrotransposon-like repeats but were depleted in long arrays of satellites. These retrotransposon-like centromeric repeats were present in the African–Asian and Australian lineage species, suggesting that they might have arisen in the common ancestor of these diploid species. Intriguingly, we observed a substantial increase and decrease in copy numbers among African–Asian and Australian lineages, respectively, for the retrotransposon-derived centromeric repeats without apparent structure or sequence variation in cotton. This result indicates that the sequence content is not a decisive aspect of the adaptive evolution of centromeric repeats or at least retrotransposon-like centromeric repeats. In addition, two active genes with potential roles in gametogenesis or flowering were identified in CENH3 nucleosome-binding regions. Our results provide new insights into the constitution of centromeric repetitive DNA and the adaptive evolution of centromeric repeats in plants.

真核生物的着丝粒由高度重复的 DNA 组成，它们进化迅速，并被认为在成熟的着丝粒中实现了有利的结构。然而，着丝粒重复序列如何进化成适应性结构尚不清楚。我们通过针对 CENH3 抗体的染色质免疫沉淀来表征异常棉的着丝粒序列。我们发现，G. anomalum着丝粒仅包含逆转录转座子样重复序列，但在长卫星阵列中被耗尽。这些反转录转座子样着丝粒重复存在于非洲-亚洲和澳大利亚谱系物种中，表明它们可能出现在这些二倍体物种的共同祖先中。有趣的是，我们观察到，在棉花中没有明显结构或序列变异的逆转录转座子衍生的着丝粒重复序列中，非洲-亚洲和澳大利亚谱系的拷贝数分别大幅增加和减少。该结果表明序列内容不是着丝粒重复或至少是逆转录转座子样着丝粒重复的适应性进化的决定性方面。此外，在 CENH3 核小体结合区域中还鉴定出了两个在配子发生或开花中具有潜在作用的活性基因。我们的研究结果为植物着丝粒重复DNA的构成和着丝粒重复的适应性进化提供了新的见解。