Background Protein phosphorylation is the best studied post-translational modification strongly influencing protein function. Phosphorylated amino acids not only differ in physico-chemical properties from non-phosphorylated counterparts, but also exhibit different evolutionary patterns, tending to mutate to and originate from negatively charged amino acids (NCAs). The distribution of phosphosites along protein sequences is non-uniform, as phosphosites tend to cluster, forming so-called phospho-islands. Methods Here, we have developed a hidden Markov model-based procedure for the identification of phospho-islands and studied the properties of the obtained phosphorylation clusters. To check robustness of evolutionary analysis, we consider different models for the reconstructions of ancestral phosphorylation states. Results Clustered phosphosites differ from individual phosphosites in several functional and evolutionary aspects including underrepresentation of phosphotyrosines, higher conservation, more frequent mutations to NCAs. The spectrum of tissues, frequencies of specific phosphorylation contexts, and mutational patterns observed near clustered sites also are different.

中文翻译：

---

磷酸岛和哺乳动物中磷酸化氨基酸的进化

背景蛋白质磷酸化是研究得最好的翻译后修饰，它强烈影响蛋白质功能。磷酸化氨基酸不仅在物理化学性质上与非磷酸化氨基酸不同，而且表现出不同的进化模式，倾向于突变为带负电荷的氨基酸 (NCA)。磷酸位点沿蛋白质序列的分布是不均匀的，因为磷酸位点倾向于聚集，形成所谓的磷酸岛。方法 在这里，我们开发了一种基于隐马尔可夫模型的磷酸岛识别程序，并研究了获得的磷酸化簇的性质。为了检查进化分析的稳健性，我们考虑了重建祖先磷酸化状态的不同模型。结果 聚集的磷酸位点在几个功能和进化方面与单个磷酸位点不同，包括磷酸酪氨酸的代表性不足、更高的保守性、更频繁的 NCA 突变。在聚集位点附近观察到的组织谱、特定磷酸化背景的频率和突变模式也不同。