Widmer Julian, Vitalis Andreas, Caflisch Amedeo
Department of Biochemistry, University of Zurich, Zurich, Switzerland.
Department of Biochemistry, University of Zurich, Zurich, Switzerland.
J Biol Chem. 2024 Dec;300(12):107998. doi: 10.1016/j.jbc.2024.107998. Epub 2024 Nov 17.
Most processes of life are the result of polyvalent interactions between macromolecules, often of heterogeneous types and sizes. Frequently, the times associated with these interactions are prohibitively long for interrogation using atomistic simulations. Here, we study the recognition of N6-methylated adenine (mA) in RNA by the reader domain YTHDC1, a prototypical, cognate pair that challenges simulations through its composition and required timescales. Simulations of RNA pentanucleotides in water reveal that the unbound state can impact (un)binding kinetics in a manner that is both model- and sequence-dependent. This is important because there are two contributions to the specificity of the recognition of the GmAC motif: from the sequence adjacent to the central adenine and from its methylation. Next, we establish a reductionist model consisting of an RNA trinucleotide binding to the isolated reader domain in high salt. An adaptive sampling protocol allows us to quantitatively study the dissociation of this complex. Through joint analysis of a data set including both the cognate and control sequences (GAC, AmAA, and AAA), we derive that both contributions to specificity, sequence, and methylation, are significant and in good agreement with experimental numbers. Analysis of the kinetics suggests that flexibility in both the RNA and the YTHDC1 recognition loop leads to many low-populated unbinding pathways. This multiple-pathway mechanism might be dominant for the binding of unstructured polymers, including RNA and peptides, to proteins when their association is driven by polyvalent, electrostatic interactions.
生命的大多数过程都是大分子之间多价相互作用的结果，这些大分子往往具有不同的类型和大小。通常，与这些相互作用相关的时间长得令人望而却步，无法使用原子模拟进行研究。在这里，我们研究了阅读结构域YTHDC1对RNA中N6-甲基腺嘌呤（mA）的识别，这是一对典型的同源配对，其组成和所需的时间尺度对模拟提出了挑战。在水中对RNA五核苷酸的模拟表明，未结合状态可以以模型和序列依赖的方式影响（非）结合动力学。这一点很重要，因为对GmAC基序识别的特异性有两个贡献：来自中央腺嘌呤相邻的序列及其甲基化。接下来，我们建立了一个简化模型，由一个RNA三核苷酸在高盐条件下与分离的阅读结构域结合组成。一种自适应采样协议使我们能够定量研究该复合物的解离。通过对包括同源和对照序列（GAC、AmAA和AAA）的数据集进行联合分析，我们得出对特异性、序列和甲基化的两个贡献都是显著的，并且与实验数据非常吻合。动力学分析表明，RNA和YTHDC1识别环的灵活性导致了许多低占据的解离途径。当非结构化聚合物（包括RNA和肽）与蛋白质的结合由多价静电相互作用驱动时，这种多途径机制可能在其中起主导作用。
