Department of Biology, College of Arts & Sciences, Saint Louis University, Saint Louis, Missouri, USA.
Program of Bioinformatics and Computational Biology, School of Science and Engineering, Saint Louis University, Saint Louis, Missouri, USA.
Protein Sci. 2023 Jan;32(1):e4531. doi: 10.1002/pro.4531.
Detection of homologous relationships among proteins and understanding their mechanisms of diversification are major topics in the fields of protein science, bioinformatics, and phylogenetics. Recent developments in sequence/profile-based and structural similarity-based methods have greatly facilitated the unification and classification of many protein families into superfamilies or folds, yet many proteins remain unclassified in current protein databases. As one of the three earliest identified RNases in biology, ribonuclease T2, also known as RNase I in Escherichia coli, RNase Rh in fungi, or S-RNase in plant, is thought to be an ancient RNase family due to its widespread distribution and distinct structure. In this study, we present evidence that RNase T2 represents a circularly permutated version of the BECR (Barnase-EndoU-Colicin E5/D-RelE) fold RNases. This subtle relationship cannot be detected by traditional methods such as sequence/profile-based comparisons, structure-similarity searches, and circular permutation detections. However, we were able to identify the structural similarity using rational reconstruction of a theoretical RNase T2 ancestor via a reverse circular permutation process, followed by structural modeling using AlphaFold2, and structural comparisons. This relationship is further supported by the fact that RNase T2 and other typical BECR RNases, namely Colicin D, RNase A, and BrnT, share similar catalytic site configurations, all involving an analogous set of conserved residues on the α0 helix and the β4 strand of the BECR fold. This study revealed a hidden root of RNase T2 in bacterial toxin systems and demonstrated that reconstruction and modeling of ancestral topology is an effective strategy to identify remote relationship between proteins.
蛋白质同源关系的检测和对其多样化机制的理解是蛋白质科学、生物信息学和系统发生学领域的主要课题。基于序列/轮廓和结构相似性的方法的最新发展极大地促进了许多蛋白质家族统一和分类为超家族或折叠,但许多蛋白质在当前的蛋白质数据库中仍然没有分类。核糖核酸酶 T2 是生物学中最早鉴定的三种核糖核酸酶之一,也称为大肠杆菌中的 RNase I、真菌中的 RNase Rh 或植物中的 S-RNase,由于其广泛的分布和独特的结构,被认为是一种古老的核糖核酸酶家族。在本研究中,我们提供的证据表明,核糖核酸酶 T2 代表了 BECR(Barnase-EndoU-Colicin E5/D-RelE)折叠核糖核酸酶的环状重排版本。这种细微的关系不能通过传统的方法如序列/轮廓比较、结构相似性搜索和环状重排检测来检测。然而,我们能够通过一种反向环状重排过程,对理论上的核糖核酸酶 T2 祖先进行理性重建,然后使用 AlphaFold2 进行结构建模和结构比较,来识别这种结构相似性。核糖核酸酶 T2 和其他典型的 BECR 核糖核酸酶,如 Colicin D、RNase A 和 BrnT,具有相似的催化位点配置,都涉及 BECR 折叠的α0 螺旋和β4 链上类似的一组保守残基,这进一步支持了这种关系。这项研究揭示了细菌毒素系统中核糖核酸酶 T2 的隐藏根源,并证明了祖先拓扑结构的重建和建模是识别蛋白质之间远程关系的有效策略。