A TALE-inspired computational screen for proteins that contain approximate tandem repeats

TAL (transcription activator-like) effectors (TALEs) are bacterial proteins that are secreted from bacteria to plant cells to act as transcriptional activators. TALEs and related proteins (RipTALs, BurrH, MOrTL1 and MOrTL2) contain approximate tandem repeats that differ in conserved positions that define specificity. Using PERL, we screened ~47 million protein sequences for TALE-like architecture characterized by approximate tandem repeats (between 30 and 43 amino acids in length) and sequence variability in conserved positions, without requiring sequence similarity to TALEs. Candidate proteins were scored according to their propensity for nuclear localization, secondary structure, repeat sequence complexity, as well as covariation and predicted structural proximity of variable residues. Biological context was tentatively inferred from co-occurrence of other domains and interactome predictions. Approximate repeats with TALE-like features that merit experimental characterization were found in a protein of chestnut blight fungus, a eukaryotic plant pathogen.转录激活因子样效应蛋白（transcription activator-like effector, TALE，复数形式TALEs）是一类由细菌分泌至植物细胞内的细菌蛋白，可作为转录激活因子发挥作用。TALEs及其相关蛋白（包括RipTALs、BurrH、MOrTL1与MOrTL2）均含有近似串联重复序列，这些重复序列在决定蛋白特异性的保守位点处存在序列差异。本研究采用PERL语言，对约4700万条蛋白质序列进行筛查，以鉴定具备TALE样结构特征的蛋白：这类结构的特征为包含长度介于30至43个氨基酸的近似串联重复序列，且保守位点存在序列变异，且无需与TALEs具有序列同源性。候选蛋白的评分依据包括其核定位倾向、二级结构特征、重复序列复杂度，以及可变残基的共变异情况与预测结构邻近性。研究人员通过其他结构域的共现情况与相互作用组预测结果，对候选蛋白的生物学背景进行了初步推断。本研究在一种真核植物病原体——板栗疫病菌的某一蛋白中，发现了具备TALE样特征、值得开展实验表征的近似串联重复序列。