Title: Toleration of Frameshift Mutations in Mrna Sequences Encoding the N-terminal Peptides of Bacterial Type III Effectors
Yejun Wang,Litao Yang,Moyang Lv,M. Y. Yu,Xinlong Wang,Ziyi Zhao,Lijun Luo,Xuxia Cai,Runhong Chen,Yueming Hu
DOI: https://doi.org/10.1101/2023.04.10.536236
2023-01-01
Abstract:Gram-negative bacteria deliver effector proteins into eukaryotic host cells through type III and type IV secretion systems, causing infections and diseases. It remains unclear about the signals guiding the specific secretion of the effectors. Here, we adopted an in silico approach to analyze the mRNA sequences encoding the putative peptides essential for effective secretion and translocation of type III and IV effectors. A surprisingly high proportion of type III effectors showed tolerance on frameshift mutations in signal-encoding mRNA sequences, and in contrast, very low percentage of type IV effectors showed the similar frameshift tolerance. The type III effectors with frameshift tolerance of secretion signals were widely distributed in effector or signal families and bacterial species. Natural frameshifts could be identified in type III effector genes, which were often remedied in time by nearby paired insertions or deletions. Frameshift-derived peptide sequences also retained the common properties present in the signal peptides of raw type III effectors. Natural language processing models were adopted to represent the common features in the mRNA sequences encoding N-terminal peptides of type III effectors or C-terminal peptides of type IV effectors, with which transfer learning models could well predict the effectors, especially type IV effectors. The observations in the study would facilitate us understand the nature and evolution of secretion signals of type III and IV effectors. Significance It has been a debate on the nature of signals for translocation of type III secreted effectors for a long time. Meanwhile, there has been no examination on the possibility of mRNA being as translocation signals for type IV or other types of secreted effectors. By computational simulation, the study demonstrated the protein nature of translocation signals for both type IV effectors and most type III effectors. Despite wide frameshift tolerance and atypical common features in mRNA sequences encoding the putative N-terminal signal sequences of type III effectors, more typical common physicochemical and amino acid composition properties between the mutation-derived and raw peptides, and the frequent self-correction phenomenon for naturally happening frameshifts supported the translocation signals at protein level of type III effectors. The common features in mRNA sequences encoding the translocation signal peptides of type III and IV effectors could also be combined in models for better prediction of the effectors respectively.