Molecular Mutation Patterns Of Angiosperm Ribosomal Internal Transcribed Spacer 2(ITS2) Under Secondary Structure Constraints

Internal transcribed spacer 2(ITS2)is one of the most used molecular markers in angiosperm molecular systematics and it is also the only widely used nuclear gene marker.It is noteworthy that ITS2 rDNA plays an important role in the form of secondary structure after transcription.Under functional constrains,a substitution on one side of a base pair of the secondary structure will be compensated by a substitution on the other side of that base pair to restore the stability of the structure.This special base mutation phenomenon is compensatory base-pair change(CBC).According to the mutation process of CBC,it can be inferred that the paired sites of ITS2 mutate once but the bases of this site change twice.The CBC phenomenon indicates that the phylogenetic analysis which based on basic hypothesis of"independence and randomness of base site mutation" is not fully applicable to ITS2 sequence.Therefore,in-depth study of the molecular mutation patterns under the secondary structural constraints of ITS2 and the construction of a new RNA evolution model can further improve the accuracy of ITS2-based phylogenetic analysis.In this study,13741 ITS2 sequences were sampled,including 55 families,80 genera and 3934 species of angiosperms,mainly covering the major orders of angiosperms.We divided these sequences by genus and use LocARNA to align the sequences and predict consensus secondary structures.Then according to the consensus structure information,the ITS2 paired regions and the unpaired regions were determined.Base on these sequence-structure matrixes,we carried out bioinformatics analysis in three aspects:(1)confirm the authenticity and process of AC-mediated CBC:RNAstat software was used to transform ITS2 sequence-structure matrixes into 28-symbol matrixes to detect accurate base-pairing stat;PHASE3.0 was used to construct the evolutionary tree,and mapped the detected AU/AC/GC base pairs on the evolutionary tree.Fitch method was used to determine the evolution direction.(2)Test the frequency and rate of each CBC process:16 RNA-specific evolution models and 2 DNA substitution models were used in PHASE3.0 software package to detect the best-fit mixed models of different groups'ITS2 sequences.According to the best-fit evolution models,PHASE was used to calculate the frequency of base pairs and the substitution rate between bases in ITS2 paired regions.(3)Verification of the accuracy of the results:4SALE was used to obtain independent ITS2 sequence-structure matrixes,by which the subsequent analyses performed following the step(2),Result differences was then compared.The ITS2 consensus secondary structures of 80 genera or lineages we obtained basically conformed to the conservative "one ring and four arms" model.The results showed that the best-fit evolution models of the ITS2 paired regions of the 80 groups were all RNA models,among which 61 groups were RNA 16C and the remaining 19 were RNA 16A.These results confirmed that the evolutionary pattern of ITS2 paired region was different from the traditional DNA sequences,indicating the necessity of using RNA model in ITS2 phylogenetic analysis.The phylogenetic relationship based on RNA evolution model clearly showed that there were CBCs with AC/CA intermediates(for example,GC?AC?AU in Aralia)in Aralia,Astilbe,Celastrus and Meconopsis.Statistical results of CBC sites based on the 28-character matrixes showed 353 CBC sites in 65/80 groups,among which 189 CBC sites with GU/UG intermediates and 26 CBC sites with AC/CA intermediates.In the remaining 15/80 groups,AC/CA or GU/UG-mediated hemi-CBCs were found.Further calculation and statistics based on evolution models showed that the average frequency of AC in ITS2 paired regions(0.0111±0.0108)was only 14%of GU(0.0786 ±0.0302),but the mutability of AC(3.2122 ±2.2888)was twice that of GU(1.6389 ± 0.9052).The overall rate of AC/CA-mediated CBCs(0.1174 ±0.0572)is half that of GU/UG-mediated CBCs(0.2264 ± 0.1087).Thus the statistics results of large datasets further provide the evidence of the existence of AC/CA-mediated CBCs and their relative frequencies and substitution rates.We used 4SALE to independently test and analyze three lineages,and found that the statistical results of CBC sites,relative base pair frequencies and substitution rates were consistent with the results by using LocARNA,indicating that our findings about AC intermediates and CBCs with AC intermediate are not significantly affected by ITS2 structure predicted models.In summary,the method of inferring the intermediate state of RNA molecules based on RNA substitution models in this paper address the difficulty of detecting unstable base pairs in previous study of RNA crystal structure,and thus provided a new approach for study of RNA structure and functions.The newly detected AU-AC-GC mutation pathway and its relative rates extend the traditionally recognized AU-GU-GC pathway,providing a theoretical basis and data support for developing new ITS2 RNA evolution models.