Widespread signatures of local mRNA folding structure selection in four Dengue virus serotypes

Abstract

BackgroundIt is known that mRNA folding can affect and regulate various gene expression steps both in living organisms and in viruses. Previous studies have recognized functional RNA structures in the genome of the Dengue virus. However, these studies usually focused either on the viral untranslated regions or on very specific and limited regions at the beginning of the coding sequences, in a limited number of strains, and without considering evolutionary selection.ResultsHere we performed the first large scale comprehensive genomics analysis of selection for local mRNA folding strength in the Dengue virus coding sequences, based on a total of 1,670 genomes and 4 serotypes. Our analysis identified clusters of positions along the coding regions that may undergo a conserved evolutionary selection for strong or weak local folding maintained across different viral variants. Specifically, 53-66 clusters for strong folding and 49-73 clusters for weak folding (depending on serotype) aggregated of positions with a significant conservation of folding energy signals (related to partially overlapping local genomic regions) were recognized. In addition, up to 7% of these positions were found to be conserved in more than 90% of the viral genomes. Although some of the identified positions undergo frequent synonymous / non-synonymous substitutions, the selection for folding strength therein is preserved, and thus cannot be trivially explained based on sequence conservation alone.ConclusionsThe fact that many of the positions with significant folding related signals are conserved among different Dengue variants suggests that a better understanding of the mRNA structures in the corresponding regions may promote the development of prospective anti- Dengue vaccination strategies. The comparative genomics approach described here can be employed in the future for detecting functional regions in other pathogens with very high mutations rates.

Evidence of translation efficiency adaptation of the coding regions of the bacteriophage lambda

“…28 This model was initially introduced in the context of enumerative combinatorics and was used by us before for studying other viruses. 16,17 We used the original source code which can be found in http://csb.cs. mcgill.ca/sparcs (7 February 2017, date last accessed).…”

“…To estimate the typical decoding time of each codon based on NFC distributions, we used a novel statistical model, 17 which takes into consideration the skewed nature of the NFC distribution. The aim is to describe the NFC histogram of each codon as an output of a random variable which is a sum of two random variables: a normal and an exponential variable.…”

“…For each gene 100 randomized MFE-profiles variants were computed basing on randomized sequences generated by the dinucleotide preserving and codon preserving randomization models (both preserving also the encoded proteins, see section above). 16,17,28 We did not change the UTRs in the randomization as in this study we are interested in the coding regions.…”

“…Local genomic regions (both in UTRs and in coding sequences) selected for strong mRNA folding may be related to various functional mRNA structures in the coding regions that effect the viral life cycle 16,17,34 ; specifically, it is possible that the mRNA structures are more important in the late genes due to less canonical regulatory aspects needed at the later phases of viral life. Interestingly, we found that in the case of the late genes there is an evidence of selection for weak folding at the 5 0 end of the coding regions, suggesting that there is a significant evolutionary pressure for improved initiation efficiency.…”

“…For example, a recent study by Liu et al 9 employed ribosome profiling to study the progress of bacteriophage lambda gene expression during phage development and showed that the known genes are expressed in a predictable fashion; in addition many previously unknown potential open reading frames were detected. Other studies focused on different aspects of viral translation (and lifecycle in general) regulation that has a significant effect on shaping viral genomes, such as: secondary structures within viral transcripts (mainly in untranslated regions 16,17 but also within coding regions [18][19][20] ) and evolutionary pressure on synonymous codons usage bias. [21][22][23][24][25] However, to the best of our knowledge, this is the first study aiming to perform a comprehensive, large scale analysis of different types of genomic synonymous information related to regulation of translation efficiency in all coding regions of a bacteriophage lambda, during different stages of its development.…”