Prediction of RNA secondary structure in NS1


Liu et al. showed that A to V amino acid substitution at position 188 in NS1 enhanced zika virus (ZIKV) infectivity (Liu et al., 2017). It was expected that A to V aa substitution affect the NS1 dimerization. However, the A to V substitution at position 188 does not influcnce NS1 dimerization. Thus the mechanism by which the A to V substitution affects NS1 phenotype remains unclear.

Many RNA viruses including flaviviruses form functional RNA secondary structure in their genomes. Thus we investigated whether the RNA secondary structure was involved in the phenotypic change of NS1 by A to V change.
A to V substitution at position 188 is caused by a single nucleotide change from C to T (Fig. 1). When the RNA secondary structure in NS1 of FSS13025 strain was predicted using RNAfold program with temperature setting 37℃, the 188 position was located in the part of stem-loop structure predicted in NS1 (Fig.2a).
However, when the stem-loop was predicted after changing the nucltotide sequence from C to T, overall structure of the stem-loop was not changed (Fig.2 c). This result was not changed even the secondary structure was predicted under the temperaure setting 18℃, which is incubation temperature for mosquito cell line.
Thus the stem-loop predicted here is unlikely to be involvement in the phenotypic change of ZIKV NS1 by C to T change.

Liu Y, Liu J, Du S, Shan C, Nie K, Zhang R, Li XF, Zhang R, Wang T, Qin CF, Wang P, Shi PY, Cheng G. (2017) Evolutionary enhancement of Zika virus infectivity in Aedes aegypti mosquitoes. Nature. 545:482-486. doi: 10.1038/nature22365.


Thanks for sharing this - it is very interesting.

Liu et al. showed that 188V produced more extracellular NS1 than 188A. Do you have any hypotheses to why this mutation, with little structural change, would then lead to enhanced release of NS1? The whole coding sequence gets transcribed as one polyprotein, then cleaved into the individual components. Would it influence any cleavage sites? So I am just wondering what mechanisms would lead to more NS1 release w/o increasing replication or translation.




Is there any additional evidence of the presence of a conserved RNA structure? dN/dS ratios across an alignment of various strains or MFED values comparing to same region scrambled (but maintaining codons and codon pair bias etc)?


This is interesting, thanks for sharing. Gotta say though, despite what RNAfold says, I still find it likely that this particular mutation could cause a structural change - experimental validation would be required since structural prediction programs are far from perfect predictors. While a single mutation almost certainly wouldn’t change the overall structure, small differences could be critically important.

I’m not sure I quite understand your figure 2, but if I run RNAfold with C or T, I do see a difference (same as your Fig. 2b > 2d).