LETTER
LETTER
Are viral small RNA regulating Dengue virus
replication beyond serotype 2?
An article in PNAS by Hussain and Asgari
(1) suggests that a viral small RNA (vsRNA)
from the 3′ UTR of the Dengue virus (DENV)
genome is important for the autoregulation
of the viral replication in mosquito cells by
acting as a microRNA (miRNA) targeting
the NS1 region of the genome. The authors
performed a sequence conservation analy-
sis on the vsRNA-5 across the four sero-
types. Unfortunately, the analysis was carried
out using only one sequence from each sero-
type, which in the case of a genome segment
from the variable region of the 3′ UTR may
misrepresent the total virus diversity and,
thus, its interpretation may be misleading.
To better examine the vsRNA-5 conserva-
tion, the sequences provided in the article
by Hussain and Asgari (1) were first mapped
to their corresponding complete DENV
genome, and all of the publicly available com-
plete DENV genome sequences (n = 3,372)
were aligned and sequence logos were gen-
erated from the vsRNA-5 and the NS1 tar-
get region. The mapping revealed that the
DENV-1 to -3–vsRNA sequences localized
to the stem loop (SL) 2 of the 3′ UTR down-
stream of the stop codon. However, in contrast,
the DENV-4–vsRNA-5 sequence mapped to
a region 1,500 nt upstream of the stop codon,
where no secondary structure has been de-
scribed. Moreover, it has been shown that the
homologous region is deleted in the DENV-4
3′ UTR (2). Therefore, this serotype should
not have been used in the conservation
analysis.
Downloaded by guest on May 11, 2021
www.pnas.org/cgi/doi/10.1073/pnas.1409972111
Interestingly, the mapping of the DENV-
2–vsRNA-5 and NS1 sequences also re-
vealed two substitutions in the strain used
by Hussain and Asgari (1), 10299U > C and
3236A > G, respectively (Fig. 1A); the latter
leads to an amino acid substitution K1047R
in the NS1 protein. The DENV-2 NS1 and
vsRNA-5 nucleotide sequence logos also
revealed that these substitutions were found
only in 2.4% and 9.3% of all published
DENV-2 genomes, respectively.
Furthermore, contrary to what we might
have predicted if vsRNA-5 works as an
miRNA, the NS1 and vsRNA-5 nucleotide
sequence logos (Fig. 1A) show that the posi-
tions involved in the seed-region binding
exhibited poor conservation across the three
serotypes. The presence of highly conserved
positions is more likely explained by the NS1
amino acid preservation within each serotype
and the requirement of these nucleotides in
the 3′ UTR for the formation of the recently
described Xrn1-resistant RNA structure (3)
(Fig. 1 B and C).
To further assess the conservation in
the complementarity between the vsRNA-
5 and its target sequence within each
serotype, the minimal free energy for the
binding in the 3,372 sequences was calcu-
lated using the ViennaRNA package (4).
The average for the binding in DENV-2
(average ± SD: −20.64 ± 3.38 kcal/mol)
was found to be significantly lower (P <
0.001) than the average minimal free en-
ergy in DENV-1 (−17.04 ± 1.43 kcal/mol)
PNAS | July 22, 2014 | vol. 111 | no. 29 | E2915–E2916
and DENV-3 (−12.01 ± 0.67 kcal/mol), sug-
gesting a lack of conservation in the com-
plementarity of the base pairing in these two
serotypes, perhaps limiting the significance
of this mechanism to DENV-2.
Taken collectively, these data—and the fact
that the production of vsRNA has not been
experimentally shown in other serotypes—
suggest that the relevance of this mechanism
might be limited to DENV-2.
ACKNOWLEDGMENTS. I thank Mariano Garcia-Blanco
and Eng Eong Ooi for critically reading this letter.
Esteban Finol1
Program in Emerging Infectious Diseases,
Duke–National University of Singapore
Graduate Medical School, Singapore 169857
1 Hussain M, Asgari S (2014) MicroRNA-like viral small RNA from
Dengue virus 2 autoregulates its replication in mosquito cells. Proc
Natl Acad Sci USA 111(7):2746–2751.
2 Shurtleff AC, et al. (2001) Genetic variation in the 3′ non-coding
region of Dengue viruses. Virology 281(1):75–87.
3 Chapman EG, Moon SL, Wilusz J, Kieft JS (2014) RNA structures
that resist degradation by Xrn1 produce a pathogenic Dengue virus
RNA. eLife 3:e01892.
4 Lorenz R, et al. (2011) ViennaRNA Package 2.0. Algorithms Mol
Biol 6:26.
5 Rother M, Rother K, Puton T, Bujnicki JM (2011) ModeRNA: A tool
for comparative modeling of RNA 3D structure. Nucleic Acids Res
39(10):4007–4022.
Author contributions: E.F. designed research, performed research,
analyzed data, and wrote the paper.
The author declares no conflict of interest.
1
Email: esteban.finol@nus.edu.sg.
 Fig. 1. Sequence conservation analysis on the regions involved in the suggested miRNA mechanism. (A) The substitutions 3236A > G and 10299U > C in the DENV-2 New
Guinea C strain used by Hussain and Asgari (1) are highlighted in colors. Nucleotide sequence logos generated from the NS1 and vsRNA5 regions are also shown. The seed region
involved in the miRNA binding is underscored in red. (B) Amino acid sequence logos for the NS1 target region show the conservation within and across serotypes. (C) Secondary
and tertiary structure of the Xrn1-resistant structure in the SL2 and SL3 of DENV-2. Secondary structure suggested by Chapman et al. (3) based on chemical probing and
conservation. The vsRNA5 sequence is highlighted in black. Positions with an identity lower than 99% across the 1,056 DENV-2 sequences are circled. The 3D RNA model was built
based on the crystallography of the homologous xrRNA structure in the Murray Valley Encephalitis virus (Protein Data Bank ID: 4PQV) using ModeRNA, a software for comparative
3D RNA modeling (5).The RNA backbone is shown in different colors according to the secondary structure.
Downloaded by guest on May 11, 2021

E2916 | www.pnas.org/cgi/doi/10.1073/pnas.1409972111 Finol