A “microRNA-like” small RNA expressed by Dengue virus?

In PNAS, Hussain and Asgari claim they identified a “microRNA-like” 23-nucleotide RNA expressed by Dengue virus 2 (DENV2), called viral small RNA (vsRNA)-5, in infected mosquito and mammalian cells (1). vsRNA-5, which could also be detected by Northern blot or RT-PCR, was reported to repress DENV2 replication in insect Aag2 cells by ∼10,000-fold, as revealed by inhibition of vsRNA-5 by “RNA oligos with reverse complementarity.” vsRNA-5 was reported to be dependent on insect Argonaute 2 (Ago2), but not Dicer1 or Dicer2, for its biogenesis and was also Ago2-associated. Finally, the authors identify the DENV2 NS1 gene as an mRNA target for vsRNA-5.

These data were surprising as deep sequencing of small RNAs expressed in DENV2-infected mosquitoes, Aag2 cells, or mammalian cells had previously failed to reveal any DENV2-encoded microRNA-like RNAs, although viral small interfering RNAs (siRNAs) were, as expected, abundant in insect cells, which use RNA interference (RNAi) as an antiviral response (2–4). This discrepancy is perhaps explained by the fact that the DENV2-derived small RNAs of 20–24 nucleotides reported by Hussain and Asgari derived from 484 different locations on the DENV2 genome represented ≤0.03% of the ∼671,000 small RNAs recovered from DENV2-infected mosquitoes. That is, vsRNA-5 was expressed at vanishingly low levels.

A key point about microRNAs is that they stoichiometrically bind to partially complementary mRNA targets to inhibit mRNA function. Therefore, microRNAs must be expressed at high levels to effectively repress cellular or viral mRNAs that show partial complementarity to the microRNA (5). Hussain and Asgari claim that vsRNA-5 inhibits DENV2 via a partially complementary target located within the NS1 ORF, and vsRNA-5 would therefore have to be expressed at a level sufficient to bind a high percentage of all DENV2 mRNAs, which are expressed at thousands of copies per infected cell, as well as host mRNAs that bear seed complementarity. However, as noted above, vsRNA-5 is expressed at levels so low that it was not even detected by previous DENV2 deep sequencing studies (2–4)! This appears inconsistent with the authors’ claim that loss of vsRNA-5 function enhances DENV2 replication by 10,000-fold, especially as microRNAs are ineffective at repressing mRNAs via partially complementary target sites located in actively translated ORFs. Moreover, the simple cRNA oligonucleotides used in this study to inhibit vsRNA-5 are not normally effective inhibitors of microRNA function. Finally, it is hard to explain why, if this small RNA is so deleterious, DENV2 mutants do not arise that have spontaneously mutated either vsRNA-5 itself or its putative target site in NS1.

Given these concerns, we feel that the evidence presented by Hussain and Asgari falls short of demonstrating that DENV2 encodes a microRNA-like small viral RNA or that this RNA functions as a virally encoded repressor of viral replication. Rather, we believe vsRNA-5 is likely part of the heterogeneous population of siRNAs of DENV2 origin, generated by the antiviral RNAi response in infected insect cells, that are known to be bound by Ago2.