Identification of novel, highly expressed retroviral microRNAs in cells infected by bovine foamy virus

Abstract

While numerous viral microRNAs (miRNAs) expressed by DNA viruses, especially herpesvirus family members, have been reported, there have been very few reports of miRNAs derived from RNA viruses. Here we describe three miRNAs expressed by bovine foamy virus (BFV), a member of the spumavirus subfamily of retroviruses, in both BFV-infected cultured cells and BFV-infected cattle. All three viral miRNAs are initially expressed in the form of an ∼ 122-nucleotide (nt) pri-miRNA, encoded within the BFV long terminal repeat U3 region, that is subsequently cleaved to generate two pre-miRNAs that are then processed to yield three distinct, biologically active miRNAs. The BFV pri-miRNA is transcribed by RNA polymerase III, and the three resultant mature miRNAs were found to contribute a remarkable ∼ 70% of all miRNAs expressed in BFV-infected cells. These data document the second example of a retrovirus that is able to express viral miRNAs by using embedded proviral RNA polymerase III promoters.

Importance: Foamy viruses are a ubiquitous family of nonpathogenic retroviruses that have potential as gene therapy vectors in humans. Here we demonstrate that bovine foamy virus (BFV) expresses high levels of three viral microRNAs (miRNAs) in BFV-infected cells in culture and also in infected cattle. The BFV miRNAs are unusual in that they are initially transcribed by RNA polymerase III as a single, ∼ 122-nt pri-miRNA that is subsequently processed to release three fully functional miRNAs. The observation that BFV, a foamy virus, is able to express viral miRNAs in infected cells adds to emerging evidence that miRNA expression is a common, albeit clearly not universal, property of retroviruses and suggests that these miRNAs may exert a significant effect on viral replication in vivo.

FIG 1

Alignment of small RNA reads to the BFV genome. (A) RNA reads that mapped to the viral genome are shown relative to their genomic origin (x axis). The number of reads is plotted as a percentage of the total number of small RNA deep-sequencing reads (y axis) from MDBK cells at 72 h postinfection (top) or from chronically infected MDBK cells (bottom). (B) Assignment of deep-sequencing reads in chronically BFV-infected MDBK cells.

FIG 2

Small RNAs expressed by BFV cluster at the size expected for miRNAs. Reads from chronically BFV-infected MDBK cells mapping to the BFV genome for a given nucleotide length (x axis) are graphed as percentages of total reads aligning to the BFV genome (y axis).

FIG 3

Origin of BFV miRNAs. (A) Predicted sequence and RNA secondary structure of the BFV pri-miRNA. Black triangles indicate the sites of cleavage during BFV pri-miRNA processing to yield the two viral pre-miRNAs. (B) Paired-end deep sequencing of BFV sequences of ≥30 nt. The y axis shows the number of reads containing a given nucleotide divided by the total number of viral reads in the library. Values obtained with the forward sequencing primer are plotted on the positive y axis, with values obtained with the reverse primer plotted on the negative y axis. (C) Proviral DNA sequence with potential Pol III transcription factor binding sites and termination sequences boxed. Numbers represent the nucleotide locations within the reference BFV strain.

FIG 4

Northern blot analysis of RNAs from uninfected (−) and chronically BFV-infected (+) MDBK cells. The antisense oligonucleotide probes used for the indicated lanes are shown at the top. The predicted migration locations of the BFV pri-miRNA, the two pre-miRNAs, and the mature miRNAs are indicated.

FIG 5

BFV miRNAs are transcribed by Pol III. Relative expression is shown for BFV miR-BF2-5p and EBV miR-BART5-5p in 293 cells transfected with the cognate miRNA expression vectors in the presence and absence of the Pol II inhibitor α-amanitin. Expression levels were determined by the ΔΔC_T method (26), using RNU48 as the internal reference, with the carrier-treated cell value set at 1.0. “Mock” refers to 293 cells transfected with the parental vector lacking any miRNA sequence and treated with carrier. Averages for three independent experiments, with standard errors of the means, are shown. ND, no detectable amplification.

FIG 6

Functional analysis of BFV miRNAs. A single perfectly complementary target sequence for each BFV miRNA was inserted into the 3′ UTR of the RLuc indicator gene and cotransfected with a BFV pri-miRNA expression vector lacking any known exogenous Pol II or Pol III promoter. RLuc values are shown normalized to the FLuc internal control and to the level seen with the parental vector lacking an inserted miRNA binding site, which was set at 1.0. Averages for three independent experiments, with standard deviations, are indicated.