Ribosome profiling of the retrovirus murine leukemia virus

Abstract

Background: The retrovirus murine leukemia virus (MuLV) has an 8.3 kb RNA genome with a simple 5'-gag-pol-env-3' architecture. Translation of the pol gene is dependent upon readthrough of the gag UAG stop codon; whereas the env gene is translated from spliced mRNA transcripts. Here, we report the first high resolution analysis of retrovirus gene expression through tandem ribosome profiling (RiboSeq) and RNA sequencing (RNASeq) of MuLV-infected cells.

Results: Ribosome profiling of MuLV-infected cells was performed, using the translational inhibitors harringtonine and cycloheximide to distinguish initiating and elongating ribosomes, respectively. Meta-analyses of host cell gene expression demonstrated that the RiboSeq datasets specifically captured the footprints of translating ribosomes at high resolution. Direct measurement of ribosomal occupancy of the MuLV genomic RNA indicated that ~ 7% of ribosomes undergo gag stop codon readthrough to access the pol gene. Initiation of translation was found to occur at several additional sites within the 5' leaders of the gag and env transcripts, upstream of their respective annotated start codons.

Conclusions: These experiments reveal the existence of a number of previously uncharacterised, ribosomally occupied open reading frames within the MuLV genome, with possible regulatory consequences. In addition, we provide the first direct measurements of stop codon readthrough efficiency during cellular infection.

Keywords: Murine leukemia virus; RNASeq; Retrovirus; Ribosome profiling; Upstream ORF; Viral translation.

Fig. 1

MuLV RNA synthesis and translation. a Map of the Moloney MuLV proviral DNA, showing the locations of the untranslated 3′ (U3), R, and U5 repeats within the LTRs, and the gag, pol and env coding regions. U3 houses the viral promoter region, while the transcription start site (TSS) corresponds to the first nucleotide of the R region. The poly-adenylation (poly-A) signal also occurs within R. Translation of pol is dependent upon readthrough of the gag gene termination codon. Two distinct mRNA species are produced from the provirus template, with the gag-pol transcript being unspliced and the env transcript being spliced at the indicated positions. b Rat2 cells were infected with MuLV and at 96 h p.i. cell lysates were prepared, separated by 10% SDS-PAGE and immunoblotted using a polyclonal anti-capsid (p30) serum. Molecular weights (MW; in kDa) are indicated on the left. GAPDH was used as a loading control. Viral proteins were detected with a green fluorescent secondary antibody, and GAPDH was detected with a red fluorescent secondary antibody. c RiboSeq CHX (red) and RNASeq (purple) densities in reads per million mapped reads (RPM), smoothed with a 31-nt sliding window. Negative-sense reads are shown in dark blue below the horizontal axes. d Bar plot showing the density of RPFs mapped to the pol gene relative to the gag gene, normalised by the equivalent density of RNASeq reads in these regions. The ratio of pol to gag translation is approximately 0.07, on average; indicating that ~ 7% of ribosomes undergo stop codon readthrough to translate pol

Fig. 2

Coverage of the 5′ region of the MuLV genomic RNA. Histograms show the number of 5′ read ends (with + 12 nt offset) mapped to viral RNA in RiboSeq HAR, RiboSeq CHX and RNASeq libraries generated from Rat2 cell infections, in reads per million (RPM). The light green rectangle represents the N-terminal extension of the Glyco-Gag-encoding sequence described in [16], while the darker green rectangle represents the standard gag coding sequence. The GUG initiation codon highlighted in green is in-frame with the gag N-terminal extension. Intervening AUC and AUG codons at which initiation was observed are shown in orange. Potential additional upstream initiation codons with coverage in the RiboSeq HAR library are shown in grey. Green, orange and blue bars correspond to reads mapping in phases 0, + 1 and + 2, respectively, relative to the gag ORF

Fig. 3

Coverage of the region surrounding the MuLV env gene splice acceptor. Bar plots show the number of 5′ read ends (with + 12 nt offset) mapped to viral RNA in RiboSeq HAR, RiboSeq CHX and RNASeq libraries generated from Rat2 cell infections, in reads per million (RPM). Blue, green and orange bars correspond to reads mapping in phases 0, + 1 and + 2, respectively, relative to the env ORF. Translation of env-uORF-1 and env-uORF-2 is initiated from CUG codons that coincide with large peaks in RiboSeq HAR data