Genomewide mapping and screening of Kaposi's sarcoma-associated herpesvirus (KSHV) 3' untranslated regions identify bicistronic and polycistronic viral transcripts as frequent targets of KSHV microRNAs

Abstract

Kaposi's sarcoma-associated herpesvirus (KSHV) encodes over 90 genes and 25 microRNAs (miRNAs). The KSHV life cycle is tightly regulated to ensure persistent infection in the host. In particular, miRNAs, which primarily exert their effects by binding to the 3' untranslated regions (3'UTRs) of target transcripts, have recently emerged as key regulators of KSHV life cycle. Although studies with RNA cross-linking immunoprecipitation approach have identified numerous targets of KSHV miRNAs, few of these targets are of viral origin because most KSHV 3'UTRs have not been characterized. Thus, the extents of viral genes targeted by KSHV miRNAs remain elusive. Here, we report the mapping of the 3'UTRs of 74 KSHV genes and the effects of KSHV miRNAs on the control of these 3'UTR-mediated gene expressions. This analysis reveals new bicistronic and polycistronic transcripts of KSHV genes. Due to the 5'-distal open reading frames (ORFs), KSHV bicistronic or polycistronic transcripts have significantly longer 3'UTRs than do KSHV monocistronic transcripts. Furthermore, screening of the 3'UTR reporters has identified 28 potential new targets of KSHV miRNAs, of which 11 (39%) are bicistronic or polycistronic transcripts. Reporter mutagenesis demonstrates that miR-K3 specifically targets ORF31-33 transcripts at the lytic locus via two binding sites in the ORF33 coding region, whereas miR-K10a-3p and miR-K10b-3p and their variants target ORF71-73 transcripts at the latent locus through distinct binding sites in both 5'-distal ORFs and intergenic regions. Our results indicate that KSHV miRNAs frequently target the 5'-distal coding regions of bicistronic or polycistronic transcripts and highlight the unique features of KSHV miRNAs in regulating gene expression and life cycle.

FIG 1

Schematic illustration of the strategy and procedures used for mapping KSHV 3′UTRs. (A) Procedures used for mapping and cloning KSHV 3′UTRs. (B) Adaptor and PCR primers used for cloning KSHV 3′UTRs. The restriction enzyme sites in the adaptor are underlined. (C) Alignment of adaptor and PCR primers with a hypothetical gene transcript. (D) Examples of 3′RACE products ORF-K2, ORF2, and ORF-K3 resolved on agarose gels. A primer from PHB was used to amplify the 3′UTR sequence from the total RNA of HL60 cells and served as a positive control. Samples without reverse transcription (RT), indicated by “non-RT,” and water were used as negative controls.

FIG 2

Annotation of 3′UTRs in the KSHV genome. 3′UTRs of KSHV genes were mapped using uninduced and TPA-induced BCBL-1 cells as described in Materials and Methods. Genes are labeled in blue with arrowheads indicating the transcription orientation. Intergenic regions are labeled in gray. 3′UTRs of viral transcripts are indicated as purple lines.

FIG 3

Features of KSHV 3′UTRs. (A) Distribution of KSHV genes with monocistronic transcripts and bicistronic and polycistronic transcripts. (B) Distribution of poly(A) signal sequences of KSHV and human transcripts. (C) Distribution of KSHV 3′UTRs by length. (D) Distribution of different lengths of KSHV and human 3′UTRs by percentage. (E) Average lengths of 3′UTRs of KSHV genes with monocistronic transcripts and bicistronic and polycistronic transcripts.

FIG 4

Newly identified viral genes targeted by KSHV miRNAs. (A) Repression of 3′UTRs by KSHV miRNAs detected by reporter assays. 3′UTR reporters from KSHV genes with bicistronic and polycistronic transcripts are labeled with an asterisk (*). The miR-K10/ORF-K2 pair, which was independently identified in a HITS-CLIP screening (40), is labeled in red type. The miR-K7/ORF50 and miR-K9/ORF50 pairs identified in previous studies are shown in “red” and used as positive controls (19, 34). (B) Distribution of 3′UTRs from KSHV genes with monocistronic transcripts, and bicistronic and polycistronic transcripts repressed by KSHV miRNAs. (C to F) Confirmation of repression of 3′UTRs by KSHV miRNAs. 3′UTRs were targeted by miR-K3 (C) miR-K8 (D), miR-K10 (E), and miR-K11 (F). 293T cells were cotransfected with 3′UTR reporter plasmids or a reporter vector control and the respective miRNA expression plasmids or a miRNA expression vector control, together with a pSV-β-galactosidase plasmid for 48 h, and the relative luciferase activities were measured after normalization. In the initial screening experiments (A), three rounds of independent screening were performed, each with three repeats. The results of the three rounds of screenings were combined and analyzed. All other experiments (C to F) were repeated at least three times. The results presented as averages and STD by setting the values of miRNA expression vector control as “1”.

FIG 5

miR-K3 targets 3′UTRs of ORF31, ORF32, and ORF33 through binding sites in the ORF33 coding region. (A) Schematic illustration of the ORF30-33 gene locus, their transcripts, 3′UTR luciferase reporters, and constructs expressing 3×FLAG fusion proteins. S1 and S2 are predicted as miR-K3-5p binding sites located in the coding region of ORF33. For the 3×FLAG fusion protein constructs, each ORF, including the downstream 3′UTR sequence, is fused in frame with the N terminus of the 3×FLAG tag. (B) Alignment of miR-K3-5p with predicted binding sites in ORF30-33 transcripts. Mutated nucleotides in the binding sites are shown in red type. Solid lines indicate the Watson-Crick base pairing, and dotted lines indicate wobble base pairing. (C) miR-K3 represses ORF31, ORF32, and ORF33 but not ORF30 3′UTRs by targeting two binding sites (S1 and S2) residing in the ORF33 coding region. miRNA expression plasmid or a miRNA expression control vector was cotransfected with the wild-type 3′UTR reporter, mutant 3′UTR reporter, or a reporter control vector into 293T cells, together with a pSV-β-galactosidase plasmid for 48 h, and the relative luciferase activities were examined and normalized for β-galactosidase activity. Experiments were repeated three times, and the results are presented as averages and STD by setting the values of miRNA expression vector control as “1”. (D and E) The mimic of miR-K3-5p does not repress the expression of fusion protein from the 3×FLAG-ORF30 construct (D). The mimic, but not the scrambled control mimic, of miR-K3-5p inhibits the expression of fusion proteins of ORF31, ORF32, and ORF33 by targeting S1 and S2 binding sites (E). Western blot analyses of 3×FLAG-ORF30 fusion protein (D) and of 3×FLAG-ORF31, -ORF32, and -ORF33 fusion proteins (E) in 293T cells were performed. The mimic of miR-K3-5p or a scrambled control mimic was cotransfected with each FLAG-tagged protein expression construct or its mutants with mutation in S1, S2, or both binding sites into 293T cells for 48 h, and the expression of the fusion protein was detected by Western blotting. A 3×FLAG-tagged firefly luciferase construct was also cotransfected as an internal control.

FIG 6

miR-K10 and variants target 3′UTRs of the ORF71-ORF73 cluster through distinct binding sites. (A) Schematic illustration of the ORF71-73 gene locus, their transcripts, 3′UTR luciferase reporters, and a 3×FLAG-ORF73 construct. S1 and S2 are predicted miR-10a-3p binding sites, while S3 and S4 are predicted miR-10b-3p binding sites. For the 3×FLAG-ORF73 construct, ORF73 and the downstream 3′UTR sequence is fused in frame with the N terminus of the 3×FLAG tag. (B) Alignment of miR-K10a-3p and miR-K10b-3p with the predicted binding sites. Mutated nucleotides in the binding sites are marked in red type. Solid lines indicate the Watson-Crick base paring, and dotted lines indicate wobble base pairing. (C) miR-K10a represses ORF71, ORF72, and ORF73 3′UTRs by targeting binding sites S1 and S2 residing in the downstream sequence and ORF71 coding region, respectively. miR-K10 expression plasmid, which expresses only miR-K10a or a miRNA expression vector control, was cotransfected with wild-type 3′UTR reporter, mutant 3′UTR reporter, or a reporter control vector into 293T cells, together with a pSV-β-galactosidase plasmid for 48 h, and the relative luciferase activities were examined and normalized for β-galactosidase activity. Experiments were repeated three times, and the results are presented as averages and STD by setting the values of miRNA expression vector control as “1”. (D) Repression of ORF71, ORF72, and ORF73 3′UTRs by miR-K10b-3p. miR-K10b-3p represses ORF73 3′UTR by targeting the S3 and possibly an unidentified site but not the S4 site. miR-K10b also weakly represses ORF71, ORF72, and ORF73 3′UTRs by targeting the unidentified binding site. Mimic of miR-K10b-3p or scrambled control mimic was transfected into 293T cells for 16 h, followed by cotransfection of the 3′UTR reporter or a reporter control vector with pSV-β-galactosidase plasmid for 48 h, and the relative luciferase activities were examined as described in panel C. (E and F) Repression of ORF73 3′UTR by miR-K10 variants. miR-K10a-3p variant miR-K10a-3p_+1_5 represses the ORF73 3′UTR by targeting both S1 and S2 sites (E), but these two sites do not mediate miR-K10b-3p variant miR-K10b-3p_+1_5 repression of the ORF73 3′UTR (F). Mimic of miR-K10a-3p_+1_5 or miR-K10b-3p_+1_5 or scrambled control mimic was transfected into 293T cells for 16 h, followed by cotransfection of the 3′UTR reporter or a reporter control vector with pSV-β-galactosidase plasmid for 48 h, and the relative luciferase activities were examined as described in panel C. (G) Mimic of miR-K10a-3p or miR-K10b-3p represses the expression of LANA from the 3×FLAG-ORF73 construct. Western blot analyses of 3×FLAG-ORF73 fusion protein in 293T cells. Mimic of miR-K10a-3p or miR-K10b-3p or scrambled control mimic was cotransfected with 3×FLAG-ORF73 expression construct into 293T cells for 48 h. A 3×FLAG-tagged firefly luciferase construct was also cotransfected as an internal control. (H) Mimic of miR-K10a-3p or miR-K10b-3p represses the expression of LANA in 293T-BAC36 cells. Mimic of miR-K10a-3p or miR-K10b-3p or scrambled control mimic was transfected into 293T-BAC36 cells for 48 h, and the expression of LANA was examined by Western blotting. β-Tubulin was used as a loading control. (I) LNA-based miR-K10-3p suppressor upregulates the endogenous protein expression of ORF73 in BCP-1 cells. BCP-1 cells transfected with LNA suppressor or scrambled control suppressor for 48 h were examined for the expression of LANA by Western blotting. β-Tubulin was used as a loading control. (J) Specific miR-K10a-3p and miR-K10b-3p sponges increase the expression of LANA, vCyclin, and vFLIP proteins in BC-3 cells. BC-3 cells transduced with lentiviral viruses expressing a miR-K10-3p sponge or a vector control for 48 h were examined for the expression of LANA, vCyclin, and vFLIP proteins by Western blotting. β-Tubulin was used as a loading control.