Epigenetic regulation of Kaposi's sarcoma-associated herpesvirus latency by virus-encoded microRNAs that target Rta and the cellular Rbl2-DNMT pathway

Abstract

Kaposi's sarcoma-associated herpesvirus (KSHV) encodes a cluster of 12 microRNAs (miRNAs) that are processed from a transcript that is embedded within the major latency control region. We have generated a deletion mutation that eliminates 10 of the 12 viral miRNAs from the KSHV bacmid by using recombineering methods. The KSHV miRNA deletion mutant (BAC36 DeltamiR) behaved similarly to wild-type (wt) BAC36 in viral production, latency gene transcription, and viral DNA copy number in 293 and dermal microvascular endothelial cells (DMVECs). However, BAC36 DeltamiR consistently expressed elevated levels of viral lytic genes, including the immediate-early transcriptional activator Rta (ORF50). At least one KSHV microRNA (miRK12-5) was capable of suppressing ORF50 mRNA, but poor seed sequence alignments suggest that these targets may be indirect. Comparison of epigenetic marks in DeltamiR KSHV genomes revealed decreases in histone H3 K9 methylation, increases in histone H3 acetylation, and a striking loss of DNA methylation throughout the viral and cellular genome. One viral miRNA, K12-4-5p, was found to have a sequence targeting retinoblastoma (Rb)-like protein 2 (Rbl2), which is a known repressor of DNA methyl transferase 3a and 3b mRNA transcription. We show that ectopic expression of miR-K12-4-5p reduces Rbl2 protein expression and increases DNMT1, -3a, and -3b mRNA levels relative to the levels for control cells. We conclude that KSHV miRNA targets multiple pathways to maintain the latent state of the KSHV genome, including repression of the viral immediate-early protein Rta and a cellular factor, Rbl2, that regulates global epigenetic reprogramming.

FIG. 1.

Deletion of KSHV miRNA causes derepression of ORF50 mRNA. (A) Schematic of KSHV genome organization surrounding the miRNA cluster and deletion boundaries for BAC36 ΔmiR. (B) Ethidium bromide stain of agarose gel for analysis of the BAC36 wt or BAC36 ΔmiR digested with XhoI. M, molecular mass marker. (C) Ethidium bromide stain (left) and Southern blot (right) for analysis of the Kan-LoxP insertion used to generate BAC36 ΔmiR. The Southern blot indicates that a single Kan-LoxP cassette was inserted into the correct EcoRI fragment. (D) RT-PCR of 293 cells infected with virus generated from the BAC36 wt or BAC36 ΔmiR. RT-PCR was performed to assay for ORF50, LANA, or K14 mRNA relative to the level for virus-encoded GFP. (E) Same as in panel D, except RT-PCR was performed to analyze for vFLIP, vCyclin, vIL-6, or PAN RNA relative to the level for GFP. (F) RT-PCR of DMVECs infected with virus generated from the BAC36 wt or BAC36 ΔmiR. RT-PCR was performed to assay for ORF50, LANA, or K14 mRNA relative to the level for virus-encoded GFP, as in panel D. (G) Same as in panel F, except RT-PCR was performed to analyze for vFLIP, vCyclin, vIL-6, or PAN RNA relative to the level for GFP. (H) DNA copy number was assayed by real-time PCR of KSHV ORF50 DNA relative to the level for cellular actin in stable 293 cell pools containing the BAC36 wt or ΔmiR.

FIG. 2.

KSHV miR-K12-5 suppresses translation of the ORF50 3′ UTR. (A) 293 cells containing the BAC36 wt (gray) or BAC36 ΔmiR (black) were transfected with pGL3-Ctrl, pGL3-Luc-ORF50, or pGL3-Luc-ORF50 3′ UTR and assayed for luciferase activity relative to the level for the Renilla internal control (relative luciferase units [RLU]). (B) BAC36 ΔmiR was transfected with control pcDNA3 vector or pcDNA3 expressing individual KSHV miRNA (as indicated). Seventy-two hours after transfection, KSHV ORF50 mRNA was assayed by RT-PCR and quantified relative to the level for β-actin. (C) 293 cells were cotransfected with pGL3-Ctrl, pGL3-Luc-ORF50, or pGL3-Luc-ORF50 3′ UTR and with either pcDNA3 (gray) or pcDNA3-miR-K12-5 (black). Luciferase units were quantified relative to the level for the Renilla internal control. (D) Two potential alignments of the miR-K12-5 seed sequence with the 3′ UTR of ORF50.

FIG. 3.

Regional loss of histone H3 K9 trimethylation in miRNA-deleted KSHV genomes. KSHV genome-wide ChIP analysis for 293 cells containing the BAC36 wt (A and C) or BAC36 ΔmiR (B and D) was performed using antibodies specific for histone H3 trimethyl K9 (A and B) or trimethyl K4 (C and D). ChIP assay results were quantified by real-time PCR as percentages of input by using the ΔΔC_T method.

FIG. 4.

Alteration in histone H3 acetylation in miRNA-deleted KSHV genomes. ChIP assays using antibodies specific for acetylated histone H3 (AcH3) (A) or AcH4 (B) were performed with 293 cells containing the BAC36 wt (gray) or BAC36 ΔmiR (black). ChIP DNA was quantified using primers specifically centered at KSHV genome position 6592, 64417, 71486 (ORF50p), or 124269, as indicated. ChIP results were quantified as percentages of input DNA by using the ΔΔC_T method.

FIG. 5.

Loss of CpG DNA methylation in miRNA-deleted KSHV genomes. (A and B) KSHV genome-wide analysis of MeDIP for the BAC36 wt (A) or BAC36 ΔmiR (B) by use of the ΔΔC_T method. (C) MeDIP assay for the BAC36 wt (gray) or BAC36 ΔmiR (black) at KSHV genome position 69580, 71486 (ORF50p), or 124269 by use of the ΔΔC_T method for quantification of PCR results. (D) MeDIP assay for quantification of the BAC36 wt (gray) or BAC36 ΔmiR (black) at cellular regions for α-globin-2 and β-actin. PCR was performed using the ΔΔC_T method, and the results are presented as percentages of input DNA.

FIG. 6.

KSHV miR-K12-4-5p targets the Rbl2 3′ UTR and activates DNMT gene transcription. (A) Alignment of the Rbl2 3′ UTR with the seed sequence of KSHV miR-K12-4-5p. (B) Western blot analysis of 293 cells transfected with pcDNA3 or pcDNA3-miR-K12-4-5p was performed using antibodies for Rbl2 (top), DNMT1 (middle), and β-actin (bottom). Quantification of Western blot signals is indicated at the right of each panel. (C) Luciferase reporter assays for the control plasmid pGL3 Ctrl or luciferase fused to the wt Rbl2 3′ UTR or a mutant (Mt) Rbl2 3′ UTR in 293 cells cotransfected with miR-K12-4-5p or a vector control. Luciferase activity was normalized to Renilla activity, and data are presented relative to each luciferase target plasmid. (D) RT-PCR analysis of DNMT1, DNMT3a, and DNMT3b mRNA in BAC36 ΔmiR-containing 293 cells transfected with pcDNA3 or pcDNA3-miR-K12-4-5p for 72 h. RT-PCR was performed for quantification relative to the level for β-actin. (E) MeDIP assay of BAC36 ΔmiR-containing 293 cells transfected with pcDNA3 or pcDNA3-miR-K12-4-5p at KSHV positions 71486 (ORF50p) and 124269 or cellular α-globin-2.