A viral microRNA down-regulates multiple cell cycle genes through mRNA 5'UTRs

Abstract

Global gene expression data combined with bioinformatic analysis provides strong evidence that mammalian miRNAs mediate repression of gene expression primarily through binding sites within the 3' untranslated region (UTR). Using RNA induced silencing complex immunoprecipitation (RISC-IP) techniques we have identified multiple cellular targets for a human cytomegalovirus (HCMV) miRNA, miR-US25-1. Strikingly, this miRNA binds target sites primarily within 5'UTRs, mediating significant reduction in gene expression. Intriguingly, many of the genes targeted by miR-US25-1 are associated with cell cycle control, including cyclin E2, BRCC3, EID1, MAPRE2, and CD147, suggesting that miR-US25-1 is targeting genes within a related pathway. Deletion of miR-US25-1 from HCMV results in over expression of cyclin E2 in the context of viral infection. Our studies demonstrate that a viral miRNA mediates translational repression of multiple cellular genes by targeting mRNA 5'UTRs.

Figure 1. RISC-IP analysis of HCMV encoded miRNA, miR-US25-1.

(a) Schematic representation of c-myc tagged Ago2 pull downs. (b) Distribution of fold enrichment and (c) enrichment levels of top 100 transcripts following RISC-IP analysis using c-myc tagged Ago2 approach.

Figure 2. RISC-IP analysis of HCMV encoded miRNA, miR-US25-1.

(a) Schematic representation of biotin pull downs. (b) Distribution of fold enrichment and (c) enrichment levels of top 100 transcripts following RISC-IP analysis using biotin approach.

Figure 3. Majority of miR-US25-1 target sites reside within 5′UTRs.

Pie chart shows the percentage of miR-US25-1 seed matches within the 5′UTR, ORF or 3′UTR of top 50 enriched transcripts.

Figure 4. miR-US25-1 represses gene expression.

(a) Schematic representation of miR-US25-1 target transcripts, cyclin E2 and (b) H3F3B. Red boxes represent the open reading frames, green boxes the UTRs. The position of the target site within the 5UTR is indicated as well as the predicted binding between miR-US25-1 and target sites. The seed region is highlighted in red. (c) 5′UTR and promoter region of cyclin E2 (CCNE2) and H3F3B were cloned upstream of a luciferase reporter construct. Cells were cotransfected with firefly luciferase (Fluc) constructs and either miR-US25-1 expression plasmid or a negative control plasmid. Following RISC-IP analysis, levels of luciferase transcript were analyzed by RT-PCR and enrichment determined by comparing IP levels with total levels. Target sites were replaced with an EcoRI site to create deletion constructs. Bars represent the fold increase in enrichment following miR-US25-1 expression compared to control. Transcript levels were normalized to GAPDH. (d) Luciferase constructs described above were cotransfected with renillin luciferase plasmid and either miR-US25-1 expression plasmid (US25-1) or the control plasmid (NEG). Luciferase activity was normalized to renillin levels then calculated as percentage of the negative control, which was set to 100%. Error bars indicate s.d. from 3 independent experiments. Equivalent western blots for each transfection is shown below luciferase graph indicating Fluc protein levels.

Figure 5. miR-US25-1 targets 5′UTR's in context of viral infection.

(a) RISC-IP analysis was conducted on uninfected human primary fibroblast cells or cells infected with HCMV using a direct Ago2 antibody. Results show levels of enrichment of cyclin E2 or TRIM28 transcript from infected cells compared to uninfected cells. RISC-IP was also conducted using pre-bleed antibody derived from rabbits before antigen inoculation. (b) miR-US25-1 was deleted from HCMV. Levels of miR-US25-1 and miR-UL112-1 were determined by RT-PCR analysis following infection of human primary fibroblast cells with either wild type HCMV or the knock out virus. RNA from uninfected cells is used as a negative control. (c) Viral growth of miR-US25-1 knock out virus was compared to wild type HCMV following low (MOI of 0.5) or high (MOI of 10) multiplicity infection of human primary fibroblast cells. Cells plus supernatant were collected at indicated times and assayed on primary human fibroblast cells by limiting dilution (d) Levels of cyclin E2 and TRIM28 protein were determined following high multiplicity infection (MOI of 10) of human primary fibroblast cells with either wild type virus or miR-US25-1 knock out virus. Cells were either grown in normal serum conditions, serum starved conditions or serum starved cells with serum replaced 10 hours prior to harvest. Cells were harvested 72 hours post infection. Relative densities of bands normalized to GAPDH are shown below each lane. Total RNA was also isolated and transcript levels for cyclin E2 (e) and TRIM28 (f) determined by RT-PCR. Transcript copy number was normalized to GAPDH levels.