Host microRNA regulation of human cytomegalovirus immediate early protein translation promotes viral latency

Abstract

Reactivation of human cytomegalovirus (HCMV) is a significant cause of disease and death in immunocompromised patients, underscoring the need to understand how latency is controlled. Here we demonstrate that HCMV has evolved to utilize cellular microRNAs (miRNAs) in cells that promote latency to regulate expression of a viral protein critical for viral reactivation. Our data reveal that hsa-miR-200 miRNA family members target the UL122 (immediate early protein 2) 3' untranslated region, resulting in repression of this viral protein. Utilizing recombinant viruses that mutate the miRNA-binding site compared to the sequence of the wild-type virus results in lytic rather than latent infections in ex vivo infections of primary CD34+ cells. Cells permissive for lytic replication demonstrate low levels of these miRNAs. We propose that cellular miRNA regulation of HCMV is critical for maintenance of viral latency.

Importance: Human cytomegalovirus (HCMV) is a herpesvirus that infects a majority of the population. Once acquired, individuals harbor the virus for life, where the virus remains, for the most part, in a quiet or latent state. Under weakened immune conditions, the virus can reactivate, which can cause severe disease and often death. We have found that members of a family of small RNAs, termed microRNAs, encoded by human myeloid progenitor cells are capable of repressing a key viral protein, thus enabling the virus to ensure a quiet/latent state. As these progenitor cells mature further down the myeloid lineage toward cells that support active viral replication, the levels of these microRNAs decrease. Together, our data suggest that host cell microRNA regulation of HCMV is important for the quiet/latent state of this pathogen.

FIG 1

In silico analysis predicts that several cellular hsa-miR-200 family members bind the UL122 3′ UTR. In silico analysis predicts that three members of the cellular hsa-miR-200 family of miRNAs are likely to bind the 3′ UTR of UL122 with perfect sequence complementarity in the seed region (shaded). A schematic of the UL122/UL123 region of HCMV, with the 3′ UTR sequence of UL122 (IE2) magnified, is shown. Predicted hybridizations between the 3′ UTR of UL122 and hsa-miR-200b, hsa-miR-200c, or hsa-miR-429 are shown along with the calculated free energy of the interaction. Ex, exon.

FIG 2

The wild-type 3′ UTR of HCMV UL122 is repressed by overexpression of hsa-miR-200b. 4T07 cells were transfected with either a control synthetic miRNA construct (gray bars) or a synthetic miRNA construct corresponding to miR-200b (white bars). These cells were then cotransfected with firefly luciferase constructs expressing either the wild-type UL122 3′ UTR, a mutant UL122 3′ UTR, a positive control (ZEB2 3′ UTR), or a negative control (empty luciferase vector). All samples were analyzed in triplicate, and the levels were adjusted to those for Renilla luciferase.

FIG 3

Cellular hsa-miR-200 inhibits wild-type infection but not infection with an UL122 (IE2) 3′ UTR mutant virus. (A) Primary human embryonic lung fibroblasts (MRC5 cells) were stably transduced with either a control retrovirus or one that overexpressed the C1 cluster of the hsa-miR-200 family. In each cell type, levels of hsa-miR-200b (light gray bars) or hsa-miR-200c (dark gray bars) were assessed by qPCR. Samples were normalized to those of cellular RNU44 and analyzed in triplicate. (B, C) MRC5 cells stably transduced with a C1-expressing lentivirus (white bars) or an empty control (gray bars) were then infected with either wild-type FixBACgfp virus (B) or FixBACgfpIE2cisΔ virus (C) at a multiplicity of 0.5 PFU/cell for 4 days. The titer of cell-free virus was then determined by a modified immunofluorescence assay for IE1. Samples were analyzed in triplicate. (D) MRC5 cells transduced with either the C1-expressing or control lentivirus were infected with either wild-type FixBACgfp virus or FixBACgfpIE2cisΔ virus at a multiplicity of 1 PFU/cell. Cell lysates were harvested at the indicated time points (hpi, hours postinfection), and IE2 levels were assessed using a monoclonal antibody (clone 3A9). α-Tubulin was used as a control.

FIG 4

Infection of Kasumi-3 cells with virus lacking the miRNA-binding site in the UL122 (IE2) 3′ UTR favors lytic replication. Kasumi-3 cells were infected with wild-type TB40/Egfp (WT), TB40/EgfpIE2cisΔ (IE2cisΔ), or TB40/EgfpIE2cisΔrep (IE2cisΔRep) virus. (A) Viral gene expression was assessed at 7 dpi by RT-qPCR with primers directed at UL123. All samples were analyzed in triplicate, and levels were normalized to the level of GAPDH gene production. (B) Extracellular virion production was assessed over a 5-day time course. Viral genomes derived from the supernatants of the infected cells were analyzed by qPCR using primers that detect UL123. All samples were analyzed in triplicate. AU, arbitrary units.

FIG 5

Infection of ex vivo-cultured primary CD34⁺ cells with a virus that lacks the miRNA seed binding sequence favors lytic gene expression. Primary human CD34⁺ hematopoietic progenitor cells were isolated from umbilical cord blood by magnetic separation. The cells were then infected with wild-type TB40/Egfp (WT), TB40/EgfpIE2cisΔ (IE2cisΔ), or TB40/EgfpIE2cisΔrep (IE2cisΔRep) viruses at a multiplicity of 2.0 PFU/cell, and then the cells were harvested at 5 dpi for IE gene transcription by RT-qPCR analysis (A) or quantification of extracellular genomes by qPCR (B). Primers directed at UL123 were used to detect IE gene transcripts as well as viral genomes. The levels of the viral transcripts in panel A were normalized to the level of the cellular GAPDH gene. All samples were analyzed in triplicate.

FIG 6

The cellular hsa-miR-200 miRNA family is highly expressed in cells that support HCMV latent infection. hsa-miR-200 levels were detected by qPCR in Kasumi-3 cells (dark gray bar) or TPA-differentiated Kasumi-3 cells (white bar) (A) or in primary human umbilical cord CD34⁺ cells (dark gray bar), primary human monocytes (light gray bar), or primary human monocyte-derived macrophages (Mϕ; white bar) (B). Samples were analyzed in triplicate, and the levels were normalized to the values for human RNU44.

FIG 7

Proposed model of host miRNA control in viral latency. Upon infection of hematopoietic progenitor cells, the HCMV MIEP is marked with repressive chromatin, thereby transcriptionally silencing the promoter. UL122 transcripts generated as a result of low-level basal transcription from this promoter are then repressed by the hsa-miR-200 family of miRNAs to inhibit IE2 translation. Together, these processes ensure viral latency in these cells. However, under conditions during which inflammatory cytokine production is increased, the MIEP is transcriptionally activated and UL122 transcripts accumulate to levels that supersede hsa-miR-200 suppression. Additionally, as the cells differentiate toward mature myeloid cells (e.g., macrophages), the levels of the cellular hsa-miR-200 family decrease such that they are not sufficient for repressing UL122 transcription. As a result, IE2 is translated and lytic reactivation commences.