Functional role of microRNAs during cytomegalovirus infection

The Grey lab applies RNA-induced silencing complex immunoprecipitation (RISC-IP) and cross-linking and sequencing of hybrids (CLASH) to study the microRNA (miRNA) interactions of human cytomegalovirus (HCMV).

The discovery of a new class of regulatory genes known as miRNAs has resulted in a paradigm shift in gene regulation research. miRNAs are small single-stranded RNA species that regulate gene expression through post-transcriptional mechanisms. The genetically economic and non-immunogenic nature of miRNAs, as well as their ability to regulate multiple genes, makes miRNAs a particularly attractive mechanism for viruses to regulate both cellular and viral gene expression. 

Currently, more than 200 viral miRNAs have been identified, encoded by multiple virus families, including herpesviruses, adenoviruses, polyomaviruses and retroviruses. HCMV encodes at least 14 pre-miRNAs corresponding to a total of 27 mature miRNA species. Each has the capacity to regulate multiple genes, both viral and cellular, with the potential to shape the virus life cycle and manipulate the host response. However, studies have only just begun to uncover the role of these small regulatory RNAs in the context of virus infection.

As one of the first groups to identify HCMV encoded miRNAs, a central aspect of the Grey lab research has been to identify the viral miRNA targets using cutting edge high throughput techniques, providing a basis for functional characterisation for the role of miRNAs in virus infection. We have shown that HCMV miRNAs target both their own viral gene expression and expression of cellular genes, creating an environment conducive to viral replication and persistence.

Herpesvirus miRNAs target the expression of their own genes to maintain and control virus latency and reactivation

Using a comparative evolutionary bioinformatic approach we identified conserved target sites for one of the HCMV miRNAs, miR-UL112-1, in the 3’UTR of the crucial viral transactivators, IE72. IE72 and IE86 are the major immediate early (MIE) transactivator genes, and are among the first transcripts expressed following virus infection and are known to be critical for the induction of subsequent viral gene expression required to drive replication of the virus. As such these genes are thought to be pivotal in regulating establishment, maintenance and reaction from latency. Our studies demonstrated that miR-UL112 was able to reduce expression of IE72 and suppress virus replication. We proposed that this regulation could contribute to the establishment and maintenance of viral latency. Subsequent studies by other groups found regulation of immediate early transactivators by viral miRNAs likely represents a conserved mechanism of latency regulation in herpesviruses.

Using cutting edge high throughput approaches for identification of viral miRNA targets

Identifying targets is a necessary, but challenging step in defining the functional role of miRNAs. Previous studies, including our own, have had success using bioinformatic prediction to identify targets of HCMV miRNAs. However, as the rules governing miRNA target interaction are not fully understood, bioinformatic prediction often results in significant numbers of false positives. Furthermore, there is little sequence conservation between virus miRNAs, meaning current prediction algorithms, which rely heavily on evolutionary conservation of target sites, are ineffective. High throughput biochemical approaches provide an unbiased approach for the identification of high confidence miRNA target interactions. In previous studies we have used RISC-immunoprecipitation approaches, that depend on immunoprecipitation of the miRNA complexed with the target transcript. Immunoprecipitated transcripts are then identified by qRT-PCR, microarray or RNAseq. Transcripts targeted by viral miRNAs are specifically enriched in pull-downs from infected cells compared to uninfected cells. Using this approach, we showed the one of the HCMV miRNAs miR-US25-1 dominates the targetome and strikingly, predominantly targets cellular transcripts in the 5’ untranslated region (UTR), rather than the conventional targeting of 3’UTRs. We also identified multiple cellular genes involved in membrane organisation targeted by miR-UL112-1, miR-US5-1 and US5-2. Targeting of these genes was demonstrated to be necessary for the efficient formation of the virus assembly compartment, a membranous complex central to virion assembly. This study illustrates how multiple viral miRNAs can co-ordinately regulate cellular genes that act within a functionally related pathway to affect virus replication.

Current work - Generation of high confidence miRNA target interactions using CLASH

As with previous RISC-IP strategies, CLASH relies on the stable interaction between the multiprotein RISC, miRNA and the associated target transcript. Cells expressing a tandem tagged version of Argonaut, a constitutive component of the RISC complex, are UV irradiated, resulting in cross-linking of the RISC proteins to closely associated RNA species. By covalently stabilising the complex, the tandem-tagged Argonaut, which consists of protein A and multiple His tags, allows for repeated wash stages in denaturing conditions. This step is crucial for removing non-specific RNA contamination. The target transcripts are trimmed with RNase, resulting in RNA clips protected by the RISC complex. An RNA-RNA ligation step then generates the hybrid sequences consisting of miRNA and trimmed target RNA, which is cloned to generate libraries for RNAseq analysis. The majority of the library consists of single target sequences, similar to high-throughput sequencing of RNA isolated by crosslinking immunoprecipitation (HITS-CLIP) or photoactivatable ribonucleoside-enhanced crosslinking and immunoprecipitation (PAR-CLIP) techniques. However approximately 2% of the library represents hybrid sequences consisting of miRNA ligated to target sequence. These hybrids form a stem loop structure representing the binding of the miRNA to its target, thereby generating precise data on miRNA target interaction. Although hybrids are not highly represented, single target sequences corresponding to the hybrid sequence are consistently and highly represented in libraries allowing the combination of data to generate highly specific and statistically powerful analysis of miRNA target interactions. In order to generate a global analysis of targets of HCMV miRNAs, we have performed CLASH on human primary fibroblast cells infected with HCMV, generating more than 18,000 hybrids. We are currently analysing and characterising the targets identified by hybrid data, including cellular and viral targets of both cellular and viral miRNAs.