Figure 1.
The majority of host and viral miRNAs begin as longer RNAs (pri-miRNAs) transcribed by host RNA Pol II that are recognized and processed by the host Microprocessor complex to produce a short stem-loop RNA (pre-miRNA). However, a minority of viruses utilize noncanonical mechanisms in the biogenesis of pre-miRNA molecules. Herpesvirus saimiri (HVS) encodes Sm class U RNAs (HSURs) that are transcribed by RNA Pol II and subsequently processed by the host Integrator complex to generate pre-miRNAs [115]. The miRNAs encoded by mouse gammaherpesvirus 68 (MHV68) and BLV are both transcribed by host RNA Pol III [35], [39], [116], [117]. The MHV68 miRNAs are processed from larger tRNA-like RNAs by host tRNase Z and possibly additional factors to generate pre-miRNAs. Pre-miRNAs are exported to the cytoplasm where they are cleaved into short ∼22 nt duplex RNAs by the host enzyme Dicer. One strand of the duplex may be incorporated into an Argonaute protein containing effector complex known as RISC. The incorporated miRNA directs RISC to target RNAs. Most commonly for animal host and virus-encoded miRNAs, imperfectly complementary base pairing occurs between miRNA and mRNA target resulting in translation inhibition and mRNA turnover. However rarely, perfect base pairing can occur resulting in siRNA-like RNA cleavage and has been reported for polyomaviruses and some herpesviruses.
Figure 2.
Model: viral latency as a simple developmental process.
(A) Some host developmental pathways may be modeled in a two-state fashion. Host miRNAs may help enforce or sharpen transitions from one developmental state to another with miRNA and mRNA target levels inverted in those states [94]. (B) Model for some miRNAs during viral latency. During viral latency, lytic mRNAs are not expressed or expressed at typically undetectable levels. Viral miRNAs are expressed at relatively high levels during latency and may suppress “leaky” lytic transcripts. Optimizing the switch from latent to lytic infection is likely a function of importance in maintaining the homeostasis of latent infection. During lytic replication, large changes in the transcriptional activity of lytic genes “overrun” the imposed inhibition by miRNAs. Note that the viral genome is represented as a circular episome even though for some viruses lytic replication can result in multiple copies of the linearized genome.
Table 1.
Known viral miRNAs and proposed functions highlighted in this review.
Figure 3.
A minority of viral miRNAs mimic host miRNAs through identical seed sequences.
(A) Some viral miRNAs share seed identity with host miRNAs, called “analogs,” while the majority of viral miRNAs do not. For each system (human, mouse, chicken), the miRBase version 18 annotated mature viral miRNA sequences were compared with respective host miRNAs for identity in nucleotides 2–7 (hexamer) or 2–8 (heptamer). Inner circles represent the number of viral miRNAs with a host seed match out of the total viral miRNAs. Percentage to the nearest whole number is presented below each diagram. (B) Models of viral miRNA function. In the host network model on the left, some viral miRNAs function as analogs of host miRNAs through seed sequence similarity, thereby targeting transcripts through the same docking sites as the mimicked host miRNAs. These docking sites for the host miRNA may represent a conserved network and allow the viral miRNA access to numerous targets working together to effect the same function. In contrast, the primary target model on the right suggests some viral miRNAs evolve to target only one or a few transcripts through novel sites not conserved for host miRNA functions. In this model, the virus may tolerate numerous neutral or disadvantageous “bystander” interactions as long as the sum total of regulation provided by the nonanalog viral miRNA is advantageous to the viral lifecycle. Additionally, host and viral miRNAs may target the same transcript through different docking sites as proposed in the convergent target model (bottom of figure).
Figure 4.
Noncanonical viral miRNA functions.
(A) The use of host miRNA biogenesis machinery to modulate mRNA in cis has been reported in two different herpesviruses. On the left, the KSHV gene Kaposin B (KapB) functions as both an mRNA encoding a protein and a pri-miRNA for two KSHV-encoded miRNAs. Each KapB transcript may function as either an mRNA or pri-miRNA, but not both since pre-miRNA biogenesis occurs in the nucleus and destroys the mRNA. The levels of the host Microprocessor component Drosha is decreased during lytic replication and thus allows a shift to the pathway favoring KapB protein production [104]. On the right, the EBV gene BHRF1 functions as both an mRNA encoding a protein and a pri-miRNA for multiple EBV-encoded miRNAs. During Latency III, Microprocessor binds the transcript and the 5′ translation inhibitory intron is retained in BHRF1 transcripts. Altered transcription start site selection during the lytic cycle results in altered splicing and expression of the BHRF1 protein [105]. The net result is little BHRF1 protein expression occurs in latency with increased expression during lytic infection. (B) Antisense miRNA/mRNA interactions have been reported in both polyomaviruses and herpesviruses. Viral miRNAs are encoded antisense to viral mRNAs and the viral miRNAs direct siRNA-like cleavage of the mRNAs transcribed in the sense orientation. This antisense arrangement may not only decrease levels of the mRNA but may also serve as a posttranscriptional insulator in cis and trans to prevent the accumulation of long antisense RNAs.