Reversible Silencing of Cytomegalovirus Genomes by Type I Interferon Governs Virus Latency

Herpesviruses establish a lifelong latent infection posing the risk for virus reactivation and disease. In cytomegalovirus infection, expression of the major immediate early (IE) genes is a critical checkpoint, driving the lytic replication cycle upon primary infection or reactivation from latency. While it is known that type I interferon (IFN) limits lytic CMV replication, its role in latency and reactivation has not been explored. In the model of mouse CMV infection, we show here that IFN blocks mouse CMV replication at the level of IE transcription in IFN-responding endothelial cells and fibroblasts. The IFN-mediated inhibition of IE genes was entirely reversible, arguing that the IFN-effect may be consistent with viral latency. Importantly, the response to IFN is stochastic, and MCMV IE transcription and replication were repressed only in IFN-responsive cells, while the IFN-unresponsive cells remained permissive for lytic MCMV infection. IFN blocked the viral lytic replication cycle by upregulating the nuclear domain 10 (ND10) components, PML, Sp100 and Daxx, and their knockdown by shRNA rescued viral replication in the presence of IFN. Finally, IFN prevented MCMV reactivation from endothelial cells derived from latently infected mice, validating our results in a biologically relevant setting. Therefore, our data do not only define for the first time the molecular mechanism of IFN-mediated control of CMV infection, but also indicate that the reversible inhibition of the virus lytic cycle by IFN is consistent with the establishment of CMV latency. Author Summary Cytomegalovirus (CMV) is a widespread herpesvirus that establishes a detente with the host immune system. Therefore, the CMV reactivates from latency in immunocompromised hosts, resulting in life-threatening disease of the vulnerable patients. However, the exact mechanism by which the immune system keeps CMV at bay remains incompletely understood. To address this question, we have used a reporter system, based on infection of cells with the mouse CMV. Our results showed that interferon (IFN), a well-known antiviral protein, blocks CMV replication at the earliest stages after the virus has entered the cell. More importantly, removing the IFN from the infected cells restarted MCMV replication, indicating that its effects are consistent with viral latency. We showed that IFN blocked virus replication by inducing the expression of proteins located in the nuclear domain 10 (ND10), a compartment in the nucleus of cells to which the incoming viral genomes are directed. Similarly, IFN was sufficient to block CMV reactivation from cells of latently infected mice. In conclusion, IFN had the ability to drive CMV into a quiescent state matching the formal definition of latency and was sufficient to prevent reactivation of bona fide latent CMV.