SUMMARY Human cytomegalovirus (HCMV) is a ubiquitous herpesvirus that persists for life. Although HCMV is generally benign in healthy individuals, the virus can cause disease in immunocompromised populations and HCMV is the leading infectious cause of congenital disease in newborns. The reasons why the immune system is able to control, but unable to eradicate HCMV are unknown. Our main hypothesis is that immunomodulatory CMV genes, particularly modulators of innate and adaptive cellular immunity, enable the establishment and maintenance of persistent infection in immunocompetent individuals. A better understanding of these immune modulatory processes will be essential for the development of vaccines against HCMV as well as for the optimization of CMV-based vaccine vectors that have recently assumed center stage in the development of vaccines against HIV and TB. Our work in non-human primate models of CMV infection revealed a series of unexpected results that indicate a complex relationship between CMV and the innate and adaptive cellular immune response involving viral evasion, viral recruitment and viral manipulation of the host's immune response. Particularly novel is our finding that CMV-encoded genes enable the virus to control the epitope specificity of the adaptive, CD8+ T cell immune response. We discovered that wildtype RhCMV and HCMV elicit MHC-I restricted CD8+ responses that are exclusively directed to sub-dominant or “non-canonical” epitopes. However, in the absence of the MHC-I inhibitors Rh189 (in RhCMV) or US11 (in HCMV), additional CD8+ T cells are induced that recognize “canonica”l MHC-I epitopes, i.e. epitopes that predominate the CD8+ T cell responses elicited by unrelated viruses or vaccines not based on CMV. Since such stringent control of T cell specificity has not been observed before for any infectious agent, we will elucidate why, how and where primate CMVs prevent the induction of canonical CD8+ T cells. In Specific Aim 1 we will test the hypothesis that canonical CD8+ T cells are particularly efficient in controlling viral dissemination and, therefore, rhesus and human CMV developed specific measures to both prevent their priming and to escape their control. In Specific Aim 2 we will investigate which structural features enable Rh189 to inhibit the induction of canonical CD8+ T cells in comparison to related MHC-I-inhibitory proteins of the same gene family that do not prevent canonical CD8+ T cell priming. In Specific Aim 3 we will test the hypothesis that spreading of Rh189-deficient RhCMV to myeloid cells is required for the induction of canonical CD8+ T cell by a direct priming mechanism. We will examine the role of viral dissemination by deleting Rh189 from single cycle viruses. Moreover, by inserting targeting sites for cell type-specific microRNAs into the 3'-UTR of Rh189 we will identify which cell types need to be infected to enable canonical CD8+ T cell priming in the absence of Rh189, The results of this work...