Cell cycle arrest by Kaposi's sarcoma-associated herpesvirus replication-associated protein is mediated at both the transcriptional and posttranslational levels by binding to CCAAT/enhancer-binding protein α and p21CIP-1

Lytic-cycle replication of Kaposi's sarcoma-associated herpesvirus (KSHV) in PEL cells causes G(1) cell cycle arrest mediated by the virus-encoded replication-associated protein (RAP) (or K8 protein), which induces high-level expression of the cellular C/EBPalpha and p21 proteins. Here we have examined the mechanism of this induction at both the transcriptional and posttranslational levels. RAP proved to bind very efficiently to both C/EBPalpha and p21 and stabilized them by up to 10-fold from proteasome-mediated degradation in vitro. Cross-linking revealed that RAP itself forms stable dimers and tetramers in solution and forms higher-order complexes but not heterodimers with C/EBPalpha.. Cotransfection of RAP with C/EBPalpha cooperatively stimulated both the C/EBPalpha and p21 promoters in luciferase reporter gene assays. Only the basic/leucine zipper region of RAP was needed for interaction with and stabilization of C/EBPalpha, but both the N-terminal and C-terminal domains were required for transcriptional augmentation. In vitro-translated RAP interfered with DNA binding by C/EBPalpha in electrophonetic mobility shift assay (EMSA) experiments but did not itself bind to the target C/EBPalpha sites or form supershifted bands. However, in endogenous chromatin immunoprecipitation (ChIP) assays with tetradecanoyl phorbol acetate-induced PEL cells, RAP proved to specifically associate with the C/EBPalpha promoter in vivo, but only in a C/EBPalpha-dependent manner,implying an in vivo piggyback interaction with DNA-bound C/EBPalpha. Expression of exogenous RAP (Ad-RAP) caused G(1)/S cell cycle arrest in human dermal microvascular endothelial cells and also induced both the C/EBPalpha and p21 proteins, which formed punctate nuclear patterns that colocalized with RAP in PML nuclear bodies. In the presence of RAP, C/EBPalpha was also efficiently recruited into viral DNA replication compartments in both infected and cotransfected cells. In support of a direct role for this interaction in viral DNA replication, three C/EBPalpha binding sites were identified by in vitro EMSA experiments within a 220-bp core segment of the duplicated KSHV Ori-Lyt region, and although RAP did not bind to Ori-Lyt DNA directly in vitro, both endogenous RAP and C/EBPalpha were found to be associated with the Ori-Lyt region by ChIP assays in lytically induced PEL cells. Finally, we found that the KSHV lytic cycle could not be triggered by either synchronizing KSHV latently infected PEL cells in G, phase or inducing p21 in a C/EBPalpha-independent process.