Abstract
Cellular CCAAT/enhancer binding protein alpha (C/EBPalpha) promotes cellular differentiation and has antimitotic activities involving cell cycle arrest at G(1)/S through stabilization of p21(CIP-1)/WAF1 and through transcriptional activation of the p21 promoter. The Epstein-Barr virus lytic-cycle transactivator protein ZTA is known to arrest the host cell cycle at G(1)/S via a p53-independent p21 pathway, but the detailed molecular mechanisms involved have not been defined. To further evaluate the role of ZTA in cell cycle arrest, we constructed a recombinant adenovirus vector expressing ZTA (Ad-ZTA), whose level of expression at a low multiplicity of infection in normal human diploid fibroblast (HF) cells was lower than or equal to the physiological level seen in Akata cells lytically induced by EBV (EBV-Akata cells). Fluorescence-activated cell sorting analysis of HF cells infected with Ad-ZTA confirmed that G(1)/S cell cycle arrest occurred in the majority of ZTA-positive cells, but not with an adenovirus vector expressing green fluorescent protein. Double-label immunofluorescence assays (IFA) performed with Ad-ZTA-infected HF cells revealed that only ZTA-positive cells induced the expression of both endogenous C/EBPalpha and p21 and blocked the progression into S phase, as detected by a lack of incorporation of bromodeoxyuridine. The stimulation of endogenous ZTA protein expression either through treatment with tetradecanoyl phorbol acetate in D98/HR1 cells or through B-cell receptor cross-linking with anti-immunoglobulin G antibody in EBV-Akata cells also coincided with the induction of both C/EBPalpha and p21 and their mRNAs, as assayed by Northern blot, Western blot, and IFA experiments. Mechanistically, the ZTA protein proved to directly interact with C/EBPalpha by coimmunoprecipitation in EBV-Akata cells and with DNA-bound C/EBPalpha in electrophoretic mobility shift assay experiments, and the in vitro interaction domain encompassed the basic leucine zipper domain of ZTA. ZTA also specifically protected C/EBPalpha from degradation in a protein stability assay with a non-EBV-induced Akata cell proteasome extract. Furthermore, both C/EBPalpha and ZTA were found to specifically associate with the C/EBPalpha promoter in chromatin immunoprecipitation assays, but the interaction with ZTA appeared to be mediated by C/EBPalpha because it was abolished by clearing with anti-C/EBPalpha antibody. ZTA did not bind to or activate the C/EBPalpha promoter directly but cooperatively enhanced the positive autoregulation of the C/EBPalpha promoter by cotransfected C/EBPalpha in transient luciferase reporter gene assays with Vero and HeLa cells as well as with DG75 B lymphocytes. Similarly, ZTA alone had little effect on the p21 promoter in transient reporter gene assays, but in the presence of cotransfected C/EBPalpha, ZTA enhanced the level of C/EBPalpha activation. This effect proved to require a previously unrecognized region in the proximal p21 promoter that contains three high-affinity C/EBPalpha binding sites. Finally, in C/EBPalpha-deficient mouse embryonic fibroblasts (MEF), Ad-ZTA was unable to induce either p21 or G(1) arrest, whereas it was able to induce both in wild-type MEF. Overall, we conclude that C/EBPalpha is essential for at least one pathway of ZTA-induced G(1) arrest during EBV lytic-cycle DNA replication and that this process involves a physical piggyback interaction between ZTA and C/EBPalpha leading to greatly enhanced C/EBPalpha and p21 levels through both transcriptional and posttranslational mechanisms.