Expression and Subcellular Localization of the Kaposi's Sarcoma-Associated Herpesvirus K15P Protein during Latency and Lytic Reactivation in Primary Effusion Lymphoma Cells

Abstract

Kaposi's Sarcoma Associated Herpesvirus (KSHV) is responsible for Primary Effusion Lymphoma (PEL), Multicentric Castleman Disease and Kaposi's Sarcoma, a multifocal angioproliferative neoplasm of endothelial cells. These clinical consequences of KSHV infection likely reflect the known in vivo reservoirs of KSHV, endothelial and B-cells. The KSHV-encoded K15P membrane protein is a key pathogenesis determinant; it interacts with multiple cellular signaling pathways and is thought to play central roles in KSHVassociated endothelial cell angiogenesis, regulation of B-cell receptor (BCR) signaling and the survival, activation and proliferation of BCR-negative primary effusion lymphoma (PEL) cells. Although the K15P gene is capable of encoding a protein of ~45 kDa in size, numerous lower molecular weight forms of K15P are actually observed in vivo; this is thought to be a consequence of differential splicing of the 8-exon K15P gene and also poorly characterized post-translational processing of the full-length K15P polypeptide. Moreover, although K15P has been reported to localize to numerous subcellular locations in heterologous expression studies there is limited data concerning the organellar targeting of K15P in KSHV-infected cells.;The purpose of this research was to investigate the relationship between the various molecular weight forms of K15P, to determine their subcellular distribution and to examine how these may change during the switch from latent to lytic KSHV replication. We used lentiviral expression vectors to stably express a hemagglutinin epitope tagged K15P reporter cDNA (K15P-HA) in the background of KSHV infected PEL cell lines, and control cells. We found that in these latently infected PEL cells the full-length 8-exon K15P cDNA, predicted to encode the full-length ~45 kDa K15P protein, nevertheless generated the 23- to 24-kDa protein normally associated with PEL cells. Subcellular fractionation of these cells by density gradient centrifugation separated the K15P-HA from markers antigens of the mitochondria and lysosomes, contradicting previous studies in heterologous or non KSHV-infected cell lines which had suggested K15 localization to those organellar compartments. In contrast, immunofluorescence analysis revealed substantial overlap between K15P-HA and markers of the cis-Golgi and trans-Golgi network (P58K and TGN46) in latent PEL cells. Such overlap was not seen upon immunocytochemical analysis of K15P-HA and antigens of the endoplasmic reticulum, mitochondria or plasma membrane.;To examine the expression pattern and localization of K15P-HA in cells undergoing KSHV lytic reactivation we stimulated lytic induction in PEL cells using sodium butyrate treatment. We found that levels of the 23- to 24-kDa K15P-HA protein diminished, and that the full-length, 45 kDa protein accumulated in these lytically induced cells. Such changes were not observed in control human B cell-lymphoma cells that express K15P-HA but are not KSHV-infected, suggesting the switch in K15P-HA molecular weight is associated with the viral lytic reactivation program. When the KSHV lytic program was induced by transfection with a plasmid expressing the KSHV lytic activator RTA (Replication and Transcription Activator), rather than by sodium butyrate treatment, similar effects upon K15P-HA were seen. Immunofluorescence analysis revealed that the shift in the molecular weight form of K15P-HA expressed upon lytic reactivation was accompanied by an increase in K15PHA expression level and a redistribution of the protein to a dispersed peripheral location that no longer overlapped with TGN46 markers, nor with any other organellar marker tested. In addition, TGN46 staining became difficult to detect in cells undergoing lytic reactivation, indicating possible fragmentation of the TGN. No such K15P-HA re-localization or disappearance of TGN46 (as detected by immunofluorescence analysis) was observed in KSHV-negative control lymphoma cells under the same conditions. K15P-HA re-localization, and increased expression of the protein, was also observed when KSHV lytic reactivation in PEL cells was stimulated by transfection with the lytic activator RTA. In addition, KSHV-infected PEL cells are known to undergo spontaneous lytic reactivation of KSHV at a low frequency (one to two percent of cell): we identified these rare cells by immunocytochemistry (using an antibody against a viral protein only expressed during lytic replication) and found that in those cells K15P-HA expression was elevated, and the protein exhibited a dispersed peripheral staining pattern. These reactivating cells also contained peripheral and diminished TGN46 staining.;In conclusion, lytic reactivation by sodium butyrate treatment, RTA transfection or by spontaneous means, led to similar changes in K15P-HA expression and intracellular localization in KSHV-infected PEL cells. These observations support the model that observed changes in K15P-HA and TGN46 staining are indeed the result of KSHV reactivation. We speculate that expression of differing molecular weight forms of K15P, in distinct cellular locations, reflects the alternative demands placed upon the protein in the latent and lytic phases.