Variable episomal silencing of a recombinant herpesvirus renders its encoded GFP an unreliable marker of infection in primary cells.

The availability of reliable recombinant reporter virus systems has been a great boon to the study of Kaposi's sarcoma-associated herpesvirus (KSHV/HHV-8). Unexpectedly, we found that expression of the ostensibly constitutive green fluorescent protein (GFP) marker was progressively lost during unselected passage in primary rat mesenchymal precursor cells (MM), despite efficient maintenance of latent viral gene expression and episomal partitioning. This repression of EF1-α promoter-driven GFP expression appeared to be passage-dependent, however, since functionally immortalized MM cells derived from long serial passage retained stable expression of GFP following rKSHV.219 infection. Chromatin analysis of cultures that we had infected in parallel demonstrated an increase in repressive H3K27 tri-methylation across the viral episome with the exception of the LANA control region in MM cells infected at early rather than late passage post-isolation. The silencing of GFP expression in the MM cells was reversible in a dose-dependent fashion by the histone deacetylase inhibitor valproic acid, further implicating cellular silencing on incoming viral genomes, and underscoring potential differences in viral gene regulation between primary and functionally immortalized cells. Furthermore, using multispectral imaging flow cytometry, we also determined that the extent of GFP expression per cell among those that were positive did not correlate with the number of LANA dots per nucleus nor the extent of overall LANA expression per cell. This suggests a more complex mode of local gene regulation, rather than one that simply reflects the relative intracellular viral copy number. In sum, we have demonstrated the significant potential for false-negative data when using a constitutive marker gene as a sole means of evaluating herpesviral infection, especially in primary cells.