Differential effects of human cytomegalovirus on integrated and unintegrated human immunodeficiency virus sequences

Human cytomegalovirus (HCMV) has been implicated as a potential cofactor in human immunodeficiency virus type 1 (HIV-1)-related disease. Previously, we reported that HCMV inhibits HIV-1 RNA and protein synthesis in cells productively infected with both viruses but, in transient assays, activates an HIV-1 long terminal repeat-chloramphenicol acetyltransferase (LTR-CAT) construct introduced into the cell by transfection (V. Koval, C. Clark, M. Vaishnav, S. A. Spector, and D. H. Spector, J. Virol. 65:6969-6978, 1991). We show here that HCMV can also activate an infectious proviral HIV-1 genome transiently transfected into a cell. To ascertain whether integration of the HIV-1 provirus plays a role in these differential effects, we generated monoclonal and polyclonal cell lines that each contain a single integrated copy of an HIV-1 LTR-CAT construct and compared the regulatory effects of HCMV and HIV-1 infection in these cells with those occurring in the same type of cell transiently transfected with the HIV-1 LTR-CAT construct. We find that HCMV activates the transfected HIV-1 promoter 230-fold but activates the integrated promoter only 2.8- to 54-fold. In contrast, HIV-1 stimulates the integrated HIV-1 promoter 2,700- to 6,000-fold but stimulates the transfected promoter only 80-fold. Thus, the relative response of the HIV-1 promoter to HCMV and HIV-1 regulatory proteins depends upon whether it is integrated. To determine if HIV-1 gene products are necessary for the HCMV-mediated repression, we constructed cell lines containing two different stably integrated HIV-1 proviruses: one is tat- and nef-minus and transcriptionally inactive, while the other is env- and nef-minus but actively expresses the other HIV-1 gene products. Upon infection with HCMV, HIV-1 antigen production was stimulated from the inactive HIV-1 genome but inhibited from the active genome. We propose that HCMV has two separate effects on HIV-1 replication during a coinfection. One is a slight stimulatory effect which would be undetectable during an active HIV-1 infection, while the other is a net inhibitory effect that is mediated by an interaction between HCMV and HIV-1 gene products.