HIV Silencing and Inducibility Are Heterogeneous and Are Affected by Factors Intrinsic to the Virus.

Transcriptionally silent HIV proviruses form the major obstacle to eradicating HIV. Many studies of HIV latency have focused on the cellular mechanisms that maintain silencing of proviral DNA. Here we show that viral sequence variation affecting replicative ability leads to variable rates of silencing and ability to reactivate. We studied naturally occurring and engineered polymorphisms in a recently identified exonic splice enhancer (ESEtat) that regulates tat mRNA splicing and constructed viruses with increased (strain M1), reduced (strain M2), or completely absent (strain ERK) binding of splicing factors essential for optimal production of tat mRNA resulting in a corresponding change in Tat activity. The mutations affected viral replication, with M1 having wild-type (WT) kinetics, M2 exhibiting reduced kinetics, and ERK showing completely abrogated replication. Using single-round infection with green fluorescent protein (GFP)-expressing viruses to study proviral gene expression, we observed progressively greater rates of silencing relating to the degree of ESEtat disruption, with the WT strain at 53%, strain M2 at 69%, and strain ERK at 94%. By stimulating infected cells with a latency reversal agent (phorbol myristate acetate [PMA], panobinostat, or JQ1), we observed that the dose required to achieve 50% of the maximum signal was lowest in the WT, intermediate in M2, and highest in ERK, indicating progressively higher thresholds for reactivation. These results suggest that the ability of silent proviruses to reactivate from latency is variable and that minor differences in the viral sequence can alter the proportion of silenced viruses as well as the threshold required to induce silenced viruses to reactivate and express. IMPORTANCE A reservoir of infected cells in which the HIV genome is transcriptionally silent is acknowledged to be the principal barrier to eradicating the virus from an infected person. A number of cellular processes are implicated in this silencing; however, the viral factors that may contribute remain underexplored. Here we examined mutations altering the correct splicing of HIV gene products as a model to study whether differences in viral sequence can affect either the proportion of viruses that are active or silent or their ability to reactivate. We found that some naturally occurring variations result in viruses that are silenced at a higher rate and require a proportionally increased stimulus for reactivation from latency. These data suggest that the silencing and reactivation behavior of HIV exists in a spectrum, influenced by factors intrinsic to the virus.