RNA splicing patterns contribute to burst size variation among HIV-1-infected Jurkat cell clones

Abstract

Accurately quantifying virus release from HIV-1-infected cells is central to predicting infection outcomes and evaluating treatment strategies. Recent studies suggest that viral shedding can vary strikingly among infected cells. To identify predictors of virus release levels, a previously developed high-throughput molecular barcoding system was used to track the expression properties of individual infected clones within a polyclonal population. Consistent with previous reports, virus release spanned four orders of magnitude among clonal integrants. While reporter gene expression correlated poorly with virus release, intracellular viral RNA levels correlated well, and levels of unspliced HIV-1 RNA correlated most closely. Comparing clones with different virus release levels showed that they varied not only in total intracellular HIV-1 RNA levels but also in levels of viral RNA splicing. Remobilizing proviruses revealed that splicing differences were largely due to cell-intrinsic properties, although splicing differences were due to heritable features of the parent provirus for at least one clone. This clone displayed high levels of reporter gene expression from an HIV-1spliced RNA, but low levels of unspliced viral RNA and virus release. This over-splicing clone, which contained a non-synonymous substitution in rev, did not respond to treatment with the latency-reversing agent JQ1 but displayed increased virus release in response to treatment with pladienolide B, a splicing inhibitor. These findings demonstrate that splicing differences can contribute to virus production levels and suggest that future studies on proviral populations that include expanded clones may benefit from assessing clone-specific splicing properties.

Importance: Many models of HIV-1 infection rely on the assumption that actively infected cells release similar amounts of virus, despite recent reports that suggest shedding differs drastically among infected cells. In this study, the expression phenotypes of hundreds of integrant clones were analyzed to identify factors contributing to burst size variation. In agreement with previous reports, virus release spanned over four orders of magnitude within the infected pool. Both proviral expression levels and variation in splicing contributed to these burst size differences. While cell-intrinsic factors appeared to be the primary contributors to heterogeneous shedding patterns, viral point mutations were also observed and, in at least one case, contributed to particle release levels. Together, these findings demonstrate that expression variation among proviruses is both large and multifaceted and suggest that clone-specific differences in HIV-1 expression properties may contribute to unpredicted responses to treatment interventions.

Keywords: HIV-1 expression parameters; HIV-1 population dynamics; RNA splicing; human immunodeficiency virus; latency-reversing agents.