HIV-1 recurrently targets active genes and integrates in the proximity of the nuclear pore compartment in CD4⁺ T cells. However, the genomic features of these genes and the relevance of their transcriptional activity for HIV-1 integration have so far remained unclear. Here we show that recurrently targeted genes are proximal to super-enhancer genomic elements and that they cluster in specific spatial compartments of the T cell nucleus. We further show that these gene clusters acquire their location during the activation of T cells. The clustering of these genes along with their transcriptional activity are the major determinants of HIV-1 integration in T cells. Our results provide evidence of the relevance of the spatial compartmentalization of the genome for HIV-1 integration, thus further strengthening the role of nuclear architecture in viral infection.

HIV-1 persists lifelong in memory cells of the immune system as latent provirus that rebounds upon treatment interruption. Therefore, the latent reservoir is the main target for an HIV cure. Here, we studied the direct link between integration site and transcription using LEDGINs and Barcoded HIV-ensembles (B-HIVE). LEDGINs are antivirals that inhibit the interaction between HIV-1 integrase and the chromatin-tethering factor LEDGF/p75. They were used as a tool to retarget integration, while the effect on HIV expression was measured with B-HIVE. B-HIVE tracks insert-specific HIV expression by tagging a unique barcode in the HIV genome. We confirmed that LEDGINs retarget integration out of gene-dense and actively transcribed regions. The distance to H3K36me3, the marker recognized by LEDGF/p75, clearly increased. LEDGIN treatment reduced viral RNA expression and increased the proportion of silent provirus. Finally, silent proviruses obtained after LEDGIN treatment were located further away from epigenetic marks associated with active transcription. Interestingly, proximity to enhancers stimulated transcription irrespective of LEDGIN treatment, while the distance to H3K36me3 only changed after treatment with LEDGINs. The fact that proximity to these markers are associated with RNA expression support the direct link between provirus integration site and viral expression.

The aim of the present study was to understand the biology of unintegrated HIV-1 DNA and reveal the mechanisms involved in its transcriptional silencing. We found that histones are loaded on HIV-1 DNA after its nuclear import and before its integration in the host genome. Nucleosome positioning analysis along the unintegrated and integrated viral genomes revealed major differences in nucleosome density and position. Indeed, in addition to the well-known nucleosomes Nuc0, Nuc1, and Nuc2 loaded on integrated HIV-1 DNA, we also found NucDHS, a nucleosome that covers the DNase hypersensitive site, in unintegrated viral DNA. In addition, unintegrated viral DNA-associated Nuc0 and Nuc2 were positioned slightly more to the 5′ end relative to their position in integrated DNA. The presence of NucDHS in the proximal region of the long terminal repeat (LTR) promoter was associated with the absence of RNAPII and of the active histone marks H3K4me3 and H3ac at the LTR. Conversely, analysis of integrated HIV-1 DNA showed a loss of NucDHS, loading of RNAPII, and enrichment in active histone marks within the LTR. We propose that unintegrated HIV-1 DNA adopts a repressive chromatin structure that competes with the transcription machinery, leading to its silencing.

Upon HIV-1 infection, a reservoir of latently infected resting T cells prevents the eradication of the virus from patients. To achieve complete depletion, the existing virus-suppressing antiretroviral therapy must be combined with drugs that reactivate the dormant viruses. We previously described a novel chemical scaffold compound, MMQO (8-methoxy-6-methylquinolin-4-ol), that is able to reactivate viral transcription in several models of HIV latency, including J-Lat cells, through an unknown mechanism. MMQO potentiates the activity of known latency-reversing agents (LRAs) or “shock” drugs, such as protein kinase C (PKC) agonists or histone deacetylase (HDAC) inhibitors. Here, we demonstrate that MMQO activates HIV-1 independently of the Tat transactivator. Gene expression microarrays in Jurkat cells indicated that MMQO treatment results in robust immunosuppression, diminishes expression of c-Myc, and causes the dysregulation of acetylation-sensitive genes. These hallmarks indicated that MMQO mimics acetylated lysines of core histones and might function as a bromodomain and extraterminal domain protein family inhibitor (BETi). MMQO functionally mimics the effects of JQ1, a well-known BETi. We confirmed that MMQO interacts with the BET family protein BRD4. Utilizing MMQO and JQ1, we demonstrate how the inhibition of BRD4 targets a subset of latently integrated barcoded proviruses distinct from those targeted by HDAC inhibitors or PKC pathway agonists. Thus, the quinoline-based compound MMQO represents a new class of BET bromodomain inhibitors that, due to its minimalistic structure, holds promise for further optimization for increased affinity and specificity for distinct bromodomain family members and could potentially be of use against a variety of diseases, including HIV infection. IMPORTANCE The suggested “shock and kill” therapy aims to eradicate the latent functional proportion of HIV-1 proviruses in a patient. However, to this day, clinical studies investigating the “shocking” element of this strategy have proven it to be considerably more difficult than anticipated. While the proportion of intracellular viral RNA production and general plasma viral load have been shown to increase upon a shock regimen, the global viral reservoir remains unaffected, highlighting both the inefficiency of the treatments used and the gap in our understanding of viral reactivation in vivo . Utilizing a new BRD4 inhibitor and barcoded HIV-1 minigenomes, we demonstrate that PKC pathway activators and HDAC and bromodomain inhibitors all target different subsets of proviral integration. Considering the fundamental differences of these compounds and the synergies displayed between them, we propose that the field should concentrate on investigating the development of combinatory shock cocktail therapies for improved reservoir reactivation.

The prediction of protein folding rates is a necessary step towards understanding the principles of protein folding. Due to the increasing amount of experimental data, numerous protein folding models and predictors of protein folding rates have been developed in the last decade. The problem has also attracted the attention of scientists from computational fields, which led to the publication of several machine learning-based models to predict the rate of protein folding. Some of them claim to predict the logarithm of protein folding rate with an accuracy greater than 90%. However, there are reasons to believe that such claims are exaggerated due to large fluctuations and overfitting of the estimates. When we confronted three selected published models with new data, we found a much lower predictive power than reported in the original publications. Overly optimistic predictive powers appear from violations of the basic principles of machine-learning. We highlight common misconceptions in the studies claiming excessive predictive power and propose to use learning curves as a safeguard against those mistakes. As an example, we show that the current amount of experimental data is insufficient to build a linear predictor of logarithms of folding rates based on protein amino acid composition.