CROI 2020 Abstract eBook

Abstract eBook

Oral Abstracts

24LB RECONSTITUTION OF HIV-1 CAPSID-DEPENDENT REPLICATION AND INTEGRATION IN VITRO Devin E. Christensen 1 , Barbie K. Ganser-Pornillos 2 , Owen Pornillos 2 , Wesley I. Sundquist 1 1 University of Utah, Salt Lake City, UT, USA, 2 University of Virginia, Charlottesville, VA, USA Background: To initiate an infection, the HIV-1 genome must be reverse transcribed and integrated into the DNA of the host cell. Despite progress in characterizing and inhibiting these viral processes, detailed mechanistic and structural studies remain challenging because they are executed by individual preintegration complexes deep within cells. Methods: To address these limitations, we have reconstituted the early stages of HIV-1 replication in a cell-free system. Starting with purified virions, membrane permeabilization, capsid stabilization, and dNTPs were used to release viral cores and initiate the process of reverse transcription. Cell-free extracts were used to facilitate efficient integration into a target plasmid. Quantitative PCR (qPCR) was used to monitor three different stages of reverse transcription (Strong Stop, First Strand Transfer, and Late RT). Integration was assayed using three different approaches: 1) a two-step PCR system designed to amplify HIV-1 integration sites coupled with qPCR, 2) deep sequencing of PCR- amplified integration sites, and 3) cloning and sequencing of target plasmids to test for concerted HIV-1 integration. Results: HIV-1 core particles released from permeabilized virions supported highly efficient, capsid-dependent endogenous reverse transcription to produce ~0.8 double-stranded DNA genomes/core. Concerted integration of the transcribed viral genome into a target plasmid then proceeded in a cell extract-dependent reaction. Controls established that, as expected, reverse transcription and integration required active RT and IN enzymes. Efficient viral replication required a stable capsid as assayed by CA mutagenesis and small molecule CA inhibitors. Sequence analyses revealed integration site preferences, and demonstrated that concerted integration occurred with the expected 5 bp target site duplication. Conclusion: Starting with purified HIV-1 virions, we have reconstituted the first half of the HIV-1 life cycle in a cell-free system. This system highlights the key role of the viral capsid and should facilitate dissection of the mechanisms and host factor contributions to HIV-1 replication. 25 STRUCTURAL BASIS OF SECOND-GENERATION HIV INTEGRASE INHIBITOR ACTION AND VIRUS ESCAPE Nicola Cook 1 , Wen Li 2 , Dénes Berta 3 , Magd Badaoui 3 , Allison Ballandras- Colas 1 , Andrea Nans 1 , Abhay Kotecha 4 , Edina Rosta 3 , Alan N. Engelman 2 , Peter Cherepanov 1 1 The Francis Crick Institute, London, UK, 2 Dana–Farber Cancer Institute, Boston, MA, USA, 3 King's College Hospital, London, UK, 4 University of Oxford, Oxford, UK Background: During integration, a multimer of integrase (IN) assembles on viral DNA ends, forming a highly stable nucleoprotein complex termed the intasome. The HIV IN strand transfer inhibitors (INSTIs) specifically target the IN active site engaged with the viral DNA end, in the context of the intasome. Previously, we adopted the intasome from the prototype foamy virus (PFV), which is amenable to X-ray crystallography, to study INSTI binding. However, scarce amino sequence identity with HIV-1 IN outside of the active site greatly limits the use of this highly tractable system in studies of drug resistance. For the same reason, the PFV structures are not ideal templates for optimization of the clinical INSTIs. Methods: To derive a robust model suited to informing INSTI development, we characterized IN proteins from a wide range of simian immunodeficiency viruses (SIVs). We discovered that IN from SIVrcm, which shares a recent common ancestor and 75% amino acid IN sequence identity with HIV-1, readily forms functional nucleoprotein complexes with viral DNA in vitro. Moreover, virus resurrected from the available sequence information was highly susceptible to the first and second-generation INSTIs. We used single-particle cryo-electron microscopy to visualize at near atomic resolution the advanced clinical INSTIs dolutegravir and bictegravir bound to the SIVrcm intasome. Results: We show that the expanded second-generation INSTI scaffolds span the active site, making critical stabilizing contacts with its boundary defined by the IN β4-α2 connector element. The Q148H/G140S mutations that pervade clinical INSTI failure perturb optimal magnesium ion coordination in the integrase active site. The expanded chemical scaffolds of the second-generation

drugs mediate novel interactions with the protein backbone, which are critical for antagonising Q148H/G140S mutant virus. Conclusion: Our results reveal that binding to magnesium ions underpins a fundamental weakness of the INSTI pharmacophore that is exploited by the virus and provide structural framework for the development of this important class of anti-HIV/AIDS therapeutics. Gerlinde Vansant 1 , Julie Janssens 1 , Heng-Chang Chen 1 , Eduard Zorita 1 , Frauke Christ 1 , Guillaume Filion 1 , Zeger Debyser 1 1 Katholieke University Leuven, Leuven, Belgium Background: Since the HIV provirus persists lifelong in memory cells of the immune system but rebounds upon treatment interruption, the latent reservoir is the main target for HIV cure. One of the less studied determinants of latency is the impact of integration site selection on HIV expression. HIV integration is catalyzed by integrase that uses the host chromatin reader LEDGF/p75 to target integration to active genes. We previously showed that inhibition of the LEDGF/ p75-IN interaction by LEDGINs retargets residual integration out of active genes. Moreover, these proviruses were more often in a latent state and refractory to reactivation. These results suggested a direct link between HIV-1 integration and transcription. Methods: We now studied the underlying mechanismwith two advanced technologies. (1) Barcoded HIV (B-HIVE) tags the HIV genome with a unique barcode that allows to determine insert-specific HIV-1 expression by simultaneously tracking the barcode in the DNA and RNA of infected cells. (2) Branched DNA (bDNA) imaging was used to visualize the effect of LEDGINs at the single cell level. bDNA is a signal amplification method for Fluorescent In Situ Hybridization (FISH) that enables simultaneous detection of viral DNA and mRNA. Results: B-HIVE confirmed that LEDGIN treatment retargets integration out of gene-dense regions. LEDGINs increased the distance to H3K36me3 (recognized by LEDGF/p75). Viral RNA expression per DNA barcode was reduced while the proportion of silent proviruses increased. Yet, at high concentrations of LEDGINs some rare residual proviruses with high RNA expression were detected. The silent proviruses after LEDGIN treatment were located further away from epigenetic marks associated with active transcription. Interestingly, while the distance to H3K36me3 changed after treatment, proximity to (super)enhancers stimulated transcription independently of LEDGF/p75. bDNA imaging of SupT1 cells infected with HIV-1 in the presence of LEDGINs showed a dose-dependent reduction in both DNA spots and RNA expression. The DNA spots obtained after treatment with LEDGINs were located at increased distance from the nuclear rim. Finally, LEDGINs hampered reactivation upon stimulation with TNFα 10 days post infection. Conclusion: Our studies reveal how the direct link between integration site selection and transcriptional status of the provirus is mediated by the epigenetic landscape surrounding the integration site. The results support block-and-lock strategies to cure HIV infection. 27 SINGLE-CELL GENOMIC ANALYSIS OF BLOOD AND CSF T CELLS IN HIV+ AND HIV– ADULTS Shelli Farhadian 1 , Ofir Lindenbaum 1 , Jun Zhao 1 , Rolando Garcia-Milian 1 , Jennifer Chiarella 1 , Michelle Chintanaphol 1 , Rachela Calvi 1 , Yuval Kluger 1 , Serena S. Spudich 1 1 Yale University, New Haven, CT, USA Background: The biology driving central nervous system T cell dysregulation in people with HIV (PWH) during antiretroviral therapy (ART) remain incompletely understood. Single cell RNAseq allows high resolution characterization of immune cells, including T cells contained in cerebrospinal fluid (CSF) and blood. We applied distinct approaches to the computational analysis of scRNAseq of T cells to identify genes distinguishing treated-HIV from the HIV-negative state. Methods: scRNA seq was performed on CSF cells and peripheral blood mononuclear cells (PBMC) from PWH on ART (plasma HIV RNA <20 cps/mL for > 1 yr, n=5) and HIV- individuals (n=4). Within each fluid, to compare T cell transcripts differentiating ART-suppressed HIV from the HIV- state we applied: 1. Standard differential expression, using the Seurat FindMarkers function, based on the Wilcoxon rank sum test; and 2. Feature selection, using logistic least absolute shrinkage (LASSO), a machine learning approach to identify genes whose variable expression is most predictive of disease state. 26 THE CHROMATIN LANDSCAPE AT THE HIV-1 INTEGRATION SITE DETERMINES VIRAL EXPRESSION

Oral Abstracts

9

CROI 2020