Background
Precise detection of HIV-1 integration sites (IS) in the human genome is crucial for studying and characterizing latent reservoirs in people living with HIV-1 (PLWH). The majority of current IS sequencing methods are expensive, require a chain of bioinformatic tools and expertise, and need relatively high numbers of HIV-1 DNA copies as input. Here, we describe a highly sensitive bulk proviral integration site mapping sequencing (PRISM-seq) and its accompanying one-click bioinformatics pipeline BulkIntSiteR, offering an economical solution for high-confidence mapping of HIV-1 IS.
Methods
Genomic DNA samples isolated from HIV-1-infected putative single-cell clones and a human donor were subjected to whole genome amplification using multiple displacement amplification followed by PCR and sequencing of viral-host junctions. Data obtained was processed using BulkIntSiteR, which performs quality control, trims adapters and LTR sequences, maps junctions to HXB2 and hg38, and annotates proviral IS.
Results
We developed and validated PRISM-seq using 14 HIV-1-infected putative single-cell clones, confirmed accurate identification of expected IS, identified low-confidence events likely resulting from PCR or sequencing artifacts, and implemented quality filters to remove such artifacts. Analysis of artificially mixed clone DNA pools showed high sensitivity of PRISM-seq, with reliable recovery of proviral IS even at single-copy input levels (~88 to 95% recovery with 10 replicates). The IS data obtained were further processed using our filtering pipeline to restrict the number of low-confidence IS and to increase assay specificity to ~90% without affecting sensitivity. Applying PRISM-seq to longitudinal clinical samples from ART-treated individuals revealed changes in IS distribution over time, with an increase in proviral integration in ZNF genes, consistent with previous findings from elite controllers and ART-treated individuals. PRISM-seq can be performed at an estimated cost of ~USD 70 per sample replicate, making it feasible for large-scale clinical studies.
Conclusions
PRISM-seq and BulkIntSiteR provide a single-copy sensitive, fully bioinformatically automated approach for high-confidence bulk HIV-1 IS mapping. By lowering technical, computational, and cost barriers, PRISM-seq enables broad adoption for studying reservoir dynamics, clonal persistence, and the genomic determinants of proviral latency, which can help in advancing HIV-1 cure research.