Fig 1.
Schematic of HIV SMRTcap pipeline.
(A) High molecular weight genomic DNA (gDNA) is extracted from either a cell- or tissue-based source. Integrated HIV-1 provirus is represented in red, with non-specific host gDNA represented in blue. Barcoded universal adapters are green and are used for non-HIV-1-specific amplification of total gDNA pre- and post-oligo-based capture. Once bound to HIV-1-specific probes, enriched with streptavidin-based beads, and amplified, the HIV SMRTcap products are then adapted into a SMRTbell sequencing template via the ligation of hairpin adapters, represented in pink. Sequencing was performed on either the Sequel IIe or Revio platform (see Table 2), and high-fidelity intramolecular consensus (“HiFi”) reads were generated on instrument. (B) Primary analysis included identification, selection, and binning of reads that contain at least 500 bp of continuous HIV-1 genomic content for further examination. (C) Characterization of HIV-1-containing reads enriched by the SMRTcap method, includes (i) identification and classification of integration sites; (ii) definition of expanded clones and filtration of PCR duplicates; and (iii) assessment of proviral integrity using a gene-by-gene mapping approach. Created in BioRender. Sadri, G. (2025) https://BioRender.com/at320g3.
Table 2.
Sequencing platforms used and read metrics for all samples presented.
Table 1.
Samples used for HIV SMRTcap validation and associated subject metadata.
Fig 2.
(A) Multi-sequence alignment of selected HIV SMRTcap reads generated from 8e5 cells carrying a known integration in chromosome 13 (chr13). The long-read, single molecule reads illustrate examples of sheared-end usage for identifying PCR duplicates, and visualization of the three HIV SMRTcap data types: (i) 5’-flanked; (ii) 3’-flanked; and, (iii) dual-flanked HIV-1 proviruses, which share the same integration site and therefore belong to the same clone. Colored bars indicate mismatches versus consensus hg38 and HIV-1 strain HXB2 sequences. Proviral sequence is denoted by the genomic content contained within the red box. (B) Evaluation of 8e5 integration sites identified by HIV SMRTcap across four serial passages: passage 2, 44, 86 and 112 are represented by pink, orange, green, and blue symbols and moving from outer to inner chords, respectively. Open symbols indicate integration within repetitive genomic regions, and filled shapes denote integration within non-repetitive regions. Triangles identify integrations within genes, and circles denote integrations in intergenic regions. The Y-axis represents the number of unique reads sharing the same integration site (clone size). The black arrow identifies the expected main integration event on chr13. (C) Orthogonal validation comparing identification of integration sites by HIV SMRTcap and LM-PCR using the same 8e5 cell-derived gDNA input, specifically from passage 86. Triangles on the outer chord indicate integration sites identified by HIV SMRTcap, while the inner chord displays integration sites detected by LM-PCR in squares. Clone size is represented on the Y-axis, and open/closed symbols again identify integration sites in repeat or non-repetitive genomic regions, respectively. (D) Molecular validation of a selected novel integration site on chromosome 15 on a TapeStation gel using PCR amplification with primers based on HIV SMRTcap data. The known chr13 integration site was included as a positive control.
Fig 3.
HIV SMRTcap Performance across four major HIV-1 subtypes.
(A) Stacked bar graphs demonstrating integration sites identified with HIV SMRTcap from cells infected with each of four major HIV-1 subtypes. Stacked bars indicate integrations sites per genomic region (black headers) with the relative height of the bar representing the proportion of reads classified as intact (blue), putatively intact (green), internal deletions (purple), or truncated (maroon). Indeterminate reads that do not inform on proviral integrity due to random shearing are orange. (B) Relative frequency of unique HIV-1 integration events occurring in non-repeat and varied genomic repeat elements per indicated subtype. (C) Detailed gene segment-by-segment evaluation of proviral integrity across HIV-1 subtypes. Each segment was classified by the extent of detected gene content when aligned to the matched viral reference. Dark blue indicates >75% expected gene content present; green indicates between 0% and 75% gene content detected; taupe indicates the gene is missing entirely, but flanking content is present indicating deletion; yellow specifies the relative frequency of reads where detectable gene content was lost due to shearing. *** indicates statistically significance (p < 0.001) in two-way comparisons between accessory gene deletions found in subtype A vs subtype B, subtype A vs subtype C, or subtype A vs subtype D, using a chi-squared tests with Bonferroni correction for multiple testing.
Fig 4.
HIV SMRTcap performance across tissues.
Circos plots represent reservoir profiles from diverse tissues captured by HIV SMRTcap using bulk gDNA as input: (A) basal ganglia, (B) heart, and (C) splenic CD4 + T cells isolated from a humanized mouse model. The Circos plot format allows visualization of the multiple features of HIV SMRTcap data concurrently: (i) the outermost hashed ring represents human chromosomes; (ii) each line within the outer ring corresponds to unique HIV-1 integration sites at their respective chromosomal locations, with line color indicating the proviral genome integrity at the corresponding integration site (see Integrity Legend); and (iii) the Y-axis of the inner ring represents the size of clonal lineages associated with each integration site, with the dot color denoting integration site genomic feature(s) (refer to Feature Legend).
Fig 5.
HIV SMRTcap limit of detection and subsampling.
(A) Subsampling of input gDNA material that occurs during the HIV SMRTcap process was modeled in the top curve, using the largest number of HIV-1 templates (10,000 HIV-1-containing cells) used as input into the limit of detection assay as an example. Stepwise subsampling is indicated by copy number decreases at each step of the protocol, concluding with the predicted final copy number used as input into sequencing is presented. The red horizontal dashed line represents 1000 ng DNA. (B) Empirical HIV SMRTcap data was collected to experimentally establish assay limits of detection. An initial spike-in of 10,000 8e5 cells was made into HIV-negative CEM for a total of 106 cells (1 HIV genome per 108 of sheared 10kb gDNA fragments). Four 10-fold serial dilutions were processed with HIV SMRTcap to generate the rarefaction curve. Green indicates the number of HIV-1 templates detected by HIV SMRTcap at each dilution point; each dilution point was assessed with four replicates to assess technical and biological rigor and reproducibility. The red line represents the predicted HIV-1 copy numbers calculated for each dilution according to the subsampling schema established in (A). (C) Empirically resolved HIV templates for each dilution and replicate compared to the predicted values. (D) Total number of CCS reads generated per replicate run allowing for evaluation of the impact of sequencing depth on HIV SMRTcap efficiency. Fig 5A Top: Created in BioRender. Sadri, G. (2025) https://BioRender.com/89hn9lt. Fig 5A Bottom: Created in BioRender. Sadri, G. (2025) https://BioRender.com/wtnujt3.
Fig 6.
HIV SMRTcap characterization of ART-suppressed individual across HIV-1 subtypes.
(A) Circos plot representation of HIV-1 reservoir in PBMCs collected from an individual (P4) infected with HIV-1 subtype B virus. (B) Clonality analysis of proviruses found in subject P4. Each dot represents a unique provirus detected by HIV SMRTcap. Connected dots filled with the same color represent proviruses belonging to a single expanded clone sharing the indicated integration. (C) Longitudinal proviral profiles from PBMC samples collected from Donor 13 infected with recombinant HIV-1 subtype A1/D. Each layer represents unique integration sites detected at 12 years (red), 14 years (orange), 15 years (yellow), and 17 years (green) after ART initiation. QVOA-based infectious units per million (IUPM) outgrowth values (at day 21) for the indicated, matched time points in Donor 13 are presented in the companion table and were sourced from Ferreira et al. (D) Clonality analysis for Donor 13 at each time point, including respective integration sites with persistent clones shown in the same color across time point graphs. Black arrows represent relevant integrations at CCR3.
Fig 7.
HIV SMRTcap characterization of ART-suppressed individual (“Donor 17”) infected with recombinant HIV subtype A/D.
(A) Longitudinal proviral profiles from PBMC samples collected from Donor 17, infected with HIV-1 subtype A at 9 (red), 11 (orange), 12 (yellow), and 14 (green) years after ART initiation. QVOA IUPM values for the indicated, matched time points in Donor 17 are presented in the companion table and were sourced from Ferreira et al. (B) Clonality graphs corresponding to each time point for Donor 17. Only shared integration sites share colors across time point. Black arrows represent integrations in BACH2.
Fig 8.
HIV SMRTcap comparison to previous methods.
(A) 5’ LTR length histogram of proviruses detected by HIV SMRTcap in PBMCs collected from subject P4. The red box highlights proviruses missing the 5’ LTR. LM-PCR nested primer binding regions clustered around x-axis position 600 are shown in green and blue bars. The pink box identifies proviral sequences harboring truncated LTRs which no longer contain the LM-PCR primer binding sites and may escape LM-PCR detection. (B-C) Stacked bar graphs show proportions of intact proviruses detected by IPDA versus HIV SMRTcap in subject P4 and Donor 17, respectively. (D) Gene-by-gene summary of viral content in subjects P4 and Donor 17, showing deletions in accessory genes (vif, vpr, vpu, nef). (E) Phylogenetic analysis of unique HiFi sequencing reads detected by HIV SMRTcap (blue) and FLIPseq (yellow) methods detected in Donor 17. The red dots indicate a bootstrap value > 0.9 (range: 0-1).
Table 3.
Full-length viral genome references used for HIV SMRTcap capture probe design.