Reference Plasmid pHXB2_D is an HIV-1 Molecular Clone that Exhibits Identical LTRs and a Single Integration Site Indicative of an HIV Provirus

Abstract

Objective: To compare long-read nanopore DNA sequencing (DNA-seq) with short-read sequencing-by-synthesis for sequencing a full-length (e.g., non-deletion, nor reporter) HIV-1 model provirus in plasmid pHXB2_D. Design: We sequenced pHXB2_D and a control plasmid pNL4-3_gag-pol(Δ1443-4553)_EGFP with long- and short-read DNA-seq, evaluating sample variability with resequencing (sequencing and mapping to reference HXB2) and de novo viral genome assembly. Methods: We prepared pHXB2_D and pNL4-3_gag-pol(Δ1443-4553)_EGFP for long-read nanopore DNA-seq, varying DNA polymerases Taq (Sigma-Aldrich) and Long Amplicon (LA) Taq (Takara). Nanopore basecallers were compared. After aligning reads to the reference HXB2 to evaluate sample coverage, we looked for variants. We next assembled reads into contigs, followed by finishing and polishing. We hired an external core to sequence-verify pHXB2_D and pNL4-3_gag-pol(Δ1443-4553)_EGFP with single-end 150 base-long Illumina reads, after masking sample identity. Results: We achieved full-coverage (100%) of HXB2 HIV-1 from 5' to 3' long terminal repeats (LTRs), with median per-base coverage of over 9000x in one experiment on a single MinION flow cell. The longest HIV-spanning read to-date was generated, at a length of 11,487 bases, which included full-length HIV-1 and plasmid backbone with flanking host sequences supporting a single HXB2 integration event. We discovered 20 single nucleotide variants in pHXB2_D compared to reference, verified by short-read DNA sequencing. There were no variants detected in the HIV-1 segments of pNL4-3_gag-pol(Δ1443-4553)_EGFP. Conclusions: Nanopore sequencing performed as-expected, phasing LTRs, and even covering full-length HIV. The discovery of variants in a reference plasmid demonstrates the need for sequence verification moving forward, in line with calls from funding agencies for reagent verification. These results illustrate the utility of long-read DNA-seq to advance the study of HIV at single integration site resolution.

“Just Saiyan: Tail-trimming Human Monkeypox Virus Assemblies Emphasizes Resolvable Regions in Inverted Terminal Repeats to Improve the Resolution of Reference and Production Genomes for Genomic Surveillance”

“…The HIV-1 genome was originally represented as a DNA provirus that was captured by restriction cloning into a molecular clone/plasmid (Gener et al, 2021). This enabled subsequent molecular biological dissection of the virus, which is still ongoing after 37+ years (Gener, 2022).…”

“…HIV-1 genomes are modeled differently depending on the reference accession, and whether the intent is to model the proviral DNA form or the viral mRNA form. The main reference genome used by the Los Alamos National Laboratory’s HIV Sequence Database is HXB2 GenBank: K03455.1, which was recently physically re-sequenced (Gener et al, 2021) and verified as a legitimate provirus with identical flanking LTRs. This sequence is important because all downstream analyses in the HIV Sequence Database are based either directly or indirectly on this historical species-specific reference.…”

Anticipating HIV drug resistance with appropriate sequencing methods

“…Thankfully, the results reported by Oluniyi et al [1] do not yet preclude the use of INSTIs in Nigeria. We recently developed a tractable method to create HIV transcript models from high-quality long-read data [10,11]. Our recent findings from reanalyzing some of the highest-quality public long-read HIV cDNA datasets [12,13] and our own [9] support the possibility of at least five previously undescribed spliceassociated HIV integrase isoforms -GENER4-integrase, GENER5-integrase, GENER6/8-integrase, GENER7integrase, GENER9-integrase -in addition to the previously well characterized p31 integrase for which modern INSTIs are designed.…”