Introduction

Retroviruses cause permanent infection in the host by integrating their reverse-transcribed viral genome into the host genome. Retroviral integration considerably impacts a wide range of biological phenomena, including the persistence of fatal human diseases and the shaping of metazoan evolution (1). Human immunodeficiency virus (HIV)-1 is a representative retrovirus that underlies the global burden of acquired immune deficiency syndrome (AIDS) (2). The chromosomal landscape of HIV-1 integration plays a critical role in proviral gene expression, persistence of integrated proviruses, and prognosis of antiretroviral therapy (3–5). Integration into the host genome is not random and displays distinct preferences for gene-dense regions, where active transcription occurs (6), by interacting host factors such as transcription activators, epigenetic marker binding proteins and super enhancers (7–13). However, transcription activity is not the sole determinant of the HIV-1 integration site landscape (10). For instance, the most favored region of HIV-1 integration is an intergenic locus, and despite the lower probability of integration, HIV-1 proviruses are observed in non-genic regions in the genomes of infected individuals (4, 6). This indicates the possibility of there being an undiscovered mechanism or determinant that composes the correct genomic environment for HIV-1 integration.

An R-loop is a three-stranded nucleic acid structure that comprises a DNA–RNA hybrid and displaced strand of DNA, and has long been considered a transcription byproduct (14, 15). R-loops in cellular genomes are enriched in actively transcribed genes as they occur naturally during transcription (14, 16), but R-loop formation is not limited to gene body regions and is widespread in the genome (14). As a result of in trans R-loop formation, R-loops are also abundant in non-genic regions, such as intergenic regions, repetitive sequences, including transposable elements, centromeres, or telomeres (14, 17–19), independently of transcription activity of the genes harboring the R-loops. Although R-loops are identified as critical intermediates and regulators in a number of biological processes (14, 15, 20), the dynamics and the role played by cellular R-loops in pathological contexts remain unrevealed.

R-loops are important contributors molding the genomic environment and spatial organization of the cellular genome, and can potentially take a novel role in host-pathogen interaction. In the cellular genome, R-loops relieve superhelical stresses and are often associated with open chromatin marks and active enhancers (21, 22), which are also distributed over HIV-1 integration sites (6, 9, 10). In the case of transcription-induced R-loop formation, a guanine-quadruplex (G4) structure can be generated in the non-template DNA strand of the R-loop (23). A recent study has shown that G4 DNA can influence both productive and latent HIV-1 integration (24). In addition, R-loops are prevalent non-canonical B-form DNA structures (25) and intermediates between B-form DNA and A-form RNA conformation (26), which have recently been disclosed to be the conformational characteristics of the target DNA during retroviral integration (26, 27). The accumulated evidence implicates that host genomic R-loops are undiscovered host factor in HIV-1 integration site selection mechanism, which dynamically interact with the host genomic environment.

Here, we showed a notable role of R-loops in the interaction between HIV-1 and its host, specifically in HIV-1 integration. HIV-1-infection induces host cellular R-loop formation and the R-loop rich regions of the host genome are preferred by HIV-1 integration. HIV-1 integrase proteins showed considerable binding affinity to nucleic acid substrate comprising R-loop structures. Our results suggest that R-loops are an important composer of host genomic environment for HIV-1 integration site determination.

Results

Host genomic R-loops accumulate by HIV-infection

To investigate the relationship between HIV-1 infection and host cellular R-loops, we first analyzed R-loop dynamics in different types of cells infected with HIV-1 at early post-infection time points using DNA–RNA immunoprecipitation followed by cDNA conversion coupled to high-throughput sequencing (DRIPc-seq) using a DNA–RNA hybrid-specific binding antibody, anti-S9.6 (28). HeLa cells, primary CD4⁺ T cells isolated from two individual donors and CD4⁺/CD8^- T cell lymphoma Jurkat cell line were infected with VSV-G-pseudotyped HIV-1-EGFP and harvested at 0, 3, 6, and 12 h post infection (hpi) for DRIPc-seq library construction (Fig. 1A and S1A-C Fig.). Our DRIPc-seq analysis yielded loci specific R-loop signals at the referenced R-loop-positive loci (RPL13A and CALM3) and an R-loop-negative locus (SNRPN) (28) that were both strand-specific and highly sensitive to pre-immunoprecipitation in vitro RNase H treatment, in HeLa cells, CD4⁺ and Jurkat T cells (Table S1-3). Notably, the number of DRIPc-seq peaks mapped to the human reference genome increased remarkably during early post infection of HIV-1 (3 and 6 hpi for HeLa cells and 6 and 12 hpi for CD4⁺ and Jurkat T cells; Fig. 1B). Most of the peaks mapped in cells harvested at 0 hpi were commonly found in all other samples, but a significant numbers of unique peaks were observed after infection (Fig. 1C).

Fig 1.

In addition to our DRIPc-seq data analysis, we used different biochemical approaches to examine R-loop accumulation after HIV-1 infection in HeLa cells. First, R-loop accumulation in HIV-1-infected cells was observed using DNA–RNA hybrid dot blots with the anti-S9.6 antibodies (Fig. 1D). The dot intensity increased significantly upon HIV-1 infection at 6 hpi and the enhanced R-loop signals on dot blots of HIV-1-infected cells were highly sensitive to in vitro treatment with RNase H (Fig. 1D). This result was highly consistent with our DRIPc-seq data analysis results in HIV-1-infected HeLa cells.

Subsequently, we observed HIV-1-induced R-loops using an immunofluorescence assay by probing HIV-1-infected or non-infected control cells with S9.6 antibody at 6 hpi (Fig. 1E, left). The nuclear fluorescence signal associated with the R-loops after subtracting the nucleolar signal was significantly enhanced in cells infected with HIV-1 (Fig. 1E, right). We validated and quantified HIV-1-infection induced R-loop formation on the host genome in a genome-site specific manner by using DRIP followed by real-time polymerase chain reaction (DRIP-qPCR). In this experiment, the S9.6 signal was determined for three and two HIV-1-induced-R-loop-positive (P1, P2, and P3) and -negative regions (N1 and N2), respectively, where were defined by DRIPc-seq data analysis (S2A-E Fig.). We detected significantly increased R-loop signals that are highly sensitive to RNase H treatment of pre-immunoprecipitates in the P1, P2, and P3 regions of HIV-1-infected cells at 6 hpi compared to those in the cells harvested at 0 hpi (S3A Fig.). However, the HIV-1-induced R-loop-negative regions, N1 and N2, did not show significant R-loop accumulations (S3A Fig.).

Importantly, the R-loop signal was enriched even in cells infected with HIV-1 when the reverse transcription or integration of HIV-1 is blocked by enzyme inhibitors like Nevirapine (NVP) or Raltegravir (RAL), respectively (S3B and S3C Fig.). This result indicates that the enrichment of R-loop signals in cells are originated from the host genome but not by DNA-RNA hybrid formation during the viral life cycle or transcriptional burst from integrated HIV-1 proviruses. In addition, we confirmed that nearly 100% of DRIPc-seq reads from HIV-1-infected HeLa, CD4⁺ and Jurkat T cells were aligned to the host cellular genome, but not on that of HIV-1 (S3D Fig.). Together, these data demonstrate that HIV-1 infection induced host genomic R-loop formation at early post-infection.

R-loops accumulation after HIV-1 infection are widely distributed in both genic and non-genic regions

To investigate the distribution of cellular genomic R-loops during HIV-1 infection, we conducted a genome-wide analysis of our DRIPc-seq data. The unique DRIPc-seq peaks observed after HIV-1 infection were not only numerous but also relative longer (Fig. 2A).

Fig. 2.

This suggests that R-loops induced by HIV-1 infection occupy a genomic region larger than that of the R-loops presents without HIV-1 infection. We observed a significant accumulation of R-loops over diverse genomic compartments at the hpi of HIV-1-infection induced R-loop formation (Fig. 2B). The presence of R-loops is often correlated with high transcriptional activity, and we found significantly high proportion of DRIPc-seq peaks enrichment upon HIV-1 infection in the gene body regions (Fig. 2B). However, we also observed enrichment of HIV-1-infection induced DRIPc-seq peaks proportion mapped to intergenic or repeat regions, including short interspersed nuclear elements (SINEs), long interspersed nuclear elements (LINEs), and long terminal repeat (LTR) retrotransposons, where transcription is typically repressed (Fig. 2B). Although the expression of repetitive elements is mostly repressed during normal cellular activities, HIV-1 infection could activate endogenous retroviral promoters (29, 30). To investigate the possibility that R-loop induction in gene-silent regions is associated with transcriptome changes during HIV-1 infection, we performed RNA sequencing (RNA-seq) for HIV-1-infected HeLa cells at 0, 3, 6, and 12 hpi. Consistent with previous reports, we observed an increase in the expression levels of repetitive elements at later time points post-infection (S4A Fig.; 12 hpi). In contrast, we found that the expression levels of SINEs, LINEs, and LTRs were even lower at both 3 and 6 hpi compared to 0 hpi while HIV-1-induced R-loops were significantly accumulated, compared to 0 hpi (S4A Fig.). We further examined the expression profile of genes containing R-loop in HeLa cells. The expression profile of genes harboring HIV-1-induced R-loops in their gene bodies showed very weak correlations with the signals of DRIPc-seq peaks at 3 hpi (Pearson’s r = 0.21, P = 1.08 × 10^-84; Fig. 2C) and at 6 hpi samples (Pearson’s r = −0.34, P = 2.40 × 10^-228; Fig. 2C), which implies that the unique R-loop peaks upon HIV-1 infection do not engage with transcriptional burst. In agreement with our DRIPc-seq and global RNA-seq data analysis, the expression level of the genes harboring HIV-1-infection induced R-loops, which were quantified by DRIP-qPCR (S3A Fig.), were not significantly affected by HIV-1 infection (S4B Fig. and Table S4). Together, our data demonstrate that host cellular R-loop accumulation upon HIV-1 infection are widely distributed in both genic and non-genic regions and are not necessarily correlate with the expression levels of the genes harboring the R-loops.

HIV-1 integration sites are enriched at systemically induced sequence-specific R-loop regions in cell model

HIV-1 completes its infection by integrating its viral genome into the host’s through dynamic interaction with the host genome (31). Besides, as HIV-1 infection induced R-loop accumulation at early post infection hours when HIV-1 genome are imported into nucleus and integration may initiate (32–34), we hypothesized that host genomic R-loops play a role in HIV-1 integration, and possibly in integration site selection. To systemically and directly assess the relationship between host genomic R-loops and HIV-1 integration in a genome-site-specific manner, we adapted and modified an elegantly designed episomal system that induces sequence specific R-loops through DOX-inducible promoters (16). To most closely mimic the presence R-loop in host cellular genome, we subcloned the R-loop-forming portion of the mouse gene encoding AIRN (mAIRN) (17) or non-R-loop-forming ECFP sequence with a DOX-inducible promoter into the piggyBac transposon vector and co-expressed the piggyBac transposase in HeLa cells. These R-loop forming (mAIRN) or non-R-loop forming sequence (ECFP) are non-human sequences. Therefore, our cell model allows us to induce and quantify R-loop formation at designated genomic region and distinguish the R-loop formation from the endogenous R-loops on the cellular genome, which are not sequence-specific and impossible to control for induction. Moreover, by using this system we can quantify R-loop-dependent site-specific HIV-1 integration events at the designated regions, which can also be distinguished from HIV-1 integration event at endogenous host genomic loci. We designated the pool of cells with the R-loop forming sequence (mAIRN) inserted into its genome as “pgR-rich (piggyBac R-loop rich)” cell line and the pool of cells with the non-R-loop forming sequence (ECFP) inserted into its genome as “pgR-poor (piggyBac R-loop poor)” cell line (Fig. 3A).

Fig. 3.

A similar number of the copies of piggyBac transposon was successfully delivered to the genome of each cell line (S5A Fig.), and DOX treatment strongly induced the transcriptional activity of mAIRN or ECFP without affecting the transcription of endogenous loci in both cell lines (S5B and S5C Fig.). Although the transcription of mAIRN or ECFP was strongly induced upon DOX treatment, the activity did not exceed that of endogenous loci in both cell lines (S5D and S5E Fig.). While two cell lines showed comparable level of DOX-inducible transcription activity at the designated sequences (Fig. 3B), only pgR-rich cells exhibited robust RNase H-sensitive stable R-loop formation upon DOX treatment (Fig. 3C, mAIRN). By contrast, R-loops were weakly formed in the pgR-poor cells where non-R-loop forming sequence (ECFP) inserted into its genome (Fig. 3C, ECFP).

To examine whether the formation of ‘extra’ R-loops in the host genome influence HIV-1-infection to the host cells, we infected both cell lines with VSV-G-pseudotyped HIV-1-luciferase viruses and examined the luciferase activity. Interestingly, we found that pgR-rich cells showed significantly high luciferase activity only when R-loops were induced by DOX treatment, whereas pgR-poor cells showed comparable luciferase activity regardless of transcription activation by DOX treatment (Fig. 3D). We conducted HIV-1 integration site sequencing in HIV-1-infected pgR-poor and pgR-rich cells to directly quantify site-specific integration events at sequence-specific R-loop regions. Remarkably, integration events were significantly higher in pgR-rich cells only when R-loops were induced by DOX treatment (Fig. 3E). However, HIV-1 integration frequency within non-R-loop forming sequence in pgR-poor cells remained very low, even with transcription activation by DOX treatment (Fig. 3E). HIV-1 integration frequency was enriched at the vicinity of R-loop forming regions in pgR-rich cell line upon DOX treatment, but the enrichment was not observed in pgR-poor cells that does not form stable R-loops even after transcription activation by DOX treatment (Figs. 3F and 3G). This cell-based R-loop inducing system with independent control over transcription and R-loop formation enabled the direct measurement of HIV-1 integration events at the defined R-loop regions, and the results indicate that host genomic R-loops are targeted by HIV-1 integration. Moreover, our data suggest that transcription activity itself is not sufficient for HIV-1 integration site determination, but the formation of R-loops accounts for HIV-1 integration site selection.

Host genomic R-loops are targeted by HIV-1 integration

We attempted to further validate the relationship between R-loops and the HIV-1 integration site selection by global analysis of HIV-1 integration sites on endogenous genomic regions of HIV-1 infected host cells. We performed HIV-1 integration site sequencing in HIV-1 infected HeLa cells, CD4⁺ and Jurkat T cells and analyzed the sequencing data combined with our DRIPc-seq data. We counted and compared the number of successfully integrated proviruses in the R-loop regions (the combined genomic regions within 30-kb windows centered on DRIPc-seq peaks from 0, 3, 6, and 12 hpi) to those in non-R-loop forming regions (the total genomic regions outside of the 30-kb windows centered on DRIPc-seq peaks). Notably, we found that approximately three to four times more integration were detected in the R-loop regions as in other genomic regions without R-loops in HeLa cells, CD4⁺ and Jurkat T cells (Fig. 4A). Interestingly, the HIV-1 integration sites preferred the center and nearby areas of the R-loops regions (Fig. 4B). We observed biases for HIV-1 integration in HIV-1-induced R-loop-positive regions, P2 and P3, where gave highly induced R-loop signal upon HIV-1 infection in DRIPc-seq analysis and DRIP-qPCR (Fig. 4C). By contrast, HIV-1 integration sites were not detected in R-loop-negative regions, N1 and N2 (Fig. 4D). Overall, our results from bioinformatics analysis using different types of naïve host cells infected with HIV-1 are consistent with the idea that the virus has a preference for targeting R-loops for integration (Fig. 3), and our data suggest R-loops as an important composer of host genomic environment for HIV-1 integration site determination.

Fig. 4.

HIV-1 integrase physically interacts with R-loops on the host genome

HIV-1 intasome tether to the host genome for its viral cDNA integration. Intasomes consist of HIV-1 viral cDNA and HIV-1 coding protein, integrases. We observed that HIV-1 preferentially integrated into R-loops on the host genome, thus we hypothesized that the HIV-1 integrase protein could directly bind and be recruited to the genomic R-loops. To test this hypothesis, we first investigated whether HIV-1 integrase proteins have physical binding affinity to nucleic acid substrates possessing R-loop structure. Although HIV-1 integrases are DNA and RNA binding proteins (35, 36), its binding ability towards such three-stranded nucleic acid structure that is composed with a DNA-RNA hybrid like R-loop has not been investigated. We carried in vitro protein-nucleic acid binding assay by electrophoretic mobility shift assay (EMSA) with Sso7d-tagged HIV-1 integrase recombinant proteins and diverse structures of nucleic acid substrates including R-loop and simple dsDNA duplex.

Interestingly, nucleic acid substrate consisted with R-loop structure bound to HIV-1 integrase proteins with greater binding affinity than simple dsDNA duplex (Fig. 5A). Additionally, R-loop composing forms of nucleic acid structures such as RNA-DNA hybrid with exposed ssDNA (R:D+ssDNA) and RNA-DNA hybrid (hybrid) also hosed high binding affinity to integrases (S6A Fig. and Fig. 5A).

Fig. 5.

We validated the interaction between cellular genomic R-loops and HIV-1 integrase proteins by DNA–RNA hybrid immunoprecipitation using S9.6 antibodies in FLAG-tagged HIV-1 integrase-expressing HeLa cells (Fig. 5B). Under our experimental conditions, R-loops were reproducibly immunoprecipitated (S6B Fig.) and HIV-1 integrase proteins co-immunoprecipitated with R-loops (Fig. 5C). DNA–RNA hybrids also co-immunoprecipitated with the positive control H3 (37) but not with the negative control LaminA/C and Actin (37) (Fig. 5C). To verify the specificity of our co-immunoprecipitation results for R-loops and HIV-1 integrases, we performed DNA–RNA hybrid immunoprecipitation with RNase H treatment (S6C Fig.). The S9.6 signal of immunoprecipitated nucleic acids was highly sensitive to RNase H treatment of pre-immunoprecipitates (Fig. 5D). Accordingly, the blotting signal of the co-immunoprecipitated HIV-1 integrase and H3 proteins was significantly reduced upon RNase H treatment (Fig. 5E). We performed reciprocal immunoprecipitation using an anti-FLAG monoclonal antibody and detected immunoprecipitated R-loops using dot blot analysis with anti-S9.6. R-loops were immunoprecipitated by HIV-1 integrase, and the S9.6 signal of immunoprecipitated nucleic acids was highly sensitive to RNase H treatment (Fig. 5F and S6D Fig.). Subsequently, we attempted to observe the interaction between the R-loops and HIV-1 integrase using a proximity-ligation assay (PLA), in HIV-1-infected cells. We used two antibodies: one that binds to R-loops (anti-S9.6) and another one that binds to GFP-tagged HIV-1 integrase. We detected PLA signals in cells infected with HIV-IN-EGFP virions and in non-infected control cells. PLA signals in non-infected cells were comparable to those in S9.6-alone and GFP-alone single antibody-negative controls; however, PLA signals significantly increased upon HIV-1 infection (Fig. 5G and S6E Fig.). Our data suggest that the HIV-1 frequently targets R-loop-rich regions for viral genome integration by physical binding of HIV-1 integrase proteins to R-loop structures on the host genome.

Discussion

In this study, we found that HIV-1 preferentially integrates into regions rich in R-loops, suggesting that R-loops are a novel host factor governing HIV-1 integration site selection. In our bioinformatics analysis, host cellular R-loops were induced by HIV-1 infection and widespread over host genomic regions. Using our R-loop-inducible cell models, R-loop formation, not necessarily transcription activity itself, was found to be important for HIV-1 integration site determination. In addition, HIV-1 integrase proteins favored physical binding with R-loops in vitro, and they interacted with host genomic R-loops in HIV-1-infected cells. These results demonstrated that HIV-1 exploits and frequently targets the host genomic R-loops for successful integration and infection.

Our data show that HIV-1 targets host genomic R-loops for viral genome integration and its integrase proteins physically interact with genomic R-loops in vitro and in cells. This may because the R-loops own an unique nucleic acid conformation of B-form DNA and A-form RNA intermediates, which possess intrinsic preferential binding ability to retroviral intasome (25–27). Another possible explanation for why HIV-1 integration shows a preference towards host genomic R-loops is that R-loops perhaps take a collaborative role with host factors governing the HIV-1 integration site selection. Cellular R-loops are recognized and regulated by numerous cellular proteins (37, 38). Besides, the correct genomic environment for HIV-1 integration site selection are composed by host proteins (9). LEDGF/p75 (9, 13, 39) and CPSF6 (7, 9) are two decisive host factors that direct HIV-1 integration by interacting with integrase or trafficking viral preintegration complex towards nuclear interior (7, 9). In fact, these host factors have recently been identified as potential R-loop binding proteins in DNA–RNA interactome analysis (37) and R-loop proximity proteomics (38), respectively. R-loops are tightly regulated by DNA damage response proteins (40) and DNA repair machineries take important roles in HIV-1 integration process (31). For example, the Fanconi anemia pathway (41, 42), a well-known R-loop regulatory pathway, has been recently proposed as an HIV-1 integration regulatory factor exploited by HIV-1 (43). Taking into account theses previous studies alongside our current findings, we propose R-loops as another pivotal host factor driving HIV-1 integration site determination and as a possible intermediate regulator of HIV-1 integration site selection by such host proteins.

Viruses often take advantage of various host factors, and targeting viral components that manipulate the host cellular environment can be an effective strategy for antiviral therapy. Our study has shown that host genomic R-loops accumulate significantly shortly after HIV-1 infection. Thus, it is possible that virion-associated HIV-1 proteins are responsible for inducing these R-loops. For instance, the HIV-1 accessory protein Vpr causes genomic damage (44) and transcriptomic changes during the early stages post infection(45), both of which can lead to in cis and in trans R-loop formation (15). Another HIV-1 accessory protein, Vif, counteracts the host antiviral factor, APOBEC3 (46, 47), which were recently found to regulates cellular R-loop levels (48). Identifying the HIV-1 components responsible for inducing host cellular R-loops, and elucidating the mechanism by which they induce genome-wide R-loop formation and contribute to successful viral integration into selective genomic regions, represents an area for further research.

Although most HIV-1 integration occurs in genic regions (4, 6), HIV-1 proviruses are also found in non-genic regions (49) and understanding these “transcriptionally silent” proviruses is critical for developing strategies to completely eliminate HIV-1. In HIV-1 elite controllers, who suppress viral gene expression to undetectable levels, HIV-1 proviruses accumulate in heterochromatic regions (5). Moreover, proviruses with lower expression level can persist in the host genome even during antiretroviral therapy (4). However, the mechanism by which HIV-1 targets gene-silent regions for “invisible” integration remains unclear. Our study has revealed that R-loops are enriched in both genic and non-genic regions during HIV-1 infection, and that the virus preferentially targets these R-loops for integration. We propose that R-loops, particularly those enriched in non-genic regions, may represent the mechanism by which the virus achieves “invisible” and permanent infection.

Materials and methods

Cell culture

HeLa and HEK293T cells were cultured in Dulbecco’s modified Eagle’s medium (Gibco) supplemented with 10% (v/v) fetal bovine serum (FBS, Cytiva), antibiotic mixture (100 units/ml penicillin–streptomycin, Gibco), and 1% (v/v) GlutaMAX-I (Gibco). Jurkat cells were cultured in Roswell Park Memorial Institute (RPMI) 1640 medium (ATCC) supplemented with 10% (v/v) FBS (Cytiva). Cells were incubated at 37°C and 5% CO₂.

Virus production and infection

VSV-G-pseudotyped HIV-1 virus stocks were prepared by performing standard polyethylenimine-mediated transfection of HEK293T monolayers with pNL4-3 ΔEnv EGFP (NIH AIDS Reagent Program 11100) or pNL4-3. Luc.R-E (NIH AIDS Reagent Program, 3418) along with pVSV-G at a ratio of 5:1. HIV-IN-EGFP virions were produced by performing polyethylenimine-mediated transfection of HEK293T cells with 6 µg of pVpr-IN-EGFP, 6 µg of HIV-1 NL4-3 non-infectious molecular clone (pD64E; NIH AIDS Reagent Program 10180), and 1 µg of pVSV-G. The cells were incubated for 4 h before the medium was replaced with fresh complete medium. Virion-containing supernatants were collected after 48 h, filtered through a 0.45-µm syringe filter, and pelleted using the Lenti-X Concentrator (631232; Clontech) according to the manufacturer’s instructions. The multiplicity of infection (MOI) of virus stocks was determined by transducing a known number of HeLa cells with a known amount of virus particles and then counting GFP-positive cells using flow cytometry. For luciferase reporter HIV-1 virus, the HIV-1 p24 antigen content in viral stock were quantified using the HIV1 p24 ELISA kit (Abcam, ab218268), according to the manufacturer’s instruction. For virus infection, HeLa cells were seeded at a density of 0.5–4 × 10⁵ cells/mL on the day before infection. The culture medium was replaced with fresh complete culture medium 2 hpi. The infected cells were washed twice with PBS and harvested at the indicated time points. Jurkat cells were seeded at a density of 1× 10⁶ cells/mL and inoculated with 300ng/p24 capsid antigen. The plates were centrifuged at 1000 g at 30°C for 1 h. The medium was replaced with fresh RPMI 2 h after infection.

Primary cell isolation, culture, T cell activation, and infection

For CD4⁺ T cells isolation, human PBMC (ST70025, STEMCELL Technologies) was mixed and incubated with MACS CD4 MicroBeads (130-045-101, Miltenyi Biotec) and FITC-conjugated mouse anti-CD4 (561005, BD Bioscience) according to the manufacturer’s instructions. Then the CD4⁺ T cells were enriched by using LS Columns (130-042-401, Miltenyi Biotec) and MidiMACS Separator (130-042-302, Miltenyi Biotec). The efficiency of magnetic separation was analyzed by using Flow-Activated Cell Sorter Canto II (BD Bioscience) and Flowjo software (Flowjo).

CD4⁺ T cells were cultured in Roswell Park Memorial Institute (RPMI) 1640 medium (Gibco), supplemented with 10% (v/v) fetal bovine serum (FBS, Cytiva), antibiotic mixture (100 units/ml penicillin–streptomycin, Gibco), 1% (v/v) GlutaMAX-I (Gibco), and 20 ng/ml of IL-2 (PHC0026, Gibco), left in resting state or activated with Dynabeads Human T-Activator CD3/CD28 (1161D, Thermo Fisher Scientific) for 72 h. CD4⁺ T cells activation efficiency was assessed by staining cells with FITC-conjugated mouse anti-CD25 (340694, BD Bioscience) and APC-conjugated mouse anti-CD69 (130-114-046, Miltenyi Biotec) and using Flow-Activated Cell Sorter Canto II (BD Bioscience) and Flowjo software (Flowjo).

Purified and activated CD4⁺ T cells were seeded at a density of 1× 10⁶ cells/mL and inoculated with 600ng/p24 capsid antigen in presence of polybrene. The plates were centrifuged at 1000 g at 30°C for 1 h. The medium was replaced with fresh RPMI 2 h after infection.

DRIP-qPCR

DRIP was performed as described for the construction of the DRIPc-seq library. After the elution of isolated complexes, nucleic acids were purified using the standard phenol-chloroform extract method and used for qPCR. S6 Table presents details of the primer sequences used for DRIP-qPCR analysis.

RNA-seq library construction

For RNA-seq, HeLa cells were infected with VSV-G-pseudotyped HIV-1 NL4-3 ΔEnv EGFP virus at a MOI of 0.6 and harvested at 0, 3, 6, and 12 hpi. Sequencing was performed with biological replicates. Total RNA was extracted using TRIzol reagent (Invitrogen), according to the manufacturer’s instructions. An mRNA sequencing library was constructed using Illumina adaptors harboring p5 and p7 sequences and Rd1 SP and Rd2 SP sequences. Sequencing was performed using the HiSeq2500 system (Illumina).

Luciferase assay

HeLa cells infected with VSV-G-pseudotyped pNL4-3.Luc.R-E HIV-1 viruses were harvested at 48 hpi, and luminescence was measured using the Dual-Luciferase Reporter Assay System (Promega) according to the manufacturer’s instructions. Briefly, 250 μl of passive lysis buffer was used to lyse cells for each sample, 20 μl of the lysate was mixed with 100 μl of the Luciferase Assay Reagent II, and the luminescence of firefly luciferase was measured using a microplate luminometer (Berthold). The luminescence signal were normalized with total protein content, measured by BCA assay.

Quantitative real-time PCR (qPCR)

For RT (reverse transcription)-qPCR, 1 μg of RNA was reverse-transcribed using the ReverTra Ace qPCR RT Kit (TOYOBO) following the manufacturer’s instructions. For qPCR, DNA extracts were prepared using a DNA purification kit (Qiagen, 51106) according to the manufacturer’s instructions. Equivalent amounts of purified gDNA from each sample were analyzed using qPCR. qPCR was performed using TOPreal qPCR PreMIX (Enzynomics, RT500M). The reactions were performed in duplicate or triplicate for technical replicates. PCR was performed using the iCycler iQ real-time PCR detection system (Bio-Rad). All the primers used for qPCR are listed in S6 Table.

DRIPc-seq library construction

DRIP followed by library preparation, next-generation sequencing, and peak calling were performed as described earlier (28). Briefly, the corresponding cells were harvested and their gDNA was extracted. The extracted nucleic acids were fragmented using a restriction enzyme cocktail with BsrB I (NEB, R0102S), HindIII (NEB, R0136L), Xba I (NEB, R0145L), and EcoRI (NEB, R3101L) overnight at 37°C. Half of the fragmented nucleic acids were digested with RNase H (New England Biolabs) overnight at 37°C to serve as a negative control. The digested nucleic acids were cleaned using standard phenol-chloroform extraction and resuspended in DNase/RNase-free water. DNA–RNA hybrids were immunoprecipitated from total nucleic acids using mouse anti-DNA–RNA hybrid S9.6 (Kerafast, ENH001) DRIP binding buffer and incubated overnight at 4°C. Dynabeads Protein A (Invitrogen, 10001D) was used to pull down the DNA-antibody complexes by incubation for 4 h at 4°C. The isolated complexes were washed twice with DRIP binding buffer before elution. For elution, the isolated complexes were incubated in an elution buffer containing proteinase K for 45 min at 55 °C. Subsequently, DNA was purified using the standard phenol-chloroform extract method and subjected to DNase I (Takara, 2270 B) treatment and reverse transcription for DRIPc-seq library construction. DRIPc-seq was performed in biological replicates. S5 Table shows details of the oligonucleotides used for DRIPc-seq library construction. DRIPc-seq libraries were analyzed using 150 bp paired-end sequencing on a HiSeqX Illumina instrument.

Immunofluorescence microscopy

For immunofluorescence assays of S9.6 nuclear signals, when indicated, the cells were pre-extracted with cold 0.5% NP-40 for 3 min on ice. Cells were fixed with 100% ice-cold methanol for 10 min on ice and then incubated with 100% ice-cold acetone for 1 min. The slides were washed three times with 1× PBS and incubated with or without 60 U/mL RNase H (M0297S, NEB) at 37°C for 36 h or left untreated. The slides were subsequently briefly rinsed thrice with 2% BSA/0.05% Tween (in PBS) and incubated with mouse anti-DNA– RNA hybrid S9.6 (Kerafast, ENH001; 1:100) and rabbit anti-nucleolin (Abcam, ab22758; 1:300) in 2% BSA/0.05% Tween (in PBS) for 4 h at 4°C. The slides were then washed three times with 2% BSA/0.05% Tween (in PBS) and incubated with goat anti-rabbit AlexaFluor-488-conjugated (Invitrogen, A-11008) and goat anti-mouse AlexaFluor-568-conjugated (Molecular Probes, A11004) secondary antibodies (1:200) for 2 h at room temperature. The slides were then washed three times with 2% BSA/0.05% Tween (in PBS) and mounted using the ProLong Gold AntiFade reagent (Invitrogen). Images were obtained using an inverted microscope Nikon Eclipse Ti2, equipped with a 1.45 numerical aperture, plan apochromat lambda 100× oil objective, and an scientific complementary metal–oxide–semiconductor camera (Photometrics prime 95 B 25 mm). For each field of view, images were obtained with DAPI395, GFP488, and Alexa594 channels using the NIS-Elements software. For quantification analysis, binary masks of nuclei and nucleoli were generated using the ROI manager and auto local thresholding using the ImageJ software. The intensity of nuclear signals for DNA–RNA hybrids and nucleolin was then quantified. The final DNA–RNA hybrid signals in the nucleus were calculated by subtracting the nucleolin signals from the DNA–RNA hybrid signals.

pgR-rich and -poor cell line generation with piggyBac transposition

We adapted and modified an elegantly designed episomal system that induces defined R-loops with controlled transcription levels (16) for R-loop-forming or non-R-loop-forming sequence subcloning into the piggyBac transposon vector. HeLa cells were seeded at a density of 5 × 10⁴ cells/ml in a 6-well plate. The next day, cells were transfected with 0.2 μg of Super PiggyBac Transposase Expression Vector (System Biosciences, PB210PA-1) and 0.2, 1, or 2 μg of transposon vectors with appropriate “cargo” sub cloned using Lipofectamine 3000 (Invitrogen) according to the manufacturer’s instructions. After 3 days, the cells were treated with 10 μg/ml blasticidin S (Gibco, A1113903) for selection. Cells with positive integrants for more than 7 days were validated using immunoblotting or RT-qPCR following treatment with DOX. Jurkat cells were seeded at a density of 8 × 10⁵ cells/ml in a 6-well plate and transfected with 0.2 µg of transposase and 1 µg of corresponding transposon vectors with Lipofectamine 3000, like HeLa cells. After 3 days, the cells were treated with 10 μg/ml blasticidin S (Gibco, A1113903) for selection. For each passage, cells were cushioned onto Ficoll-Pacque (Cytiva, 17144002) to separate live cells from dead cell debris. The cells over the cushion were washed with PBS and incubated in cell culture medium with 10 µg/ml of blasticidin for further selection for at least 14 days. Cells with positive integrants were validated by immunoblotting after treatment with DOX. Quantification of successfully integrated piggyBac transposons was performed using a piggyBac qPCR copy number kit (System Biosciences, PBC100A-1) according to the manufacturer’s instructions.

HIV-1 integration site sequencing library construction

HIV-1 integration site sequencing library construction was performed as described earlier (7, 9). Summarily, HeLa cells were infected with VSV-G-pseudotyped HIV-1 NL4-3 ΔEnv EGFP virus at a MOI of 0.6 and harvested 5 days post infection. gDNA was isolated using a DNA purification kit (Qiagen, 51106), according to the manufacturer’s instructions. gDNA (10 µg) was digested overnight at 37°C with 100 U each of the restriction endonucleases MseI (NEB, R0525L) and BglII (NEB, R0144L). Linker oligonucleotides, which were compatible for ligation with the MseI-generated DNA ends, were ligated with gDNA overnight at 12°C in reactions containing 1.5 μM ligated linker, 1 μg fragmented DNA, and 800 U T4 DNA ligase (NEB, M0202S). Viral LTR–host DNA junctions were amplified using semi-nested PCR with a unique linker-specific primer and LTR primers. The second round of PCR was carried out with primers binding to the LTR and the linkers for next-generation sequencing. Two PCRs were performed in parallel for the first round of PCR and five PCRs were performed in parallel for the second round of PCR to enhance library diversity. S7 Table presents details of the oligonucleotides used for HIV-1 integration site sequencing library construction. HIV-1 integration site sequencing was performed in biological replicates. Integration site libraries were analyzed using 150 bp paired-end sequencing on a HiSeqX Illumina instrument.

Recombinant Sso7d-IN protein purification

Sso7d-integrase active site mutant E152Q was expressed in Escherichia coli BL21-AI and purified essentially as previously described (50). Briefly, Sso7d-IN (E152Q) expressed BL21-AI cells were lysis in lysis buffer (20 mM HEPES pH 7.5, 2 mM 2-mercaptoethanol, 1 M NaCl, 10% (w/v) glycerol, 20 mM imidazole, 1 mg RNase A, and 1000 U DNase I) and purified by nickel affinity chromatography (Qiagen, 30210). Protein were first loaded on HeparinHP column (GE Healthcare) equilibrated with equilibrated with 20 mM Tris, pH 8.0, 0.5 mM TCEP, 200 mM NaCl, 10% glycerol for anion exchange chromatography prior to the size exclusion chromatography. Proteins were eluted with a linear gradient of NaCl from 200 mM to 1 M. Eluted fractions were pooled and then separated on a Superdex-200 PC 10/300 GL column (GE Healthcare) equilibrated with 20 mM Tris pH 8.0, 0.5 mM TCEP, 500 mM NaCl and 6% (w/v) glycerol. The purified protein was concentrated to 0.6 mg/ml using an Amicon centrifugal contentrator (EMD Millipore), flash-frozen in liquid nitrogen and stored at −80°C.

Electrophoretic mobility shift assay for R-loop binding of Sso7d-IN

To test the binding affinity of Sso7d-tagged HIV-1 IN to different types of nucleic acid substrates, we prepared R-loop, dsDNA, RNA-DNA hybrid with exposed ssDNA (R:D+ssDNA), RNA-DNA hybrid (Hybrid), ssDNA, and ssRNA by annealing different combinations of Cy3, Cy5 or non-labeled oligonucleotides following the previous protocol (51). 10 nM of DNA substrate was incubated with Sso7d-IN at different concentrations in assembly buffer (20 mM HEPES pH 7.5, 5 mM CaCl₂, 8 mM 2-mercaptoethanol, 4 uM ZnCl₂, 100 mM NaCl, 25% (w/v) glycerol and 50 mM 3-(Benzyldimethylammonio) propanesulfonate (NDSB-256)), for 1 h at 30°C then incubated for 15 min on ice. All the reactants were run on 4.5% non-denaturing PAGE in 1× TBE and then Cy3 or Cy5 fluorescence signal was imaged by ChemiDoc MP imaging system (Bio-Rad). S8 Table presents details of the oligonucleotide sequence used for EMSA.

Co-immunoprecipitation of DNA–RNA hybrid

DNA–RNA hybrid immunoprecipitation was performed as described earlier (37). Summarily, non-crosslinked HeLa cells transfected with the pFlag-IN codon-optimized plasmid were lysed in 85 mM KCl, 5 mM PIPES (pH 8.0), and 0.5% NP-40 for 10 min on ice, and then, the lysates were centrifuged at 750 g for 5 min to pellet the nuclei. The pelleted nuclei were resuspended in sodium deoxycholate, SDS, and sodium lauroyl sarcosinate in RSB buffer and were sonicated for 10 min (Diagenode Bioruptor). Extracts were then diluted (1:4 in RSB + T buffer) and subjected to immunoprecipitation with the S9.6 antibody overnight at 4°C. Antibody-bound complexes were incubated with Protein A Dynabeads (Invitrogen) for 4 h at 4°C for immunoprecipitation. Normal mouse IgG antibodies (Santa Cruz, sc-2025) were used as negative controls. RNase A (Thermo Scientific, EN0531) was added during immunoprecipitation at 0.1 ng RNase A per µg gDNA. Beads were washed four times with RSB + T; twice with RSB, and eluted either in 2× LDS (Novex, NP0007), 100 mM DTT for 10 min at 70°C (for western blot), or 1% SDS and 0.1 M NaHCO₃ for 30 min at room temperature (for DNA–RNA hybrid dot blot).

For co-immunoprecipitation of DNA–RNA hybrids with RNase H treatment, gDNA containing RNA–DNA hybrids was isolated from HeLa cells transfected with a pFlag-IN codon-optimized plasmid using a QIAmp DNA Mini Kit (Qiagen, 51304). gDNA was sonicated for 10 min (Diagenode Bioruptor) and then treated with 5.5 U RNase H (NEB, M0297) per µg of DNA overnight at 37 °C. A fraction of gDNA was stored as “nucleic acid input” for dot blot analysis. gDNA was cleaned using standard phenol-chloroform extraction, resuspended in DNase/RNase-free water, enriched for DNA–RNA hybrids using immunoprecipitation with the S9.6 antibody (overnight at 4°C), isolated with Protein A Dynabeads (Invitrogen; 4 h at 4°C), washed thrice with RSB+T. The immunoprecipitated complexes were incubated with nuclear extracts of HeLa cells transfected with the pFlag-IN codon-optimized plasmid for 2 h at 4°C with diluted HeLa nuclear extracts. The cell lysate containing proteins were pre-treated with 0.1 mg/ml RNase A (Thermo Scientific, EN0531) for 1 h at 37°C to degrade all RNA–DNA hybrids, and the excess of RNase A was blocked by adding 200 U of SUPERase in RNase inhibitor (Invitrogen, AM2694) for immunoprecipitation. In addition, 100 μL fraction of diluted and RNase A pre-treated extracts prior to immunoprecipitation was stored as “protein input” for western blotting. Beads were washed four times with RSB + T; twice with RSB, and eluted either in 2× LDS (Novex, NP0007), 100 mM DTT for 10 min at 70°C (for western blot), or 1% SDS, and 0.1 M NaHCO₃ for 30 min at room temperature (for DNA–RNA hybrid dot blot).

PLA

For PLA, HeLa cells were grown on coverslips and infected with HIV-IN-EGFP virions. At 6 hpi, cells were pre-extracted with cold 0.5% NP-40 for 3 min on ice. The cells were fixed with 4% paraformaldehyde in PBS for 15 min at 4 °C. The cells were then blocked with 1× blocking solution (Merck, DUO92102) for 1 h at 37°C in a humidity chamber. After blocking, cells were incubated with the following primary antibodies overnight at 4°C for S9.6-HIV-1-IN_PLA: mouse anti-DNA–RNA hybrid S9.6 (1:250; Kerafast, ENH001) and rabbit anti-GFP (1:500; Abcam, ab6556). The following day, after washing with once with buffer A twice (Merck, DUO92102), cells were incubated with pre-mixed Duolink PLA plus (anti-mouse) and PLA minus probes (anti-rabbit) antibodies for 1 h at 37°C. The subsequent steps in the proximal ligation assay were performed using the Duolink PLA Fluorescence kit (Sigma) according to the manufacturer’s instructions. To obtain images, the mounted specimens were visually scanned and representative images were acquired using a Zeiss LSM 710 laser scanning confocal microscope (Carl Zeiss). The number of intranuclear PLA puncta was quantified using the ImageJ software. For each biological replicate and experiment, a PLA with a single antibody was performed as a negative control under the same conditions.

DRIPc-Seq data processing and peak calling

DRIPc-seq reads were quality-controlled using FastQC v0.11.9 (52), and sequencing adapters were trimmed using Trim Galore! v0.6.6 (53) based on Cutadapt v2.8 (54). Trimmed reads were aligned to the hg38 reference genome using bwa v0.7.17-r1188 (55). Read deduplication and peak calling were performed using MACS v2.2.7.1 (56). Because R-loops appear as both narrow and broad peaks in DRIPc-seq read alignment owing to its variable length, two independent “MACS2 callpeak” runs were performed for narrow and broad peak calling. The narrow and broad peaks were merged using Bedtools v2.26.0 (57). To increase the sensitivity of DRIPc-seq peak identification, peaks were called after pooling the two biological replicates of the DRIPc-seq sequencing data for each condition.

Consensus R-loop peak calling

The R-loop peaks at 0, 3, 6 and 12 hpi were first merged using “bedtools merge” to create a universal set of R-loop peaks across time points (n = 46542). Then, each of the universal R-loop peaks was tested for overlap with the R-loop peaks for 0, 3, 6 and 12 hpi using “bedtools intersect”. In all, 9,190, 21,403, 33,544, and 9,941 peaks overlapped with 0, 3, 6, and 12 hpi R-loop peaks, respectively. For CD4 cells, we identified a universal R-loop set consisting of 3,928 R-loops, and among them, 737, 722, 1,796 and 2,766 peaks overlapped with 0, 3, 6 and 12hpi R-loop peaks.

HIV-1 integration site sequencing data processing

Quality control of HIV-1 integration site-sequencing reads was performed using FastQC v0.11.9. To discard primers and linkers specific for integration site-sequencing from reads, we used Cutadapt v2.8 with the following option: “-u 49-U 38--minimum-length 36--pair-filter any--action trim-q0,0 –a linker-A TGCTAGAGATTTTCCACACTGACTGGGTCTGAGGG-A GGGTCTGAGGG--no-indels--overlap 12”. This allowed the first position of the read alignment to directly represent the genomic position of HIV-1 integration. Processed reads were aligned to the hg38 reference genome using bwa v0.7.17-r1188, and integration sites were identified using an in-house Python script. Genomic positions supported by more than five read alignments were regarded as HIV-1 integration sites. For Jurkat cells, we adopted integration site sequencing data of HIV-1 infected wild type Jurkat cells from SRR12322252 (58).

Co-localization analysis of R-loops and integration sites

Enrichment of integration sites near the R-loop peaks was tested using a randomized permutation test. Randomized R-loop peaks were generated using “bedtools shuffle” command, thus preserving the number and the length distribution of the R-loop peaks during the randomization process. Similarly, integration sites were randomized using the “bedtools shuffle” command. Randomization was performed 100 times. ENCODE blacklist regions (59) were excluded while shuffling the R-loops and integration sites to exclude inaccessible genomic regions from the analysis. For each of the observed (or randomized) integration sites, the closest observed (or randomized) R-loop peak and the corresponding genomic distance were identified using the “bedtools closest” command. The distribution of the genomic distances was displayed to show the local enrichment of integration sites in terms of the increased proportion of integration sites within the 30-kb window centered on R-loops compared to their randomized counterparts.

DNA plasmid construction and transfection

R-loop-forming mAIRN and non-R-loop forming ECPF sequences were subcloned from pSH26 and pSH36 plasmids, which were generously provided by Prof. Karlene A. Cimprich, into the piggyBac transposon vector, where the tet operator sequences were located upstream of the minimal CMV promoter. The pFlag-IN codon-optimized plasmid and pVpr-IN-EGFP were kindly provided by Prof. A. Engelman and Prof. Anna Cereseto, respectively. Lipofectamine 3000 (Invitrogen) transfection reagent was used for the transfection of all plasmids into cells, according to the manufacturer’s protocol.

DNA–RNA hybrid dot blotting

Total gDNA was extracted using the QIAmp DNA Mini Kit (Qiagen, 51304) according to the manufacturer’s instructions. gDNA (1.2 μg) was treated with 2 U RNase H (NEB, M2097) per µg of gDNA for 4 h at 37°C, with half of the sample left untreated but denatured. Half of the DNA sample was probed with S9.6 antibody (1:1000), and the other half was probed with an anti-ssDNA antibody (MAB3034, Millipore, 1:10000).

Immunoblotting

Cells were lysed using RIPA buffer (50 mM Tris, 150 mM sodium chloride, 0.5% sodium deoxycholate, 0.1% SDS, and 1.0% NP-40) supplemented with 10 μM leupeptin (Sigma-Aldrich) and 1 mM phenylmethanesulfonyl fluoride (Sigma-Aldrich) and boiled at 98°C for 10 min with SDS sample buffer prior to SDS-PAGE. The primary antibodies used were mouse monoclonal anti-FLAG M2 (Sigma, F3165), monoclonal mouse anti-HSC70 (Abcam, ab2788), polyclonal rabbit anti-histone H3 (tri methyl K4) antibody (Abcam, ab8580), monoclonal mouse anti-HIV-1 Integrase (Santa Cruz, sc-69721), rabbit anti-LaminA/C antibody (Cell Signaling, 2032), and monoclonal mouse anti-Actin (Invitrogen, MA1-744). All primary antibodies were used at a dilution of 1:1000 for western blotting. Peroxidase-conjugated anti-mouse IgG (115-035-062) and anti-rabbit IgG (111-035-003; both Jackson Laboratories) were used as secondary antibodies at 1:5000 dilution. Signals were detected using the SuperSignal West Pico chemiluminescence kit (Thermo Fisher Scientific).

RNA-seq data processing

RNA-seq reads were quality-controlled and adapter-trimmed as in DRIPc-seq processing. To quantify the expression levels of protein-coding genes, processed reads were aligned to the hg38 reference genome with GENCODE v37 gene annotation (60) using STAR v2.7.3a (61). Gene expression quantification was performed using RSEM v1.3.1. To quantify the expression levels of transposable elements (TEs), we used TEtranscripts v2.2.1 (62).

Processed reads were first aligned to the hg38 reference genome using GENCODE v37 and RepeatMasker TE annotation using STAR v2.7.3a. In this case, STAR options were modified as follows to utilize multimapping reads in downstream analyses: “--outFilterMultimapNmax 100--winAnchorMultimapNmax 100--outMultimapperOrder random--runRNGseed 77--outSAMmultNmax 1--outFilterType BySJout--alignSJoverhangMin 8--alignSJDBoverhangMin 1--alignIntronMin 20--alignIntronMax 1000000--alignMatesGapMax 1000000”. Expression levels of TEs were quantified as read counts with the “TEcount” command.

Genome annotations

All bioinformatic analyses were performed using the hg38 reference genome and GENCODE v37 gene annotation. Promoters were defined as a 2-kb region centered at the transcription start sites of the APPRIS principal isoform of protein-coding genes. TTS regions were defined as the 2-kb region centered at the 3′ terminals of protein-coding transcripts. CpG island annotations were downloaded from the UCSC table browser. CpG shores were defined as 2-kb regions flanking CpG islands, excluding the regions overlapping with CpG islands. Similarly, CpG shelves were defined as 2-kb regions flanking the stretch of CpG islands and shores while excluding the regions overlapping with CpG islands and shores. Annotations for LINE, SINE, and LTR were extracted from the RepeatMasker track in the UCSC table browser.

Identification of viral sequencing reads in DRIPc-seq

To identify sequencing reads originating from the viral genome, we aligned DRIPc-seq reads to a composite reference genome consisting of the human and HIV1 genome (Genbank accession number: AF324493.2) and computed the proportion of the reads mapped to HIV1 genome.

Code availability

Bioinformatics pipelines and scripts used in this study are accessible from https://github.com/dohlee/hiv1-rloop.

Acknowledgements

We are grateful to Prof. Karlene A. Cimprich (Standford University) for providing the pSH26 and pSH36 plasmids, Prof. A. Engelman (Harvard Medical School) for providing pFlag-IN codon optimized plasmid and Prof. Anna Cereseto (University of Trento) for providing pVpr-IN-EGFP. The NL4-3 ΔEnv EGFP and pNL4-3.Luc.R-E-viral plasmids were obtained through the NIH HIV Reagent Program, Division of AIDS, NIAID, NIH. We thank Dr. Sungchul Kim (IBS center for RNA Research) and Seongjin An (Korea University) for their technical support in recombinant protein purification.

Author contributions

K.P. and K.A. designed experiments. K.P., J.J and S.L. performed experiments. D.L. performed the bioinformatical and statistical analyses. K.P., D.L., K.A. and S.K. analyzed the data. K.P., D.L., and K.A. wrote the manuscript.

Funding

This work was supported by the Institute for Basic Science of the Ministry of Science Grant (IBS-R008-D1) and the National Research Foundation of Korea (NRF) grant funded by the Korea government (NRF-2020R1A2C3011298) (to K. A.) and (NRF-2020R1A5A1018081) (to K.A.). The funders had no role in the study design, data collection, analysis, decision to publish, or preparation of the manuscript.

Competing interests

The authors have declared that no competing interests exist.

Supplemental Information

S1 Fig.

S2 Fig.

S3 Fig.

S4 Fig.

S5 Fig.

S6 Fig.

S1 Table.

S2 Table.

S3 Table.

S4 Table.

S5 Table.

S6 Table.

S7 Table.

S8 Table.

S9 Table.