HIV-1 Vpr induces ciTRAN to prevent transcriptional repression of the provirus

The functional consequences of circular RNA (circRNA) expression on HIV-1 replication are largely unknown. Using a customized protocol involving direct RNA nanopore sequencing, here, we captured circRNAs from HIV-1–infected T cells and identified ciTRAN, a circRNA that modulates HIV-1 transcription. We found that HIV-1 infection induces ciTRAN expression in a Vpr-dependent manner and that ciTRAN interacts with SRSF1, a protein known to repress HIV-1 transcription. Our results suggest that HIV-1 hijacks ciTRAN to exclude serine/arginine-rich splicing factor 1 (SRSF1) from the viral transcriptional complex, thereby promoting efficient viral transcription. In addition, we demonstrate that an SRSF1-inspired mimic can inhibit viral transcription regardless of ciTRAN induction. The hijacking of a host circRNA thus represents a previously unknown facet of primate lentiviruses in overcoming transmission bottlenecks.


Ribosomal RNA (rRNA) depletion
We used the NEBNext rRNA depletion kit v2 (NEB #E7405) to deplete the rRNA, which works based on the RNaseH mediated degradation of rRNA bound to probes.For efficient removal, 50 µg RNA was divided into two parts and used in 20 µL rRNA depletion reactions separately, using the manufacturer's protocol.Initially, the RNA probe master mix was prepared by adding 1 μL of NEBNext rRNA depletion solution and 2μL of probe hybridization buffer to the PCR tube containing 12 μL of the total RNA sample.Further, the reaction was mixed thoroughly by pipetting up and down at least ten times.Next, the reaction was immediately transferred to the thermocycler and set up as follows: 95°C for 2 min, 95-22 °C with a ramp rate of 0.1°C/min, and hold at 22°C for 5 min.The sample was spun down briefly and immediately proceeded for the RNaseH digestion.For RNaseH digestion, five microliters master mix containing 2 μL RNaseH, 2μL RNaseH reaction buffer, and 1 μL of nuclease-free water was added to the previous reaction and after mixing incubated sample at 37°C for 30 min.Post incubation, the sample was spun down and kept on ice.Immediately, the sample proceeded next to DNaseI digestion to degrade the rRNA probes.For DNaseI digestion, 30 μL of master mix containing 5 µL DNaseI reaction buffer, 2.5 μL DNaseI, and 22.5 μL nuclease-free water was added to the RNA from the previous step and incubated at 37°C for 30 min, and immediately proceeded to the next for RNA purification.

RNA purification using magnetic beads
RNA purification was done using the standard NEBNext RNA purification module; briefly, 110 μL RNAClean XP beads were added to the RNA sample in the 1.5 mL DNA LoBind tube (Cat#0030108051, Eppendorf) and mixed thoroughly by pipetting up and down 8-10 times and then incubated on ice for 15 min.Next, the tube was transferred to the magnetic rack and incubated until the solution became clear.Further, the supernatant was removed carefully, and the pellet was washed twice using 80% freshly prepared ethanol for 30 seconds while keeping the tube on the rack.After removing the tube from the stand, RNA was eluted from the pellet using 8 μL nuclease-free water and mixed by pipetting 8-10 times.After that sample was incubated further for 2 min, and the tube was transferred back to the magnetic rack, and the clear supernatant was collected in 0.2 mL RNase-free PCR tubes (Axygen).The RNA was supplemented with Ribolock RNase inhibitor (Thermo Fisher Scientific, EO0381) to preserve its integrity and quantified using Qubit 4.0 using RNA quantification HS kit.rRNA depletion was confirmed by RT-qPCR using rRNA-specific primers.Primer sequences are provided in the Data Table S1.RNA clean and concentrator TM -5 (R1016, Zymo Research) was used for RNA cleanup and concentration purposes.

RT-qPCR analysis and validation for rRNA depletion
Purified RNA was further used for cDNA synthesis using SuperScript® III First-Strand Synthesis System (Cat#18080044, Invitrogen) using manufacturer's protocol in a 20 μL reaction containing RNA, 1x first-strand synthesis buffer, dNTPs, and random hexamer primers, DTT and reverse transcriptase (Thermo Fisher Scientific).The cDNA reaction mixture was then incubated at 25°C for 5 min followed by 60 min at 50°C; the RT enzyme was then inactivated by heating at 70°C for 15 min.After cDNA synthesis, cDNA was diluted 10-fold and used in PCR to confirm rRNA depletion using specific primers (Data Table S1).After the rRNA depletion, RNA was then used for further processing.

Polyadenylation of the linear RNA for depletion
Poly(A) tail was added to remaining linear RNAs after rRNA depletion, including fragmented mRNAs, residual rRNAs, and short 3' overhangs containing snRNAs using E. Coli poly(A) polymerase (Cat#M0276L, NEB).A 20 μL reaction containing RNA, 1x E. coli Poly(A) Polymerase Reaction Buffer, 1mM ATP, Ribolock RNase inhibitor, and 5 U E. Coli poly(A) polymerase was incubated at 37°C for 30 min.Immediately, post-incubation RNA was processed further for purification using an RNA clean and concentrator kit (Zymo Research).

Depletion of poly(A) RNA
Following polyadenylation of the RNA, the sample was processed for poly(A) RNA depletion using NEBNext poly(A) mRNA Magnetic Isolation Module (New England Biolabs).RNA was diluted with nuclease-free water to make a final volume of 50 μL in a 0.2 mL RNase-free PCR tube.In a second PCR tube, 20 μL of NEBNext Magnetic Oligo d(T)25 Beads were aliquoted and washed twice by adding 100 µl of RNA binding buffer and mixed thoroughly by pipetting up and down at least six times.The tube was placed on the magnetic rack until the solution became clear.The supernatant was then discarded, the tube was removed from the magnetic rack, and magnetic beads were resuspended using 50 μL RNA binding buffer and then mixed with 50 μL RNA reaction from tube one.The reaction was mixed thoroughly using a pipette and then incubated on a thermal cycler at 65°C for 5 min and then further held at 4°C for denaturation and binding of poly(A) RNA to the magnetic beads.Once samples reached 4°C, the sample was removed and mixed using a pipette 8-10 times and incubated at room temperature for 5 min.The sample was again mixed and incubated and repeated further for five more minutes.Then the sample was kept on a magnetic stand until it became clear.This time, the supernatant containing the RNA pool of interest was collected carefully and transferred to a fresh 1.5 mL DNA LoBind tube (Eppendorf).The supernatant was subjected to RNA isolation using RNA clean and concentration (Zymo Research).

RNaseR treatment and validation of poly(A) mRNA depletion
Despite efficient removal of rRNA and poly(A) RNAs, including mRNA, possible contamination of small fractions of fragmented RNAs is always possible.To enrich the highly pure fraction of circular RNAs, we treated the sample using RNaseR.RNaseR is quite efficient in degrading the residual linear fraction, provided the sample complexity is low.Therefore, treating samples at this stage using RNaseR will enrich the presence of circular RNAs as it will degrade all the linear fractions efficiently.A 20 μL reaction was set up using RNA which was supplemented with 1x RNaseR buffer and 20U of RNaseR.Ribolock RNase inhibitor (Thermo Fisher Scientific) was also supplied in the reaction to prevent circRNA degradation.The complete reaction was then incubated at 37 °C for 30 min.Post incubation, RNA was isolated from the reaction using RNA clean and concentrator following the manufacturer's (Zymo research) protocol and eluted in 10 μL of nuclease-free water.Isolated RNA was used further for cDNA synthesis using reverse transcriptase (Superscript III) with random hexamer primers.To confirm the depletion of poly(A) mRNA and validation for circular RNAs enrichment, PCR followed by 2% agarose gel electrophoresis was performed using circular RNAs and GAPDH-specific primers.At this step, obtained RNA is highly enriched in circular RNAs and can also be used further for sequencing studies and library preparation.

Circular RNA fragmentation
Further, we used a NEBNext Magnesium RNA Fragmentation Module (New England Biolabs) to linearize the circular RNA in preparation for RNA polyadenylation for direct RNA sequencing using Nanopore.In order to generate fragments of different sizes, RNA was divided into three parts, and RNA fragmentation was done for 30 sec, 1 min, and 1.5 min, respectively.All three reactions were pooled together, and RNA was concentrated in the same way as described previously.Since fragmentation produces ends with 3' phosphate, the 3' phosphate group was removed and added to the 5' ends using T4 polynucleotide Kinase (New England Biolabs) by first incubating for 30 minutes each without/with ATP to remove and add the phosphate group, respectively.

Polyadenylation of Circular RNA fragments
Fragmented circular RNAs were next polyadenylated using E.coli poly(A)polymerase (New England Biolabs) for 30 minutes.Post polyadenylation RNA was concentrated using RNA clean and concentrator TM (Zymo Research).Polyadenylated RNA was further quantified using Qubit 4.0 using RNA HS kit and proceeded for Nanopore library preparation.

RT adaptor ligation and Reverse transcription
For nanopore library preparation, 9µl of poly(A) RNA, 3.0 µl NEBNext Quick Ligation Reaction Buffer, 0.5 µl RNA CS (RCS), 1.0 µl RT adapter, and 1.5 µl T4 DNA ligase were added in a 0.2 ml thin-walled PCR tube.The reaction was mixed thoroughly using pipetting.The reaction mixture was incubated for 10 minutes.After that, a reverse transcription master mix (9.0 µl nuclease-free water, 2.0 µl 10 mM dNTPs, 8.0 µl 5x first-strand buffer, and 4.0 µl 0.1 M DTT) was added to the RT adapter-ligated RNA reaction and mixed by pipetting.Finally, 2µl of Superscript III reverse transcriptase was added to the reaction, and cDNA first-strand synthesis was performed with thermocycler using the following protocol: one cycle at 50 0 C for 50 minutes, one cycle at 70 0 C for 10 min for heat inactivation, and finally cool down to 4 0 C.

RNA cleaning and purification
After first-strand synthesis, the sample was transferred to a clean 1.5 mL DNA LoBind tube (Eppendorf), well-suspended 72 µl of Agencourt RNAClean XP beads were added to each tube and mixed thoroughly by pipetting.After mixing, tubes were incubated at room temperature for 5 minutes in a rotator mixer.Post incubation, tubes were spun down and kept on a magnetic rack and washed with freshly prepared 70% ethanol.Pellet was resuspended in 20 µl nuclease-free water.After 5 minutes of incubation, RNA was collected from the magnetic rack into a fresh DNA LoBind tube.Further, this RNA was used for RNA adapter ligation.

RNA adaptor ligation and RNA Elution
Eluted RNA was mixed with the 8.0 µl of NEBNext Quick Ligation Reaction Buffer, 6.0 µl of RNA adapter (RMX), 3.0 µl of nuclease-free water, and 3.0 µl T4 DNA Ligase.The reaction mixture was mixed using pipetting thoroughly.The reaction mixture was incubated again at room temperature for 15 minutes.Post incubation, 40 µl of well-resuspended RNAClean XP beads were added to the reaction and incubated on a rotor mixer for the next 5 minutes.After incubation, the sample was spun down and pelleted down using a magnetic rack.The pellet was washed twice using wash buffer (WSB), and finally, RNA was eluted in 21 µl elution buffer in a DNA LoBind tubes.

Flow cell priming and sample loading
Flow cell priming solution was made using the standard manufacturer's protocol.Precisely, 30µl of thawed and mixed Flush Tether (FLT) was directly added to the tube of thawed and mixed Flush Buffer (FB), and mixed by pipetting up and down.800 µl of priming mix was added to the priming port and left for 5 minutes.After 5 minutes, 20 µl RNA was mixed with 17.5 µl of nuclease-free water and 37.5 µl of RNA Running Buffer (RRB), and 75 µl sample was loaded on the spot on Nanopore, and run was performed for 48h.

Nanopore Basecalling and Data Analysis
Data was basecalled using the guppy software, and for quality control, we used FastQC to check all the parameters in both control and Infected samples.To check the circRNA in nanopore Direct RNA seq (DRS) we developed a custom script that converts the linear junction reads into the backspliced junction reads.Nanopore library was loaded onto the flow cell (FLO-MIN106) and data was generated.Nanopore raw data was basecalled using Guppy (v 3.2.10),keeping high-accuracy basecalling.The filtered data were mapped to the human genome (hg38) using a custom command-line Blat software package.For capturing the circRNA from the DRS Nanopore reads, we designed a virtual backspliced junction library of all circRNA present in the circBase and circAtlas databases.The circRNA sequences were downloaded from circBase (hg19) and circAtlas(hg38) databases, respectively.Fasta sequences were converted into backsplice fasta sequence using the circDR-seq custom script.Then these backsplice sequence are shortened to 100bp, 50bp upstream, and 50bp downstream to the backsplice junction.This shortening is done mainly to easily find the backsplice read in the Nanopore reads.Further, using pblat, backsplice sequences were aligned to the sequencing reads (generated in this study).pblat is able to map the sequences across the reads, but when it encounters a backsplicing junction, it gets split into two segments that are individually mapped to the different regions of the same gene.Two segments of a read that match upstream and downstream of a splice site signifies a backsplice junction and, therefore, can be taken as a circRNA.To qualify as a circRNA, the Blat score has to reach 60 and should be appropriately aligned 40 bp across the junction (20 upstream and 20 downstream).The circRNAs analysis so far was limited to circbase and circAtlas, which makes it restricted to the annotated circRNAs.We, therefore, expanded our analysis beyond the public database (circBase and circAtlas) which we may have missed.To comprehensively understand the novel circRNA, we prepared a library of all possible exonic combinations of a gene that may lead to backspliced junction.However, this library was limited to genes possessing up to 20 exons.We downloaded all the exonic coordinates from the Ensembl biomart, which then were converted into fasta sequence using bedtools.Next, we joined all possible combinations of exons in a gene and prepared an exonic circRNA library with the help of circDR-seq custom script to identify the backsplice junctions from our DRS Dataset.

Comparison of different pipelines with circDR-seq
Motivated by the fact that we can detect novel backsplice junctions in their native form from DRS Dataset that we generated, we decided to check the robustness of this approach by using a publically available dataset.For this, we circDR-seq (this paper) compared with a most recently developed tool called "CIRI-long" (13) and circNICK-LRS (15) and checked the overlap between the pipelines.

Fig. S1 |
Fig. S1 | Procedures and checks for circRNA enrichment from the infected and mock Jurkat T cells.(A) Flow-cytometry of mock and HIV-1 zsGreen infected E6.1 T cells.Schematics depicting sequential steps for depleting (B) rRNAs by recruiting ASO followed by RNAseH mediated degradation (Refer to supplementary text on circDR-seq for more details), (C) linear RNAs by polyadenylation, oligo(dT) magnetic bead-based removal, and RNaseR treatment to enrich circRNA fraction.(D) Validation of circRNA fragmentation and addition of polyA by oligo(dT) primed cDNA synthesis and PCR in mock(red) and infected(green) samples.

Fig. S2 |
Fig. S2 | Quality checks for Nanopore library and sequencing.(A, B) the representative cumulative yield is shown as a parameter of flow cell performance over time in mock-treated (A) and infected (B) cells.(C, D) Read-length distribution analysis of mock-treated (C) and infected cells (D).(E, F) qualitative analysis of base called sequencing reads over time obtained from mock-treated (E) and Infected cells (F).(G, H) qualitative analysis of read length and read quality of sequencing reads obtained from mock-treated (G) and Infected cells (H).

Fig. S3 |
Fig. S3 | Nanopore data analysis, pipeline validation and identification of previously unannotated circRNAs.(A) Qualitative analysis of various libraries using precision, recall, and F1 score after simulation.(B) Unannotated (UN) and annotated (AN) circRNAs in mock and infection.(C, D) Validation of unannotated circRNAs (from B) for the presence of novel back-spliced junction by Sanger sequencing.(E, F) cataloguing of circRNAs according to the RNA classes (G) circRNAs length distribution obtained herein and its comparison with those reported in circBase.(H) circRNA length-distribution correlation in mock infection with circBase database.(I, J, K) Strand assignment of detected circRNAs (I, J) and its comparison with the circBase database for occurrence (K).

Fig. S4 |
Fig. S4 | Detection of m6A, protein-coding potential of circRNAs, and comparison of circDR-seq with published pipelines.(A, B) Detection of M6A among the reads obtained and M6A presence in the circRNAs.(C, D) m6A profile of randomly selected circRNAs from mock and infection revealed by m6A pulldown followed by RT-qPCR.(E, F) The protein-coding potential of circRNAs assumed by IRES presence.(G) An overlap between M6A modified circRNA with circRNAs having coding potential as estimated using CPC2.(H, I) A comparison of circDRseq with circNICK-LRS (Ref#15) and the percentage overlap between circRNA estimation from the data generated for circNICK-LRS (Ref#15).(J, K), A comparison of circDRseq with CIRI-long (Ref#13) and the percentage overlap between circRNA estimation from the data generated for the CIRI-long (Ref#13).

Fig. S7 |
Fig. S7 | Effects of ciTRAN knockdown and overexpression on viral and host RNA splicing.(A)Effect of ciTRAN overexpression and concomitant SRSF1 sponging on host RNA splicing.(B, C) Effect of ciTRAN overexpression on viral RNA expression (gag, vpr, vpu) and alternative splicing from 2KB multispliced RNA.(D, E, F) Effect of ciTRAN knockdown on HIV-1 unspliced, single spliced and multispliced RNA under ciTRAN or RFP knockdown conditions.

Fig. S9 |
Fig. S9 | Vpr association, primary cell purification, infectivity and SRSF1 and SMARCA5 level during Vpr expression.(A) Flow cytometry of CD3 + /CD4 + cells purified from PBMCs of three donors.Infectivity in PBMCs (B), and primary CD4 + T cells (C) infected using VSVG pseudotyped NLBN zsGreen reporter virus at MOI 5 for 48 h.(D) Effect of Vpr expression on SMARCA5 mRNA and SMARCA5 protein.(E) Effect of Vpr expression on SRSF1 mRNA and SRSF1 protein.(F) details of Vpr and its mutants describing various functions (G) Induction of ciTRAN by indicated Vpr mutants and immunoblotting showing expression of respective Wild-type Vpr and its mutant counterparts.(H) ciTRAN level induced by Vpr across different conditions and CRISPR-mediated knockdown validation.(I) THP-1 infection at different MOIs and ciTRAN induction.

Fig. S10 |
Fig. S10 | Conservation of SRSF-1 binding site across LTRs of different clades and ciTRAN copy number analysis.(A) SRSF1 binding site in the LTRs across different clades and transmitted-founder viruses.(B) LTR-luciferase reporter minigene construct (PGL3 backbone).(C) Effects of infection on SRSF1 expression levels.Actin served as a loading control.(D) copy number estimation of ciTRAN using splint-based ligation in mock, infection, and overexpression conditions.

Table :
reagents and software source table.