Fig 2: Nuclear retention of US HIV-1 RNA in PBMCs from 22 ART-treated PWH.a PBMCs from healthy donors were left untreated or were treated with PHA [1.66 µg/ml] and after 24 h collected for immunoblotting with the use of anti-MATR3, anti-MTR4, anti-αtubulin and anti-GAPDH antibodies. αtubulin and GAPDH are loading controls. b–e Freshly isolated PBMCs from 16 ART-treated PWH were subjected to nucleocytoplasmic fractionation for subsequent nucleic acid isolation and RT-qPCR quantification of (b) US-short HIV-1 RNA, (c) gapdh mRNA, and (d) US-long HIV-1 RNA. Percentages of each of the measured RNAs in the cytoplasmic fraction are shown in (e). f–k Ex vivo cultures of CD8+-depleted PBMCs from 6 ART-treated PWH were mock-treated, treated with SAHA [0.5 µM], disulfiram [0.5 µM], romidepsin [17.5 nM] or PHA [1.66 µg/ml] as a positive control in the presence of ARV [280 nM ritonavir, 180 nM azidothymidine, 200 nM raltegravir, 100 nM efavirenz]. Three days post-treatment cells were subjected to nucleocytoplasmic fractionation for subsequent nucleic acid isolation and RT-qPCR quantification of (f, g) US-short HIV-1 RNA, (h, i) gapdh mRNA, and (j, k) US-long HIV-1 RNA. Percentages of each of the measured RNAs in the cytoplasmic fraction are shown in (g, i, k). b–k Medians are indicated as horizontal lines, and for (e, g, i, k) interquartile ranges are indicated as well. The RNA copy numbers were normalized to the cell numbers measured by the beta-actin qPCR assay. Open symbols in (b, d, f, j) indicate undetectable measurements of US RNA that were assigned the values corresponding to 50% of the corresponding assay detection limits. The detection limits depended on the amounts of the normalizer (input cellular DNA) and, therefore, differed between samples. Open symbols in e, g, k indicate samples where US RNA in the cytoplasmic fraction was undetectable and censored to 50% of the detection limits; because US RNA in the nuclear fraction in these samples was detectable, these circles depict the upper limits of the percentages of US RNA in the cytoplasm. Open crossed circles in e, g, k indicate samples where US RNA was undetectable in the nuclear fraction and censored to 50% of the detection limits; because US RNA in the cytoplasmic fraction in these samples was detectable, these circles depict the lower limits of the percentages of US RNA in the cytoplasm. Percentages of US RNA in the cytoplasmic fraction are not shown if US RNA was undetectable in both nuclear and cytoplasmic fractions. b–d RNA levels were compared between nuclear and cytoplasmic fractions using a two-sided Wilcoxon signed-rank test. **, 0.001 < p < 0.01. ns, not significant. l A schematic view of the post-transcriptional regulation of US HIV-1 RNA by MATR3, MTR4, and Rev. Limiting levels of MATR3 and Rev cause nuclear retention of US HIV-1 RNA that is subjected to MTR4 for further degradation (left panel). During reactivation, MATR3 competes with MTR4 to stabilize the RNA. Rev hijacks the pathway to export the RNA (right panel). Created in BioRender. Wadas, J. (2025) https://BioRender.com/w90s107. Source data are provided as a Source Data file.

Fig 3: Rev determines the MATR3 and MTR4 binding to viral RNA.Primary CD4+ T cells were isolated from blood of healthy donors and infected with WT pNL4.3 HIV particles or Rev-mutated pNL4.3 HIV particles and collected after 3 days and used for (a) DNA extraction to quantify total HIV DNA by qPCR, for (b) total RNA extraction used for RT-qPCR quantification of RRE-containing RNA and Tat HIV-1 RNA, for (c) nuclear-cytoplasmic fractionation protocol followed by quantitative real-time RT-PCR for Tat and (d) RRE HIV-1 RNA, for (e) RNA immunoprecipitation using anti-MATR3, anti-MTR4 and anti-IgG isotype antibodies. RNA was purified from IPs and subjected to RT-qPCR targeting RRE-containing HIV-1 RNA. Values were normalized to input and shown as fold enrichment of IgG control. f–h Jurkat T cells were infected with WT pNL4.3 HIV particles or Rev-mutated pNL4.3 HIV particles and collected after 3 days and used for (f) DNA extraction to quantify total HIV DNA by qPCR, for (g) total RNA extraction used for RT-qPCR quantification of RRE-containing and Tat HIV-1 RNA and (h) RNA immunoprecipitation using anti-MTR4 and anti-IgG isotype antibodies. RNA was purified from IPs and subjected to RT-qPCR targeting RRE-containing HIV-1 RNA. All results are shown as mean values ± SEM from n = 5 (a), n = 7 (b), n = 4 (c, d, f, g), n = 6 (e), n = 3 (h) biological replicates in duplicates. Statistics were performed using a two-sided Wilcoxon signed-rank test for HIV integration (a), a two-way ANOVA test (b–g), and a two-tailed unpaired Student’s t test (h). Significant p values are indicated by the asterisks above the graphs p ≤ 0.05 [*], p ≤ 0.01 [**], ≤0.001 [***], ≤0.0001 [****]. Source data are provided as a Source Data file.

Fig 4: HP1-dependent CTCF and MTR4 at promoters.a Number of MTR4 ChIP-seq peaks in Het32 (blue bars) as in Fig. 3b, colocalizing or not (None) with the other indicated ChIP-seq peaks from Figs. 3b, 4b, compared to the mean number of colocalizations +/- SD of the same list of MTR4 ChIP-seq peaks randomized (200 times shuffle) on their genomic locations (gray bars). Statistical significance of colocalizing peaks was assessed using a rank-based permutation test (n = 200 randomizations), and empirical p values were computed from the null distribution. **, p = 0.005; ns, p > 0.05. b Number of CTCF ChIP-seq peaks versus IgG defined by MACS2 analysis in Het32 (blue) and TKO2 (brown). c Genome view of chromatin-enriched RNA (as in Fig. 2a), CTCF and MTR4 densities at a locus showing two regions bearing TKO-only CTCF peaks (on the left), or peaks common to both cell types (on the right). d Heatmaps of genome-wide CTCF ChIP-seq densities in Het32 and TKO2 centered on all CTCF ChIP-seq peaks. e Heatmaps of ChIP-seq densities of the indicated proteins on 23808 gene bodies defined as TSS to TES intervals, in Het32 (dark blue), in TKO2 (dark red). Source data are provided as a Source Data file.

Fig 5: Protein-protein association between HP1 and the RNA exosome.a STRING analysis of known physical associations of proteins among factors co-precipitating with ZC3H18 proteins, identified by interaction profiling of ZC3H1843. Pink and blue network edges correspond to protein associations based on experiments or curated databases, respectively. b Genotype and subcellular fractionation scheme of the six BMEL clones expressing (Het) or not (Triple Knock-out, TKO) HP1 proteins used in this study. c Protein levels in total extracts, and (d) subcellular distribution in the three fractions, cytosolic (C), nucleosoluble (N), and chromatin (X), depicted in Fig. 1b of indicated exosome subunits and cofactors in Het and TKO cells by Western blot analysis and Ponceau staining of total proteins. e–g Co-immunoprecipitation of HP1γ by ZC3H18, EXOSC10, MTR4, or EXOSC9 in BMEL nuclear extracts. % IP, quantitation of HP1γ IP efficiency versus 100% input. Both input and IP panels in Fig. 1g are split lanes from the same gel and same membrane. Blots are representative of n = 3 (c–e, g) or 4 (f) independent experiments. Source data are provided as a Source Data file.