CPSF6 knock-out primary CD4+ T cells are more permissive to HIV-1 infection

To better understand the role of CPSF6 in HIV-1 infection, we used CRISPR-Cas9 ribonucleoproteins (crRNPs) to knock-out CPSF6 in primary CD4+ T cells from 4 independent donors (Fig 1a). Briefly, bulk CD4+ T cells isolated from the peripheral blood of seronegative human donors were activated and electroporated with crRNPs targeting our genes of interest, resulting in polyclonal pools of knock-out cells. Two independent guide RNA targeting CPSF6 were used in each experiment alongside a non-targeting (NT) negative control guide and two previously validated [3] positive control guides targeting the early-acting host dependency factors CXCR4 and CYPA. Efficient knock-out of CYPA and CPSF6 in each donor was validated by immunoblotting, and quantification of the CPSF6 bands showed substantial reduction in CPSF6 protein levels across all 4 donors (Figs 1b and S1a). CPSF6 depletion persisted in the polyclonal knock-out cell pool over the course of the experiment (S1b Fig). There was no significant difference in percent viable cells between the NT control and any of the knock-outs as measured by amine dye staining and flow cytometry 4 days post-electroporation (Fig 1c). Daily monitoring of viable cell count by flow cytometry and/or by a luminescent ATP detection assay in 3 additional donors likewise revealed only modest differences in cell proliferation or ATP levels over time, with the CPSF6 knock-out cells showing only slight decreases in each metric by 10 days post-electroporation (S1c Fig).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. CPSF6 knock-out in primary CD4+ T cells leads to increased HIV-1 infection rates and dampened innate immune induction.

a, Experimental workflow for CRISPR-Cas9 editing of primary CD4+ T cells and downstream assays. CD4+ T cells were isolated from human blood, activated, and edited via electroporation of crRNPs. Edited cells were activated, expanded, and split into replica plates for knock-out validation, viability stain, and spreading infection assays. For spreading infection assays, cells were infected with HIV-1 NL4-3 nef:IRES:GFP in technical triplicate per condition and harvested for analysis of infection rates by flow cytometry at days 2 and 5 post-infection. b, Immunoblot (left) shows knock-out of CPSF6 and CYPA in primary CD4+ T cell protein lysates harvested at day 4 post-editing in one representative biological replicate (donor C). Chart (right) shows CPSF6 band density divided by β-actin band density in blots from Figs 1b and S1a in 4 biological replicates (donors A-D). Density measurements are normalized to NT controls per donor, horizontal line shows average of biological replicates. c, Knock-out primary CD4+ T cells exhibit similar viability to NT controls at day 4 post-editing as assessed by amine dye stain and flow cytometry. Dots represent cell viability (% Ghost Red negative cells) per condition, horizontal lines represent the average of viability measurements in 4 biological replicates (donors A-D). Statistics were calculated relative to the NT control by two-way ANOVA with Dunnet’s test for multiple comparisons; ns = not significant. d, HIV-1 infectivity (% GFP positive cells normalized to NT control per donor) at days 2 (left) and 5 (right) post-challenge with HIV-1 NL4-3 nef:IRES:GFP in indicated knock-out primary CD4+ T cells from 4 biological replicates (donors A-D) as assessed by flow cytometry. Cells were treated with protease inhibitor (saquinavir) or DMSO control at 24 hours post-challenge as indicated. Each bar represents the average of technical triplicates + /- SD with individual data points shown. Statistics were calculated relative to the NT control per condition by one-way ANOVA with Dunnet’s test for multiple comparisons; * = p ≤ 0.05, ** = p ≤ 0.01, *** = p ≤ 0.001, **** = p ≤ 0.0001. e, Immunoblot showing expression of representative ISGs (IFIT1 and MX1) in protein lysates harvested from HIV-1-infected primary CD4+ T cells at day 5 post-infection in 2 biological replicates (donors A and B).

https://doi.org/10.1371/journal.ppat.1013745.g001

To determine the impact of CPSF6 knock-out on HIV-1 infection, cells were subsequently challenged with cell-free, replication-competent HIV-1 NL4-3 nef:IRES:GFP at 4 days post-electroporation in 6 technical replicates. The following day, 3 replicates were treated with the protease inhibitor Saquinavir (SQV, final concentration 5 µM) to limit infection to a single round of replication, while the other 3 replicates were treated with a matched volume of DMSO as a vehicle control. Percent infected (GFP+) cells were monitored by flow cytometry at 2 and 5 days post-infection and normalized to the NT control (Figs 1d, raw infectivity data S1d). NT cells reached an infection rate of ~1% by day 5 post-infection. As expected, CXCR4 and CYPA knock-out strongly inhibited infection at days 2 and 5, including in the SQV-treated samples, confirming their role in the early phase of the lifecycle. CPSF6 knock-out, on the other hand, resulted in a 2- to 4-fold increase in infection at day 2, and a 3- to 9-fold increase in infection at day 5. We did not observe a change in GFP mean fluorescence intensity, suggesting that the effect is not being driven by a change in the GFP transcript itself. In the presence of SQV, CPSF6 knock-out cells maintained a 1.5- to 3-fold increase in infection at both days 2 and 5, suggesting that this increased permissivity is due, at least in part, to changes in the early stage of the lifecycle.

However, unlike knock-out of the CXCR4 or CYPA dependency factors, which resulted in a consistent drop in infection over a 40-fold range of input virus, the effect of CPSF6 knock-out was dose dependent (S1e Fig). Lower doses of virus resulted in the largest fold changes in infection relative to NT at day 5 while the highest dose challenges resulted in no significant difference. As more virus is added and more cells are infected in the NT condition, the denominator of our normalized fold change calculation increases, which impacts the maximum fold change we can observe. The CPSF6 knock-out infection rate can only increase until it starts to reach the maximum infection rate achievable in the model. Thus, as we increase the NT infection rate, the overall fold change will eventually approach 1 (S1e Fig). This suggests that CPSF6 knock-out lowers a surmountable barrier to infection as opposed to facilitating a rate-limiting step that cannot be overcome by mass action.

Upon induction of the innate immune response, interferon (IFN) signaling results in the expression of a myriad of IFN-stimulated genes (ISGs), many of which have antiviral functions [33]. To determine if CPSF6 knock-out could be altering the innate immune response to infection, we assessed expression of two representative ISGs, IFIT1 and MX1, in primary CD4+ T cells from 2 of the above donors at 5 days post-infection (Fig 1e). IFIT1 and MX1 expression varied by donor in the NT cells. In the donor with higher basal levels of ISGs, CXCR4 knock-out decreased expression and CYPA knock-out increased expression of both ISGs; neither knock-out influenced ISG expression in the donor with lower basal levels. On the contrary, CPSF6 knock-out cells had decreased expression of both ISGs in both donors compared to the NT control. Thus, while there are some donor-to-donor differences, these results demonstrate that CPSF6 knock-out enhances HIV-1 infectivity and downregulates ISG induction in primary CD4+ T cells.

CPSF6 knock-out primary CD4+ T cells exhibit broad transcriptional reprogramming

Given the observed downregulation of ISGs and the known role of CPSF6 in post-transcriptional mRNA processing, we hypothesized that CPSF6 knock-out could be altering innate immune induction and viral infectivity through transcriptional reprogramming. To explore this hypothesis, we performed bulk RNA-sequencing (RNA-Seq) on uninfected CPSF6 knock-out primary CD4+ T cells from 3 independent donors (Fig 2a). CPSF6 knock-out primary CD4+ T cells were generated via CRISPR-Cas9 gene editing as above alongside NT and TRIM5α knock-out controls (2 independent guide RNA per condition). TRIM5α is an HIV-1 restriction factor that inhibits the early stage of the viral lifecycle through binding and premature disassembly of incoming cores [5,6]. It has been previously implicated in innate immune signaling upon core binding and oligomerization [34,35]. After allowing 5 days for protein turnover, RNA was extracted from the edited cells and processed for RNA-Seq. CPSF6 and TRIM5α knock-out were confirmed in each donor by immunoblot (Figs 2b and S2a), and each cell population retained high viability (S2b Fig). These CPSF6 knock-out cells exhibited enhanced permissivity to HIV-1 infection, in line with our previous observations (Fig 1d), while knock-out of TRIM5α had no impact on wild-type infection as we previously reported [3] (Fig 2c).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. CPSF6 knock-out primary CD4+ T cells exhibit broad transcriptional reprogramming.

a, Schematic shows experimental workflow for RNA-Seq. Primary CD4+ T cells were isolated from the blood of 3 independent donors and edited via electroporation of 2 crRNPs targeting CPSF6 and TRIM5α alongside NT controls. RNA was extracted at day 5 post-electroporation and used for bulk RNA-Seq. In parallel, additional cells were used for viability assays, protein knock-out verification, and spreading infection assays. b, Immunoblot shows knock-out of CPSF6 and TRIM5α in primary CD4+ T cell protein lysates harvested at day 5 post-editing in one representative biological replicate (donor E). c, HIV-1 infectivity (% GFP positive cells normalized to NT control per donor) at day 5 post-challenge with HIV-1 NL4-3 nef:IRES:GFP in indicated knock-out primary CD4+ T cells from 3 biological replicates (donors E-G) as assessed by flow cytometry. Each bar represents the average of technical triplicates + /- SD with individual data points shown. Statistics were calculated relative to the NT control per condition by one-way ANOVA with Dunnet’s test for multiple comparisons; ns = not significant, * = p ≤ 0.05, ** = p ≤ 0.01, *** = p ≤ 0.001, **** = p ≤ 0.0001. d, Volcano plot shows differentially expressed genes in RNA-Seq data from uninfected CPSF6 knock-out primary CD4+ T cells as compared to NT controls in 3 biological replicates (donors E-G). Dots show significantly downregulated (blue) and upregulated (red) genes, dotted line shows threshold of 0.05 for adjusted p-value. e, Volcano plot shows differentially expressed genes in RNA-Seq data from uninfected TRIM5α knock-out primary CD4+ T cells as compared to NT controls in 3 biological replicates (donors E-G). Dots show significantly downregulated (blue) and upregulated (red) genes, dotted line shows threshold of 0.05 for adjusted p-value. f, Charts show functional enrichment (Metascape) analysis of top 20 downregulated (left, blue) and upregulated (right, red) pathways in differentially expressed genes from CPSF6 knock-out primary CD4+ T cells as compared to NT controls in 3 biological replicates (donors E-G). g, Histograms show fluorescence intensity of CD4 (first), CXCR4 (second), CCR5 (third), and CD25 (fourth) in primary CD4+ T cells from 1 representative biological replicate (donor I) at day 5 post-editing as measured by immunostaining and flow cytometry. Y-axis is normalized to the mode, histogram visualized using FlowJo version 10.10.

https://doi.org/10.1371/journal.ppat.1013745.g002

Principal component analysis of the RNA-Seq data revealed that while the NT and TRIM5α knock-out specimens clustered tightly together, the CPSF6 knock-out specimens clustered separately (S2c Fig). After normalization to account for donor-to-donor variation, both guides from each donor were treated as replicates (n = 6 replicates per condition). Differential gene expression analysis revealed broad transcriptional reprogramming of the CPSF6 knock-out cells as compared to NT controls with 4,694 upregulated and 3,960 downregulated genes (Figs 2d and S2d). In contrast, the transcriptional profile of the TRIM5α knock-out cells remained nearly unchanged, with only 51 upregulated and 24 downregulated genes (Figs 2e and S2d).

Functional enrichment analysis revealed that several genes involved in cytokine signaling and T cell activation were upregulated in the CPSF6 knock-out cells (Fig 2f), including CD25 and CXCR4. To determine if these transcriptional changes were reflected in changes to cell surface expression, we repeated our knock-outs of CPSF6 and TRIM5α using 2 independent guides in 3 additional donors alongside NT, CXCR4-targeting, and CYPA-targeting controls. At 5 days post-editing, we performed cell surface immunostaining for the HIV-1 receptor CD4, co-receptors CXCR4 and CCR5, and activation marker CD25, quantifying mean fluorescence intensity (MFI) by flow cytometry. While TRIM5α knock-out had no significant impact on surface receptor expression, CPSF6 knock-out cells displayed a slight, but significant increase in the cell surface expression of CD4 and larger increases in cell surface expression of CXCR4 and CD25 (Figs 2g and S2h). CPSF6 knock-out did not have a significant impact on cell surface expression of CCR5 (Figs 2g and S2h). Thus, increased co-receptor expression and enhanced T cell activation may contribute to the enhanced HIV-1 replication observed in the CPSF6 knock-out primary CD4+ T cells.

Several pathways involved in the innate immune response were simultaneously observed to be downregulated in CPSF6 knock-out cells, including response to virus and interferon signaling (Fig 2f). Key components of the type I and type II interferon signaling pathway were identified as being downregulated in each donor, including IFNAR2, STAT3, IRF9, and JAK2, while negative regulators of the pathway were upregulated, including SOCS1 and SOCS3 (S2e Fig and S2f). Several ISGs and known antiretroviral restriction factors were also downregulated, including SAMHD1 and TRIM5α (S2e and S2f Fig). Indeed, decreased TRIM5α protein levels were observed in all 3 donors upon CPSF6 knock-out in the knock-out validation immunoblots (Figs 2b and S2a, quantification in S2g). These results suggest that CPSF6 knock-out leads to broad transcriptional reprogramming in primary CD4+ T cells that results in upregulation of the viral receptors and downregulation of the innate immune response, which ultimately yields increased permissivity to HIV-1 infection.

HIV-1 infection phenotypes in CPSF6 knock-out cells vary by viral strain

Given the increase in CXCR4 cell surface expression in the CPSF6 knock-out cells, we next wanted to assess the impact of tropism on the observed infection phenotypes. To test this, we again repeated our knock-outs of CPSF6 using two independent guides in 3 independent donors alongside NT, CXCR4-targeting, and CYPA-targeting controls (validated by immunoblot, Fig 3a). Cells were subsequently challenged in technical triplicate at 4 days post-electroporation with p24-normalized amounts of either replication-competent HIV-1 NL4-3 nef:IRES:GFP (as used before) or VSV-G pseudotyped, replication-incompetent HIV-1 NL4-3 dEnv nef:IRES:GFP. Percent infected (GFP+) cells were monitored by flow cytometry at 2 days post-infection and normalized to the NT control (Fig 3b). While the replication-competent virus again showed a 2- to 3-fold increase in replication in CPSF6 knock-out cells, there was no significant difference in infection with the VSV-G pseudotyped virus (as previously reported [21]). This suggests that enhanced co-receptor expression is at least in part responsible for the increased permissivity.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. HIV-1 infection phenotypes in CPSF6 knock-out cells vary by viral strain.

a, Immunoblots shows knock-out of CPSF6 and CYPA in primary CD4+ T cell protein lysates harvested at day 4 post-editing in 3 biological replicates (donors K-M). b, HIV-1 infectivity (% GFP positive cells normalized to NT control per condition) at day 2 post-challenge with HIV-1 NL4-3 nef:IRES:GFP (top) or VSV-G pseudotyped HIV-1 NL4-3 nef:IRES:GFP (bottom) in indicated knock-out primary CD4+ T cells from 3 biological replicates (donors K-M) as assessed by flow cytometry. Cells were treated infected with virus across a range of concentrations standardized by p24 content. Each bar represents the average of technical triplicates + /- SD with individual data points shown. Statistics were calculated relative to the NT control per condition by one-way ANOVA with Dunnet’s test for multiple comparisons; * = p ≤ 0.05, ** = p ≤ 0.01, *** = p ≤ 0.001, **** = p ≤ 0.0001. c, HIV-1 infectivity (% GFP positive cells normalized to NT control per condition) at day 5 post-challenge with HIV-1 JR-FL iGFP in indicated knock-out primary CD4+ T cells from 1 biological replicate (donor N) as assessed by flow cytometry. Each bar represents the average of technical triplicates + /- SD with individual data points shown. Statistics were calculated relative to the NT control by one-way ANOVA with Dunnet’s test for multiple comparisons; *** = p ≤ 0.001, **** = p ≤ 0.0001. d, HIV-1 infectivity (% GFP or FITC positive cells normalized to NT control per condition) at day 5 post-challenge with HIV-1 NL4-3 nef:IRES:GFP or CH040 in indicated knock-out primary CD4+ T cells from 1 biological replicate (donor O) as assessed by flow cytometry. Each bar represents the average of technical triplicates + /- SD with individual data points shown. Statistics were calculated relative to the NT control per condition by one-way ANOVA with Dunnet’s test for multiple comparisons; ns = not significant.

https://doi.org/10.1371/journal.ppat.1013745.g003

To determine whether the HIV-1 infection phenotype in CPSF6 knock-out cells is strain specific, we repeated our virus challenge experiments with the lab-adapted, CCR5-tropic HIV-1 JR-FL iGFP strain (Fig 3c) and the CCR5-tropic transmitted founder HIV-1 CH040 strain (Fig 3d). As before, we observed increased infection with our CXCR4-tropic NL4–3 strain at day 5 post-infection in the CPSF6 knock-out cells (Fig 3d). However, we observed no significant impact of CPSF6 knock-out on infection with the CCR5-tropic JR-FL strain. Surprisingly, we observed a strong decrease in infection with the CCR5-tropic transmitted founder strain upon CPSF6 knock-out. These results indicate that the CPSF6 knock-out phenotype in primary CD4+ T cells is strain-dependent and that the increased infection observed with the CXCR4-tropic HIV-1 NL4-3 strain is at least in part due to the increased cell surface expression of CXCR4.

The interferon response is dampened in CPSF6 knock-out primary CD4+ T cells

Besides enhanced co-receptor expression, CPSF6 knock-out cells also demonstrated broad transcriptional downregulation of the interferon signaling pathway. To test if this downregulation was functionally related to the observed increase in HIV-1 permissivity, primary CD4+ T cells from three independent donors were electroporated with 2 independent crRNPs targeting CPSF6 alongside NT, CXCR4-targeting, and CYPA-targeting controls as well as a multiplexed pool of 5 crRNPs targeting the type I interferon receptor gene (IFNAR1). Cells were then pre-treated with media alone or 10 U/mL IFNα, IFNβ, Universal type I IFN, or IFNγ for 16 hours prior to challenge with replication-competent HIV-1 NL4-3 nef:IRES:GFP. At 5 days post-challenge, in the media only condition, CXCR4 and CYPA knock-out decreased infection while CPSF6 knock-out increased infection as expected (Figs 4a and S3a). Knock-out of IFNAR1 resulted in a roughly two-fold increase in infection over the NT control compared to the 4- to 5-fold increase observed in the CPSF6 knock-out cells. IFNα, IFNβ, and Universal type I IFN treatment decreased infection in the NT cells while IFNγ had no significant impact. The IFNAR1 knock-out cells were not impacted by IFN treatment as expected. Notably, CPSF6 knock-out cells showed a dampened response to IFNα, IFNβ, and Universal type I IFN treatment with a partial rescue of HIV-1 replication in these cells. Protein lysates collected after interferon pre-treatment, but prior to infection, showed decreased levels of representative ISGs IFIT1 and MX1 upon CPSF6 knock-out, though not to the same extent as observed upon IFNAR1 knock-out (Figs 3b and S3c).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. CPSF6 regulates the interferon pathway and expression of HIV-1 restriction factor TRIM5α in primary CD4+ T cells.

a, HIV-1 infectivity (Raw % GFP positive cells normalized to NT control per condition) at day 5 post-challenge with HIV-1 NL4-3 nef:IRES:GFP in indicated knock-out primary CD4+ T cells from 2 biological replicates (donors P and Q) as assessed by flow cytometry. Cells were pre-treated with 100 U/mL IFNα, IFNβ, IFNγ, Universal Type I IFN, or media only control. Each bar represents the average of technical triplicates + /- SD with individual data points shown. Statistics were calculated relative to the NT control per condition by one-way ANOVA with Dunnet’s test for multiple comparisons; * = p ≤ 0.05, ** = p ≤ 0.01, *** = p ≤ 0.001, **** = p ≤ 0.0001. b, Immunoblot (left) shows expression of representative ISGs (IFIT1 and MX1) in knock-out primary CD4+ T cells treated with 10 U/mL IFNα, IFNβ, Universal Type I IFN, IFNγ, or media-only control for 16 hours in 1 representative biological replicate (donor P). Chart (right) shows IFIT1 band density divided by β-actin band density in adjacent blot. Density measurements are normalized to NT controls per condition. c, Diagram depicts methods of altering TRIM5α-mediated restriction in CD4+ T cells. Left: Wild-type HIV-1 NL4-3 capsid binds CYPA, which shields from TRIM5α restriction. Middle: The P90A capsid mutant HIV-1 derived from NL4-3 nef:IRES:GFP is deficient in CYPA binding, leading to TRIM5α-mediated restriction. Right: Cyclosporin A (CsA) inhibits CYPA activity, leading to TRIM5α-mediated restriction. d, HIV-1 infectivity (% GFP positive cells normalized to NT control per donor) at day 5 post-challenge with HIV-1 NL4-3 nef:IRES:GFP P90A capsid mutant in indicated knock-out primary CD4+ T cells from 5 biological replicates (donors R-V) as assessed by flow cytometry. Each bar represents the average of technical triplicates + /- SD with individual data points shown. Statistics were calculated relative to the NT control per condition by one-way ANOVA with Dunnet’s test for multiple comparisons; * = p ≤ 0.05, ** = p ≤ 0.01, *** = p ≤ 0.001, **** = p ≤ 0.0001. e, HIV-1 infectivity (% GFP positive cells normalized to NT control per condition) at day 5 post-challenge with HIV-1 NL4-3 nef:IRES:GFP in indicated knock-out primary CD4+ T cells from 3 biological replicates (donors W-Y) as assessed by flow cytometry. Cells were pre-treated with cyclosporin A or a DMSO control 4 hours prior to infection. Each bar represents the average of technical triplicates + /- SD with individual data points shown. Statistics were calculated relative to the NT control per condition by one-way ANOVA with Dunnet’s test for multiple comparisons; * = p ≤ 0.05, ** = p ≤ 0.01, *** = p ≤ 0.001, **** = p ≤ 0.0001.

https://doi.org/10.1371/journal.ppat.1013745.g004

To determine if the dampened IFN sensitivity in CPSF6 knock-out cells depended on the cell entry pathway, we repeated the IFN treatment experiment using VSV-G pseudotyped, replication-incompetent HIV-1 NL4-3 dEnv nef:IRES:GFP alongside replication-competent HIV-1 NL4-3 nef:IRES:GFP and a wider range of doses of IFNα and IFNβ (1-1000 U/mL). Replication-competent HIV-1 NL4-3 showed enhanced infection in CPSF6 knock-out cells compared to the NT control across all tested IFNα doses and the majority of IFNβ doses (S3b Fig). In contrast, VSV-G pseudotyped HIV-1 infection rates in CPSF6 knock-out cells were similar to infection rates in NT control cells across all tested conditions (S3b Fig). This reaffirms our prior data that the increase in CXCR4 expression is a major driver of the observed increase in permissivity (Figs 2g and 3b), though it is also possible that this result stems from differential sensitivity of VSV-G pseudotyped viruses to inhibition by IFN. Taken together, these data suggest that CPSF6 knock-out cells have a reduced responsiveness to IFN, which may in part explain the enhanced permissivity of CPSF6 knock-out cells to HIV-1 infection.

Our RNA-Seq data also showed downregulation of a number of innate immune effectors and known HIV-1 restriction factors upon CPSF6 knock-out, including TRIM5α, which binds to and restricts incoming cores. To protect itself from TRIM5α-mediated restriction, incoming cores bind CYPA, which masks the TRIM5α binding site [2,3]. Disruption of CYPA binding either by CYPA knock-out (as demonstrated previously [2,3] and shown above), mutation of the capsid (i.e., through a capsid P90A mutation), or through treatment with the CYPA inhibitor cyclosporin A (CsA) decreases HIV-1 infection due to TRIM5α-mediated restriction (Fig 4c). To determine if CPSF6 knock-out cells were functionally defective in TRIM5α, we electroporated primary CD4+ T cells from 5 independent donors with 2 independent crRNPs targeting CYPA, CPSF6, or TRIM5α alongside a NT control (immunoblot validation in S3d Fig). Cells were then challenged with HIV-1 NL4-3 nef:IRES:GFP containing a P90A capsid mutant in technical triplicate. Across all donors, CYPA knock-out had no significant effect on P90A virus infection, consistent with the inability of P90A capsids to recruit CYPA. Knock-out of TRIM5α led to a 10- to 30-fold increase in infection of the P90A virus, roughly correlating with knock-out efficiency as assessed by immunoblot (Figs 4d, raw infectivity data in S3e, immunoblot in S3d), consistent with prior reports [3]. Knock-out of CPSF6 led to a decrease in TRIM5α steady-state levels across all donors (S3d Fig) and rescued P90A virus replication to similar levels as TRIM5α knock-out itself (Fig 4d).

To further confirm that CPSF6 knock-out leads to restoration of infectivity under conditions that enhance TRIM5α-mediated restriction, we knocked out CPSF6 and TRIM5α using 2 independent guides in 3 additional donors alongside NT, CXCR4-targeting, and CYPA-targeting controls (immunoblot validation in S3f Fig). Each population was challenged with HIV-1 NL4-3 nef:IRES:GFP in the presence and absence of CsA in technical triplicate. In the absence of CsA, CYPA knock-out decreased infection, CPSF6 knock-out increased infection, and TRIM5α knock-out had no effect (Fig 4e). CsA strongly inhibited replication in the NT cells, though this was rescued in the TRIM5α knock-out cells with the degree of rescue correlating with knock-out efficiency (infection data in Figs 4e, knock-out efficiency in S3f). CPSF6 knock-out partially rescued replication in the presence of CsA, consistent with the P90A capsid mutant data. These data suggest that the broad transcriptional reprogramming upon CPSF6 knock-out leads to enhanced co-receptor expression, decreased interferon signaling, and reduced restriction factor expression, all of which partially contribute to enhanced HIV-1 replication in these cells.

CPSF6 knock-out in primary CD4+ T cells leads to changes in alternative polyadenylation

Given that CPSF6 regulates polyadenylation site selection through its role in the CFIm complex, we hypothesized that the broad transcriptional changes in primary CD4+ T cells upon CPSF6 knock-out could be driven by changes in APA. To test this hypothesis, we re-analyzed our RNA-Seq data from uninfected CPSF6 knock-out primary CD4+ T cells using Regression of PolyAdenylation Compositions (REPAC) analysis [36]. Briefly, this algorithm compares RNA-Seq reads on either side of known, annotated poly(A) sites and calculates the compositional Fold Change (cFC) for each gene as a measure of 3’ UTR length changes between an experimental (CPSF6 knock-out) and control (NT) condition. Shortening of 3’ UTRs is indicated by a cFC value < -0.25, and lengthening by a value > 0.25. REPAC analysis revealed that CPSF6 knock-out primary CD4+ T cells exhibited an overall bias towards shortening of 3’ UTR lengths (n = 2,536 shortened, n = 198 lengthened 3’ UTRs) as compared to NT controls, consistent with what has been observed in other cell types [16] (Fig 5a). These changes in 3’ UTR lengths can be readily observed in the transcript read coverage tracks, as shown at the transforming growth factor-beta receptor type 1 (TGFBR1) locus, for example (Fig 5b). Functional enrichment analysis of genes with significantly shortened 3’ UTRs identified broad changes in APA at genes associated with immune response signaling, suggesting that shortening of these transcripts may be associated with their downregulation (Fig 5c). To validate this result, we plotted the cFC values for all genes previously identified in the functional enrichment analysis associated with the term ‘Interferon Signaling’ (Fig 2f) and found these genes showed clear evidence of 3’ UTR shortening (Fig 5d).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. CPSF6 knock-out leads to changes in alternative polyadenylation in primary CD4+ T cells a, Volcano plot shows REPAC analysis of changes in 3’ UTR lengths for CPSF6 knock-out primary CD4 + T cells as compared to NT controls in 3 biological replicates (donors E-G, RNA-Seq data from Fig 2).

3’ UTRs are designated shortened (blue) for Compositional Fold Change (cFC) < -0.25 and -log₁₀(adjusted p-value) < 0.05 and lengthened (red) for cFC  > 0.25 and -log₁₀(adjusted p-value) < 0.05. Dots represent individual altered 3’ UTRs. b, Tracks show read coverage of TGFBR1 in RNA-Seq data from CPSF6 knock-out and NT control cells from one representative biological replicate (donor E), visualized using Integrative Genomics Visualizer. 3’ UTRs show evidence of shortening in response to CPSF6 knock-out. c, Charts show functional enrichment (Metascape) analysis of top pathways enriched in genes with shortened 3’ UTRs (left, blue) and lengthened 3’ UTRs (right, red) in response to CSPF6 knock-out in primary CD4+ T cells as assessed by REPAC analysis in 3 biological replicates (donors E-G). d, Violin plot shows comparison of cFC values from REPAC analysis in all tested genes as compared to genes in the indicated pathways (GO:0042110:T Cell activation, R-HSA-1280215: Cytokine Signaling in Immune system, R-HSA-913531: Interferon Signaling) in 3 biological replicates (donors E-G). e, Scatterplot shows cFC of all tested APA sites in REPAC analysis of CPSF6 knock-out primary CD4+ T cells in as compared to NT controls plotted against differential expression data as shown in Fig 2 in data aggregated from 3 biological replicates (donors E-G). 3’ UTRs are designated shortened (blue) for cFC < -0.25 and -log₁₀(adjusted p-value) < 0.05 and lengthened (red) for cFC > 0.25 and -log₁₀(adjusted p-value) < 0.05. Dots represent individual altered 3’ UTRs. f, Immunoblot shows knock-out of CPSF5 in protein lysates harvested at day 4 post-editing in 3 biological replicates (donor Z-AB). g, HIV-1 infectivity (% GFP positive cells normalized to NT control per donor) at day 5 post-challenge with HIV-1 NL4-3 nef:IRES:GFP in indicated knock-out primary CD4+ T cells from 3 biological replicates (donors Z-AB) as assessed by flow cytometry. Each bar represents the average of technical triplicates + /- SD with individual data points shown. Statistics were calculated relative to the NT control per condition by one-way ANOVA with Dunnet’s test for multiple comparisons; * = p ≤ 0.05, ** = p ≤ 0.01, *** = p ≤ 0.001, **** = p ≤ 0.0001.

https://doi.org/10.1371/journal.ppat.1013745.g005

A decrease in 3’ UTR length was not universally associated with decreased transcript abundance. Indeed, similar decreases in cFCs were observed for transcripts associated with the ‘Cytokine Signaling in Innate Immune Signaling’ and ‘T Cell Activation’ pathways, both of which were enriched pathways among the upregulated genes (Figs 5c and 2f). To assess the relationship between changes in 3’ UTR length and transcript abundance more globally, we plotted cFC values from all analyzed poly(A) sites against the associated log₂(fold change) values from the differential gene expression analysis in Fig 2. While there was a bias towards genes with shortened 3’ UTRs exhibiting downregulation, changes in 3’ UTR length were associated with variable changes in transcript abundance, which suggests multiple modes of gene-dependent regulation (Fig 5e). Additionally, several genes with significant changes in abundance did not exhibit a significant change in 3’ UTR length, suggesting there are additional and/or secondary effects of CPSF6 knock-out on the transcriptome by this timepoint (Fig 5e). CXCR4 and CD25, for example, were upregulated and had increased cell surface expression (Fig 2g), but only CD25 showed a significant decrease in 3’ UTR length (S4a Fig).

Several different algorithms have been developed to assess poly(A) site usage, all of which use slightly different approaches. The REPAC algorithm uses only well-annotated poly(A) sites, but as a result does not give information on every locus. Therefore, to validate these results across the transcriptome, we leveraged an alternative method of APA analysis, Dynamic analysis of Alternative PolyAdenylation (DaPars) [37], that infers poly(A) site usage and calculates changes in the distal polyadenylation usage index (ΔPDUI) as a proxy for changes in 3’ UTR length for each gene (S4b Fig). As with the REPAC analysis, ΔPDUI analysis showed that CPSF6 knock-out led to global 3’ UTR shortening, with genes involved in the interferon pathway, cytokine signaling, and cellular activation showing preferential 3’ UTR shortening (S4c Fig).

These data suggest that changes in APA are induced by CPSF6 knock-out and are at least partially responsible for the broad transcriptional reprogramming and increased permissivity to HIV-1 observed in these cells. If this is true, we would hypothesize that knock-out of the other CFIm complex member, CPSF5, would result in a similar phenotype, as CPSF5 knock-out has also previously been reported to trigger changes in APA [38]. To test this hypothesis, we knocked out CPSF5 in primary CD4+ T cells using 2 independent guide RNA in 3 independent donors and conducted HIV-1 spreading infection assays as before alongside NT, CXCR4 knock-out, and CYPA knock-out controls. CPSF5 knock-out had minimal impact on cell viability (S4d Fig). While CPSF5 knock-out efficiency wasn’t as high as that achieved for CPSF6 as monitored by immunoblot (Fig 5f), it still resulted in a significant increase in HIV-1 infection (Fig 5g). Taken together, these data suggest that the induction of APA through CFIm complex disruption results in broad transcriptional reprogramming and enhanced CD4+ T cell permissivity to HIV-1 infection.

HIV-1 capsid binding to CPSF6 triggers transcriptional rewiring

Previously described roles of CPSF6 in HIV-1 infection, including its roles in nuclear import and integration site selection, are dependent on CPSF6 binding to HIV-1 cores. Several studies have shown that this interaction can also lead to relocalization of CPSF6 to nuclear speckles, with the extent of relocalization dependent on the cell type and timepoint [39–42]. This is dependent on the interaction with the core as infection with the N74D capsid mutant, which does not bind CPSF6, does not lead to CPSF6 relocalization to nuclear speckles [42]. It is unclear, however, if the interaction of CPSF6 with HIV-1 capsid and its subsequent relocalization to nuclear speckles also disrupts CFIm function and triggers changes in APA.

To test this, we first used the HT1080 cell line model as recruitment of CPSF6 to nuclear speckles had previously been reported following HIV-1 infection in these cells [40]. To more readily track changes in CPSF6 localization, we engineered an endogenous monomeric NeonGreen (mNGreen) C-terminal tag into the CPSF6 locus in HT1080 cells via CRISPR-Cas9 knock-in (HT1080-CPSF6-mNGreen). We challenged these HT1080-CPSF6-mNGreen cells with VSV-G pseudotyped HIV-1 NL4-3 dEnv encoding either a wild-type (WT) capsid or a capsid mutant that is deficient in CPSF6 binding (N74D) at a multiplicity of infection (MOI) of 10 in technical triplicate. At 6 hours post-challenge, we performed immunofluorescent staining for a marker of nuclear speckles, SC35, in tandem with harvesting RNA and protein for analysis by bulk RNA-Seq and immunoblotting, respectively. Widefield deconvolution fluorescent microscopy imaging confirmed that infection with WT HIV-1 led to the formation of CPSF6 puncta in cell nuclei, which were not observed in the mock or N74D infected cells (S5a Fig). The CPSF6 puncta in the WT HIV-1 infected cells were largely colocalized with SC35-containing nuclear speckles, consistent with prior reports (S5a and quantification in S5b Fig). Immunoblotting of protein lysates harvested from the infected cells demonstrated that CPSF6 protein levels were unchanged between the mock, WT, and N74D infection conditions (S5c Fig).

We next sought to profile the transcriptional changes in our WT or N74D infected HT1080 cells relative to our mock infected controls at 6 hours post-challenge. Challenge with WT HIV-1 resulted in 3,770 differentially expressed genes (Fig 6a), whereas challenge with the N74D virus resulted in slightly fewer changes with 3,450 differentially expressed genes (Fig 6b). While a majority of differentially expressed genes are shared in both infection conditions, 1,362 differentially expressed genes were found only in the WT, but not N74D, infected condition (Fig 6c). Functional enrichment analysis of this set of genes revealed upregulation of several pathways involved in organelle organization and membrane trafficking and downregulation of several pathways involved in RNA metabolism and actin filament regulation (Fig 6d).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 6. HIV-1 infection induces APA in HT1080 cells a, Volcano plot shows differentially expressed genes in RNA-Seq data from HT1080-CPSF6-mNGreen cells infected with VSV-G pseudotyped WT HIV-1-NL4-3 at an MOI of 10 at 6 hours post-infection as compared to mock-infected controls in 3 technical replicates.

Dots show significantly downregulated (blue) and upregulated (red) genes, dotted line shows threshold of 0.05 for adjusted p-value. b, Volcano plot shows differentially expressed genes in RNA-Seq data from HT1080-CPSF6-mNGreen cells infected with VSV-G pseudotyped N74D capsid mutant HIV-1-NL4-3 at an MOI of 10 at 6 hours post-infection as compared to mock-infected controls in 3 technical replicates. Dots show significantly downregulated (blue) and upregulated (red) genes, dotted line shows threshold of 0.05 for adjusted p-value. c, Venn diagrams shows overlap of upregulated (top) and downregulated (bottom) differentially expressed genes between WT infected and N74D infected HT1080-CPSF6-mNGreen cells in 3 technical replicates. d, Charts show functional enrichment (Metascape) analysis of top 15 upregulated (top, red) and downregulated (bottom, blue) pathways in differentially expressed genes from WT infected HT1080-CPSF6-mNGreen cells that are not differentially expressed in N74D infected HT1080-CPSF6-mNGreen cells at 6 hours post-infection in 3 technical replicates. e, Volcano plot shows REPAC analysis of changes in 3’ UTR lengths for WT VSV-G-pseudotyped HIV-1-NL4-3-infected versus mock-infected HT1080-CPSF6-mNGreen cells from 3 technical replicates. Cells were infected at an MOI of 10 and RNA was extracted at 6 hours post-infection. 3’ UTRs are designated shortened (blue) for Compositional Fold Change (cFC) < -0.25 and -log10(adjusted p-value) < 0.05 and lengthened (red) for cFC > 0.25 and -log10(adjusted p-value) < 0.05. Dots represent individual altered 3’ UTRs. f, Volcano plot shows REPAC analysis of changes in 3’ UTR lengths for N74D VSV-G-pseudotyped HIV-1-NL4-3-infected versus mock-infected HT1080-CPSF6-mNGreen cells from 3 technical replicates. Cells were infected at an MOI of 10 and RNA was extracted 6 hours post-infection. 3’ UTRs are designated shortened (blue) for cFC < -0.25 and -log10(adjusted p-value) < 0.05 and lengthened (red) for cFC > 0.25 and -log10(adjusted p-value) < 0.05. Dots represent individual altered 3’ UTRs. g, Volcano plot shows REPAC analysis of changes in 3’ UTR lengths for WT HIV-1 NL4-3 nef:IRES:GFP infected versus uninfected primary CD4+ T cells from 3 biological replicates (donors AC-AE). 3’ UTRs are designated shortened (blue) for cFC < -0.25 and -log₁₀(adjusted p-value) < 0.05 and lengthened (red) for cFC > 0.25 and -log₁₀(adjusted p-value) < 0.05. Dots represent individual altered 3’ UTRs. h, Volcano plot shows REPAC analysis of changes in 3’ UTR lengths for N74D capsid mutant HIV-1 NL4-3 nef:IRES:GFP infected versus uninfected primary CD4+ T cells from 3 biological replicates (donors AC-AE). 3’ UTRs are designated shortened (blue) forcFC < -0.25 and -log₁₀(adjusted p-value) < 0.05 and lengthened (red) for cFC > 0.25 and -log₁₀(adjusted p-value) < 0.05. Dots represent individual altered 3’ UTRs.

https://doi.org/10.1371/journal.ppat.1013745.g006

To assess the impact of infection on APA, we next performed REPAC analysis on the RNA-Seq data from the HT1080 cells. The HT1080 cells challenged with WT HIV-1 demonstrated a bias towards an overall shortening of 3’ UTR lengths as compared to mock-infected controls (333 shortened, 71 lengthened; Fig 6e). On the other hand, HT1080 cells challenged with the N74D capsid mutant HIV-1 virus did not show a directional bias in 3’ UTR lengths as compared to mock-infected controls (108 shortened, 84 lengthened; Fig 6f). These results suggest that binding of CPSF6 to HIV-1 capsid cores and/or its subsequent relocalization is sufficient to partially perturb CFIm function and result in changes to APA.

To determine whether HIV-1 infection induces changes in APA and transcriptional reprogramming through recruitment of CPSF6 to capsid cores in a more physiologically relevant cell type, we next challenged primary CD4+ T cells from 3 independent donors with HIV-1 NL4-3 nef:IRES:GFP encoding either a WT or N74D mutant capsid. The N74D capsid mutant is known to have a severe replication defect in primary CD4+ T cells [22,23]. To account for this, we performed live cell sorting on infected (GFP+) cells at day 2 post-challenge to obtain pure populations of infected cells, thus mitigating differences in infection rates between the two viruses. RNA was then extracted from the sorted cell populations alongside uninfected controls and subjected to RNA-Seq (S6a Fig). In parallel, we generated CPSF6 knock-out and NT control primary CD4+ T cells and extracted RNA at 2 days post-electroporation to directly compare the transcriptional profiles between the infected cells at day 2 post-infection and CPSF6 knock-out cells at day 2 post-editing (immunoblot validation of knock-out efficiency in S6b Fig).

As reported in Fig 2, CPSF6 knock-out resulted in broad transcriptional rewiring of primary CD4+ T cells with 3,579 differentially expressed genes (S6c Fig), including upregulation of genes involved in cytokine signaling and downregulation of those involved in the innate immune response (S6d Fig). As observed in the HT1080 cells, infection with WT virus resulted in broad transcriptional changes with 4,537 differentially expressed genes (S6e Fig) while infection with the N74D virus resulted in slightly fewer changes with 3,787 differentially expressed genes (S6f Fig). Comparing between conditions, the WT infected and CPSF6 knock-out cells shared 757 misregulated genes, while the N74D infected and CPSF6 knock-out cells shared 640 misregulated genes (S6g Fig). While the two infected conditions were more similar, the cells with WT infected virus had 1,650 differentially expressed genes not shared by the cells infected with N74D mutant virus. Functional enrichment analysis of these genes showed upregulation of RNA metabolism and cell cycle pathways and downregulation of membrane trafficking and organelle localization pathways, contrary to what was observed in the infected HT1080 cells (S6h Fig).

To assess if these changes were associated with changes in APA, we next performed REPAC analysis on the RNA-Seq data from the CPSF6 knock-out and infected primary cells. The CPSF6 knock-out cells, as before, showed a strong bias towards overall shortening of 3’ UTR lengths with 928 shortened and only 50 lengthened 3’ UTRs (S6c Fig). The primary T cells infected with WT virus showed a similar, but much smaller, bias towards overall 3’ UTR shortening as compared to uninfected controls (133 shortened, 60 lengthened; Fig 6g, individual gene names provided in S6i Fig). On the contrary, cells infected with the N74D virus did not show a directional bias in 3’ UTR lengths as compared to uninfected controls (64 shortened, 60 lengthened; Fig 6h). These data were independently confirmed using the ΔPDUI algorithm, which likewise confirmed a bias towards 3’ UTR shortening in the cells infected with the WT virus (S6j Fig), but not the N74D virus (S6k Fig). Similar to the CPSF6 knock-out data, a majority of gene expression changes cannot be directly linked to a change in 3’ UTR length at this timepoint, though this doesn’t rule out secondary or indirect effects of APA.

If transcriptional rewiring triggered by CPSF6 recruitment to capsid cores enhances CD4 + T cell permissivity to infection, we would expect that mutations that prevent CPSF6 binding would decrease infectivity, as has been reported with the N74D capsid mutant virus [22,23], and that knock-out of CPSF6 would help rescue infection of those viruses. To test this, we electroporated primary CD4+ T cells from 5 independent donors with 2 independent crRNPs targeting CPSF6 and TRIM5α alongside a NT and CYPA-targeting control (immunoblot validation in S3c Fig). Cells were then challenged with HIV-1 NL4-3 nef:IRES:GFP containing the N74D capsid mutant in technical triplicate. At day 5 post-challenge, CPSF6 knock-out cells showed a 4- to 6-fold increase in infection relative to the NT controls, indicating that knock-out of CPSF6 enhances infectivity of the N74D capsid mutant (S6l Fig). Previous reports have shown the N74D capsid mutant also results in reduced binding of CYPA [21], and that knock-out of TRIM5α was sufficient to rescue infection, which we also observe here (S6l Fig). The pleiotropic nature of the N74D capsid mutant, the previously demonstrated impact of CPSF6 on TRIM5α levels (Fig 4), and the large number of transcriptional changes induced by CPSF6 knock-out complicates interpretation of this experiment. It may be that the downregulation of TRIM5α is primarily responsible for the observed rescue of N74D replication, though more would need to be done to follow up on this phenotype.

Taken together, these data are consistent with a model in which CPSF6 recruitment and relocalization by incoming capsid cores can trigger transcriptional reprogramming of infected cells through changes in APA to enhance permissivity to infection. The strain and cell type specificity of the observed infectivity changes upon CPSF6 knock-out suggests a complex interplay between the viral and host determinants which should be carefully considered when interpreting prior and future experimental data on this host-pathogen interaction.

Cell Lines

Human embryonic kidney (HEK) 293T cells were obtained from ATCC (#CRL-3216) and cultured in 1x Dulbecco’s Modified Eagle Medium (DMEM) (Corning, #10–013-CV) supplemented with 10% heat inactivated fetal bovine serum (Gibco, #16140–071) and 1% penicillin/streptomycin (Cytiva, #SV30010). 293AAV (Cell Biolabs, Cat# AAV-100) and HT1080 (ATCC, #CVCL_0317) cell lines were maintained in 1x DMEM (ThermoFisher, 11995065), supplemented with 10% fetal calf serum (Sigma, #F092) and gentamicin (Gibco, #15750060). Cells were maintained in a humidified cell culture incubator at 37°C with 5% CO₂.

Plasmid Constructs

An HIV-1 NL4-3 molecular clone with GFP cloned behind an IRES cassette following the viral nef gene (nef:IRES:GFP) was used as the base vector for all viral plasmids (BEI Resources, #11349). N74D and P90A capsid mutants were generated from the original plasmid using site-directed mutagenesis (SDM). Briefly, a portion of the original NL4-3 nef:IRES:GFP capsid region between the SfoI and SpeI restriction sites was amplified and cloned into a pJet1.2/blunt vector (Thermo Scientific, #K1231). The pJet1.2 plasmid with the capsid insert was used for SDM to make the desired point mutations with the following primers:

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

https://doi.org/10.1371/journal.ppat.1013745.t001

The PCR reactions were transformed into chemically competent E. coli (Takara Bio, #636763), and colonies were screened for the desired mutation by Sanger sequencing using the following primer sequences: 5’-CGACTCACTATAGGGAGAGCGGC-3’, 5’-AAGAACATCGATTTTCCATGGCAG-3’. Capsid region inserts bearing the desired mutation were isolated from the pJet1.2 vector and re-inserted into the original NL4-3 nef:IRES:GFP vector by Gibson assembly. Final plasmid sequences were confirmed via Sanger sequencing with the following primers: 5’-AGCGTCGGTATTAAGCGGGG-3’, 5’-ATTCCCTGGCCTTCCCTTGT-3’.

Primary CD4+ T Cell Isolation and Culture

Primary CD4+ T cells were isolated from human peripheral blood leukopaks sourced from de-identified healthy donors (STEMCELL Technologies, #200–0092). Leukopak blood products were diluted to 150 mL in Dulbecco’s Phosphate-Buffered Saline (DPBS) (Corning, #21–031-CV) with 2mM EDTA (Corning, #46–034-Cl) per donor. PBMCs were isolated through Ficoll-Paque centrifugation (Cytiva, #17-5442-03) followed by washing with DPBS with 2mM EDTA. PBMCs were resuspended in MACS buffer (DPBS with 0.5% bovine serum albumin (BSA) (Fisher Scientific, #BP9706100), and 2mM EDTA), and CD4+ T cells were then isolated from PBMCs using an EasySep Human CD4+ T Cell Isolation Kit (STEMCELL Technologies, #17952) per the manufacturer protocol. Cells were resuspended at 2.5x10⁶ cells/mL in complete Roswell Park Memorial Institute (RPMI) medium consisting of RPMI 1640 (Corning, #10–040-CV) supplemented with 10% heat inactivated fetal bovine serum (Gibco, #16140–071), 12.5 mL HEPES (Cytiva, #SH30237.01), 5 mL sodium pyruvate (Corning, #25–000-CI), and 1% penicillin/streptomycin (Cytiva, #SV30010). Media was supplemented with interleukin-2 (IL-2) at 20 IU/mL immediately before use (Miltenyi, #130097744). Cells were activated on α-CD3-coated plates (Tonbo Biosciences, UCHT1) incubated at 4°C overnight before use, with soluble α-CD28 added at 5 mg/mL to media before plating cells (Tonbo Biosciences, CD28.2). Cells were stimulated for 72 hours at 37°C with 5% CO₂ prior to downstream applications.

5-10x10⁴ PBMCs, unstimulated CD4+ T cells, and stimulated CD4+ T cells were taken for CD4 and CD25 cell surface marker staining (Miltenyi Biotec, #130-113-225, 1130-115-535, respectively) to assess cell population purity and activation for each donor. Stained samples were fixed in 1% formaldehyde in DPBS for analysis by flow cytometry within 1 week post-fixation.

CRISPR-Cas9 Knock-Outs in Primary CD4+ T cells

CRISPR-Cas9 knock-outs were generated following previously published protocols [50]. To synthesize CRISPR-Cas9 ribonucleoprotein complexes (crRNPs), lyophilized CRISPR RNA (crRNA) and trans-activating CRISPR RNA (tracrRNA) (Dharmacon) were spun down and resuspended at 160 µM in buffer of 10 mM Tris-HCl (7.4 pH) and 150 mM KCl. For bulk crRNP synthesis, 10 µL of 160 µM crRNA was mixed with 10 µL of 160 uM tracrRNA, then incubated for 30 minutes at 37°C. The resulting crRNA:tracrRNA complexes were gently mixed with 20 µL of 40 µM S. Pyogenes Cas9 (UC-Berkeley Macrolab) then incubated at 37°C for 15 minutes. crRNPs were aliquoted into ten sets of 3.5 µL each and stored at -80°C. For pooled guide conditions, 5 independent crRNA targeting the same gene were mixed in equal volumes prior to tracrRNA addition. crRNAs were obtained from the Dharmacon predesigned Edit-R library or as custom sequences (see table of sequences below).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

https://doi.org/10.1371/journal.ppat.1013745.t002

To generate polyclonal knock-out cells, 1x10⁶ cells per electroporation reaction were spun down at 400xg for 5 minutes and supernatant was removed by aspiration. Cells were resuspended in electroporation buffer consisting of 16.4 µL P3 Nucleofector solution with 3.6 µL supplement per reaction. 20 µL of cell suspension was mixed with 3.5 µL of each preformed crRNP and transferred to 96-well electroporation cuvettes for electroporation with the 4D Core Unit using pulse code EH-115 (Lonza). 100 µL of warm cRPMI was added to each well following electroporation, and cuvettes were placed in cell culture incubator to recover at 37°C for 30 minutes. Cells were moved to 96-well flat bottom tissue culture plates prefilled with 100 µL complete RPMI, 2.5 µL T cell activation/expansion beads (Miltenyi Biotec, #130-091-441), and 0.2 µL of IL-2 (Miltenyi, #130097744). Cells were incubated at 37°C with 5% CO₂.

CRISPR-Cas9 Knock-In in HT1080 Cell Line

Endogenous monomeric NeonGreen (mNGreen) C-terminal tagged CPSF6 in HT1080 cell lines were generated by CRISPR-Cas9 knock-in. Briefly, antisense guide RNA (CACACATTTAACAGGGAACA) targeting sequences overlapping the stop codon of the CPSF6 gene locus was designed using the Benchling guide design online tool (Zang lab, MIT) and synthesized as lyophilized crRNA along with tracrRNA (IDT). The crRNA and tracrRNA were duplexed according to the manufacturer’s guidelines (IDT). 1 μM of the complexed sgRNA was mixed gently with 1µM of Alt-R S. Pyogenes Cas9 enzyme (IDT) in Opti-MEM (ThermoFisher, #11058021), incubated for 5 minutes at room temperature and stored in 4^oC as an RNP complex. A template of mNGreen gene, flanked by 800 base pairs homology arms complementary to the CPSF6 loci, was PCR amplified and cloned into Adeno-associated virus (AAV) using the NheI and XhoI restriction sites in the AAV backbone (Cell Biolabs) flanked by the AAV-ITRs to generate the pAAV-CPSF6-mNGreen construct. AAV particles containing CPSF6-mNGreen template were produced as previously described [51]. Briefly, plasmids encoding the recombinant AAV DJ capsid and replication genes (pAAV-DJ Rep-Cap, Cell Biolabs), helper (pAAV-Helper, Cell Biolabs), and knock-in template (pAAV-CPSF6-mNGreen) were transfected in 293AAV cells in a 1:1:1 ratio using 1 mg/mL polyethylenimine (PEI, Polysciences, #23966). 72 hours post-transfection, the viral particles were harvested, filtered, aliquoted, and stored at -80^oC. The sgRNA-Cas9 RNP complex was reverse transfected into 4x10⁴ HT1080 cells using RNAiMAX (ThermoFisher, #13778150) in a 96 well plate format and infected with 100 µL of AAV-containing CPSF6-mNGreen template. 48 hours post-transfection, cells were sub-cultivated in 48 well plate and sorted for single cell clones. Homozygous and heterozygous fluorescent knock-in CPSF6 cell clones were validated by PCR, sanger sequencing, western blotting and fluorescent widefield microscopy.

Virus Stock Preparation

Replication competent HIV-1 was prepared using the original CXCR4-tropic HIV-1 NL4-3 nef:IRES:GFP plasmid (BEI Resources, #ARP-11349) in addition to the N74D and P90A capsid mutant plasmids described in the plasmid construction methods. Additional replication-competent virus was generated using the CCR5-tropic JR-FL iGFP plasmid and the transmitted founder CH040 plasmid (BEI Resources, #HRP-11740). For single round virus, a wild-type or N74D capsid mutant HIV NL4-3 dEnv nef:IRES:GFP plasmid was pseudotyped with VSV-G. For virus production, low passage (<15) HEK293T cells were plated at a density of 5x10⁶ cells per 15 cm tissue culture dish 24 hours prior to transfection. 10 µg of plasmid DNA, 500 µL serum-free DMEM, and 30 µL of PolyJet (SignaGen Laboratories, # SL100688) were combined per plate of virus and transfected according to manufacturer protocol. 25 mL of viral supernatant was collected at 48 hours post-transfection, replaced with fresh media, and a second set of supernatants were collected at 72 hours post-transfection. Media was combined, filtered through 0.2 µM filters (Fisherbrand, #FB12566504), and incubated with 8.5% polyethylene glycol (Sigma Aldrich, #81260–5 KG) and 0.3 M sodium chloride for 8 hours at 4°C. Viral precipitant was centrifuged at 3500 rpm for 20 minutes, then resuspended in 250 µL DPBS resulting in 100x concentration. Virus aliquots were stored at -80°C prior to p24 viral titer quantification and use in infection assays.

p24 ELISA for Viral Titer Quantification

Viral stocks were heat inactivated at 60°C for 30 minutes, then quantified using an HIV-1 Gag p24 Quantikine ELISA kit (R&D Systems, #DHP240B) according to manufacturer protocol. Samples were run in duplicate at dilutions of 1:500 and 1:2500. Viral stock concentrations were determined using a standard curve.

HIV-1 Spreading Infection Assays

CRISPR-Cas9 knock-out primary CD4+ T cells were generated as described above. All experiments included knock-outs for HIV-1 co-receptor CXCR4, host factor CYPA, and NT controls alongside knock-out of other host factors of interest. At day 4 post-electroporation, cells were split into replica plates of 1x10⁵ cells per well to generate samples for protein lysates for knock-out validation, viability assays, genomic DNA, HIV-1 infected protein lysates, and HIV-1 infection in triplicate per condition. Cells designated for protein lysates and viability assays were processed as described below. Genomic DNA was extracted using QuickExtract DNA Extraction Solution (Lucigen, #QE09050) according to manufacturer protocols and stored at -20°C. Plates designated for infection were infected in triplicate per condition with HIV-1 NL4-3 nef:IRES:GFP, JR-FL iGFP, or CH040 viral stocks diluted to a concentration of 0.2 ng p24 per 50 µL of complete RPMI with IL-2, and 50 µL of diluted virus was added to each well. Proper viral dose was determined by virus titration to optimize percent infection and cell viability (see S1e Fig). Infected cells were incubated in a cell culture incubator at 37°C with 5% CO₂. 24 hours post infection, half of the cells received 50 µL of 25 µM Saquinavir (BEI Resources, #NIH-ARP 4658) diluted in complete RPMI with IL-2. The remaining cells received an equal volume of dimethyl sulfoxide (DMSO) (Sigma Aldrich, #472301–100ML) diluted in media. At days 2 and 5 post-infection, 75 µL of infected cell suspension was removed per well and fixed with 75 µL of 2% formaldehyde in DPBS for later analysis of infection by flow cytometry within 1 week post-fixation. For CH040-infecteted wells, cells were stained with anti-HIV-1 KC57 antibody and fixed according to the intracellular staining protocol below. The remaining cells were fed with an additional 75 µL of complete RPMI supplemented with IL-2. Protein lysates were collected from the appropriate plate on day 5 post-infection.

For spreading infections conducted in the presence of cyclosporin A, cells were treated with cyclosporin A (Sigma-Aldrich, #SML1018) at a final concentration of 5 µM 4 hours prior to infection.

Single Round Infection Assays

CRISPR-Cas9 knock-out primary CD4+ T cells were generated as described above using cells from 3 independent blood donors. All experiments included knock-outs for HIV-1 co-receptor CXCR4, host factor CYPA, and NT controls alongside knock-out of other host factors of interest. At day 4 post-electroporation, cells were split into replica plates of 1x10⁵ cells per well to generate samples for protein lysates for knock-out validation, viability assays, and HIV-1 infection in triplicate per condition. Cells designated for protein lysates and viability assays were processed as described below. Plates designated for infection were infected in triplicate per condition with viral stocks (HIV-1 NL4-3 nef:IRES:GFP or VSV-G-pseudotyped NL4-3 dEnv nef:IRES:GFP) at a concentration of 0.02, 0.05, 0.1, 0.2, 0.5, or 1 ng p24 per well suspended in 50 µL complete RPMI supplemented with IL-2. At day 2 post infection, 75 µL of cell suspension was fixed with an equal volume of 2% formaldehyde in DPBS for later analysis of infection by flow cytometry within 1 week post-fixation.

Interferon Titration and Sensitivity Assays

Primary CD4+ T cells from 3 independent donors underwent CRISPR-Cas9 editing using a NT control and guides targeting CXCR4, CYPA, IFNAR1, and CPSF6. Cells were replica plated as previously described. Additional replicates of 1x10⁵ edited cells per well in 96 well U-bottom plates were generated for the interferon titration assay. Plates for HIV-1 infection were treated with 100 U/mL (Donors P, Q, and AN) universal type I interferon (PBL Assay Science, #11200–2), IFNα (PBL Assay Science, #11100–1), IFNβ (PeproTech, #300–02BC), or IFNγ (PeproTech, #300–02) or a media-only control 16 hours prior to infection. Donor AO was treated with a range (1–1000 U/mL) of IFNα and IFNβ for 16 hours prior to infection. Following incubation with interferon, cells were infected with HIV-1 NL4-3 nef:IRES:GFP as described in spreading infection assay protocol. Infection timepoints were collected at days 2 and 5 post-infection, and day 5 HIV-1 infected protein lysates were collected as described below. To assess the impact of IFN treatment on ISG production, replica plated cells were treated with 10 U/mL of universal type I interferon, IFNα, IFNβ, IFNγ, or media-only controls. 16 hours post-interferon treatment, protein lysates were harvested according to protocol in immunoblotting methods for later analysis of ISG expression.

Receptor/Co-receptor and Activation Marker Surface Staining

CXCR4, CCR5, CYPA, CPSF6, and TRIM5α knock-out cells and NT control cells were generated using CRISPR-Cas9 editing as previously described using primary CD4+ T cells isolated from the blood of three independent human donors. 1x10⁵ cells per condition were used for cell surface marker staining for CXCR4 (APC, Miltenyi Biotec, #130-120-708 [clone REA649]), CD4 (PE, Miltenyi Biotec, #130-113-225 [clone REA623]), CD25 (APC, Miltenyi Biotec, #130-115-535 [clone REA945]), and CCR5 (APC, Miltenyi Biotec, #130-120-708 [clone REA245]) at day 5 post-editing. Antibodies were diluted in MACS buffer at a dilution of 1:50 and samples were stained for 15 minutes at 4°C in the dark. After washing with MACS buffer, samples were fixed in 1% formaldehyde in DPBS for analysis by flow cytometry within 1 week post-fixation. Flow analysis was performed using the gating strategy shown in S7 Fig.

Anti-HIV-1 Intracellular Staining

At days 2 and 5 post-infection, 75 µL of infected cell suspension from cells infected with virus lacking a fluorescent reporter (CH040) was removed per well and transferred to a new 96-well U-bottom plate. Cells were spun down at 400xg for 5 minutes, then the supernatant was aspirated and cells were fixed with 250 µL of 1% formaldehyde in DPBS for 30 minutes at room temperature. The fixed cells were spun down at 400xg for 5 minutes, the supernatant was aspirated, and then the cells were resuspended in 100 µL of a permeabilization buffer consisting of 1% BSA and 0.1% saponin (Sigma, #47036) in DPBS per well. Cells were blocked by incubating the cells in the permeabilization buffer for 20 minutes at room temperature. After blocking, the cells were spun down, the supernatant was aspirated, and the cells were stained with 100 µL of antibody solution consisting of a 1:100 dilution of anti-HIV-1 KC57-FITC antibody (Beckman Coulter, #6604665) in permeabilization buffer. Cells were incubated for 30 minutes in the dark at room temperature for staining. Cells were then spun down, washed with 200 µL of DPBS with 1% BSA, then fixed with 150 µL of 1% formaldehyde in DPBS for later analysis of infection by flow cytometry within 1 week post-fixation.

Viability Assays

Cell viability was assessed at specified timepoints by Ghost Red amine dye stain and/or CellTiter-Glo assay. For Ghost Red stain, 1x10⁵ cells per condition were pelleted in 96-well U-bottom plates by centrifuging at 400xg for 5 minutes. Following a wash with DPBS, cells were resuspended in 100 µL of Ghost Red dye (Tonbo Biosciences, #13–0871-T100) diluted 1:1000 in DPBS. Cells were incubated in the dark at 4°C for 30 minutes, then spun down and supernatant was aspirated. Cells were washed with MACS buffer (DPBS (Corning, #21–031-CV) with 0.5% BSA (Fisher Scientific, BP9706100) and 2mM EDTA (Corning, #46–034-Cl)), spun down, and resuspended in 150 µL 1% formaldehyde in PBS for later analysis by flow cytometry.

For CellTiter-Glo Assay, cells in 96 well U-bottom plates were resuspended in 100 µL media and moved to an opaque, white 96 well plate (Corning, #3912). Cells were incubated at room temperature for 30 minutes to equilibrate to room temperature. Cell suspension was mixed with 100 µL of room temperature CellTiter-Glo reagent (Promega, #G7570) and placed on an orbital shaker for 2 minutes to complete cell lysis. Plate was incubated 10 additional minutes at room temperature to stabilize luminescence signal before reading luminescence on an Omega plate reader (BMG LabTech), with a consistent gain setting used across readings.

Immunoblotting

Primary CD4+ T cell protein lysates were generated by washing 1x10⁵ cells per condition in DPBS, removing supernatant, and resuspending cells in 50 µL 2.5x Laemmli Sample Buffer (1.9 mL 0.5 M Tris-HCl pH 6.8 (Fisher Bioreagents, #BP153-1), 6 mL 50% glycerol (Fisher Bioreagents, #BP229-1), 3 mL 10% SDS (Corning, #46–040-CI), 250 µL β-mercaptoethanol (Fisher Chemical #O3666I-100), 50 µL 1% bromophenol blue (Fisher Bioreagents, #BP115–25), 18.8 mL DPBS (Corning, #21–031-CV)). Lysates were heated at 98°C for 20 minutes, then stored at -20°C prior to use for immunoblotting. Lysates were run on 4-20% Criterion Tris-HCl 18 or 26 well gels (Bio-Rad, #3450033 and # 3450034, respectively) at 90V for 30 minutes followed by 150V for 70 minutes. Proteins were transferred to PVDF membranes (Bio-Rad, #1620177) for 2 hours at 90V. Membranes were blocked in 4% milk in PBS (Sigma-Aldrich, #P5493-4L) with 0.1% Tween-20 (Fisher Bioreagents, #BP337–500) or 5% BSA in PBS with 0.1% Tween-20 according to antibody manufacturer recommendations for 1 hour, then incubated with primary antibody overnight at 4°C with shaking. Primary antibodies used in this study are as follows: CPSF6 (Rabbit, 1:3000 in 5% BSA, Novus, #NBP1–85676), CPSF5 (Rabbit, 1:5000 in 4% milk, Proteintech, #10322–1-AP), CYPA (Rabbit, 1:12000 in 4% milk, Cell Signaling, #2175), IFNAR1 (Rabbit, 1:1000 in 4% milk, abcam, #ab124764 [clone EPR6244]), MX1 (Rabbit, 1:1000 in 5% BSA, Cell Signaling, #37849S [clone D3W7I]), IFIT1 (Rabbit, 1:1000 in 4 milk, Cell Signalling #14769 [clone D2X9Z], TRIM5α (1:1000 in 4% milk, Cell Signaling, #14326 [clone D6Z8L]), and β-actin (Mouse, 1:10000 in 4% milk, Cell Signaling, #3700S [clone 8H10D10]). Membranes were washed with PBS with 0.1% Tween-20, then incubated with either anti-mouse (Peroxidase AffiniPure Goat Anti-Mouse IgG (H + L), Jackson ImmunoResearch, #115-035-003) or anti-rabbit (Peroxidase AffiniPure Goat Anti-Rabbit IgG (H + L), Jackson ImmunoResearch, # 111-035-003) HRP-conjugated secondary antibodies at 1:10000 in 4% milk in PBS with 0.1% Tween-20 for 1 hour, washed with PBS with 0.1% Tween-20, then detected using Immobilon Western Chemiluminescent HRP Substrate (EMD Millipore, #WBKLS0500) and imaged using an iBright imaging system (Thermo Fisher). Blots were incubated in antibody stripping solution (EMD Millipore, #2502) prior to re-probing. Blots were quantified by measuring the rolling ball-corrected density of each analyzed band using iBright Image Analysis Software (version 5.2.1.0). Density values for the target protein bands (CPSF6 or IFIT1) were divided by density values for β-actin loading control bands, and the resulting ratio was normalized to the NT control ratio for each donor.

Flow Cytometry

Flow cytometry analysis was performed using an Attune NxT acoustic focusing cytometer and CytKick Max Autosampler attachment (Thermo Fisher Scientific). Samples were suspended in 150 µL DPBS with 1% formaldehyde and were subjected to one 150 µL mixing cycle. All events in 50 µL of sample were recorded. FCS3.0 files were exported using Attune NxT Software (version 5.3.0) and analyzed using a FlowJo (version 10.10) template shown in S7 Fig for HIV-1 infection (BL1:GFP and FITC), Ghost Red stain (RL-2), or cell surface receptor stain (RL-1: CXCR4-APC, CCR5-APC, CD25-APC, BL-2: CD4-PE).

HIV-1 Infection and Live Cell Sorting for RNA-Seq

21x10⁶ stimulated primary CD4 + T cells from donors AC-AE were infected with wild-type or N74D capsid mutant HIV-1 NL4–3 nef:IRES:GFP alongside uninfected controls. Briefly, cells were resuspended in complete RPMI supplemented with IL-2, concentrated HIV-1 stocks for the appropriate virus were added to the cell suspension at a dose of 210 ng per condition, and the final volume of media was brought to 42 mL. Cells were then distributed into 96 well U-bottom plates, with each well containing 200 µL media, 1x10⁵ cells, and 1 ng p24. 1x10⁶ cells per donor were distributed into 96 well U-bottom plates at a concentration of 1x10⁵ cells per well in 200 µL complete RPMI with IL-2 to generate uninfected control cells. All cells were incubated in a cell culture incubator at 37°C with 5% CO₂. On day 2 post-infection, infected cells were pooled per donor and virus strain. 1x10⁶ uninfected cells per donor immediately underwent RNA extraction according to protocol below to generate the uninfected RNA samples. The remaining infected cells (~20 million per condition) were spun down at 400xg for 5 minutes, resuspended in 1 mL MACS buffer, and transferred to MACSQuant Tyto Cartridges (Miltenyi Biotec, #130-104-791). Cells in cartridges were immediately sorted using a MACSQuant Tyto sorter (Miltenyi Biotec) with MACSQuantify Tyto Software (version 2.0). Cells were sorted by gating for infection (GFP+) status as described in S7 Fig The cartridge mixer was set at a speed of 800 rpm, and the cell sorting pressure was set to 150 hPa. Following sorting, cells were resuspended in 200 µL MACS buffer and retrieved from the sorted cartridge chamber. Samples were spun at 400xg for 5 minutes, supernatant was aspirated, and RNA was extracted according to the protocol below.

RNA Extraction

RNA was extracted from donors E-G (Fig 2) at day 5 post electroporation and donors AC-AE (Fig 6) at day 2 post-infection or day 2 post-electroporation depending on the condition. RNA was extracted from cell pellets using RNeasy Mini Kit (QIAGEN, #74104) according to manufacturer protocol. Samples were treated with DNAse I according to manufacturer protocol (QIAGEN, # 79254) and eluted in RNase-free water. RNA concentration was determined by NanoDrop (Thermo Scientific), and samples were stored at -80°C prior to analysis.

HT1080-CPSF6-mNGreen HIV-1 Infection, RNA Extraction, Immunofluorescence Staining, and Imaging

1x10⁵ HT1080-CPSF6-mNGreen cells per condition were seeded in 12-well plate and adhered to plates overnight prior to infection. In parallel, 3x10⁴ HT1080-CPSF6-mNGreen cells were seeded in 8-well chamber slides (Ibidi, #80806) coated with 0.1 mg/ml poly-D lysine hydrobromide (Advanced Biomatrix, #5049). These cells were infected with VSV-G pseudotyped wild-type or N74D capsid mutant HIV-1-NL4-3 at an MOI of 10 or mock-infected by spinoculation at 16°C for 1 h at 1200xg. Cell cultures were washed once with prewarmed DMEM and returned to a tissue culture incubator at 37°C with 5% CO₂ for 6 hours. RNA from wild type, N74D capsid mutant HIV-1-NL4-3, and mock infected cells in the 12-well plate was then extracted using a NucleoSpin RNA Mini kit (Machery-Nagel, # 740955.50) according to manufacturer’s protocol. RNA concentration was determined by NanoDrop (Thermo Scientific), and samples were stored at -80°C prior to analysis. Protein lysates were harvested from HIV-1-NL4-3 wild-type, N74D and mock infected HT1080-CPSF6-mNGreen cells infected under the same conditions described above for analysis of CPSF6 protein levels by immunoblot. In parallel, HIV-1-NL4-3 wild-type, N74D and mock infected HT1080-CPSF6-mNGreen cells cells on the 8-well chamber slides were washed with 1x DPBS and fixed with 4% paraformaldehyde (PFA) (Sigma, #P6148-1 KG) for 15 minutes. The PFA fixation was quenched with 5mM glycine for 10 minutes and washed twice with 1x DPBS. The cells were permeabilized and blocked with 0.1% Triton- X100 and 10% goat serum respectively for 30 minutes. Mouse monoclonal anti-SC35 antibody (abcam, #ab11826) was subsequently used to stain the cells at a 1:250 dilution for 1 hour, washed twice with 1x DPBS, followed by goat anti-mouse secondary antibody conjugated to janelia fluor 646 (Novus Biologicals, #72739JF646) at a 1:500 dilution for 1 hour. Mock infected control cells were stained with purified mouse IgG1 kappa isotype control antibody (Biolegend, #401401) at a 1:250 dilution. Following the secondary antibody staining, the cells were washed twice and stained with Hoechst 33258 (abcam, #ab145596) for DNA at a 1:1000 dilution for 30 minutes. Images for each experimental condition were acquired with widefield deconvolution fluorescent microscope (GE, DeltaVision OMX SR Imaging System) equipped with 60x objective with 1.42 NA, using GE oil emersion with 1.518 refractive index, complementary metal-oxide semiconductor (CMOS) cameras and 405 nm, 488 nm, 568 nm and 640 nm Diode lasers. Acquisitions were performed with the following conditions, 10% transmission and 100-ms exposure for the green channel (CPSF6), 10% laser power and 100ms exposure for the far-red channel (SC35), and 5% laser power and 100ms exposure for the blue channel (Hoechst 33258). All images were acquired with the same microscope settings and deconvolved using the ratio conservative algorithm method SoftWorx (version 7.2.2).

Quantification of CPSF6 Puncta and CPSF6/SC35 Colocalization

Images in the CPSF6, SC35 and Hoechst channels were acquired deconvoluted as described above. Images were then opened in Imaris (version 10.2) and CPSF6 puncta were detected using the Imaris Spots module (estimated XY diameter 0.25 µm, estimated Z diameter 0.5 µm, quality > 300, region growing option was used to account for varying spot sizes). SC35 nuclear speckles were segmented using the Imaris Surfaces module (surface grain size = 0.1 um, manual threshold = 3000). The detected CPSF6 puncta were then classified into colocalizing (class A) and non-colocalizing (class B) with SC35. An Imaris pipeline was created with the above parameters and applied to all images in the batch. The sum fluorescence intensity of each spot in conditions (Mock, WT, and N74D) was exported as a.csv file.

RNA-Sequencing

RNA samples from donors E-G (Fig 2) were submitted to the NUSeq core for next-generation sequencing. Libraries were prepared using TruSeq Total RNA-Seq Library Prep (Illumina) and sequenced using an Illumina NovaSeq 6000 with 150 bp paired-end reads. RNA samples from donors AC-AE (Fig 6) were submitted to Novogene for next-generation sequencing. Libraries were prepared using NEBNext Ultra II kit with poly(A) selection according to manufacturer protocols (New England Biolabs). Samples were sequenced using an Illumina NovaSeq 6000 with 150 bp paired-end reads. RNA samples from HT1080 cells (Fig 6) were submitted to Novogene for next-generation sequencing. Libraries were prepared using NEBNext Ultra II kit with poly(A) selection according to manufacturer protocols (New England Biolabs). Samples were sequenced using a NovaSeq X Plus with 150 bp paired-end reads.

RNA-Sequencing Data Processing

The RNA-Seq data processing pipeline was adapted from published protocols [52]. Briefly, we first trimmed the paired-end reads in FASTQ files using Trimmomatic (version 0.39) [53], targeting known adapters (TruSeq3-SE.fa:2:30:10), and also trimming the first and last 3 base pairs if their quality fell below set thresholds. Additionally, we employed a 4-base sliding window technique, trimming reads when the average window quality dropped below 15. Reads shorter than 36 base pairs were excluded. Next, the trimmed reads were aligned to the human reference genome GRCh38 (https://www.ncbi.nlm.nih.gov/datasets/genome/GCF_000001405.26/) using HISAT2 [54] (version 2.1.0) and converted into BAM format using samtools (version 1.6) [55]. The alignment data was then processed by StringTie (version 2.1.3) [56] to estimate gene expression levels. The StringTie merge function merged samples’ gene structures to create a unified set of transcripts across all samples. Finally, read counts were obtained using prepDE.py (https://ccb.jhu.edu/software/stringtie/index.shtml?t=manual) to prepare the data for DESeq2.

Differential Gene Expression Analysis

The raw read counts were processed in R (version 4.2.3). Initially, transcript names were converted to gene symbols, eliminating duplicate symbols and dropping transcripts that did not map to known gene symbols. The resulting reads were used to construct a DESeqDataSetFromMatrix object using DESeq2 (version 1.38.3) [57] for investigating differences across conditions, accounting for donor variability. Genes with a minimum count of 10 in at least 3 samples were retained for analysis. To visualize treatment effects, counts were transformed using the variance stabilizing transformation (VST) function from DESeq2 and subjected to principal components analysis (PCA), visualizing the first two principal components [58].

Differential gene expression was assessed using the DESeq function from DESeq2, which estimates size factors and dispersions, fitting a negative binomial generalized linear model. P-values were adjusted using the Benjamini-Hochberg method [59], considering adjusted p-values below 0.05 as significant. Subsequently, complete cases were filtered, and the lists of differentially expressed genes were input into Metascape (version 3.5.20250101) for identifying pathways associated with up- or downregulated gene sets [60]. The log fold change for genes identified by Metascape that were associated with the interferon gamma signaling pathway were visualized for all 3 donors (individually and collectively) in a heatmap. The top 2000 significant genes with highest log fold changes were also z-score transformed and clustered using hierarchical clustering that was visualized in a heatmap [61].

Alternative Polyadenylation Analysis

APA analysis was conducted by analyzing RNA-Seq data using the REPAC pipeline (version 0.99.0) to determine changes in 3’ UTR length in treated (CPSF6 knock-out or infected) versus control (non-targeting or uninfected) conditions [36]. Polyadenylation sites that had less than 20 counts assigned to them were excluded from further analysis. Differential polyadenylation site usage was then calculated between experimental (CPSF6 knock-out or HIV-1 infected) versus control (non-targeting or uninfected) conditions. 3’ UTRs were designated shortened for Compositional Fold Change (cFC) < -0.25 and -log₁₀(adjusted p-value) < 0.05 and lengthened for cFC > 0.25 and -log₁₀(adjusted p-value) < 0.05. Lists of genes with significant changes in 3’ UTR length were input into Metascape (version 3.5.20250101) for identifying pathways associated with genes sets with shortened or lengthened 3’ UTRs, and the list of all genes with tested APA sites was provided as a background gene set [60].

Additional validation of 3’ UTR length changes were carried out using DaPars (Dynamic analysis of Alternative PolyAdenylation from RNA-Seq) (version 0.9.1) [37,62]. Distal polyadenylation sites (PAS) for each gene were extracted from the GRCh38/hg38 refseq gene model downloaded from UCSC (https://genome.ucsc.edu/). A proximal PAS for each gene was determined using DaPars by identifying the optimal fitting point of a linear regression model using alignment wiggle files for each sample created using DeepTools (version 3.5.6). Differential PAS usage was then quantified as a change in Percentage of Distal PAS Usage Index (ΔPDUI) for each gene in treated (CPSF6 knock-out or infected) versus control (non-targeting or uninfected) conditions, using default parameters. Genes were considered significantly shortened or lengthened if they had a false discovery rate (FDR) of < 0.05.

RNA-Seq read tracks were visualized using Integrated Genomics Viewer (version 2.17.4) to show changes in 3’ UTR read coverage for genes of interest (TGFBR1, IL2RA/CD25, CXCR4) using RNA-Seq data from Donor E. Track heights were kept consistent across samples. Violin plots depicting changes in 3’ UTR length among gene sets of interest (Cytokine Signaling in Immune system (R-HSA-1280215), Interferon signaling (R-HSA-913531), and T cell activation (GO:0042110)) were generated by extracting gene names associated with the gene sets using biomaRt (version 2.58.2) and annotating the output of the REPAC and ΔPDUI analyses to indicate which of the tested genes were a member of the gene sets.

Statistics

All statistical analysis was performed using GraphPad Prism version 10.2.1 (395) for Windows, GraphPad Software, Boston, Massachusetts USA, www.graphpad.com and R (version 4.2.3) stats package (version 3.6.2). All statistical analysis performed in this manuscript is available in S1 Table.