Exploiting a rodent cell block for intrinsic resistance to HIV-1 gene expression in human T cells

ABSTRACT HIV-1 virion production is inefficient in cells derived from mice and other rodents reflecting cell-intrinsic defects to interactions between the HIV-1 auxiliary proteins Tat and Rev and host dependency factors CCNT1 (Cyclin T1) and XPO1 (exportin-1, also known as CRM1), respectively. In human cells, Tat binds CCNT1 to enhance viral RNA transcription and Rev recruits XPO1 to mediate the nuclear export of intron-containing viral RNA. In mouse cells, Tat’s interactions with CCNT1 are inefficient, mapped to a single species-specific residue Y261 instead of C261 in humans. Rev interacts poorly with murine XPO1, mapped to a trio of amino acids T411/V412/S414 instead of P411/M412/F414 in humans. To determine if these discrete species-specific regions of otherwise conserved housekeeping proteins represent viable targets for inhibiting HIV-1 replication in humans, herein, we employed CRISPR/Cas9 to recode the relevant regions of CCNT1 and XPO1 in human CD4+ T cells. While efforts to modify XPO1 were inconclusive, we generated isogenic CCNT1.C261Y cell lines exhibiting remarkable resistance to HIV-1 Tat, exhibiting near total inactivation of viral gene expression for all X4- and R5-tropic HIV-1 strains tested, as well as the more distantly related primate lentiviruses HIV-2 and SIVagm. Induction of viral reactivation using latency reversal agents (LRAs) was also restricted in CCNT1.C261Y cells. These studies validate a minor and naturally occurring, species-specific difference in a conserved human host factor as a compelling potential target for achieving broad-acting cell-intrinsic resistance to HIV’s post-integration phases. Importance Unlike humans, mice are unable to support HIV-1 infection. This is due, in part, to a constellation of defined minor, species-specific differences in conserved host proteins needed for viral gene expression. Here, we used precision CRISPR/Cas9 gene editing to engineer a “mousified” version of one such host protein, cyclin T1 (CCNT1), in human T cells. CCNT1 is essential for efficient HIV-1 transcription, making it an intriguing target for gene-based inactivation of virus replication. We show that isogenic cell lines engineered to encode CCNT1 bearing a single mouse-informed amino acid change (tyrosine in place of cysteine at position 261) exhibit potent, durable, and broad-spectrum resistance to HIV-1 and other pathogenic lentiviruses, and with no discernible impact on host cell biology. These results provide proof of concept for targeting CCNT1 in the context of one or more functional HIV-1 cure strategies.

. Allogeneic bone marrow transplant using cells homozygous for CCR5Δ32 has contributed to the only known instances of HIV/AIDS functional cure (4)(5)(6)(7).These cases, known as the Berlin, London, and New York patients, demonstrated that drugfree remission or cure is possible through targeted inactivation of CCR5.However, the potential drawbacks to gene-targeted CCR5-based cures include the potential for viral evolution of altered receptor tropism (8) (e.g., outgrowth of CXCR4-tropic HIV-1) and the ∆32 allele being linked to increased risk of symptomatic flavivirus infections (9)(10)(11).These issues underscore the need to consider additional gene targets in the context of cell-based functional cures.
In this context, studies of HIV-1 replication in non-human mammals have revealed a trove of species-specific gene barriers to infection (12,13).For example, rhesus macaques resist HIV-1 due to the activities of endogenous APOBEC3G and TRIM5 proteins that disrupt intact viral genome delivery to targeted cells (14,15).Other forms of HIV-1 resistance can be attributed to species-specific differences to so-called hostdependency factors (HDFs); cellular proteins hijacked by the virus to carry out specific life cycle stages.For example, mice and other rodents lack HIV-compatible versions of both CCR5 and the primary attachment receptor CD4, thereby resisting viral entry.Exogenous expression of human CD4 and CCR5 was shown to alleviate viral entry blocks in some rodent cell lines (16)(17)(18)(19)(20)(21)(22)(23)(24)(25)(26)(27)(28).However, these provisions were insufficient to restore virus replication due to additional rodent cell blocks affecting HIV-1's post-integration stages .
Of the post-integration rodent cell blocks, those affecting the essential HIV-1 accessory proteins Tat and Rev are best characterized.In human cells, HIV-1 Tat drives viral transcription by recruiting the positive transcription elongation factor b (p-TEFb) complex, consisting of a heterodimer of the CDK9 kinase and its regulatory partner CCNT1, to the trans-activation response element (TAR) structure present in the 5′ portion of initiated viral transcripts (51).CDK9 phosphorylates the C-terminal domain of RNA polymerase II to promote processivity and synthesis of full-length viral RNAs (51).Human CCNT1 encodes a cysteine at residue 261 (C261) thought to coordinate a zinc ion that stabilizes the functional Tat/p-TEFb/TAR complex (34)(35)(36).The murine ortholog of CCNT1 encodes a tyrosine at the equivalent position (Y261) and is associated with poor Tat-dependent transcription (34)(35)(36).
By contrast, in human cells, HIV-1 Rev enables the nuclear export of unspliced and partially spliced viral RNAs that would otherwise remain retained in the nucleus (51).In the nucleus, Rev binds to and multimerizes on the Rev response element (RRE) structure present on unspliced and partially spliced transcripts and engages the XPO1 (exportin-1, also called CRM1) nuclear export machinery (51).Human XPO1 features a species-specific surface element defined by a proline, methionine, and phenylalanine at residues 411, 412, and 414, respectively (P411, M412, and F414 in humans and T411, V412, and S414 in mice), that we and others have shown is needed for efficient Rev activity (48)(49)(50).This "patch-like" element is likely important for stabilizing the formation of functional Rev/RRE/XPO1 export complexes consisting of multiple XPO1 proteins (52,53).An alternative but nonmutually exclusive model suggests that Rev/RRE complexes bind with greater affinity to human XPO1 compared to the murine counterpart (54).
Based on this knowledge, in this study, we asked if these discrete, naturally occurring, rodent-specific HDF blocks to HIV-1 Tat and Rev function could be engineered into human CD4+ Jurkat T cells, taking advantage of precision CRISPR/Cas9 genome editing.We provide strong evidence that human CCNT1 can be "mousified, " i.e., modified to engender a rodent cell-like block to Tat function, and as proof-of-concept in the context of an antiviral target, isolated isogenic CCNT1.C261Y cell lines that were effectively resistant to HIV-1 Tat activity and viral gene expression despite the C261Y edit having no discernible effects on CCNT1-regulated host cell functions.Furthermore, we demonstrate that CCNT1.C261Y cells are likely to exhibit cell-intrinsic resistance to all known HIV-1 strains and subtypes, inducing a state of forced viral latency consistent with a "blockand-lock" provirus inactivation scenario.By contrast, we were unable to validate effects on Rev function attributable to the species-specific trio of XPO1 residues.Collectively, these results highlight the potential value of a species-specific domain of human host factor CCNT1 as a target for anti-HIV intervention.

RESULTS
A dual-reporter strategy to screen for cell-intrinsic blocks to early-(Tatdriven) and late-stage (Rev-driven) HIV-1 gene expression So that we could screen for cell-intrinsic blocks to Tat-and Rev-dependent HIV-1 gene expression, we first devised a flow cytometric assay to capture fluorescence profiles of cell populations infected with a two-color HIV-1 reporter virus (E-R-/2FP, Fig. 1A); E-R-/2FP encodes mCherry from the nef locus to monitor early, Tat-dependent but Rev-independ ent gene expression and mVenus as a Gag fusion (matrix [MA]-mVenus-capsid [CA]) reporting on late, both Tat-and Rev-dependent gene expression.Inactivating mutations in env and vpr genes (E-/R-) limited virus replication to a single round and reduced cytopathic effects (Fig. 1B).
To validate our reporter, we infected Jurkat T cells with our E-R-/2FP virus or control versions engineered to carry inactivating mutations in tat (E-R-Tat-/2FP) or rev (E-R-Rev-/ 2FP) (Fig. 1C).Compared to mock-infected cells (gray dots), E-R-/2FP-infected cells (orange dots) yielded a robust population of bright double-positive cells (orange dots in dashed oval), consistent with efficient early and late viral gene expression phases.E-R-Rev-/2FP-infected cells (red dots) were mCherry-bright but mVenus-dim, consistent with the activity of Tat, but not Rev.As expected, E-R-Tat-/2FP-infected cells (gold dots) lacked both mCherry and mVenus fluorescence but, like the E-R-Rev/2FP cells, were moderately mVenus-positive due to cellular uptake of the Gag-mVenus fluorescent virus particles.In sum, E-R-/2FP control variants confirmed our dual reporter as capable of quantifying the relative activities of Tat and Rev with single-cell resolution.
For further validation, we next tested our E-R-/2FP assay using a panel of informative cell lines predicted to encode permissive or non-permissive species-specific determi nants encoded by CCNT1 and XPO1, respectively (Fig. 1D, shown in blue).Expectedly, primate cell lines including 293T, HeLa, Cos7, and WES exhibited fluorescence profiles similar to infected Jurkat T cells, confirming robust Tat and Rev activity (Fig. 1E).In contrast, mouse cell lines including 3T3, L, and C2C12 exhibited deficiencies to Tat-and Rev-dependent gene expression, yielding few to no late stage, bright double-positive cells.Interestingly, M. dunni, Rat1 and Chinese hamster ovary (CHO) cells were largely deficient in late gene expression but exhibited a moderate degree of early, Rev-inde pendent gene expression, illustrating that the transcription block was not absolute in these cell types despite their encoding CCNT1 Y261.Infected DF1 chicken cells, that encode a serine at the equivalent position of CCNT1 C261, were also only modestly permissive for early viral gene expression.CRFK cat cells were interesting in that, despite encoding a XPO1 patch region containing rodent-like V412 and S414 residues, these cells exhibited strong Rev activity, potentially highlighting the relative importance of the XPO1 P411 residue or other yet to be mapped determinants.For all lines, treatment with the reverse transcriptase inhibitor nevirapine (gold dots) served as a control to demarcate productively infected populations (orange dots) from uninfected cells (gray dots) (Fig. 1E).Taken together, we concluded that our E-R-/2FP assay faithfully detected cell-intrinsic differences to Tat and Rev activities and would, therefore, be useful to screen modified human T cell populations for equivalent defects.

Host-targeted CRISPR/Cas9 knock-in strategy to abrogate HIV-1 gene expression in human T cells
Next, we designed CRISPR/Cas9 gene knock-in schemes to determine if we could engineer rodent-informed Tat (CCNT1)-or Rev (XPO1)-dependent HIV-1 gene expres sion blocks into human CD4+ T cells, using the Jurkat E6-1 cell line as a tractable model system.Gene knock-in modifications were performed by co-delivering singlestranded oligodeoxynucleotide (ssODN) donor templates along with recombinant Cas9 in complex with guide RNAs bearing complementary sequences to targeted loci.For human CCNT1, the endogenous "TGC" codon for cysteine at residue position 261 was replaced with "TAC" in the ssODN, directing a substitution for the mouse-encoded tyrosine (Fig. 2A, C261Y, modifications in red).Codon mutations were similarly made for the combined proline, methionine, and phenylalanine residues at positions 411, 412, and 414 encoded by human XPO1, with the goal of replacing them with the mouse-encoded threonine, valine, and serine, respectively (Fig. 2B, P411T, M412V, and F414S, modifications in red; collectively termed, "mPatch").Silent mutations in ssODNs encoding novel restriction enzyme sites (BsiWI for CCNT1 and PvuII for XPO1) allowed for knock-in detection using restriction enzyme digest (Fig. 2A and B, modifications in blue).Protospacer adjacent motifs (PAMs) were also mutated to prevent Cas9-mediated cleavage of introduced template sequences (Fig. 2A and B, modifications in green).
After editing, Jurkat cell populations were screened using the E-R-/2FP assay descri bed for Fig. 1.We hypothesized that, if refractory, CCNT1or XPO1-modified cells would exhibit profiles similar to those observed for rodent cell lines (see Fig. 1E).To better resolve changes, we further augmented the assay to include fluorescence-associated cell sorting (FACS) using seven cell gates (Fig. 2C) allowing us to separate cell populations to represent the progression of viral gene expression beginning with mCherry-positive, Tatdependent but Rev-independent stages (P1-P4) and ending with dual positive, latestage Rev-dependent stages (P5-P7).Comparing CCNT1-targeted cells to both wild type (WT) and XPO1-edited cells, we observed marked increases to the number of cells in P2 and P3 gates, consistent with blocks to Tat function in a significant subset of CCNT1edited cells (Fig. 2E).By contrast, the number of cells in any gate of the XPO1-targeted cell population was not significantly altered (Fig. 2E).
Because spontaneous insertion or deletion events can manifest off-target effects, we also analyzed each sorted pool to confirm allele knock-in frequency using restriction digests (P1-P7).Because the CCNT1 knock-in was predicted to manifest a Tat block, we hypothesized that alleles bearing the CCNT1 knock-in signature would be enriched in cells within gates P1-P4 compared to those in the P5-P7 gates, confirmed by BsiWI sensitivity in the heterogeneous amplicon pools (Fig. 2F).Analysis of XPO1 knock-in by PvuII digestion sensitivity confirmed modifications; however, we observed no allele enrichment in any single gated population (Fig. 2G).Overall, these results suggested that a mouse-encoded CCNT1 HIV-1 block to Tat function could be engineered into a human T cell population with relatively high efficiency; however, that modifying XPO1 to achieve a Rev block was less feasible.

CCNT1.C261Y Jurkat T cells suppress HIV-1 Tat activity
Data indicating strong and stable blocks to Tat function in CCNT1-edited Jurkat cells prompted us to isolate two CCNT1.C261Y cell lines, named C1 and C2, encoding CCNT1 loci confirmed to be entirely sensitive to BsiWI digestion (Fig. 3A) and seamlessly edited, evidenced by the absence of unintended mutations at the targeted locus (Fig. 3B and C).CCNT1.C261Y protein levels were equivalent to non-edited CCNT1 in the parental cell line (Fig. 3D), and we observed no effects on cell proliferation (Fig. 3E).The E-R-/2FP reporter assay confirmed that the modified cells exhibited a fluorescence profile consistent with loss of Tat function (Fig. 3F, compare panels i-iv and vii) and, importantly, significantly rescued by exogenous expression of human CCNT1 but not a CCNT1.C261Y mutant transgene, delivered using retroviral vectors (Fig. 3F and G).
To address Tat activity independently of infection, we next employed a transient reporter assay wherein a firefly luciferase gene under transcriptional control of the HIV-1 LTR (LTR-Luc) was expressed in the presence or absence of Tat in WT and both C261Y cell lines (Fig. 3H) (55).When Tat was expressed in the presence of the LTR-Luc construct, we observed a luciferase signal >30-fold higher in the parental cell line relative to either C261Y clone (Fig. 3H, compare lane 10-11 and 12) confirming severe but not absolute disruption to Tat-driven gene expression, and again significantly rescued when coexpressed with WT human CCNT1 but not the CCNT1.C261Y mutant (Fig. 3H, compare lane 14-17; compare lane [15][16][17][18].Together, these results demonstrated that otherwise permissive human T cells can be rendered resistant to Tat function engendered by a single amino acid substitution at position 261.

The CCNT1.C261Y modification potently inhibits diverse HIV-1 strains as well as HIV-2 and SIV agm
Having confirmed a Tat-specific block using viral reporter assays, we next examined the effects of C261Y on bona fide HIV-1 in cell cultures, under conditions of continuous virus replication and spread.We first infected each cell line with either of two CXCR4 (X4)tropic HIV-1 strains, NL4-3 or IIIB, at a low multiplicities of infection (MOI 0.5) and maintained the cultures for 10 days to allow for multiple rounds of virus replication.After 10 days, compared to parental cells, there was no evidence of infectious virion produc tion in infected CCNT1.C261Y cultures based on measuring levels of viral p24 Gag protein and infectivity in the supernatants, correlating with a striking, commensurate lack of viral RNA expression (Fig. 4A, compare panels ii-v and viii and compare panels iii-vi and ix, and Fig. 4B).To test if there was potential for viral escape, we performed infections with a higher viral dose (HIV-1.NL4-3, MOI ~10) and maintained the cultures for 30 days to allow greater time for viral adaptation and outgrowth.No signs of virus amplification following initial infection were evident, confirming a strong and persistent block (Fig. 4C).Because most transmitted strains of HIV-1 are R5-tropic, we also tested transmitted/founder strains HIV-1.CH106.c/2633,HIV-1.RHPA.c/2635,and HIV-1.SUMA.c/2821 in CCNT1 WT and C261Y lines engineered to stably express CCR5 (Fig. 4D) (56).Again, compared to the R5-positive parental cells, we observed no evidence of virus replication in R5-positive CCNT1.C261Y cells monitored over 20 days post infection.
Because primate lentiviruses code for divergent Tat proteins that similarly drive transcription elongation (57), we tested if CCNT1.C261Y cells would also inhibit transcrip tion using validated GFP-encoded viral vector systems derived from HIV-1, HIV-2, and simian immunodeficiency virus from the African green monkey (SIV agm ) (58,59) while also including the Tat-independent murine leukemia virus (MLV) as a control.For each vector except for MLV, GFP fluorescence was significantly diminished in the CCNT1.C261Y cell lines compared to the parental cells, consistent with the notion that most if not all primate lentiviruses rely on CCNT1 C261 to achieve efficient transcription (Fig. 4E with quantitative comparisons shown in Fig. 4F).Taken together, the C261Y substitution appeared to be sufficient to achieve remarkably broad-spectrum resistance to primate lentiviral gene expression.

Infected CCNT1.C261Y cells establish an intrinsic "block-and-lock" provirus inactivation scenario
Given the canonical role of CCNT1 residue C261 in HIV-1 infection, we hypothesized that CCNT1.C261Y cells would be susceptible to HIV-1 infection but would yield a forced latency phenotype downstream of proviral integration due to Tat inactivation.To test this hypothesis, we infected WT and C261Y cell lines with an established latency-tracking HIV-1 reporter virus, R7GeMC, that encodes two reporter cassettes engineered into the nef-coding position: an egfp gene responsive to the viral LTR promoter and Tat and a mcherry gene under control of a constitutively-active EF1α promoter (Fig. 5A; assay depiction in Fig. 5B) (60).At 24 h post infection, we observed equivalent numbers of mCherry-positive (~15%) cells for both the CCNT1.C261Y and parental control cells, confirming that HIV-1 entry is not altered in CCNT1.C261Y cell lines and that the antiviral effects occur post-integration (Fig. 5C).
A recently described HIV-1 functional cure strategy termed "block-and-lock" aims to use drugs to silence HIV-1 Tat-driven gene expression ("block") long enough for cell epigenetic modifications to progressively restrict the capacity of HIV-1 to reactivate ("lock") should drug therapy be discontinued (61).To test the hypothesis that, should CCNT1.C261Y cells become infected, they would effectively represent a cell-intrinsic "block-and-lock" scenario, we tested the responsiveness of R7GeMC-infected WT and C261Y cells to latency reversal agents that act by potentiating basal, Tat-independent viral gene expression.One day following R7GeMC infection, cultures were washed to remove input virus and treated with the LRAs phorbol 12-myristate 13-acetate (PMA; 10 ng mL −1 ), tumor necrosis factor alpha (TNFα; 10 ng mL −1 ), or with DMSO as a vehicle control.In the absence of LRAs, the wild-type cell population featured >28-fold more eGFP-positive cells relative to either infected CCNT1.C261Y cell line, consistent with CCNT1.C261Y cells harboring a significantly greater proportion of inactive proviruses (Fig. 5D, compare "active" gate in panels i-iv and vii; quantification in Fig. 5E; conversely, compare "inactive" gate in panels i-iv and vii; quantification in Fig. 5F).PMA and TNFα treatment increased the proportion of mCherry-and eGFP-positive CCNT1.C261Y cells (compare "active" gate in panels iv-v and vi and panels vii-viii and ix; quantification in Fig. 5E).However, the net viral gene expression was highly attenuated relative to untreated wild-type cells (compare "active" gate in panel ii-v and viii; and iii to vi and ix; quantification in Fig. 5E through G).In sum, the CCNT1.C261Y modification enforced a state of viral latency in human T cells consistent with the inhibition of HIV-1 Tat and severely restricted the potential for viral reactivation in the presence of LRAs, collectively achieving the defining characteristics of a "block-and-lock" scenario even at early time points post-infection and in the absence of suppressive agents.

Editing CCNT1.C261Y has no discernible effects on p-TEFb activity or major effects on transcription in HIV-1-negative cells
Finally, the capacity of CCNT1.C261Y cells to achieve cell-intrinsic HIV-1 restriction in the absence of apparent effects on cell viability prompted us to investigate the potential for impacts of the CCNT1.C261Y substitution on p-TEFb complex formation or activity in the context of the cellular host (and independent of HIV-1 infection).The p-TEFb complex is conserved in eukaryotes and regulates transcription elongation for a substantial proportion of RNA polymerase II-dependent genes (62)(63)(64).While an important role for CCNT1 residue 261 in HIV-1 transcription was obvious, any potential effects of targeting C261 on cellular gene expression would need to be considered should this residue/ region be considered in the context of an antiviral strategy.
To do so, we first compared CCNT1-CDK9 interactions in WT and CCNT1.C261Y cells using co-immunoprecipitation (co-IP) analysis.Comparable amounts of CDK9 were pulled down by CCNT1 or CCNT1.C261Y from respective cell lysates, indicating that p-TEFb complexes were unperturbed by the CCNT1 C261Y substitution (Fig. 6A, represen tative blot in Fig. 6B).Previous studies had demonstrated that treatment of cells with an RNA polymerase II inhibitor, 5,6-dichlorozo-1-β-D-ribofuranosylbenzimidazole (DRB), leads to a shift of p-TEFb from a high molecular weight complex to a low molecular weight complex, reflecting the release of p-TEFb from its transcriptionally repressive 7SK RNA scaffold (65,66).Upon DRB-treatment, we observed an induction of p-TEFb in CCNT1.C261Y cells (i.e., a decrease in the inactive form of p-TEFb and increase in the active form) comparable to that observed in parental cells (Fig. 6C, representative blot in Fig. 6D).Thus, p-TEFb responses to signal-induced stimuli appeared to be intact in CCNT1.C261Y cells.
To employ a more global approach, we next sequenced and compared the transcrip tomes of uninfected parental cells to those of either the C1 or C2 CCNT1.C261Y cell clones, performing differential gene expression (DGE) analysis using DESeq2 with a low stringency threshold (adjusted P-value, P* ≤ 0.05, baseMean count value of ≥5).This analysis indicated that 12.7% of differentially expressed genes (DEGs) were poten tially linked to the C261Y substitution.However, the magnitude of these differences was very low (typically <twofold) compared to the hyperstimulatory effects of C261 on Tat-dependent gene expression (typically >100-fold) observed in our replication experiments (Fig. 6F).These data underscored the relative specificity of CCNT1.C261 in regulating viral versus cellular transcription.
Analogous to HIV-1 Tat, the cellular bromodomain-containing protein 4 (BRD4) recruits p-TEFb to cellular promoters to exert control at a transcription elongation step (Fig. 7A) (67), with Tat thought to compete with BRD4 for p-TEFb binding during HIV-1 infection (67,68).Therefore, we surveyed the published literature to compile lists of experimentally defined genes reported to be regulated by either CCNT1 (69)(70)(71) or BRD4 (72)(73)(74)(75)(76)(77)(78)(79) and asked if either of these gene sets were more likely to be linked to the C261Y modification (Fig. 7B), using the same low stringency statistical threshold described above.Based on these criteria, only two potential CCNT1-associated genes (DTX4 and PCOLCE2) and two BRD4-associated genes (MCTP2 and IL15RA) were differentially expressed greater than twofold in both CCNT1.C261Y cell clones.Accord ingly, the vast majority of genes associated with BRD4 or CCNT1 were unaffected in CCNT1.C261Y cells, further supporting the conclusion that, unlike Tat-dependent viral gene activation, cellular gene regulation in Jurkat T cells is largely unaffected by CCNT1 C261Y modification.

DISCUSSION
In this study, we engineered human T cells using CRISPR/Cas9 genome editing to test the hypothesis that minor and "rodent-inspired" gene modifications can create a human T cell environment exhibiting intrinsic resistance to HIV-1 early and/or late-stage viral gene expression.We provide proof of concept for manifesting a block to Tat-dependent HIV-1 gene expression in human Jurkat T cells by editing a relevant species-specific locus (C261) encoded within the endogenous CCNT1 gene.Furthermore, we validated the strength and specificity of the Tat block by isolating two HIV-1-refractory cell clones (C1 and C2) that contained prescribed, homozygous knock-in edits encoding a single nonsynonymous mutation in endogenous CCNT1 resulting in the C261Y substitution.
Because our study relied heavily on measuring changes to HIV-1 Tat-and/or Revdependent gene expression in infected cell populations, we created the E-R-/2FP dual reporter assay to rapidly detect and measure defects to Tat-and Rev-dependent gene expression.In Fig. 1, we validated this assay by screening non-human mammalian cell lines predicted to pose genetic blocks to HIV-1 Tat and Rev activity (38,80).As expected, rodent cell lines were largely refractory to HIV-1 Tat and/or Rev function.However, that Rat1 and CHO cells exhibited substantial early, Rev-independent viral gene expression was surprising considering that they encode the non-permissive CCNT1 tyrosine residue at the 261-equivalent position.The competency for early viral gene expression in these cell lines raises the possibility that there are other cell-intrinsic factors that overcome or offset the HIV-refractory effects of the Y261 allele, at least in some cell types.Cat (CRFK) cells are also of interest to us going forward because they supported Rev function despite the cat XPO1 "patch" domain differing from that of the human ortholog at two of the three known relevant species-specific residues.
Recent advances in precision CRISPR/Cas9 genome editing provided us with the means to investigate the potential to introduce species-specific Tat and Rev blocks into human T cells.Unlike HIV-refractory ∆32 CCR5 alleles, the C261Y and P411T/M412V/ F414S substitutions we engineered had not been documented in primates.Thus, at the outset, it was unclear whether human cells would tolerate these changes.Indeed, CCNT1 and XPO1 have only experienced modest changes throughout mammalian speciation, with the mouse and human orthologs of CCNT1 and XPO1 exhibiting 89% and 98% amino acid identity, respectively.Considering such strong negative selection, the significance of these HIV-relevant species-specific residues to basic cellular functions (if any) remains unknown.Using our CRISPR/Cas9 knock-in designs, we successfully achieved targeted editing of CCNT1, observing viral gene repression consistent with disruption of Tat function in viable cell populations.We also readily isolated seamlessly edited cell clones encoding exclusively CCNT1.C261Y, allowing us to confirm unequivo cally that the modification is tolerated in this context.However, attempts to isolate comparably edited homozygous XPO1.P411T/M412V/F414S cells were not successful, suggesting that these modifications are deleterious.
Using the CCNT1.C261Y cell clones, we confirmed that the C261Y modification imposed a severe block specific to Tat-dependent viral gene expression rescued by exogenous delivery of wild-type CCNT1 and potently inhibited spreading HIV-1 replication (Fig. 3 and 4).Given the apparent tolerance and anti-HIV effects of the C261Y substitution in human T cells, we considered this knock-in approach in the context of a "block-and-lock" functional cure scenario (81), employing a latency reporter system (60) to demonstrate that the CCNT1.C261Y cells remain susceptible to HIV-1 infection but are markedly refractory to drug-induced viral reactivation (Fig. 5).Akin to this block, a recently described small molecule, didehydro-cortistatin A (dCA) has shown promise as a viral transcription inhibitor disrupting Tat/TAR interactions (82).Valente and colleagues recently isolated dCA-resistant HIV-1 clones with dCA resistance mapped to mutations in the viral LTR, nef or vpr genes; all of which were suggested to promote basal, Tat-inde pendent transcription of viruses that then lost their ability to enter viral latency (83).To date, we have no indication of viral resistance to the CCNT1 C261Y modification in virus replication assays (Fig. 4).As such, we propose that CCNT1.C261Y cell lines will serve as a useful platform for longitudinal studies of latency in the context of HIV-1 LTR silencing and the net potential for LRA-driven, Tat-independent infectious virion production (84).
In Fig. 6, we showed that human T cells bearing CCNT1.C261Y limit the activity of not only HIV-1 Tat but also the Tat proteins of HIV-2 and SIV.These results empha size how fundamental the Tat-C261 interaction must be to the replicative success of primate lentiviruses and suggest that any intervention strategy that targets the crucial CCNT1.C261 residue will be broad-spectrum, likely affecting all HIV-1 and HIV-2 strains and subtypes.Moreover, because this mechanism extends to SIV, the block may be testable in non-human primate model systems.Interestingly, a subset of non-primate lentiviruses is thought to be less dependent on CCNT1 C261 (e.g., bovine immuno deficiency virus) (85,86) so that additional comparative structure-function studies of Tat-CCNT1 interactions seem warranted.
Aside from its crucial role in HIV-1 Tat function, little is known regarding the relevance, if any, of CCNT1 residue 261 to cellular processes; a key consideration should this protein domain be targeted in the context of antiviral strategies.CCNT1.C261Y clones exhibited no significant changes to rates of cell proliferation or basal CCNT1 expression (Fig. 5), confirming that the C261Y modification is not profoundly detrimental, at least in cell lines.In Fig. 6, we further investigated the C261Y modification with experiments designed to interrogate endogenous p-TEFb complexes and their response to stimuli and performed a comprehensive transcriptomic analysis of CCNT1.C261Y cell lines.CDK9 co-IP experiments indicated that the steady-state levels and interactions between CCNT1 and CDK9 were unperturbed relative to the parental cells.CCNT1 knock-out cells and mice depleted of CCNT1 have been reported with notable changes to the expres sion of CDK9 and other transcription factors (87,88).However, our glycerol gradient analysis indicated similar expression of CDK9 for both wild-type CCNT1 and CCNT1.C261Y cells, as well as successful release of CDK9 from high-molecular-weight complexes in both cell backgrounds in response to DRB treatment (65,66).Further, transcriptomic analysis using RNA-seq demonstrated that the majority of CCNT1-and BRD4-dependent genes were unperturbed in the CCNT1.C261Y background.Taken together, the effects of CCNT1.C261Y in these cell lines were largely virus-targeted, being far more severe on Tat-dependent gene expression relative to any subset of cellular genes.That said, potential functional roles for CCNT1 C261 in primary human lymphocytes and in vivo, if any, are pending investigation.
It remains highly interesting to us and unexplained why rodents code for variation in these otherwise conserved CCNT1 and XPO1 genes, with these variations prohibiting HIV-1 gene expression in a synergistic fashion.The functional consequences of this constellation of polymorphisms on modern lentiviruses including not only HIV-1 but also HIV-2 and SIV agm (Fig. 4) are remarkable.We propose that harnessing these species-spe cific polymorphisms to engineer cell-intrinsic resistance barriers to Tat (and potentially Rev, if not both in combination) in human T cells will provide targeted means to study these selective pressures and should also be considered in the context of functionalcure-based antiviral strategies.

Cell culture
Human T lymphocyte (Jurkat E6-1), human embryonic kidney (293T), human cervical carcinoma (HeLa), mouse fibroblast (3T3), and CHO cell lines were obtained from the American Type Culture Collection (ATCC).African green monkey osteosarcoma (Cos7), chicken fibroblast (DF-1), cat kidney (CRFK), mouse fibroblast (L cells, and M. dunni), and mouse myoblast (C2C12) cell lines were a kind gift of Dr. Michael Malim (King's College London).The Rat fibroblast (Rat1) cell line was a kind gift of Dr. Robert Kalejta (University of Wisconsin-Madison).The chimpanzee skin fibroblast (WES) cell line was a kind gift of Dr. Beatrice Hahn (University of Pennsylvania).The TZM-bl reporter cell line was a kind gift of Dr. John Kappes and Dr. Xiaoyun Wu and obtained from the HIV Reagent Program (NIH).CHO, Jurkat, and Jurkat-derived cell lines were cultured in Roswell Park Memorial Institute (RPMI) 1640 supplemented to 10% fetal bovine serum (FBS), 1% penicillin-streptomycin, and 1% L-glutamine (Sigma).All other established cell lines were cultured in Dulbecco's Modified Eagle's Medium supplemented to 10% FBS, 1% L-glutamine, and 1% penicillin-streptomycin. Cell cultures were maintained at 37°C and 5% CO 2 in a humidified incubator.

Plasmids
The E-R-/2FP dual reporter virus construct is derived from the E-R-/mCherry construct described elsewhere (89).Using overlapping PCR and standard molecular cloning procedures, cDNA sequences encoding the mVenus fluorescent protein with flanking glycine-rich linker segment at the 5′-(GGSGGTR) and 3′-(TRGGSGG) ends were fused in frame with the HIV-1 gag sequence between matrix and capsid coding regions (90).E-R-Tat-/2FP was generated by overlapping PCR using the primers 5′-CCT AGA CTA GAG CCC TGG AAG CAT CCA TGA AGT TAA CCT AAA ACT G-3′ and 5′-GGC TCT AGT CTA GGA TCT ACT GGC TCC GTT TCT TGC TC-3′.E-R-Rev-/2FP was generated by subcloning the modified Rev coding region from E-R-Rev-/YFP described elsewhere (91).The lentiviral packaging plasmid psPax2 was a kind gift from Dr. Didier Trono (plasmid #12260, Addgene).The pVSV-G expression construct is described elsewhere (92).Retroviral vector plasmids encoding CCNT1 variants were generated by subclon ing modified CCNT1 cDNA sequences fused to cDNA encoding the FLAG-epitope tag (DYKDDDDK) into the pELE88 vector (80) using BglII and NotI restriction enzyme sites.Retroviral vector plasmids encoding GFP or CCR5 were generated by subcloning cDNAs into MIGR1-derived vectors (93) using vector-encoded BglII and XhoI restriction enzyme sites.The retroviral packaging plasmid pMD.GagPol was a kind gift from Dr. Richard Mulligan.For the Tat/LTR reporter assay, the construction of the gaussia luciferase (gLuc) HIV-1 LTR reporter construct is described elsewhere (55).The pEV280 Tat-FLAG expression construct was a kind gift of Dr. Melanie Ott.Additional plasmids used for experimental controls include pmCherry-C1 (Clontech Laboratories Inc., Mountain View, CA, USA), cpTK-Cluc (New England Biolabs, Ipswich, MA, USA), pBluescript (Stratagene, La Jolla, CA, USA), and pSEAP (Clontech Laboratories Inc.).Plasmids encoding the infectious molecular clones of HIV-1, pCH106.c/2633,pRHPA.c/2635,and pSUMA.c/2821were kind gifts of Dr. John Kappes and Dr. Christina Ochsenbauer and obtained from the HIV Reagent Program (NIH).The R7GEmC reporter virus (60) was a kind gift of Dr. Vincenzo Calvanese and Dr. Eric Verdin and was obtained from the HIV Reagent Program (NIH).The GFP-encoding and env-deficient HIV-1.NL4-3, HIV-2.ROD, and SIV agm .Tan-1 reporter virus clones are described elsewhere (58,59) and were a kind gift of Dr. Paul Bieniasz.

E-R-/2FP infection assays
For infections with suspension cell lines, 1 × 10 6 cells were plated in each well of a 6-well dish and infected immediately with equivalent doses of E-R-/2FP virus stock.For infections with adherent cell lines, the indicated cell line was plated to 30%-50% confluency in each well of a 6-well dish and infected 24 h post plating as above.Nevirapine stock was obtained from the HIV Reagent Program (NIH) and added immediately prior to infection to a final concentration of 1 µM where indicated.Fortyeight hours post infection, cells were harvested, washed in PBS, fixed with 4% reconstitu ted paraformaldehyde, and analyzed by flow cytometry.Flow cytometry was performed using either a LSRII (BD Biosciences, Mississauga, ON, Canada) or an Attune NxT (Thermo Fisher Scientific, Waltham, MA, USA).Flow cytometry data analysis was performed using FlowJo software (FlowJo, LLC, Ashland, OR, USA).For CCNT1 rescue experiments, 5 × 10 5 cells were transduced by spinoculation (94) with retroviral vectors encoding CCNT1, CCNT1.C261Y, or vector only control in the presence of 5 µg/mL polybrene (Sigma) prior to plating.Following transduction, cell cultures were diluted 1:1 with fresh culture medium volume, plated in 6-well tissue culture-treated dishes, and cultured for 3 days.Cells were then infected, harvested, and analyzed as described above.

E-R-/2FP FACS analysis
For infections of heterogeneous populations of CRISPR-treated Jurkat cells, equivalent numbers of unfixed and infected cells from each gate (>10,000 per gate; gate scheme shown in Fig. 2C) were analyzed by flow cytometry and sorted by FACS using a FACSAria cell sorter (BD Biosciences) instrument.Bulk-sorted cells were screened for CRISPR knock-in as described above.Corresponding flow cytometry measurements were analyzed as described above.

SDS-PAGE and immunoblotting
To measure endogenous CCNT1 protein levels, equivalent numbers of each Jurkat cell line were pelleted, washed with PBS, and lysed using radioimmunoprecipitation assay buffer (10 mM Tris-HCl [pH 7.5], 150 mM NaCl, 1 mM EDTA, 0.1% sodium dodecyl sulfate [SDS], 1% Triton X-100, 1% sodium deoxycholate) containing complete protease inhibitor cocktail (Roche).Cell lysates were prepared for immunoblot by sonication and clarified supernatants were boiled with 2× dissociation buffer (62.5 mM Tris-HCl [pH 6.8], 10% glycerol, 2% SDS, 10% β-mercaptoethanol) at a 1:1 ratio prior to polyacryla mide gel electrophoresis (PAGE) and immunoblot.Prepared cell lysates were subjected to SDS-PAGE using 10% polyacrylamide gels and a conventional Tris-glycine buffering system.Following electrophoresis, resolved lysates were transferred to 0.2 µm pore size nitrocellulose membranes (GE Healthcare) for immunoblot.Nitrocellulose membranes were blocked in blocking solution (PBS containing 0.1% Tween-20 (PBS-T) and recon stituted non-fat dry milk at a 2% final concentration).Following incubation in the blocking solution, membranes were incubated in primary antibody solutions consisting of fresh blocking solution containing CCNT1 antisera (Santa Cruz Biosciences, Dallas, TX, USA, sc-398695; Cell Signaling Technologies, Danvers, MA, USA D1B6G), CDK9 antisera (Bethyl Laboratories, Montgomery, TX, USA, A303-493A), or HSP90 antisera (Santa Cruz Biosciences, sc-7947; sc-13119).Following incubation in primary antibody solution, membranes were washed three times with PBS-T and incubated in secondary antibody solutions consisting of fresh blocking solution containing either anti-mouse or anti-rab bit secondary antisera conjugated to either IRDye680 or IRDye800 (LI-COR Biosciences, Lincoln, NE, USA) infrared fluorophores.Following incubation in secondary antibody solution, membranes were washed three times in PBS-T and analyzed by quantitative immunoblot using an Odyssey Fc instrument (LI-COR Biosciences).

Cell proliferation assay
A total of 5 × 10 4 cells from each Jurkat line was plated in each well of a 12-well tissue culture treated dish containing 1 mL of cell culture medium.At 4 and 6 days post plating, cultures for each cell line were resuspended, diluted, and mixed at 1:1 concentration with a 0.4% Trypan blue (Sigma) solution.Cells lacking Trypan blue uptake were counted manually using a hemocytometer.

Transient HIV-1 LTR reporter assays
For each Jurkat cell line, 5 × 10 5 cells were transfected using the Neon electroporation system using the electroporation parameters described above per the manufacturer's instructions.Plasmid mixtures contained 75 ng of HIV-1 LTR reporter constructs gaussia luciferase (gLuc) and nanoluciferase (nLuc, Promega), 250 ng pmCherry-C1 plasmid, 250 ng pTK-cLuc plasmid, 1,200 ng CCNT1 or control plasmid (pBluescript or pSEAP), and 25 ng pEV280 Tat expression plasmid or control carrier DNA.Total DNA concentrations were maintained at 2,500 ng by vector control plasmids or calf thymus DNA.Twenty-four hours post transfection, 10 µL culture media was removed, diluted with 40 µL PBS, and assayed for secreted gaussia luciferase (gLuc) activity by injecting 30 µL coelenterazine solution (Promega) followed by 1.6 s incubation period and a luminescence integration read time of 1.0 s.Secreted cypridina luciferase (cLuc) activity from the internal control plasmid was similarly measured using the cypridina luciferase kit (New England Biolabs) according to the manufacturer's instructions.Fold activation was calculated as the ratio of gLuc:cLuc signals in each well compared to the average of the pBluescript and pSEAP control transfections.

HIV-1 infection assays
A total of 2.5 × 10 6 cells from each Jurkat cell line was plated in each well of a 48-well plate and infected with HIV variants, NL4-3 or IIIB, at a multiplicity of infection of 0.5.The following day, the inoculum was removed, cells were washed twice with PBS and then replenished with fresh culture medium, and maintained at normal culture conditions.At 4 and 10 days post infection, cells and supernatant were collected for viral RNA quantification, viral outgrowth assays, and flow cytometry staining analysis.

Viral RNA quantification
Viral RNA measurements were performed using the method described previously (95).From 100 µL of supernatant collected at each time point, viral RNA was extracted using the QIAamp viral RNA mini kit (Qiagen, Hilden, Germany) following the manufacturer's protocol.SYBR green real-time PCR assay was carried out in a 20-µL PCR mixture volume consisting of 10 µL of 2× Quantitect SYBR green RT-PCR Master Mix (Qiagen) containing HotStarTaq DNA polymerase, 0.5 µL of 500 nM each oligonucleotide primer, 0.2 µL of 100× QuantiTect RT Mix (containing Omniscript and Sensiscript RTs), and 8 µL of RNA extracted from samples or standard HIV-1 RNA (from 5 × 10 5 to 5 copies per 1 mL).Highly conserved sequences on the gag region of HIV-1 were chosen, and specific HIV-1 gag primers were selected.The sequences of HIV-1 gag primers are 5′-CAA TGG CAG CAA TTT CAC CA-3′ and 5′-GAA TGC CAA ATT CCT GCT TGA-3′.Amplification was done in an Applied Biosystems 7,500 real-time PCR system (Applied Biosystems, Waltham, MA, USA), and it involved activation at 45°C for 15 min and 95°C for 15 min followed by 40 amplification cycles of 95°C for 15 s, 60°C for 15 s, and 72°C for 30 s.For the detection and quantification of viral RNA, the real-time PCR of each sample was compared with threshold cycle value of a standard curve.

Viral outgrowth assay in TZM-bl cells
Twenty-four hours prior to infection, 20,000 TZM-bl cells were plated in each well of a 48-well tissue culture-treated plate.The cells were incubated with 10 µL superna tant collected from infected Jurkat cell cultures for overnight incubation at 37°C.The following day, infected cells were washed and fresh culture media was added.Forty-eight hours post infection, luciferase activity in cell lysates was measured using a luciferase assay kit, following the manufacturer's protocol (Promega).
p24 Gag staining HIV-infected cells for each Jurkat cell lines were permeabilized using the Cytofix/Cytoperm Fixation/Permeabilization Kit (BD Biosciences) and stained intracellularly using PE-conjugated mouse anti-p24 mAb (clone KC57; Beckman Coulter, Brea, CA, USA, 1:100 dilution).Flow cytometry measurements were performed using a FACSCalibur (BD Biosciences) and data analysis was performed as above.

HIV-1 replication curves
A total of 5 × 10 6 cells from each Jurkat cell line was infected with 10 ng p24 Gag -equiv alent volumes for the indicated HIV-1 variants in 6 mL antibiotic-replete RPMI.Twentyfour hours post infection, cultures were pelleted, washed three times with 1× PBS, and resuspended in fresh culture medium.Each 6 mL culture was maintained under normal culture conditions and split by removal of 4 mL culture (media containing cells) and refreshed with an equal volume of fresh media every 3 days.Where indicated and prior to infection, cells were transduced with a CCR5-encoding retroviral vector to confer susceptibility R5-tropic HIV-1 variants.Harvested cultures were processed by 0.45 μm filtration to remove cells and debris and aliquots were stored at −80°C.As previously described, p24 Gag levels were quantified by ELISA periodically (every 1-2 weeks) to monitor supernatant virus particle levels over time.

R7GeMC infection assays
A total of 1 × 10 6 cells was plated in each well of a 6-well dish and infected immediately with equivalent doses of R7GeMC virus stock.Twenty-four hours post infection, cells were pelleted and washed thrice in PBS and then resuspended in culture medium containing phorbol 12-myristate 13-acetate (PMA, Sigma) or tumor necrosis factor alpha (EMD Millipore, 80054-834) at 10 ng mL −1 or vehicle control (DMSO) where indicated.Two days post treatment, cells were harvested, processed, and analyzed by flow cytometry as described above.

Retroviral GFP reporter assays
A total of 1 × 10 6 cells was plated in each well of a 6-well dish and infected immediately with equivalent doses of the indicated GFP-encoding reporter virus stock.Twenty-four hours post infection, cells were harvested, processed, and analyzed by flow cytometry as described above.

Co-immunoprecipitation
An amount of 5 × 10 6 Tat-expressing cells for each Jurkat cell line was lysed in 500 µL ice cold Tris buffered saline NP-40 (TBSN) (150 mM NaCl, 50 mM TRIS [pH 7.5], 1 mM EDTA, 2 mM β-mercaptoethanol, 0.5% NP40, 1× HALT protease inhibitor cocktail [Invitrogen, Carlsbad, CA]) buffer.Cells were lysed on ice for 10 min, nuclei were pelleted (20,000 × I for 10 min at 4°C), and supernatant was transferred to a new tube.Fifty microliters (10%) of the input material was removed for SDS-PAGE.The remaining sample was immunoprecipitated with anti-CCNT1 antibody (Bethyl Laboratories, A303-499A) or non-specific normal rabbit IgG (Cell Signaling Technologies, 2,729S) in the presence of 2,000 gel units micrococcal nuclease (New England Biolabs) for 1 h.Fifty microliters of magnetic protein A/G beads (Pierce, 88803) was added, and the samples were incubated for an additional hour.The samples were washed thrice in TBSN, and the beads were resuspended in 50 µL SDS-PAGE loading buffer.The samples were boiled, and equivalent volumes of input and IP elution were separated by SDS-PAGE and immunoblotted with mouse anti-CDK9 (Santa Cruz Biosciences, sc-13130) and mouse anti-CCNT1 (Santa Cruz Biosciences, sc-271348) as described above.

Transcriptome analysis
For each Jurkat cell line, 1.5 × 10 6 cells in 10 mL fresh culture medium were plated in T75 tissue culture-treated flasks and maintained under normal culture conditions.Forty-eight hours post plating, cells were pelleted from suspension (400 × g for 5 min at 4°C) and delicately resuspended in 1 mL chilled PBS and transferred to 1.5 mL microcentrifuge tube.Cell suspensions were pelleted (400 × g for 5 min at 4°C), and clarified PBS was aspirated.Cell pellets were stored at −80°C.Total RNA was extracted from cell pellets of three biological replicates (i.e., time-separated cultures) using RNeasy Plus Mini Kit (Qiagen) with genomic DNA eliminator columns and resuspended in nuclease-free water according to the manufacturer's instructions.RNA extracts were stored at −80°C prior to submission to the University of Wisconsin-Madison Biotechnology Center for quality control assessment, library preparation, and sequencing.RNA sample concentrations and quality assessments were measured using a NanoDrop One spectrophotometer (software version 1.4.0) and an Agilent 2100 Bioanalyzer (G2939A; firmware: C.01.069; Assay: Eukaryote Total RNA Nano).Poly(A)-enriched mRNA-sequencing libraries were prepared using an Illumina TruSeq stranded mRNA Library Prep Kit (Illumina, San Diego, CA, USA) according to the manufacturer's instructions.Samples were loaded onto a MiSeq NanoCell (Illumina), and paired-end RNA sequencing was performed on a NovaSeq 6,000 sequencing platform (Illumina).Sequencing was carried out with a read depth target of approximately 30 million reads per sample.Sequencing reads were mapped to human genome assembly hg38 using HISAT2 (96) (version 2.0.5).Annotated feature count tables were prepared using the htseq-count tool within HTSeq (97) (version 0.6.0).DGE analyses were performed using DESeq2 (version 1.26.0)(98).Computational resources for RNA sequencing analysis were kindly provided by the Bioinformatics Resource Center at the University of Wisconsin-Madison. the Malim (King's College London), Bieniasz (Rockefeller University), Verdin (University of California-Los Angeles), Kappes (University of Alabama), Hahn (University of Pennsyl vania), and Kalejta (University of Wisconsin-Madison) laboratories for valuable contribu tions.

FIG 1
FIG 1 Development of a dual-reporter assay to screen for blocks to early-stage (Tat-driven) and late-stage (Rev-driven) HIV-1 gene expression.(A) Viral genome diagram of the two-color, HIV-1.NL4-3-derived viral reporter construct, E-R-/2FP.Early, Rev-independent viral gene expression is reported by mCherry, encoded within the nef locus.Late, Rev-dependent viral gene expression is reported by mVenus, encoded in gag as an in-frame insertion between the matrix (MA) and capsid (CA) protein-coding regions.Gray rectangles denote inactivated env, vpr, and nef genes to limit virus replication to a single cycle.(B) E-R-/2FP reporter assay schematic.VSV-G pseudotyped virus particles containing E-R-/2FP genomes are delivered to target cells.Prior to infection, target cells (gray) can become mVenus-positive due to virion-associated MA-mVenus during virus particle uptake (gold).Following proviral insertion, early (red) and late (orange) viral gene expression is reported by de novo mCherry and mVenus expression, respectively.(C) Fluorescence profiles of Jurkat E6-1 cells transduced with wild type (WT), Tat-or Rev-deficient E-R-/2FP, or untransduced control cells.Cells exhibiting robust early and late viral gene expression are indicated (dashed circle).Dot plot is an overlay of the four indicated cell populations.(D) Species-specific CCNT1 or XPO1 residues that govern HIV-1 Tat and Rev activity, respectively (blue).CCNT1 residue 261 governs Tat activity.XPO1 residues 411, 412, and 414 govern Rev activity.Rodents predicted from species-specific residue identity to exhibit poor Tat and Rev activities are indicated (boxed).(E) E-R-/2FP screen of various mammalian cell lines.Dashed circles highlight cells exhibiting robust early and late viral gene expression.Nevirapine treatment to block E-R-/2FP infection was to facilitate the identification of productively infected cell populations.

FIG 2
FIG 2 CRISPR/Cas9 knock-in strategy to abrogate Tat or Rev function in human T cells.(A) Design schematic for CCNT1 gene knock-in editing.Approximate sequence lengths of the donor template (pink) and PCR amplicon (black) were used to analyze the genomic locus.DNA sequences relevant to the edit design are shown; the PAM (bold) and guide RNA target (underlined) specify the Cas9 cleavage site (arrows).Mutations engineered into the donor template are color-coded according to their intended use.CCNT1 protein-coding sequences are shown with substitutions color-coded to match the relevant DNA mutation.(B) Design schematic for XPO1 gene knock-in editing labeled as in Fig. 2A.(C) FACS gating legend for sorted populations P1-P7.Gates P1-P4 contain cells exhibiting early, Tat-dependent viral gene expression.Gates P5-P7 contain cells exhibiting late, Tat-/Rev-dependent viral gene expression.(D) Fluorescence profiles of CCNT1or XPO1-edited or control cells infected with E-R-/2FP with (bottom row) or without (top row) P1-P7 gate overlays.P411T, M412V, and F414S substitutions collectively termed "mPatch." (E) Percentage of cells inoculated with E-R-/2FP within gates P1-P7 for the CCNT1-(pink), XPO1-(orange), and control-edited (black) cell populations.Error bars represent the standard deviation from the mean for three independent experiments.A two-tailed Student's t-test with Welch's correction was used to compare the indicated populations (*P ≤ 0.05).(F) Sensitivity of PCR amplicons from E-R-/2FP-inoculated, CCNT1-edited or control cells before and after FACS.PCR amplicons (top row) were digested with BsiWI (middle row) or SacI (bottom row) to screen for amplicons containing gene knock-in for each sorted population (P1-P7, lanes 8-14 and 17-23).Unsorted E-R-/2FP-inoculated or no infection control cell populations (US, lanes 1-5) were analyzed to demonstrate input amplicon sensitivities to the indicated restriction enzymes.(G) Sensitivity of PCR amplicons from E-R-/2FP-inoculated XPO1-edited or control cells before and after FACS.PCR amplicons (top row) were digested with PvuII (middle row) or XbaI (bottom row) to screen for amplicons containing gene knock-in for each sorted population (P1-P7, lanes 8-14 and 17-23).Unsorted E-R-/2FP-inoculated or no infection control cell populations (US, lanes 1-5) were analyzed to demonstrate input amplicon sensitivities to the indicated restriction enzymes.

FIG 3 7 FIG 4
FIG 3 Homozygous CCNT1.C261Y Jurkat T cells potently suppress HIV-1 Tat activity.(A) Analysis of the CCNT1 locus by PCR amplicon restriction digest for CCNT1.C261Y clones, C1 and C2, compared to WT parental cells.(B) Sequence analysis summary for CCNT1.C261Y cells.(C) Sequence read chromatograms for CCNT1.C261Y cells.Read peaks of loci targeted for mutation are color-coded by intended purpose.(D) Western blot analysis of CCNT1 protein levels in CCNT1.C261Y cells.(E) Cell proliferation analysis of CCNT1.C261Y cells.A two-tailed Student's t-test with Welch's correction was used to compare C1 or C2 to wild-type parental cells at the specified time point (ns, not significant).(F) Exogenous hCCNT1, but not hCCNT1.C261Y, restored the E-R-/2FP gene expression and the wild-type fluorescence profile in CCNT1.C261Y clones, C1 and C2.Vectors encoding wild-type hCCNT1 (middle column) or hCCNT1.C261Y (right column) or control (left column) were delivered to the parental Jurkat cell line (top row, panels i-iii) or CCNT1.C261Y clones C1 (middle row, panels iv-vi) and C2 (bottom row, panels vii-ix) then infected with E-R-/2FP virus (orange) and analyzed by flow cytometry.(G) Quantitative summary of late gene expression was presented in Fig. 3F.A two-tailed Student's t-test with Welch's correction was used to compare CCNT1.C261Y cells expressing exogenous hCCNT1 to those expressing hCCNT1.C261Y.(H) Exogenous hCCNT1, but not hCCNT1.C261Y, restores Tat activity in an infection-independent Tat/LTR-luciferase reporter assay.Plasmids encoding luciferase under control of the HIV-1 LTR and HIV-1 Tat were delivered in combination with either hCCNT1, hCCNT1.C261Y, or a vector control by transfection.A two-tailed Student's t-test with Welch's correction was used to compare CCNT1.C261Y cells expressing exogenous hCCNT1 to those expressing hCCNT1.C261Y.

FIG 5
FIG 5 CCNT1.C261Y can be infected by HIV-1 but effectively establish an intrinsic "block-and-lock" provirus inactivation scenario.(A) Virus genome diagram of R7GeMC, an established two-color reporter virus (60) designed to discern transcriptionally active or inactive HIV-1 proviruses.LTR-dependent viral gene expression is reported by eGFP.Inactive proviruses are reported by mCherry, encoded downstream of the constitutively active EF1α promoter.Gray rectangles denote inactivated env and nef genes to limit virus replication to a single cycle.(B) R7GeMC reporter assay schematic.Each cell line was inoculated with VSV-G pseudotyped R7GeMC.Twenty-four hours post inoculation, cells were sampled for fluorescence measurements and the remaining were washed to remove input virus.Cells were split into equivalent cultures and treated with HIV-1 latency reversal agents (LRAs).Seventy-two hours post inoculation, cells were harvested for fluorescence measurements.(C) Wild-type and CCNT1.C261Y cells are similarly susceptible to R7GeMC.Twenty-four hours post inoculation, eGFP-and mCherry-positive cells and mean fluorescence intensities were measured.A two-tailed Student's t-test with Welch's correction was used to compare measurements for C1 or C2 to wild type (ns, not significant).(D) CCNT1.C261Y cells inhibit integrated R7GeMC proviruses from escaping viral latency, even in the presence of LRAs.Overlaid dot plots of the four indicated cell populations harvested 72 h post inoculation.(E) Percentage of eGFP-positive infected cells, indicating cells exhibiting active, LTR-dependent viral gene expression (Fig. 5D, "active" gate percentages).(F) Percentage of infected mCherry-positive, but transcriptionally inactive eGFP-negative cells, indicating cells with integrated proviruses that lack LTR-dependent viral gene expression (Fig. 5D, "inactive" gate percentages).(G) eGFP mean fluorescence intensity (MFI) in the transcriptionally active infected cell populations (Fig. 5D, eGFP MFI in "active" gates).

FIG 6
FIG 6 Analysis of p-TEFb and cellular transcriptomes in CCNT1.C261Y cells.(A) CCNT1.C261Y does not alter CDK9-CCNT1 interactions.Protein complexes purified from lysates using CCNT1 antisera were analyzed for CDK9 protein levels.Recombinant IgG was used to control for immunoprecipitation (IP) specificity.CDK9 levels in IP complexes were measured by semi-quantitative western blot.Error bars represent the standard deviation from the mean for three independent experiments.(B) Representative western blot for data was presented in Fig. 6A.(C) CCNT1.C261Y does not change DRB-induced p-TEFb activation responses.Parental and CCNT1.C261Y clone C1 were cultured in the presence or absence of 10 µM DRB and lysed.Cellular lysates were subjected to glycerol gradient fractionation (5%-45%) and the levels of CCNT1 and CDK9 protein were measured by semi-quantitative western blot.Error bars represent the standard deviation from the mean for three independent experiments.(D) Representative western blot for data presented in Fig. 6C.(E) Venn diagram of the summarized differential gene expression (DGE) analysis comparing CCNT1.C261Y and wild-type transcriptomes (DESeq2).Total expressed genes (16,038) were defined by a baseMean count value of ≥5.Differentially expressed genes (DEGs, 5,168) were defined by an adjusted P-value (P*) ≤ 0.05.DEGs unique to C1 (1,274), C2 (1,851), or shared in both (2,043) are defined.DEGs that are up-regulated (1,100) or down-regulated (943) in both C1 and C2 compared to wild-type cells are defined.(F) Relative levels of cellular and Tat-dependent RNA in wild-type and C261Y cell lines.Relative RNA abundance (normalized read counts from DESeq2 analysis) for all expressed genes (16,038) in uninfected cells are plotted (C1, circles; C2, crosses).Tat-dependent HIV-1 transcript levels (RT-qPCR copy number; from Fig. 4B) are embedded for context (red, NL4-3; blue, IIIB).
Restriction enzyme digests of PCR amplicons were carried out following the manufacturer's recommended reaction composition with BsiWI or SacI for CCNT1 or PvuII or XbaI for XPO1.Predicted CCNT1 SacI digestion products: 804 and 196 bp.Predicted CCNT1 BsiWI digestion products: 712 and 288 bp.Predicted XPO1 XbaI digestion products: 679 and 298 bp.Predicted XPO1 PvuII digestion products: 497 and 480 bp.