HIV-1 Transcription Inhibitor 1E7-03 Decreases Nucleophosmin Phosphorylation

Transcription activation of latent human immunodeficiency virus-1 (HIV-1) occurs due to HIV-1 rebound, the interruption of combination antiretroviral therapy, or development of drug resistance. Thus, novel HIV-1 inhibitors, targeting HIV-1 transcription are needed. We previously developed an HIV-1 transcription inhibitor, 1E7-03, that binds to the noncatalytic RVxF-accommodating site of protein phosphatase 1 and inhibits HIV-1 replication in cultured cells and HIV-1–infected humanized mice by impeding protein phosphatase 1 interaction with HIV-1 Tat protein. However, host proteins and regulatory pathways targeted by 1E7-03 that contribute to its overall HIV-1 inhibitory activity remain to be identified. To address this issue, we performed label-free quantitative proteome and phosphoproteome analyses of noninfected and HIV-1–infected CEM T cells that were untreated or treated with 1E7-03. 1E7-03 significantly reprogramed the phosphorylation profile of proteins including PPARα/RXRα, TGF-β, and PKR pathways. Phosphorylation of nucleophosmin (NPM1) at Ser-125 residue in PPARα/RXRα pathway was significantly reduced (>20-fold, p = 1.37 × 10−9), followed by the reduced phosphorylation of transforming growth factor-beta 2 at Ser-46 (TGF-β2, >12-fold, p = 1.37 × 10−3). Downregulation of NPM1’s Ser-125 phosphorylation was further confirmed using Western blot. Phosphorylation mimicking NPM1 S125D mutant activated Tat-induced HIV-1 transcription and exhibited enhanced NPM1–Tat interaction compared to NPM1 S125A mutant. Inhibition of Aurora A or Aurora B kinases that phosphorylate NPM1 on Ser-125 residue inhibited HIV-1, further supporting the role of NPM1 in HIV-1 infection. Taken together, 1E7-03 reprogrammed PPARα/RXRα and TGF-β pathways that contribute to the inhibition of HIV-1 transcription. Our findings suggest that NPM1 phosphorylation is a plausible target for HIV-1 transcription inhibition.


In Brief
Small molecule HIV-1 transcription inhibitor, 1E7-03, that binds to the noncatalytic RVxF accommodating site of protein phosphatase 1 reprogramed protein phosphorylation in the PPARα/ RXRα, TGF-β, and PKR pathways. Phosphorylation of nucleophosmin (NPM1) at Ser-125 residue was significantly reduced. NPM1 S125D mutant activated Tat-induced HIV-1 transcription and exhibited enhanced NPM1-HIV-1 Tat interaction. Inhibition of Aurora A or Aurora B kinases that phosphorylate NPM1 inhibited HIV-1. Our findings suggest that NPM1 phosphorylation is a plausible target for HIV-1 transcription inhibition.
Previous quantitative phosphoproteome analyses focused on the phosphorylation of host proteins modulated during HIV-1 infection (26,27). Here, we profiled global protein phosphorylation changes in CEM T cells induced by 1E7-03. Proteins potentially involved in HIV-1 transcription activation were identified using quantitative phosphoproteome and proteome analyses. The top targets were further validated using immunoblotting, site directed mutagenesis, and the use of specific kinase inhibitors.

Experimental Design and Statistical Rationale
Noninfected CEM T cells untreated or treated with 1E7-03 were compared to CEM T cells infected with VSV-G-pseudotyped HIV-1 without or with 1E7-03 treatment. The sample size for each condition n = 1. Each sample was divided to three parts for phosphoenrichment. N = 12 samples were processed in instrumental triplicates and that gave N = 36 individual files to analyze. Trypsindigested protein peptides, either nonenriched or enriched on Fe-NTA or TiO 2 columns, were analyzed by tandem liquid chromatography-Fourier transform mass spectrometry (LC-FT/MS) followed by label-free quantitative analysis and Ingenuity Pathway Analysis (IPA). Biological functions and biological pathways were accessed by IPA with cutoff >1.5 fold and p-value <0.05.

Chemicals
1E7-03 (purity above 98%) was synthesized by Enamine, as previously described (22). 1E7-03 was dissolved in dimethyl sulfoxide (DMSO) to obtain 10 mM stock solutions and stored at −20 • C. Acetonitrile and water-containing 0.1% formic acid (FA) were Optima LC/MS grade (Thermo Fisher Scientific). High-purity nitrogen (99.9%) was purchased from Roberts Oxygen Co, Inc. Other reagents were of analytical grade. DMSO and acetone were from Thermo Fisher Scientific.

Cell Lysis and Protein Digestion
Cells were collected and washed three times with PBS. To obtain cell lysates, the cells were suspended in 1 ml cold whole cell lysis buffer (50 mM Tris-HCl, pH 7.5, 0.5 M NaCl, 1% NP-40, 0.1% SDS), supplemented with protease inhibitors (P8340-1 ML, Sigma Aldrich) and phosphatase inhibitors (P044-1 ML, Sigma Aldrich). The cytosolic protein fraction was isolated by centrifugation at 13,000g for 30 min at 4 • C to remove cellular debris. Protein concentration was determined using Bradford assay (Bio-Rad). An aliquot of 500 μg proteins from each sample was mixed with 4-fold volume cold acetone to precipitate proteins. All samples were centrifuged at 13,000g for 5 min. Precipitated proteins were collected and dried in SpeedVac concentrator. The pellet was resuspended in 500 μl of sodium phosphate buffer (pH 8.0), reduced in 10 mM DTT (1 h at 60 • C), alkylated with 30 mM iodoacetamide (20 min, room temperature in the dark), and digested with 10 μg trypsin (Promega) at 37 • C on orbital shaker overnight.

Phosphopeptide Enrichment and Purification
An aliquot of 150 μl tryptic solution was processed by three different methods following manufacturer's instructions. In nonenriched group (No ENR), peptide mixture was directly purified by Pierce Graphite Spin Columns (88302, Thermo Fisher Scientific). Briefly, columns were prepared by washing with 100 μl of 1 M NH 4 OH (×2), activating with 100 μl of acetonitrile, followed by 100 μl of 1% TFA washing (×2). Peptide mixtures were loaded into columns and incubated for 10 min with periodic vortex mixing. After incubation, each sample was cleaned by 200 μl of 1.0% TFA (×2) and eluted with 100 μl of 0.1% FA in 50% acetonitrile (×3). The elution was gently dried using a SpeedVac concentrator. In the Fe-NTA-enriched group, peptide mixtures were dried, resuspended in 200 μl of binding buffer, added to Fe-NTA spin columns (88300, Thermo Fisher Scientific), and incubated for 20 min at room temperature with end-over-end rotation. Samples were in turn washed with 100 μl of wash buffer A (×2), wash buffer B (×2), and ultrapure water. Then, samples were eluted with 50 μl of elution buffer (3-5 min incubation each time, ×2). The elution fractions were pooled together, acidified by adding 200 μl of 2.5% TFA, and purified using Graphite Spin Columns as described above. In TiO 2 -enriched group, TiO 2 columns were prepared by washing with 20 μl of 0.4% TFA in 80% acetonitrile and 20 μl of 25% lactic acid. Peptide mixtures were dried, resuspended in 150 μl of 25% lactic acid, added to TiO 2 spin columns (88301, Thermo Scientific), and incubated for 10 min at room temperature with end-over-end rotation (×2). Samples were washed with 20 μl of 25% lactic acid and 20 μl of 25% lactic acid (×2). Then, samples were eluted using 50 μl of 1.5% ammonium hydroxide and 50 μl of 5% pyrrolidine. The elution fractions were combined, acidified by adding 100 μl of 2.5% TFA, and purified using Graphite Spin Columns. Note: (×number) is repeated times of washing and spin speed for centrifuge for each step, please follow manufacturer's instructions. We used one biological sample for each condition (four), which was processed using three different methods. Twelve samples were analyzed by mass spectrometry, using instrumental triplicates (in total, 36 samples).
(LTQ) XL Orbitrap mass spectrometer (Thermo Fisher Scientific) with the installed Xcalibur software (version 2.0.7, Thermo Fisher Scientific, https://www.thermofisher.com/order/catalog/product/OPTON-30965). Enriched and/or purified peptides were resuspended in 50 μl of water with 0.1% FA (v/v). A total of 10 μl of sample was loaded and washed for 6 min on a C 18  Charge state rejection (charge state 1 was rejected) as well as dynamic exclusion (repeat counts, 2; repeat duration, 10 s; exclusion duration, 10 s) was enabled.

Phosphoproteomic Data Analysis
LC-FT/MS raw data were searched by Proteome Discoverer 1.4 (PD 1.4) using SEQUEST search engine (Thermo Fisher Scientific), against the Uniprot Human database (11/26/2019, 140705 sequences) at a false discovery cut off ≤1%. A maximum of two missed cleavage sites was allowed with trypsin full cleavage. The mass tolerance for the precursor ion was 30 ppm and for the fragment -0.1 Da. Phosphorylation of serine, threonine, and tyrosine residues were enabled as dynamic modifications, while carbamidomethylation of cysteine was set as fixed modification. Filters were set for peptides with different charges as follows: charge 2 = 1.5 (XCorr score), charge 3 = 2.0, and charges >4 =2.5 for highly confident peptides; charge 2 = 0.5, charge 3 = 0.8, and charges >4 =1 for modestly confident peptides. The labelfree quantification of phosphopeptides eluting between 10 min and 70 min was performed with SIEVE 2.1 software (Thermo Fisher Scientific, https://tools.thermofisher.com/content/sfs/manuals/Man-XCALI-97695-SIEVE-22-Proteomics-ManXCALI97695-A-EN.pdf). Briefly, the chromatographic peaks detected by Orbitrap were aligned and the peptides peaks were detected with a minimum signal intensity of 1 × 10 5 ; quantitative frames were determined based on m/z (width: 10 ppm) and retention time (width: 2.5 min). The PD 1.4.-identified phosphopeptides were uploaded as framing seeds. Statistical filters were set to assess the quality of the data. The CV raw MS intensities of the triplicates had to be within 30%. This helped minimizing the effect of run-to-run variability. Nonphosphorylated peptides were filtered out to eliminate their interference on the quantitative result. SIEVE 2.1 generated lists of differentially expressed proteins with a cut off >1.5-fold or <0.66 and p-value<0.05. The raw files and PD1. 4 search files can be accessed at https://massive.ucsd.edu/, The dataset access number is MSV000090850. The public access password is nekhai@1. The download ftp link is ftp://massive.ucsd.edu/ MSV000090850/.
Biological Function and Pathway Analysis SIEVE 2.1 generated lists of differential phosphoproteins or expressed proteins with >1.5-fold of upregulation or downregulation and p-value<0.05 were uploaded to IPA (Ingenuity Systems) server for a Core analysis to investigate the protein function and biological networks. Canonical pathway analysis was carried out. PPARα/RXRα activation and TGF-β signaling networks were shown using Cytoscape 3.8.0.

NanoBiT Assay
LgBiT-fused PRKA2A and SmBiT-fused PRKACA were obtained from Promega. 293T cells were plated in 96-well white/clear culture plates with 40% confluence and transiently transfected with the indicated constructs (1:1 ratio of interacting pairs) using Lipofectamine 3000 Plus in OPTI-MEM as directed by the manufacturer's protocol. At 24 h post transfection, cells were treated with serial concentrations of 1E7-03 for the indicated amount of time. Nano-Glo Live Cell Substrate (N2012, Promega) was added, and luminescence was measured using a GloMax-Multi Detection System (Promega).

Western Blot Analysis
CEM T cells or HEK293T cells expressing NPM1-EGFP were treated with 10 μM 1E7-03 or DMSO for 3 h or 24 h as indicated. The cells were also treated where indicated with 3 nM or 100 nM okadaic acid (OA) for 2 h. Total proteins were extracted using the whole cell lysis buffer supplemented with protease and phosphatase inhibitors (described above). Equal amounts of total protein were separated by electrophoresis and transferred to polyvinylidene fluoride membrane. The membrane was probed with antibodies for NPM1 (sc-529252, Santa Cruz Biotechnology, Inc), anti-p-S125 NPM1 antibodies (ab109546, Abcam, 1:1000 dilution), or GFP antibodies (ab290, Abcam, 1:1000 dilution).

Transient Transfections
HEK293T cells were cultured in 24-well plates in Dulbecco's modified Eagle's medium containing 10% FBS. Cells were transfected at 30% confluence with Lipofectamine 3000 (Invitrogen), according to the manufacturer's recommendations. Cells were cultured 48 h posttransfection, and luciferase activity was analyzed using Luclite plus Reporter Gene Assay (PerkinElmer) measured by Glo-Max Microplate Multimode reader (Promega).

Coimmunoprecipitation
HEK293T cells were transfected with the indicated FLAG-Tat and GFP-NPM1 vectors, as described above. The whole cell extracts were prepared using the whole cell lysis buffer supplemented with protease and phosphatase inhibitors (described above). About 300 μg of whole cell extract supplemented with 8 μg of anti-FLAG antibodies was incubated with 5% bovine serum albumin preblocked protein A/Gagarose beads in TNN buffer (50 mM Tris-HCl, pH 7.5, 0.5% NP-40, 150 mM NaCl) at room temperature for 4 h with rocking. The agarose beads were precipitated and washed with TNN buffer. Proteins were resolved on 10% Bis-Tris SDS-PAGE, transferred to polyvinylidene fluoride membrane, and probed with the indicated antibodies.

Colocalization Analysis
HEK293T cells were transfected with vectors expressing EGFPtagged Tat and RFP-tagged NPM1, WT, and mutants. At 24 h post transfection, the cells were photographed on Olympus IX51 using a filter for FITC and Texas Red fluorescence at 400× magnification. Quantification of the colocalization was conducted in Image J using JACoP plug-in that allows Pearson's correlation analysis. The image colors were split, and threshold parameters were adjusted prior to the correlation analysis. CO 2 . The cells were treated with serial dilutions of 1E7-03, Aurora A inhibitor, Barasertib (Aurora B inhibitor), or vehicle (DMSO). At 24 h posttreatment, 100 μl of reconstituted luciferase buffer (Luclite Kit, PerkinElmer) was added to each well, incubated for 10 min, and luminescence was measured using Glo-Max Microplate Multimode reader (Promega).

Cell Viability Assays
Cell viability was measured with MTT assay. CEM T cells (3 × 10 5 ) were grown in 96-well plates and treated overnight with serially diluted compounds or DMSO control. Post treatment, 10 μl of MTT solution (ATCC 30-1010K) was added to each well, and the samples were incubated at 37 • C for 2 h. Then, 100 μl of lysis buffer (0.03% HCl with 10% SDS) was added to each well. The cells were incubated for 20 min in the dark. Absorbance was measured at 490 nm at a microplate reader (Bio-Rad Model 680). Each point was measured in triplicate, and serum-free medium with MTT solution was used as a negative control.

Statistical Analysis
All graphs were prepared using GraphPad prism 6 software. Data are presented as mean ± SD or SEM as indicated in the figure legends. Means were compared with Student t-tests.

Phosphopeptide Enrichment
To gain an insight on the host proteins phosphorylation induced by 1E7-03, we compared noninfected CEM T cells, untreated or treated with 1E7-03 (  Table S1 and Proteome Discoverer 1.4 for identified proteins in supplemental Table S2). We also detected approximately 3500, 800, and 2200 nonphosphorylated peptides in the nonenriched, TiO 2 -, and Fe-NTA-enriched individual groups (Fig. 1C). Number of specific serine-and threonine-phosphorylated residues were equal in the nonenriched and Fe-NTA-enriched peptide groups, whereas TiO 2 -enriched peptide group contained primarily serine phosphorylated peptides (Fig. 1D).

Phosphorylation Profile of Host Proteins Reprogramed by
1E7-03 CEM T cells, noninfected and untreated or treated with 1E7-03 ( Fig. 2, groups i and ii) and HIV-1-infected CEM T cells treated with 1E7-03 (Fig. 2, group iv) were compared to CEM T cells infected with HIV-1 and not treated with 1E7-03 that was designated as a reference group in SIEVE 2.1 label-free quantification analysis (cutoff >1.5-fold and p value<0.05) (Fig. 2, group iii). SIEVE 2.1 ion integration data are shown in supplemental Tables S3-S8 and summarized in supplemental Table S9. The overall phosphorylation patterns induced by 1E7-03 treatment, that we observed, fell into two categories: 1) increased phosphorylation by 1E7-03 and 2) decreased phosphorylation by 1E7-03 when groups i and ii, and groups iii and iv were compared (Fig. 2). In Fe-NTA-enriched groups ( Fig. 2B and supplemental Table S9), 1E7-03 increased protein phosphorylation to higher levels in noninfected cells than HIV-1-infected cells, whereas 1E7-03 reduced protein phosphorylation in HIV-1-infected cells to higher extent than the noninfected cells. In nonenriched and TiO 2 -enriched groups (Fig. 2, A and C, and supplemental Table S9), 1E7-03 similarly increased and decreased protein phosphorylation in HIV-1 infected, whereas in the noninfected cells, many peptides remain phosphorylated regardless of 1E7-03 treatment. Despite only 14% of the cells being infected, 1E7-03 induced more significant phosphorylation changes in HIV-1-infected cells comparing to non-infected cells suggesting potential interactions between the host, the virus, and the drug.
We compared proteins upphosphorylated and downphosphorylated in CEM T cells infected with HIV-1 and treated with 1E7-03 with the previously set of PP1-interacting proteins (28) that were identified in a broad in silico screening. Only one protein, MAP kinase kinase kinase 4 (MAP3K4), was found to overlap between the predicted PP1 interactors and the phosphoproteins in 1E7-03-treated CEM T cells (supplemental Fig. S4A). We also assessed levels of HIV-1 peptides in the infected CEM cells treated with 1E7-03 which showed overall reduction (supplemental Fig. S4B).
Taken together, modulation of protein phosphorylation in multiple pathways including PPARα/RXRα and TGF-β pathways is likely to contribute to HIV-1 inhibition by 1E7-03.

1E7-03 Does Not Affect the Overall Host Proteins Expression Level
We next analyzed the effect of 1E7-03 on host protein expression levels by conducting label-free quantitative proteome analysis in HIV-1-infected CEM T cells with or without 1E7-03 treatment. No changes in overall protein expression levels were observed (Fig. 4A). Volcano map analysis of total 5311 detected proteins (1.5-fold cutoff and p-value < 0.05) showed 21 upregulated proteins (green dots, Figs. 4B) and 29 downregulated proteins (cyan dots, Fig. 4B and supplemental Table S11). This observation indicated that expression of most host proteins was not significantly affected by 1E7-03 treatment. These differentially expressed proteins were further analyzed by IPA which showed that no major regulatory pathway was affected (Fig. 4C and Supplemental Table S12). Analysis of individual proteins also did not show any differentially expressed proteins related to HIV-1 replication regulation except heat shock proteins (1.56-2.049-fold reduction) and HLA-G (2.062-fold reduction) (supplemental Table S11). Comparison of the 50 differentially expressed proteins (supplemental Table S11) with the 212 proteins that had changes in their phosphorylation (Table 1) pointed to four overlapping proteins, including Hsp90AA1 (Fig. 4D). Thus, phosphorylation of Hsp90AA1 may be affected indirectly through the changes in its expression level. Taken together, 1E7-03 is likely to achieve antiviral effect by inducing protein phosphorylation and dephosphorylation rather than affecting the expression level of host proteins.

Phosphorylation of NPM1 Ser-125 Residue Contributes to HIV-1 Transcription Activation
We next validated NPM1 phosphorylation and assessed its biological effect. We chose NPM1 as its dephosphorylation was the strongest among the phosphorylation changes that were induced by 1E7-03 (Fig. 3A). To confirm that PP1 was involved in NPM1 phosphorylation, we analyzed endogenous NPM1 phosphorylation in CEM T cells treated with 10 μM 1E7-03 (24 h) or 100 nM OA (2 h), which served as positive control as OA inhibits PP1 at higher 10 nM concentrations and PP2A at lower 3 nM concentration (29). NPM1 expression levels and its Ser-125 phosphorylation were analyzed by immunoblotting using antibodies against NPM1 and NPM1 phosphorylated on Ser-125 (p-S125 NPM1). Compared to DMSO control, both 1E7-03 and OA treatments significantly reduced NPM1 phosphorylation on Ser-125 residue without affecting its protein expression level (*p < 0.015 for 1E7-03 and **p < 0.002, Fig. 5, A-C). We also tested the effect of short-term 3 h treatment with 1E7-03 comparing to 24 h treatment. HEK293T cells were transfected with vectors expressing NPM1-GFP or NPM1 S125A-GFP. While 24 h treatments led to decreased NPM1 phosphorylation, there were no changes in NPM1 phosphorylation at 3 h treatment (Fig. 5, D and E), We also analyzed the effect of lower concentration of OA, inhibitory for PP2A, on NPM1 phosphorylation using NPM1-GFP expressed in HEK293T cells. Comparing to the reduced NPM1 phosphorylation at 100 nM OA concentration, there was no effect at 3 nM OA treatments (Fig. 5, F and G).
Previously, NPM1 was shown to facilitate nuclear localization of HIV-1 Tat, thus enhancing viral transcription (30). HIV-1 transcription activation was evaluated by coexpressing GFPtagged NPM1 WT, NPM1 S125A mutant, or NPM1 S125D mutant along with HIV-1 long terminal repeat-luciferase reporter without or with Tat-expressing vector in HEK293T cells. NPM1 phosphorylation did not affect its nuclear colocalization with Tat as NPM1, NPM1 S125A, and NPM1 S125D mutants showed similar nuclear colocalization with Flag-Tat, when quantified with Pearson's correlation analysis (supplemental Fig. S5). Coexpression of NPM1 S125A or NPM1 S125D mutants increased HIV-1 basal transcription (2.1-fold, Fig. 6A). Coexpression of NPM1 S125D mutant induced Tat-mediated HIV-1 transcription (1.77-fold increase, Fig. 6B) compared to the expression of WT NPM1. In contrast, expression of nonphosphorylation-mimicking NPM1 S125A mutant had no effect on Tat-activated HIV-1 transcription (Fig. 6B). As NPM1 was shown to interact with Tat, we tested their binding by coimmunoprecipitation of GFP-tagged NPM1 (WT, S125A, and S125D) and Flag-tagged HIV-1 Tat. NPM1 was migrating as 62-kDa monomeric form and a larger 250-kDa form that might represent an oligomeric form of NPM1 or its covalently modified species (Fig. 6C). Both protein bands were excised and further analyzed by nano LC-MS/MS that confirmed the presence of NPM1 with 69% coverage (supplemental Fig. S6,  A and B). Phosphorylation of Ser-70 and Ser-125 residues was detected in 250-kDa species (supplemental Fig. S6B). Analysis of NPM1 acetylation showed that 62-kDa species was both acetylated and phosphorylated, whereas the 250-kDa species was phosphorylated but not acetylated (supplemental Fig. S7). Analysis of Tat interaction with NPM1 showed that 62-kDa form interaction was disrupted when nonphosphorylated NPM1 S125A mutant was used, whereas the phosphorylation mimicking NPM1 S125D mutant bound similar to the WT NPM1 (Fig. 6, D, and E). In contrast, the 250-kDa species of NPM1 showed equal interaction of WT and S125A mutant with Tat but enhanced interaction of NPM1 125D mutant (Fig. 6, F and G). These observations suggest that dephosphorylated and acetylated monomeric form of NPM1 might not bind Tat efficiently unless NPM1 is phosphorylated. In the oligomeric nonacetylated form, NPM1 phosphorylation enhances Tat binding. Taken together, these results show that downregulation of NPM1 Ser-125 phosphorylation by 1E7-03 is likely to contribute to HIV-1 transcription inhibition.

Aurora Kinases Inhibition as a Drug Target for HIV-1 Replication
NPM1 is phosphorylated on Ser-125 by several kinases including Aurora A, Aurora B, and casein kinase 2 (CK2) (31). To further confirm the role of Aurora kinases in HIV-1 replication, Aurora A inhibitor I and Aurora B inhibitor, Barasertib, were tested in one round HIV-1 infection and on cellular cytotoxicity using 1E7-03 as a reference. Aurora A inhibitor I and Barasertib suppressed one round HIV-1 infection (IC 50 = 7.8 μM and IC 50 = 13.7 μM, respectively) (Fig. 7A). The effects of Aurora kinase inhibitors were less pronounced comparing to the effect of 1E7-03 (IC 50 = 5.3 μM) (Fig. 7A). Cell viability analysis showed minimal toxicity for 1E7-03 and Barasertib in CEM T cells (CC 50s >100 μM, Fig. 7B), whereas Aurora A inhibitor I showed significant toxicity (CC 50 = 10 μM, Fig. 7B).
Taken together, this study shows that HIV-1 inhibition by 1E7-03 might be the outcome of complex attenuation of host protein phosphorylation within the PPARα/RXRα and TGF-β pathways that involves dephosphorylation of NPM1. DISCUSSION HIV-1 replication takes advantage of posttranslational modifications of many cellular proteins, including NPM1 (schematics depicted in Fig. 7C). NPM1 Ser-125 phosphorylation is likely to be regulated by various kinases and phosphatases including PP1, Aurora A/B, and CK2. As we showed here, PP1-targeting 1E7-03 blocks HIV-1 replication in part by downregulating NPM1 phosphorylation that may impede NPM1-Tat interaction (Fig. 7C). Decreased NMP1 phosphorylation in the cells treated with 1E7-03 or high concentration of OA indicate that PP1 is involved in the control of NPM1 phosphorylation. However, the effect of PP1 on NPM1 is indirect as NMP1 phosphorylation decreases upon PP1 inhibition. The NPM1 phosphorylation is mediated by one of the NPM1 kinases, and PP1 is likely affecting an upstream kinase. As NPM1 phosphorylation contributes to the inhibition of HIV-1 transcription, targeting NPM1 phosphorylation, indirectly with 1E7-03 or directly with a kinase inhibitor, such as Barasertib, can serve as a future avenue for HIV-1 transcription inhibition.
PP1 has traditionally been overlooked as a drug target because of the conserved catalytic site of PP1 that makes it nearly impossible for achieving selective enzymatic inhibition without having global cellular effect (32). The existing enzymatic PP1 inhibitors, such as microcystin, nodularin, OA, and tautomycin, are toxic to the host cells. However, as the substrate specificity of PP1 is mediated by its regulatory subunits, there is an opportunity for developing noncatalytic PP1 inhibitors that might specifically target individual PP1 holoenzymes (17,18). In this study, we conducted for the first time a comprehensive analysis of host protein phosphorylation in cells treated with PP1-targeting 1E7-03 molecule using quantitative mass spectrometry. Different enrichment strategies were employed including Fe-NTA and TiO 2 enrichment and compared to the nonenriched peptide samples. While only 14% of the cells were infected with HIV-1, 1E7-03 seems to have more robust effect in HIV-1-infected cells, implying intricate interactions among virus, host, and the drug. The expected effect of 1E7-03 is the increased protein phosphorylation as the compound affects PP1-mediated dephosphorylation and may enhance phosphorylation of PP1-targeted proteins. Label-free quantitative phosphoproteome analysis showed that 1E7-03 induced phosphorylation of several proteins from several pathways including PPARα/ RXRα and TGF-β pathways. Also, unexpectedly, we observed decreased phosphorylation of several proteins that also were part of the same pathways. Both pathways have been implicated in the regulation of HIV-1 replication. In the PPARα/ RXRα canonical pathway, proteins with increased phosphorylation included HLA-B/C, LSM14A, and ASPM. In the PPARα/ RXRα pathway, proteins with decreased phosphorylation included RGSL1, p53 (coded by TP53), HSP90AA1, TGF-β2, and NPM1. HLA-B*27, HLA-B*57, and HLA-B*58 are wellknown protective alleles that contribute to the suppression of HIV-1 replication and prevent progression toward AIDS (33). LSM14A initiates cellular antiviral response in the early phase of viral infection by regulating MITA level in a cell-specific manner (34,35). ASPM is a binding partner of PP1 and contains RVxF motif (KVSF) which is phosphorylated by Aurora B kinase during mitosis (36). TGF-β superfamily of cytokines include three isoforms (β1, β2, and β3) that elicit similar biological responses through the same receptor (37). HIV-1infected patients exhibit higher levels of TGF-β1 which is considered a major driver of immunosuppression (38). TGF-β1 might also induce apoptosis of CD4+ T lymphocytes following macrophage-tropic (R5) HIV-1 infection, by reducing levels of anti-apoptotic factors as well as by increasing apoptosisinducing factors (39). HIV-1 Tat protein induces mRNA expression and secretion of TGF-β1, which is also induced by HIV-1 glycoprotein (gp) 160 (38). TGF-β2 has been shown to exert suppressive effects on IL-2-dependent T-cell growth (40). TGF-β isoforms can exert their effects through either SMAD-dependent or SMAD-independent signaling pathways (41).
We validated the role of NPM1 phosphorylation in HIV-1 transcription as this was the most significantly affected protein whose phosphorylation decreased in 1E7-03-treated cells. NPM1 has been reported to play a role in nuclear localization of HIV-1 Tat and Rev proteins and its acetylation was shown to be required for HIV-1 transcription activation (30,42). NPM1 Ser-125 residue is phosphorylated by cell cycle-related Ser/Thr protein kinases such as Aurora A, Aurora B, and CK2 (43). NPM1 is acetylated by p300 (30,42). To the best of our knowledge, NPM1 phosphorylation on Ser-125 has not been linked to HIV-1 transcription regulation nor connected to PP1. NPM1 phosphorylation having effect on the interaction with HIV-1 Tat has not yet been described either. Our results indicate that monomeric form of NPM1 might be acetylated but Tat binding of this form is not affected by phosphorylation. In contrast, oligomeric form of NPM1 was not acetylated and its phosphorylation enhanced Tat binding. It is possible that acetylation and phosphorylation are two independent mechanisms that both may contribute to HIV-1 transcription activation. Further analysis is needed to analyze crosstalk of phosphorylation and acetylation and whether both modifications contribute independently to HIV-1 activation.
SMAD7, a potent inhibitor of TGF-β-dependent signaling, was previously shown to be phosphorylated on Ser-206 and Ser-249 residues (61,62). The Ser-206 phosphorylation, adjacent to the PPxY motif, enhances SMAD7 binding to WW4 domain of WWP2, an E3 ubiquitin ligase (62). SMAD7 interacts with GADD34 and facilitates recruitment of PP1 catalytic subunit to dephosphorylate TGF-β type I receptor (63). TGF-β promotes formation of the ternary complex between PP1, GADD34, SMAD7, and TβRI. Hypoxia-inducible factor 1α (HIF-1α) expression is increased in HIV-1 infection (64), and TGF-β induces HIF-1α accumulation and activity by increasing HIF-1α protein stability. TGF-β specifically decreases both mRNA and protein levels of a HIF-1α-associated PHD2 (encoded by EGLN1 gene), through the SMAD signaling pathway (65). Phosphorylation of PHD2 on Ser-125 by rapamycin (mTOR) downstream kinase P70S6K increases its activity, and dephosphorylation by protein PP2A reduces it (66). We observed increased PHD2 phosphorylation in HIV-1infected cells, suggesting that PP1 might also be involved in the regulation of PHD2 activity in addition to PP2A. Treatment with 1E7-03 might block PHD2 Ser-246 dephosphorylation and this may lead to HIV-1 inhibition. We also observed increased phosphorylation of TAF4. As HIV-1 Tat facilitates recruitment of TBP in the absence of TAFs (67), the significance of increased TAF4 phosphorylation by 1E7-03 is yet to be clarified.
In conclusion, our study advanced the understanding of HIV-1 transcription regulatory mechanism by PP1 and pointed to PP1-targeting molecules such as 1E7-03 as viable drugs to be further developed for HIV-1 inhibition. The phosphoregulation of host proteins such as NPM1 by phosphatase or FIG. 4. 1E7-03 has no effect on host proteins expression. A, the overall protein abundance in HIV-1-infected cells with or without 1E7-03 treatment. Blue color represents DMSO control group, and orange color represents 1E7-03-treated group. B, volcano map of host protein expression whose expression was changed by 1E7-03. Cyan and green dots represent host proteins significantly downregulated or upregulated by 1E7-03 with cutoff >1.5 fold and p < 0.05. C, canonical pathways analysis of host protein expression affected by 1E7-03. Color (gray) indicate that pathways had no activity. D, Venn diagram showing comparison of changes in expression induced by 1E7-03. HIV-1, human immunodeficiency virus-1; DMSO, dimethyl sulfoxide.
FIG. 5. Effect of PP1 inhibitors on NPM1 Ser-125 phosphorylation. A, 1E7-03 downregulates NPM1 Ser-125 phosphorylation. CEM T cells were treated with 10 μM 1E7-03 for 24 h and 100 nM okadaic acid (OA) for 2 h as a positive control. The protein expression and phosphorylation of NPM1 Ser-125 was analyzed by Western blotting with antibodies against NPM1 and Ser-125 phospho-specific antibodies (p-S125 NPM1). B and C, quantification of Ser-125-phosphorylated NPM1 and NPM1 performed using Prism 6 from three independent experiments. Asterisks indicate p < 0.05 (*) and p < 0.01 (**). D, effect of short incubation with 1E7-03 on NMP1 phosphorylation. HEK293T cells were transfected with vectors expressing GFP-tagged WT NPM1 and NPM1 S125A mutant and treated with 10 μM 1E7-03 for 3 h (short treatment) and 24 h (long treatment). The phosphorylation of NPM1 Ser-125 and protein expression was analyzed by Western blotting with antibodies for p-S125 NPM1 and GFP. E, quantification of results from panel D by Prism 6 from three independent experiments. F, effect of low concentration of OA on NPM1 phosphorylation. HEK293T cells were transfected as in panel D and treated with 3 nM or 100 nM OA for 2 h. Phosphorylation of NPM1 Ser-125 and protein expression were analyzed with antibodies for p-S125 NPM1 and GFP. G, quantification of the data form panel F was performed using Prism 6 from three independent experiments. HIV-1, human immunodeficiency virus-1; NPM1, nucleophosmin.
kinase-targeting small molecules may serve as new promising avenue for HIV-1 transcription inhibition.

DATA AVAILABILITY
The mass spectrometry raw data and Proteome Discoverer processed files are uploaded to UCSD Massive Server (https://massive.ucsd.edu/). Supplemental data -This article contains supplemental data.
Funding and additional information -This work was supported by the research grants from NIH to S. N. (1R01HL125005, 5U54MD007597, and 1UM1AI26617). This research was also supported by a P30 AI117970 grant from FIG. 6. Effect of NPM1 Ser-125 phosphorylation on HIV-1 transcription. A and B, effect of expression of NPM1 and its phosphorylation mutants (S125A and S125D) on basal (panel A) and Tat-mediated HIV-1 transcription (panel B). HEK293T cells were transfected with vectors expressing the indicated GFP-tagged NPM1 mutants and cotransfected with vectors expressing HIV-1 LTR-luciferase reporter (panel A) or HIV-1 LTR-luciferase reporter and FLAG-tagged Tat expression vector (panel B). Luciferase activity was analyzed using Luclite plus Reporter Gene Assay and normalized to GFP intensity. All data were shown in triplicates as transcription activation fold related to the WT NPM1. C, effect of NPM1 Ser-125 phosphorylation on the interaction with Tat protein. GFP-tagged NPM1 (WT, S125A, and S125D) and Flag-tagged HIV-1 Tat were expressed in 293T cells. Tat was precipitated with anti-Flag antibody, resolved on SDS-PAGE, and immunoblotted with anti-GFP to detect NPM1 and anti-Flag antibodies to detect Tat. Lane 1, IgG control; Lane 2, No Tat; Lane 3, Tat+NPM1 WT; Lane 4, Tat+NPM1 S125A; Lane 5, Tat+NPM1 S125D. D-G, quantification of the coimmunoprecipitation data from three independent experiments. NPM1's 62 kDa and 250 kDa isoforms were normalized to NPM1 and Tat input, respectively. Asterisks indicate p < 0.05 (*) and p < 0.01 (**). HIV-1, human immunodeficiency virus-1; LTR, long terminal repeat; NPM1, nucleophosmin.