Epigenetics, N-myrystoyltransferase-1 and casein kinase-2-alpha modulates the increased replication of HIV-1 CRF02_AG, compared to subtype-B viruses

HIV subtypes distribution varies by geographic regions; this is likely associated with differences in viral fitness but the predictors and underlying mechanisms are unknown. Using in-vitro, in-vivo, and ex-vivo approaches, we found significantly higher transactivation and replication of HIV-1-CRF02_AG (prevalent throughout West-Central Africa), compared to subtype-B. While CRF02_AG-infected animals showed higher viremia, subtype-B-infected animals showed significantly more weight loss, lower CD4+ T-cells and lower CD4/CD8 ratios, suggesting that factors other than viremia contribute to immunosuppression and wasting syndrome in HIV/AIDS. Compared to HIV-1-subtype-B and its Tat proteins(Tat.B), HIV-1-CRF02_AG and Tat.AG significantly increased histone acetyl-transferase activity and promoter histones H3 and H4 acetylation. Silencing N-myrystoyltransferase(NMT)-1 and casein-kinase-(CK)-II-alpha prevented Tat.AG- and HIV-1-CRF02_AG-mediated viral transactivation and replication, but not Tat.B- or HIV-1-subtype-B-mediated effects. Tat.AG and HIV-1-CRF02_AG induced the expression of NMT-1 and CKII-alpha in human monocytes and macrophages, but Tat.B and HIV-1-subtype-B had no effect. These data demonstrate that NMT1, CKII-alpha, histone acetylation and histone acetyl-transferase modulate the increased replication of HIV-1-CRF02_AG. These novel findings demonstrate that HIV genotype influence viral replication and provide insights into the molecular mechanisms of differential HIV-1 replication. These studies underline the importance of considering the influence of viral genotypes in HIV/AIDS epidemiology, replication, and eradication strategies.

This differential geographic distribution of HIV clades is likely associated with differences in viral fitness, replication capacity, and viral adaptation in a given environment. In fact, the transactivator of transcription (Tat) binds to the trans-activating response element (TAR) to modulate the expression of viral genes and HIV replication; and genetic variations in the Tat region has been shown to affects its interaction with TAR and viral replication 15,16 . The host genetics, ethnicity and immune response also drive HIV genetic changes and play a major role in the control of viral replication, mutations, immune response, and disease progression 17,18 .
Viral genotype can influence Tat-induced blood-brain barrier inflammation 19,20 , cytokine expression and chemotactic activities 20,21 , and can affect the progression to AIDS 22,23 . Ex-vivo and in-vitro studies also showed increased replication of HIV-1 CRF02_AG isolates compared to subtypes A and G viral isolates 24,25 . The molecular mechanisms modulating this subtype-based differential viral replication are not known. In the present study, using cell lines, primary HIV-1 isolates, human peripheral blood mononuclear cells (PBMC), monocytes-derived macrophages (MDM), and HIV/AIDS animal models, we show significantly higher replication of primary HIV-1 CRF02_AG (AG) isolates in-vitro and in-vivo, compared to subtype-B viral isolates. Mechanistic studies demonstrate that this increased replication of AG isolates was associated with increased histone acetyl transferase (HAT) activity, increased acetylation of histone H3 and H4 in the viral long-terminal repeat (LTR) region, increased viral promoter transactivation by Tat proteins derived from HIV-1 CRF02_AG (Tat.AG) and primary CRF02_AG viral isolates (but not Tat.B or HIV-1 subtype-B). We demonstrate that Tat.AG and HIV-1 CRF02_AG (but not Tat.B or HIV-1 subtype-B) increased N-myrystoyltransferase (NMT)-1 and casein kinase-II-alpha (CKIIα) expression in human monocytes and MDM, and silencing NMT1 and CKIIα (but not NMT2 or CKIIβ) genes blocked HIV-1 CRF02_AG (but not subtype B) infection of human macrophages. These mechanistic studies provide insights into the molecular mechanisms modulating the increased replication of CRF02_AG viruses, and have implications for the transmission, adaptation, and predominance of this recombinant virus in Sub-Saharan Africa.

Results
Increased replication of primary HIV-1 CRF02_AG isolates, compared to HIV-1 subtype-B, in human blood cells. HIV replicative capacity determines its fitness and adaptation in a given environment.

Infection of NOD/scid-IL-2Rγ c null (NSG) mice with HIV-1 subtype-B resulted in higher weight loss, compared to animals infected with HIV-1 CRF02_AG.
To validate our in-vitro findings, we performed in-vivo experiments using NSG mice engrafted with human peripheral blood lymphocytes (PBL). Animals were monitored and weighed every 2-3 days before (pre-E) and after (post-E) engraftment, before and after infection. All engrafted animals (mock-infected controls, AG-infected, and clade-B-infected mice) had similar body weight pre-and post-engraftment and at week-1 p.i., mean mouse weight: 23-25.6 grams(g) ( Fig. 2A,B). At week-2 p.i., compared to controls (mean weight 27.38 ± 1.77(SD) g/mouse), the weight of clade-B (20.38 ± 3.14 g/mouse) and AG (22.22 ± 1.74 g/mouse) infected mice were significantly lower (P < 0.0001, Fig. 2A,B), but there was no significant difference between AG-and clade-B-infected animals' weight. At week-3 p.i., compared to controls (mean weight 28 ± 2.2 g/mouse), the weight of clade-B (14.73 ± 1.5 g/mouse) and AG (18.5 ± 2.55 g/mouse) infected mice were significantly lower (P < 0.0001); and the weight of clade-B infected mice was significantly lower than that of AG-infected animals (P < 0.0001, Fig. 2A,B).
Increased LTR transactivation with Tat.AG and HIV-1 CRF02_AG compared to Tat.B and HIV-1 clade-B. Efficient and increased HIV-1 promoter activity correlate with increased LTR transcription and viral replication, and this is modulated by Tat binding to TAR 15,16 . Therefore, using U38 cells, we assessed the effects of Tat.AG and Tat.B on LTR promoter activity. Both Tat.AG and Tat.B (10-1000 ng/ml) significantly increased LTR transcriptional activity compared to cells exposed to heat-inactivated (HI) Tat proteins (Fig. 5A). LTR transcriptional activity in cells exposed to Tat.AG was 1.4-1.76-fold higher than LTR activity in cells exposed to Tat.B (P < 0.05, Fig. 5A).  www.nature.com/scientificreports www.nature.com/scientificreports/ and Tat.B significantly increased HAT activity in U38 cells compared to cells exposed to heat-inactivated (HI)-Tat (Fig. 5B). HAT activity in cells exposed to Tat.AG was 1.4-2.53-fold higher than HAT activity in cells exposed to Tat.B (Fig. 5B). Because macrophages are some of the major HIV cellular reservoirs in humans, we assessed the effect of CRF02_AG and clade-B HIV-1 (3 different isolates for each subtype) on HAT activity in primary human monocytes and MDM. Compared to controls, exposure of human monocytes to AG isolates increased HAT activity by 2.66-4.34-fold whereas exposure of monocytes to clade-B HIV-1 isolates increased HAT activity by 1.5-2.5-fold (Fig. 5C,D); HAT activity in monocytes exposed to HIV-1 CRF02_AG isolates was 1.2-1.73-fold higher than HAT activity in monocytes exposed to clade-B isolates (Fig. 5C,D). Compared to non-infected MDM, exposure of human MDM to HIV-1 CRF02_AG isolates increased HAT activity by 2.6-5.24-fold whereas exposure of MDM to clade-B isolates increased HAT activity by 1.8-3.37-fold (Fig. 5E,F); HAT activity in MDM exposed to HIV-1 CRF02_AG isolates was 1.45-1.55-fold higher than HAT activity in MDM exposed to clade-B isolates (Fig. 5E,F).

Increased viral transactivation with Tat.AG and HIV-1 CRF02_AG is associated with increased
Increased HIV-1 transactivation and HAT activity with Tat.AG are associated with increased acetylation of histone H3 and H4 in the LTR promoter. Acetylation of histone proteins correlate with increased HAT activity, and HIV-1 transactivation is often associated with increased histone acetylation 29,30 . We www.nature.com/scientificreports www.nature.com/scientificreports/ performed Chromatin Immunoprecipitation (ChIP) assays to determine whether Tat.AG and Tat.B differentially induce acetylation of histones in the HIV-1 LTR promoter region(s) (Fig. 6A). Tat increased acetyl-H3 and acetyl-H4 levels in nucleosome (Nu)-0 and Nu-1 regions of the HIV-1 promoter. Compared to Tat.B, Tat.AG induced higher levels of acetyl-H3 and acetyl-H4 in Nu-0 (Fig. 6B). No acetylated histone was detected in the nucleosome-free regions (NFRs). Acetyl-H3 and acetyl-H4 levels in cells treated with similar concentrations of HI-Tat proteins were similar to levels in controls (C, Fig. 6B).

NMT1 and CKIIα are involved in Tat.AG but not Tat.B mediated HIV-1 transactivation.
Our previous analyses of Tat sequences from over 100 human plasma samples showed conserved post-translational modifications (PTMs) in Tat functional domains, including N-myristoylation and CKII domains, and demonstrated subtype-based differences in these Tat PTMs 12 . To determine whether these subtype-based differences in Tat PTMs can influence its function, we silenced two NMT (NMT1 and NMT2) and two CKII (CKIIα and CKIIβ) genes in U38 (  7F) (P < 0.0001), but had no effect on Tat.B-induced LTR transcription (Fig. 7C,F). Silencing the NMT2 or CKIIβ genes had no effect on Tat.AG or Tat.B -induced LTR transcription (Fig. 7C,F).

NMT1 and CKIIα mediate HIV-1 CRF02_AG, but not clade-B, replication in human macrophages.
To validate our data and determine whether CKII and/or protein myristoylation are involved in HIV-1 replication and whether there are subtype differences, we assessed the effects of NMT1, NMT2, CKIIα, and CKIIβ on HIV-1 CRF02_AG and clade-B replication in primary human macrophages. Silencing of NMT1, NMT2, CKIIα, and CKIIβ genes in MDM was confirmed by RT-PCR (Fig. 8A) and Western blot (Fig. 8B); 5 × 10 4 infectious units of each shRNA induced maximal gene silencing without causing cellular toxicity. From day-5 to day-21 p.i., RT activity in MDM infected with AG viruses was consistently higher than RT activity in MDM infected with clade-B viruses (P < 0.0001, Figs 8 and 9). Silencing NMT1 blocked MDM infection by AG isolates but had no effect on infection by clade-B isolates (Fig. 8C,D). Compared to scrambled shRNA, NMT1 shRNA reduced HIV-1 CRF02_AG replication in MDM by 71.4-86.7% (P < 0.0001, Fig. 8C,D) but had no effect on the replication of clade-B viruses. From day-5 to day-21 p.i., the replication of clade-B viruses in MDM with silenced NMT1 gene was 1.8-4.1-fold higher than the replication of AG viruses in MDM with silenced NMT1 gene (P < 0.0001, Fig. 8C,D). Silencing the NMT2 gene had no effect on the replication of AG or clade-B viruses (Fig. 8E,F).
Additional experiments also confirmed the involvement of CKIIα in the replication of HIV-1 CRF02_AG, but not clade-B viruses. Silencing the CKIIα gene blocked MDM infection by HIV-1 CRF02_AG isolates, but had no effect on infection by clade-B isolates (Fig. 9A,B). Compared to scrambled shRNA, CKIIα shRNA reduced HIV-1 CRF02_AG replication in MDM by 71-87.8% (P < 0.0001, Fig. 9A,B) but had no effect on the replication of clade-B viruses. From day-5 to day-21 p.i., the replication of clade-B viruses in MDM with silenced CKIIα www.nature.com/scientificreports www.nature.com/scientificreports/ gene was up to 4-fold higher than the replication of AG viruses in MDM with silenced CKIIα gene (P < 0.0001, Fig. 9A,B). Silencing the CKIIβ gene had no effect on the replication of AG or clade-B viruses (Fig. 9C,D).
Tat.AG and HIV-1 CRF02_AG, but not Tat.B or clade-B viruses, induced NMT1 expression and secretion in human monocytes and MDM. Compared to cells exposed to HI-Tat, Tat.AG increased NMT1 expression in monocytes by 5.5-14.88-fold (Fig. 10A) and in MDM by 6-13-fold (Fig. 10B). Tat.B did not induce much NMT1 in monocytes or MDM. NMT1 levels in monocytes and MDM exposed to Tat.AG were respectively 7.6-19.4-fold (Fig. 10A) and 3.75-16.4-fold (Fig. 10B) higher than NMT1 levels in monocytes and MDM treated with similar concentrations of Tat.B. Time-course experiments confirmed these findings; whereas Tat.B treatment of monocytes or MDM for 6-24 hours(h) did not induce NMT1, Tat.AG time-dependently increased NMT1 expression and secretion: treatment with Tat.AG for 6 h, 12 h, and 24 h increased NMT1 expression respectively by 3-fold, 9.74-fold, and 17.09-fold in monocytes (Fig. 10C) and by 3-10.4-fold in MDM (Fig. 10D), compared to cells exposed to similar concentrations of HI-Tat. Tat.AG also significantly increased NMT1 expression in U38 cells compared to Tat.B (data not shown).

Discussion
The epidemiology of HIV/AIDS is characterized by distinct and differential geographic subtypes distribution, with HIV-1 CRF02_AG prevalent throughout West and Central Africa 5-7 . This differential geographic distribution of HIV clades is likely associated with differences in viral fitness, transmission efficiency, replication capacity, www.nature.com/scientificreports www.nature.com/scientificreports/ and viral adaptation in a given environment. In the present study, we use cell lines, primary HIV-1 isolates, primary human PBMC, MDM, and in-vivo HIV/AIDS animal models to demonstrate significantly increased replication of HIV-1 CRF02_AG compared to clade-B isolates. In-vivo and ex-vivo data also showed higher plasma www.nature.com/scientificreports www.nature.com/scientificreports/ viremia and significantly higher viral load in the heart, kidney, liver, spleen, and lung tissues of AG-infected animals compared to animals infected with clade-B viruses; indicating a generalized increased replication of AG isolates throughout animals' blood and tissues.
Previous studies in activated T-cells 25 and PBMC 24 also showed increased ex-vivo replicative fitness of CRF02_ AG isolates compared to its parental HIV-1 subtypes A and G 24,25 and independently of patient's CD4 counts or co-receptor use 25 . It is therefore likely that the increased replication capacity of HIV-1 CRF02_AG is responsible for its predominance throughout West and Central Africa. The associated mechanisms are not known; elucidating the mechanisms responsible for this increased replication of HIV-1 CRF02_AG strains is critical to understanding viral epidemiology, including its transmission and propagation in Sub-Saharan Africa. In the present study, we show that in addition to the increased replication of CRF02_AG isolates in primary human cells and in-vivo, these AG isolates showed significantly higher LTR transactivation compared to clade-B viruses. Tat proteins are expressed early in the HIV life cycle and transactivate the LTR to induce the expression of HIV genes and viral replication. Our data showing significantly higher transactivation of HIV-1 LTR with Tat.AG, compared to Tat.B, further confirm that viral genotype influence the magnitude of LTR transactivation and viral replication.
Increased HIV-1 transactivation and viral replication are associated with histone acetylation and increased HAT activity; whereas decreased HIV-1 transactivation, lower viral replication, and viral latency are associated with histone deacetylation and increased histone deacetylase activity [29][30][31] . Our current study shows that increased transactivation and replication of CRF02_AG isolates correlated with increased HAT activity and acetylation of histone H3 and H4, with higher levels of acetylated H3 and H4 observed with Tat.AG, compared to Tat.B. These data suggest that the increased replicative capacity of CRF02_AG isolates involves epigenetic modifications, especially acetylation of histones H3 and H4 in the Nu-0 and Nu-1 of the viral promoter.
PTMs of proteins modulate their structure and function, including signaling and interactions with other molecules and co-factors. Our previous analyses of Tat sequences in plasma from over 100 HIV-1-infected subjects showed the presence of conserved PTMs in Tat functional domains, including CKII and N-myristoylation domains 12 , with CRF02_AG Tat sequences having two N-myristoylation domains (in the Core and N-terminal regions) and a CKII domain (in the glutamine-rich region) compared to Tat sequences from other HIV-1 subtypes 12 . Therefore, we investigated whether N-myristoylation and/or CKII play a role in the differential transactivation and viral replication observed. Analyses using two different cell types containing integrated silent copies of the HIV-1 promoter showed increased viral transactivation with Tat.AG compared to Tat.B, and further showed that silencing the NMT1 or CKIIα genes blocked Tat.AG-induced HIV transactivation but had no effect on Tat.B-induced viral transactivation; whereas silencing the NMT2 or CKIIβ genes had no effect on Tat.AG-or Tat.B-induced viral transactivation. Significantly, we showed that silencing the NMT1 or CKIIα genes blocked human macrophage infection by HIV-1 CRF02_AG isolates but had no effect on infection by clade-B isolates, and silencing the NMT2 or CKIIβ genes had no effect on macrophage infection by both viral subtypes.
CKII is a serine/threonine selective protein kinase that preferentially phosphorylate serine and threonine residues 32,33 , and has been shown to play a role in HIV infection and HIV/AIDS pathogenesis 34 . CKII-mediated phosphorylation of Rev, Vpu, and protease at serine residues facilitates viral infection, syncytia formation, and disease progression [34][35][36][37][38] . CKII-mediated phosphorylation of Rev serine residues also influences Vpu-CD4 interactions 39,40 ; and mutations in the Vpu CKII site significantly altered its biological activity 41 . Our current data shows that CKIIα differentially mediates Tat transactivation and HIV replication based on viral genotype, with AG viruses more susceptible to CKIIα-mediated viral transactivation and replication compared to clade-B viruses. In the NFκB pathway, IκB phosphorylation and degradation results in NFκB translocation to the nucleus where it activates target genes [42][43][44] . HIV-1 has NFκB binding sites and this pathway play a major role in viral transactivation and replication [45][46][47] . CKII associates with IκBα, is required for basal and HIV-1-induced IκBα degradation 48 , and CKIIα phosphorylates IκBα at serine and threonine residues to induce IκBα degradation and NFκB signaling 48 . Our current study shows that this CKIIα subunit modulates the transactivation and replication of CR02_AG HIV-1 isolates, whereas the CKIIβ subunit has no effect, suggesting that targeting this alpha subunit could help abrogate the replication and expansion of AG isolates.
Myristoylation is a lipidic modification whereby a myristoyl group is covalently attached to an N-terminal glycine residue 49 . Myristoylation can occur co-translationally or post-translationally, and is catalyzed by NMT using myristoyl-coenzyme-A as substrate 49 . Higher eukaryotes have two NMT isozymes (NMT1 and NMT2) encoded by two genes that share about 70% similarity 50 . Myristoylation increases protein-protein interactions, guides protein sub-cellular localization, intra-cellular trafficking, signaling, and immune responses 49,51 . Myristoylation has been shown to play an important role in cell viability, cell survival, innate and adaptive immune response, and HIV infection 49,52,53 . Myristoylation is critical for HIV-1 Nef function and is involved in Nef-CD4 interactions 54,55 . Myristoylation also play a major role in Gag function and disrupting Gag myristoylation block HIV-1 replication, formation and release of new virions [56][57][58] . In fact, NMT targets Gag on the cell membrane 57,58 and this Gag myristoylation enables protein-protein and protein-lipid interactions, resulting in the recruitment of Gag, viral RNA and proteins to the host' cellular membrane and budding of new virions 56,57,59 . It has also been shown that NMT1 is the predominant isoform involved in Nef-CD4 interactions 60 and HIV-1 production 61 . Our current study shows that NMT1 modulates the transactivation and replication of CR02_AG HIV-1 isolates, suggesting that targeting NMT1 could help abrogate the transactivation and replication of AG isolates.
The potential implications of this higher replicative fitness of HIV-1 CRF02_AG are significant, because higher AG replicative fitness can influence viral transmission and HIV/AIDS epidemiology. In fact, it has been shown that increased AG replicative fitness also influence its transmission in humans. Monitoring of at-risk seronegative women in West-Africa showed that among subjects who became seropositive, those infected with HIV-1 CRF02_AG had significantly higher viral loads during the early stages of infection compared to those infected with non-AG HIV-1 62 . Studies of HIV-infected humans in Sub-Saharan Africa showed that higher HIV replication capacity was associated with differential inflammatory states and increased T-cells activation 63 , increased (2019) 9:10689 | https://doi.org/10.1038/s41598-019-47069-9 www.nature.com/scientificreports www.nature.com/scientificreports/ risk of HIV transmission to other humans, including increased risk of mother-to-child transmission 64 , and that there was a positive correlation between high ex-vivo HIV replicative capacity and the magnitude of viral burden in-vivo 63 .
Paradoxically, higher viral loads in HIV-1 CRF02_AG-infected animals did not correlate with increased immunosuppression or wasting syndrome, as animals infected with HIV-1 subtype-B showed significantly more weight loss, lower hCD4+ T-cells and lower hCD4/hCD8 ratios. This suggest that factors other than viremia may be playing a role in HIV-associated immunosuppression and weight loss. Inflammation could be one such factor, as we previously showed differential inflammatory effects of Tat based on viral subtypes, including significant increase in inflammatory cytokines, chemokines, and matrix metalloproteinases, as well as increase and activation of complement factors in primary human brain endothelial cells exposed to Tat.B, compared to cells exposed to Tat.AG 19,20 . Our subsequent studies will investigate the association between inflammation, immunosuppression, and wasting syndrome between these viral subtypes.
Taken together, our current data showed significantly higher replication of HIV-1 CRF02_AG in-vitro and in-vivo, compared to HIV-1 subtype-B, and provide insights into the molecular mechanisms by which epigenetics, CKII-alpha, and NMT1 modulates this increased replication of CRF02_AG viruses. The fact that silencing the NMT1 and CKIIα genes blocked the infection of human cells by HIV-1 CRF02_AG but not clade-B viruses, suggests that the additional N-myristoylation and CKII domains present in Tat.AG sequences are functional and directly modulate the replication of CRF02_AG viruses. Such genetic differences should be considered when designing strategies to curb HIV replication. Our current studies have implications for viral eradication strategies and HIV/AIDS epidemiology, including the transmission, adaptation, and predominance of this recombinant virus across West-Central Africa.

Materials and Methods
Human monocytes, MDM, PBMC, PBL cultures. Monocytes, PBMC, and PBL were obtained by countercurrent centrifugal elutriation of leukopheresis packs from HIV-1, -2 and hepatitis-B seronegative donors (obtained from the Omaha American Red Cross), and cultured as previously described 65,66 . To obtain MDM, freshly elutriated monocytes (2 million cells per well in 6-well plates) were differentiated into MDM by culture for 7 days in Dulbecco's Modified Eagle's Medium (DMEM, Sigma, St. Louis, MO) supplemented with 10% heat-inactivated pooled human serum, 1% glutamine, 50 µg/ml gentamicin, 10 µg/ml ciprofloxacin (Sigma), 1000 U/ml highly purified recombinant human macrophage colony stimulating factor (hMCSF), and cultured as we previously described 65,66 . All reagents were prescreened for endotoxin (<10 pg/ml, Associates of Cape Cod, Woods Hole, MA) and mycoplasma contamination (Gen-probe II, Gen-probe, San Diego, CA). For cells from each donor, each experimental condition was performed in triplicate, and experiments were repeated using cells from two other donors (minimum of 3 different donors). Primary HIV-1 strains were propagated in phytohemagglutinin-stimulated human PBMC and titrated as previously described 47,69 . For infections, cells were cultured in media containing HIV-1 for 4 h (each viral strain at MOI of 0.01; each experimental condition tested in triplicate), washed 3 times with serum-free media and cultured for up to 21 days, with culture media changed every 2 or 3 days. RT activity was quantified as we previously described 47 . Tat.AG and Tat.B. Recombinant Tat proteins from a subtype-B HIV-1 isolate (Tat.B) (amino acids 1 to 86; accession number: P69697) were purchased from Diatheva (Viale Piceno, Fano, Italy). Recombinant Tat proteins from HIV-1 CRF02_AG (Tat.AG) (amino acids 1 to 86; accession number: AY371128) were made by Diatheva under a custom-order agreement with our laboratory, using similar procedures as for Tat.B 19,20 . For controls, HIV-1 Tat proteins were heat-inactivated at 100 °C for 30 minutes (min), cooled down and centrifuged at 78 g for 5 min to recover all of the solution. Protein aliquots were stored at −80 °C.

U38 and TZM-bl cells.
Monocytic U38 cells are derivative of U937 cells that contains stably integrated, silent copies of the HIV-1 LTR linked to the chloramphenicol acetyltransferase (CAT) gene. TZM-bl cells stably express large amounts of CD4 and CCR5 and contain integrated copies of the luciferase and β-galactosidase genes under the control of the HIV-1 promoter. U38 and TZM-bl cells were obtained from the NIH AIDS Research and Reference Program and cultured as previously described 47 . Histone acetyltransferase (HAT) assay. Following infection and cellular treatment, nuclear extracts were prepared using the Epigentek nuclear extraction kit (Epigentek, Farmingdale, NY) and total HAT activity in nuclear extracts (20 μg protein) quantified using the EpiQuik TM HAT Activity /Inhibition Assay Kit (Epigentek) as previously described 47 . Chromatin immunoprecipitation (ChIP) analysis. ChIP assays were performed using the ChromaFlash TM Chromatin Extraction Kit and ChromaFlash TM One-Step ChIP kit (Epigentek) as we previously described 47 . Briefly, following treatment, cells pellets were resuspended in ice-cold phosphate-buffered saline (PBS) containing 0.5 mM phenylmethylsulphonyl fluoride, and cellular chromatin was sheared using a closed system ultrasonic cell disruptor (Microson TM , Qsonica LLC, Newtown, CT). Sheared samples were centrifuged www.nature.com/scientificreports www.nature.com/scientificreports/ (17,606 g, 10 min at 4 °C) and chromatin aliquots (supernatants) immunoprecipitated with acetylated H3 and H4 antibodies (Active Motif, Carlsbad, CA) using ChromaFlash TM One-Step ChIP kit (Epigentek), per manufacturer's protocol. Controls included aliquots (10%) of each sheared chromatin ("input" DNA control) and chromatin samples immunoprecipitated with isotype-matched control IgG. Immunoprecipitated samples were amplified by PCR using LTR-specific primers. PCR cycle was as follows: 95 °C, 3 min denaturation; followed by 30 cycles of 95 °C, 20 sec, 55 °C, 20 sec, and 72 °C, 8 sec; 72 °C, 1 min, and hold at 10 °C. Amplified samples were analyzed by agarose gel electrophoresis using 2.5% agarose, and images captured using the G-BOX gel-doc system (Syngene).
Chloramphenicol acetyltransferase (CAT) and iuciferase assays. Following treatment, U38 cells were washed with PBS, lysed using the lysis buffer of the CAT-ELISA kit (Roche Diagnostics Indianapolis, IN), and protein levels quantified using the bicinchoninic acid assay (BCA) assay 65 . The amount of CAT enzyme in each sample (150 μg protein) was quantified as we previously described 47 , using the CAT-ELISA kit and standards (Roche) per manufacturer's protocol.
The luciferase reporter assay was performed as previously described 47 using the Luciferase Assay System (Promega, Madison, WI), per manufacturer's protocol. Briefly, following treatment, cells were washed with PBS, and lysed using the Luciferase Assay Reporter System lysis reagent. Cell lysates (20 μl containing 50 μg proteins) were mixed with 100 μl Luciferase Assay Reagent and luciferase activity was measured using SpectraMax-M5 microplate reader (Molecular Devices, Sunnyvale, CA).

RNA extraction.
Total RNA was extracted from treated cells or animal tissues using the Trizol reagent (Life Technologies-Ambion, Austin, TX), per manufacturer's protocol. The RNA was further cleaned using Total RNA cleanup kit (Qiagen, Valencia, CA). RNA yield and quality were checked using a NanoDrop spectrophotometer (NanoDrop Technologies, Wilmington, DE) and for all samples, absorbance ratio of 260/280 was ≥2.
Real-time PCR. Each experimental condition was tested in triplicate and for each replicate sample, cDNA was generated from 1 μg RNA in a 20 μl reaction volume, using the Verso cDNA kit (Thermo Fisher) per manufacturer's protocol. Reverse transcription and qPCR were performed as previously described 47,70 , using LightCycler ® 480 II (Roche) Real-Time PCR System. Human CD45 and CD4 qPCR was used as internal controls to normalize gene expression. These controls were used at the same primer-probe ratio as the target genes (900 nM of each primer and 250 nM TaqMan MGB probe).
ACH-2 cells were used for normalization based on hCD45+ cells levels in tissues. ACH-2, an HIV-1 latent T-cell clone containing one integrated copy of proviral DNA per cell, was obtained from the NIH AIDS Reagent Program. Standard curves from ACH-2 qPCR were used to quantify HIV-1 LTR, pol, tat, and gag copy numbers in each sample, and results were further normalized to levels of human CD45 cells in each tissue sample. For normalization based on hCD4+ cells levels in tissues, qPCR for hCD4 and HIV-1 LTR, pol, tat, gag was performed for each tissue sample and gene expression levels were quantified using the cycle threshold (ΔC T ) method as described in the software user manual of LightCycler ® 480 II Real-Time PCR System. Each viral gene expression was normalized to the sample hCD4. All PCR reagents, primers, and probes were from Applied Biosystems, and primers' IDs were as follows: LTR (AIWR3QG), pol (AIY9Z2W), tat (AIX01W0), gag (AIo1X84), CD45 (Hs04189704), and CD4 (Hs01058407).
Hu-PBL-NSG mice model. Four-week old NOD/scid-IL-2Rγ c null (NSG) mice were purchased from the Jackson Laboratory (Bar Harbor, ME), maintained in sterile microisolator cages under pathogen-free conditions in accordance with UNMC and NIH ethical guidelines for care of laboratory animals, and bred at the UNMC animal facility to expand the colony. This study was performed under a protocol approved by the UNMC Institutional Animal Care and Use Committee. Mice, 4 to 6 weeks old male, were engrafted by intra-peritoneal (i.p.) injection of human PBL (30 × 10 6 cells/mouse). One week after PBL injection, levels of human CD45 cells in each mouse's blood sample were quantified by FACS to confirm engraftment. Engrafted animals were randomly assigned into 7 groups (10-12 mice per group): a non-infected control group, 3 groups infected with 3 different HIV-1 CRF02_AG strains (AG-3, AG-4, AG-5), and 3 groups infected with 3 different HIV-1 clade-B strains (Ba-L, US1, BX08). For infection, a single dose of 10 4 tissue culture infectious doses-50 (200 μl) of the corresponding HIV-1 strain was injected (i.p.) to animals. Controls were mock-infected by i.p. injection of PBS (200 μl). Animals' blood samples were collected and analyzed at week-1, -2, -3 p.i. Animals were sacrificed at week-3 p.i. and tissues samples harvested and analyzed. FACS analysis. Human (hCD4+, hCD8+, hCD3+, and hCD45+) cells in animal's blood were quantified by FACS, using the following anti-human antibodies: CD45-PE-Cy7, CD8-APC, CD4-FITC and CD3-Pacific blue (Biolegend, San Diego, CA). Briefly, blood (200 μl) collected in EDTA-tubes were centrifuged (543 g, 8 min), plasma collected and cryopreserved, cell pellets resuspended in 50-200 μl FACS buffer (PBS containing 2% fetal bovine serum) and transferred into 5 ml polypropylene round-bottom tubes (BD Falcon, Franklin Lakes, NJ). Antibody cocktail including CD45-PE-Cy7, CD8-APC, CD4-FITC and CD3-Pacific blue was added to each sample and the mixture incubated 1 h on ice. One ml red blood cells lysis buffer (Roche) was then added to each sample, followed by 5 min incubation at RT and centrifugation (377 g, 5 min). Cells pellets were washed 4 times using the FACS buffer, and resuspended in PBS containing 2% paraformaldehyde and analyzed using BD LSRII and FACSDiva 8.0.