TZM-bl cells were obtained from the NIH HIV Reagent Program (Derdeyn et al., 2000; Wei et al., 2002). TZM-bl and 293T cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% heat-inactivated fetal bovine serum (HI-FBS), 2mM glutamine, 100U of penicillin per ml, and 100μg of streptomycin per ml (p/s; DMEM complete). The immortalized PTM CD4⁺ T cells, obtained from Dr. Hans-Peter Kiem (Fred Hutchinson Cancer Research Center), were maintained in Iscove’s Modified Dulbecco’s Medium (IMDM) containing 10% HI-FBS, 2mM glutamine, p/s, and 100U/ml human interleukin 2 (IL-2; Roche; Munoz et al., 2009). CEM×174 were obtained from the American Type Culture Collection and cultured in Roswell Park Memorial Institute (RPMI) media with 10% HI-FBS, 2mM glutamine, 100U of penicillin per ml, and 100μg of streptomycin per ml (RPMI complete).

Total CD4⁺ T cells were isolated from 1×10⁷ PBMCs recovered at 196 and 200wpi of pigtailed macaque M08009 by negative selection using the Miltenyi nonhuman primate CD4⁺ T cell isolation kit (Miltenyi Biotech). The cells were isolated according to the manufacturer’s protocol. The M08009 CD4⁺ T cells were cocultured with the human T cell-B cell hybrid cell line, CEMx174 for up to 16days. Supernatants after 7 and 16days were assayed for HIV-1 p24 antigen by ELISA (Advanced Bioscience Laboratories). If positive, supernatants were passed through 0.45-μm syringe filters, aliquoted, and frozen at −80°C. The infectious titers of the stocks were determined by limiting dilution infection analysis using TZM-bl reporter cells and luciferase assay as described (Misra et al., 2018). Infectious virus recovered from CD4⁺ T cells at 196 and 200wpi was named C/196 and C/200, respectively.

TZM-bl cells (1×10⁴ cells per well) were plated in wells of a 96-well plate in DMEM complete with 30μg/ml DEAE-dextran. In triplicate cultures, cells were treated with either the CXCR4 inhibitor, AMD3100, or CCR5 inhibitor, Maraviroc, such that after adding 250 infectious units of C/196, C/200, or control viruses HSIV-vif_NL4-3 or HSIV-vif_AD8 the final concentrations of inhibitors were 1μM, 500nM, or 250nM, and the final concentration of DEAE-dextran was 20μg/ml. After 2days of infection, the cells were washed once with PBS and lysed with Promega Glo Lysis buffer. Lysates were assayed for luciferase activity using the Promega luciferase assay system and tube luminometer according to the manufacturer’s instructions (Promega).

Neutralizing antibody titers in serum specimens from PTMs infected with HSIV-vif_NL4-3 were determined using a TZMbl-based neutralization assay as described previously (Wu et al., 2006). Serum samples were heat-inactivated at 56°C for 30min prior to use.

Construction of the HSIV-vif clones based on NL4-3, NL-AD8, and Bru-Yu2 has been reported before (Thippeshappa et al., 2011). To generate Vpr⁺ HSIV-vif_NL4-3 and HSIV-vif_AD8 clones, SphI to SalI fragment of HSIV-vif_NL4-3 encompassing HIV gag, pol, SIV vif, and HIV-1 vpr genes was cloned into pCR2.1 TOPO vector (Thermofisher). SIV vpx start codon and two additional ATG codons upstream of the HIV-1 vpr start codon were mutated by Quickchange mutagenesis (Stratagene) and the sequence between the stop codon of vif and start codon of vpr were deleted. After mutagenesis, SphI and SalI fragment was cloned back into HSIV-vif_NL4-3 and HSIV-vif_AD8. Similarly, SphI to SalI fragment of HSIV-vif_Yu2 was cloned into SacI site removed pUC19 vector (Thermofisher), ATG codons upstream of vpr were mutated, and cloned back into HSIV-vif_Yu2 (Supplementary Figure 1). HSIV-vif-Vpx_Yu2 clone was generated by cloning SacI to NcoI fragment of SIVmne027 vpx gene into pUC19 vector containing SphI to SalI fragment of HSIV-vif_Yu2. SphI to SalI fragment containing full length vpx was confirmed by sequencing and cloned back into HSIV-vif_Yu2 (Supplementary Figure 2).

Virus stocks were generated by transfection of 293T cells with each plasmid clone of HSIV-vif using Fugene 6 or X-tremeGENE 9 DNA transfection reagent according to the manufacturer’s protocol (Roche). Infectious titers were determined by limiting dilution infection analysis using TZM-bl indicator cells, and the amount of virus in supernatants was measured by HIV-1 p24^gag antigen ELISA (Advanced Bioscience Laboratories).

293T cells, seeded the day before, in 6-well plates were transfected with HSIV-vif plasmids using Fugene 6 or X-tremeGENE 9 DNA transfection reagent (Roche/Sigma). At 48h post-transfection, cell culture supernatants were used to concentrate viral particles by centrifugation at 23,600×g for 1h at 4°C. Viral particles were mixed with 2X SDS samples buffer and separated by SDS-PAGE using Tris-HCl ready gels (Bio-Rad). Cell lysates were prepared as previously described (Thippeshappa et al., 2011). Proteins were transferred to either nitrocellulose or PVDF membrane and probed with rabbit antiserum to Vpr, Nef, or Vpx. Goat anti-rabbit IgG HRP (Promega) was used as secondary antibody. Antiserum to Vpr (Catalog # ARP-11836), Nef (Catalog # ARP-2949), Vpx (Catalog # ARP-2609), and anti-HIV-1 p24 Gag monoclonal (ARP-6458) were obtained from NIH HIV reagent program.

For viral replication assays, PTM PBMCs and immortalized PTM CD4⁺ T cells were infected as previously described (Thippeshappa et al., 2011, 2013). PTM PBMCs were isolated using Ficoll hypaque gradient method. PBMCs were activated with concanavalin A (7μg/ml) for 3days. Cells were then washed twice and resuspended in RPMI complete media containing 40U/ml IL-2 (Sigma) and cultured for 2days. Approximately 1×10⁶ activated PBMCs were infected in duplicate at a multiplicity of infection (MOI) of 0.01. To compare viral replication in PTM CD4⁺ T cells, approximately 2.5×10⁵ or 5×10⁵ cells in 100U/ml IL-2 containing IMDM were infected in duplicate with viruses at a MOI of 0.01 or 0.05. Human monocyte-derived macrophages (MDMs) were generated from PBMCs using previously described methods (Kimata et al., 1998, 2004; Biesinger et al., 2010). PBMCs of anonymous donors were isolated from leukopacks purchased from the Gulf Coast Blood Center, Houston, TX, Unites States. Briefly, human monocytes were isolated from PBMCs by plate adherence methods. Approximately 4×10⁶ PBMCs were plated into each well of a 24-well plate and monocytes were allowed to adhere to the plate for 1h. Monocytes were then stimulated with RPMI complete containing 10U/ml GM-CSF (Invitrogen) for 7–10days to generate MDMs. Human MDMs were infected with the Vpr⁺ and Vpr-HSIV-vif in duplicate. Infection experiments were conducted at least 2–3 times. PBMCs from different donors were also used. After 3h of incubation, the cells were washed twice with phosphate buffered saline (PBS) or complete medium to remove unbound virus. Infected cells were then resuspended in RPMI complete media containing IL-2. To study the effect of IFNα, 200U/ml of Interferon-αA/D (IFN-αA/D or IFNα, Sigma) was added to the culture media. To monitor replication of HSIV-vif clones, supernatants were harvested every 2–4days for measurement of HIV-1 p24^gag antigen using ELISA kit (Advanced Bioscience Laboratories or ExpressBio). Statistical analysis in GraphPad Prism was performed to compare groups using the Mann-Whitney test.

Four PTMs specific pathogen free for simian T lymphotropic virus type 1, SIV, simian retrovirus type D, and herpes B virus were enrolled for the study. All animals were housed and cared for in accordance with the guidelines of the American Association for Accreditation of Laboratory Animal Care and the Animal Care and Use Committee of the University of Washington. In passage 1: two PTMs were inoculated IV with a mixture of CXCR4 and CCR5-tropic viruses. In passage 2: pooled peripheral blood from passage 1 PTMs collected at 14wpi was used for inoculation of 1 PTM. In passage 3, peripheral blood from passage 2 PTM at 8wpi was used for transfusion into an additional PTM. At several time points post-inoculation, peripheral blood was drawn for CD4⁺ T cell count determinations and isolation of plasma, sera, and PBMC.

Plasma viral load measurements were determined by using the Roche Amplicor HIV-1 monitor test, version 1.5 according to the manufacturer’s protocol. CD4⁺ T cell counts were determined as previously described (Polacino et al., 2007). HIV-1-specific antibodies were measured by ELISA as previously described, using gradient-purified and disrupted whole HIV-1 virions as the capture antigen (Hu et al., 1989; Polacino et al., 2007).

The following steps were performed to generate IMCs (Supplementary Figure 3).

Sequences of HSIV-P3-114, HSIV-P3-161, and HSIV-P3-284 are deposited under accession numbers MZ146778, MZ146779, and MZ146780, respectively.

We previously reported HSIV-vif_NL4-3 could persistently infect PTMs (Thippeshappa et al., 2011). We continued to monitor the viral loads in two of those HSIV-vif_NL4-3-infected PTMs for nearly 4years (Figure 1A). Although viral RNA was below the detection limit at most of the late time points measured, viral DNA could be detected in PBMCs through 200wpi. Additionally, we recovered infectious virus from PBMCs at 196 and 200wpi, suggesting that the virus had been replicating in the animals for nearly 4years. Interestingly, one of the animals (M08009), despite low or undetectable viral loads, showed a gradual decline and then stably depressed CD4⁺ T cell counts, suggesting disease progression (Figure 1B).

We recovered infectious virus by coculturing peripheral blood CD4⁺ T cells from infected PTM (M08009, Figure 1) at 196 and 200wpi with CEMx174 cells (biological isolates C/196 and C/200). C/196 and C/200 were susceptible to inhibition by AMD3100 but not Maraviroc, suggesting that viral isolates were only CXCR4-tropic (Supplementary Figure 4). We next determined the replication capacity of C/196 and C/200 in PTM CD4⁺ T cells. Although the differences were not statistically significant, C/196 replicated to higher levels in PTM CD4⁺ T cells in both the presence (22-fold) and absence (17-fold) of IFNα relative to the parental clone, HSIV-vif_NL4-3 (Figure 2). While C/200 displayed more limited replication, it was not affected by the addition of IFNα (Supplementary Figure 5). Partial genome sequencing analysis of PCR amplified fragments of C/196 and C/200 indicated that both isolates were clonal (Supplementary Figure 6). We also observed that both C/196 and C/200 were neutralization resistant to sera from M08009, the animal in which it evolved, as well as sera from the other PTM, F08003 that had been infected with HSIV-vif_NL4-3 (Table 1). The emergence of immune escape variants of HSIV-vif_NL4-3 suggest that it had persistently replicated in PTMs.

An explanation for the low but persistent replication of HSIV-vif_NL4-3 in vivo is that the virus is attenuated because it does not express accessory proteins necessary for more robust viral replication. Two possible proteins that could enhance replication of HSIV-vif_NL4-3 are the HIV-1 Vpr or SIV Vpx. Since we introduced SIV vif, which includes a partial open reading frame (ORF) for vpx into HIV-1 backbone, we determined whether it had affected the expression of HIV-1 vpr. Indeed, HIV-1 Vpr protein was not observed in virions of HSIV-vif clones, HSIV-vif_NL4-3 or HSIV-vif_AD8 (Figure 3). We determined that this is because singly spiced HIV-1 vpr RNA from HSIV-vif is generated using the splice acceptor site within the SIV vif gene (Supplementary Figure 1). The transcript, therefore, includes the partial ORF for the SIV Vpx protein upstream of the translational initiation site for HIV-1 Vpr, which may interfere with its expression (Supplementary Figure 1).

Vpr is a small 96 amino acid (14kDa) protein that is not required for HIV-1 replication in vitro. However, it is conserved among all primate lentiviruses, indicating its importance for pathogenesis (Tristem et al., 1992, 1998). Therefore, we modified HSIV-vif clones to express Vpr. We deleted the sequence between the stop codon of vif and start codon of vpr and disrupted the vpx translational start codon and other ATG codons upstream of the vpr initiation site by site directed mutagenesis (Supplementary Figures 1 and 8). Mutation of these three ATG codons, one of which results in M181L in SIV Vif, resulted in expression of Vpr, which is incorporated into progeny virions (Vpr⁺ HSIV-vif_NL4-3 and Vpr⁺ HSIV-vif_AD8, Figure 3). Antibody to Nef was used as a control for incorporation of a virion-associated protein in HSIV-vif clones. Additionally, since Vpx counteracts the function of SAMHD1 (Hrecka et al., 2011; Laguette et al., 2011) and is essential for replication of SIV in macaques (Hirsch et al., 1998; Belshan et al., 2012; Shingai et al., 2015) and because HSIV-vif already has a partial ORF for vpx, we also generated an HSIV-vif derivative carrying the full-length vpx gene (Supplementary Figure 2). We used HSIV-vif_Yu2, which is IFN-resistant (Thippeshappa et al., 2013), to generate HSIV-vif-vpx carrying the full-length vpx gene (named HSIV-vif-vpx_Yu2). By Western blot using rabbit anti-serum to HIV-2_Rod Vpx protein, Vpx could be detected in the cell lysates of SIV, HIV-2, and HSIV-vif-vpx_Yu2, but not SIVΔVpx. However, it was not detected in virion lysates of HSIV-vif-vpx_Yu2 (Supplementary Figure 2) as HIV-1 does not have determinants in p6 Gag required for virion incorporation of Vpx (Sunseri et al., 2011).

We tested the effect of Vpr and Vpx expression on HSIV-vif replication in an immortalized PTM CD4⁺ T cell line (Munoz et al., 2009). PTM CD4⁺ T cells were infected with Vpr- (HSIV-vif_NL4-3, HSIV-vif_AD8, and HSIV-vif_Yu2), Vpr⁺ (Vpr⁺HSIV-vif_NL4-3, Vpr⁺HSIV-vif_AD8, and Vpr⁺HSIV-vif_Yu2), HSIV-vif-vpx_Yu2, or wild type HIV-1 (NL4-3, NL-AD8, and Bru-Yu2) viruses at an MOI of 0.01. Vpr⁺ HSIV-vif viruses, and HSIV-vif-vpx_Yu2 replicated in PTM CD4⁺ T cells to similar levels as Vpr- viruses (Figure 4). Expectedly, wild type HIV-1 (NL-AD8 and Bru-Yu2) failed to replicate (Figure 4).

We also determined the replication capacity of vpr and vpx carrying HSIV clones in human MDMs. Human MDMs were generated using previously described methods (Biesinger et al., 2010) and infected with the Vpx⁺, Vpr⁺, and Vpr- HSIV-vif clones at an MOI of 0.01. Vpr⁺ HSIV- vif_AD8 (Figure 5A) and Vpr⁺ HSIV-vif_Yu2 (Figure 5B) replicated to similar levels as Vpr- HSIV-vif_AD8 and Vpr⁺ HSIV-vif_Yu2, but slightly less than wild type HIV-1 NL-AD8 and Bru-Yu2 (Figures 5A,B), although it was not statistically significant. Additionally, HSIV-vif-vpx_Yu2 replicated as well as the parental HIV-1 Bru-Yu2 (Figure 5B).

Because HSIV-vif_NL4-3 replication in the initial PTM experiment was low with peak viremia <10⁵ copies/ml (Figure 1; Thippeshappa et al., 2011), we conducted animal to animal transfer of infected PTM peripheral blood to adapt HSIV-vif to PTMs. For this experiment, the initial inoculum contained a mixture of CXCR4- (C/196 and C/200) and CCR5-tropic viruses (Vpr⁺ HSIV-vif_AD8, Vpr⁺ HSIV-vif_Yu2, and HSIV-vif-vpx_Yu2). At 14wpi, pooled blood from infected PTMs (Z09080 and Z09067) was used to inoculate a naïve macaque (Z13086). At 8wpi, peripheral blood from Z13086 was transferred into an additional PTM (Z13098). Interestingly, all the macaques showed a peak viremia close to or above 10⁵ copies/ml and the viral loads persisted for at least 20wpi (Figure 6A). Furthermore, increases in antibody titer over time suggest that all PTMs were persistently infected with HSIV-vif (Figure 6B). However, CD4⁺ T cell decline was not observed in the infected PTMs (Figure 6C).

We generated IMCs from proviral DNA isolated from PBMCs of the passage 3 macaque (Z13098). Near full-length proviral genomes (NFLG) were amplified using nested PCR as described by Hiener et al. (2017) with slight modifications in PCR conditions. NFLGs was cloned into a vector PCR product containing 5′ LTR sequences (amplified from HSIV-vif_NL4-3 plasmid) using NEBuilder HiFi assembly mix (Supplementary Figure 3). Briefly, ends of NFLG PCR product and vector PCR product containing 5′LTR sequences overlap with each other, which can be assembled using NEBuilder HiFi assembly mix. We screened nearly 300 colonies to identify 54 plasmids with full length genomes. To determine whether they could produce infectious virus, full length clone plasmids were transfected into 293T cells to generate virus. Infectious nature of the supernatants was determined by infecting TZM-bl cells. Three of the 54 plasmid clones tested (HSIV-P3-114, HSIV-P3-161, and HSIV-P3-284) produced measurable infectious virus. DNA sequencing showed that the three IMCs were closely related to the C/196 and C/200 biological clones of HSIV-vif_NL4-3 (Table 2; Supplementary Figures 5, 6). Importantly, the three IMCs had more nonsynonymous mutations (Table 2) than synonymous mutations (Supplementary Table 1) throughout the genome suggestive of adaptation to PTMs. Most of the mutations in Env and Nef were seen in the biological clones of HSIV-vif_NL4-3 (C/196 and C/200) recovered from PTM M08009 (Table 2), suggesting that these mutations have persisted through three additional in vivo passages. In the Vpr⁺ HSIV-vif clones, we had mutated SIV vpx ATG codon to ACG (silent mutation) and ATG codon at amino acid position 181 in SIV vif to TTG, which codes for leucine (M181L substitution; Supplementary Figure 1). However, SIV vpx start codon was present in the recovered IMCs. Additionally, ATA codon was present at amino acid position 181, which codes for isoleucine (M181I substitution). Further, all the recovered IMCs had the deletion of bases between SIV vif stop codon and vpr start codon (Supplementary Figure 8), suggesting that these clones could express Vpr protein. Third, recombination analysis using RAPR program suggested that the recovered IMCs are recombinants that consist of biological clone C/196 with an insertion of 737bp region spanning the 3′ end of vif to 5′ end of vpu (nucleotides 5,447–6,171 according to HXB2 reference sequence) from Vpr⁺HSIV-vif_AD8. We confirmed the expression of Vpr from HSIV-P3 IMCs by western blot using rabbit anti-sera against Vpr protein (Figure 7). Interestingly, virion-associated Vpr was higher than that for the VPR⁺ HSIV-vif_NL4-3 clone.

We determined whether HSIV-P3 IMCs (HSIV-P3-114, HSIV-P3-161, and HSIV-P3-284) replicate in PTM PBMCs. PBMCs isolated from different donor PTMs were activated with concanavalin A for 3days and maintained in IL-2 containing media for 2days. Activated PBMCs were infected with HSIV-P3 IMCs and Vpr⁺ HSIV-vif_NL4-3 at an MOI of 0.01. Viral supernatants were collected at various days post-infection to assay for p24 levels. We observed that HSIV-P3 IMCs replicated with different efficiency in PBMCs from different donor PTMs (Figure 8). Among the three HSIV-P3 IMCs, HSIV-P3-284 replicated to similar levels as Vpr+HSIV-vif_NL4-3 in activated PBMCs.