Patients and samples
The samples and host/viral genotype data analyzed in this study were obtained from three independent sources: (i) the IMSUT cohort at the Institute of Medical Science, the University of Tokyo, Japan, (ii) the baseline (pre-therapy) cross-section of the HOMER cohort in British Columbia, Canada ( and unpublished), and (iii) a longitudinal multicenter cohort of acute/early infected individuals . The IMSUT cohort, which consists primarily of Asian patients with chronic infection, was used to determine the viral sequences and immune responses.
Forty-six HLA-A*24:02-positive, antiretroviral-naïve, chronically HIV-infected subjects were selected from among patients participating in an ongoing HIV-1-immunopathogenesis study at an HIV outpatient clinic affiliated with the Institute of Medical Science, the University of Tokyo (IMSUT). Study procedures included routine collection of blood samples for virologic and immunologic testing. Peripheral blood mononuclear cells (PBMCs) and plasma samples were separated and preserved in liquid nitrogen or at -80°C, respectively, until use. The study was approved by the internal review board of the Institute of the Medical Science of the University of Tokyo (No. 11–2), and all subjects were adults and provided written informed consent.
A total of 1038 patients from the HAART Observational Medical Evaluation and Research (HOMER) cohort, an open cohort of initially antiretroviral-naïve chronically HIV-infected individuals in British Columbia, Canada of predominantly Caucasian ethnicity, were analyzed in the present study. Plasma HIV-1 RNA sequencing and HLA class I sequence-based typing were performed as previously described . We applied phylogenetically-corrected methods  to determine the strength of association between amino acid variants at Nef codons 133 and 135 in this dataset.
Longitudinal acute/early infection cohort
Kaplan-Meier analysis was used to investigate the time course of selection of specific immune escape mutations at Nef codons 133 and 135 among 16 HLA-A*24 expressing individuals from a longitudinal, multicenter, acute/early HIV-1 infection cohort . “Time to escape” was defined as the number of days elapsed between estimated infection date and first detection of the escape variant (as a full or partial amino acid change).
Plasma viral RNA sequences
Viral RNA was extracted from 140 μl of plasma using the QIAamp viral RNA Mini kit (QIAGEN). Using 4 μl of RNA as starting material, reverse transcription and first PCR were carried out according to the manufacturer’s protocol with SuperScript III One-Step RT-PCR System with Platinum Taq High Fidelity (Invitrogen). Two μl of the first PCR product was subjected to nested PCR, performed using Ex-Taq HS (Takara) with 35 cycles of 30s at 94°C, 30s at 55°C, 60s at 72°C and a final extension for 7 min at 72°C. The primer sets were as follows (Nucleotide positions are those of the published HIV-1 SF2 strain (GenBank accession number: K02007). For the first PCR, primers Nef-1F (5’-GTAGCTGAGGGGACAGATAGGGTTAT-3’) (nt 8,688 to 8,731) and Nef-1R (5’-GCACTCAAGGCAAGCTTTATTGAGGC-3’) (nt 9,632 to 9,607) were used; and for the nested PCR, primers Nef-2F (5’-CGTCTAGAACATACCTAGAAGAATAAGACAGG-3’) (nt 8,746 to 8,777) and Nef-2R (5’-CGGAATCCGTCCCCGCGGAAAGTCCCTTGTA-3’) (nt 9,477 to 9,444) were used. The PCR products were purified with a PCR purification kit (QIAGEN) before sequencing. DNA sequencing was performed using an ABI Prism dye terminator cycle sequencing-ready reaction kit (Applied Biosystems) on a Perkin-Elmer ABI-377 sequencer.
To construct the HLA-A*24:02 expression vector, pcDNA3.1-A24-DsRedm, the HLA-A*24:02 sequence was amplified by PCR using cloned HLA-A*24:02 cDNA as a template [29, 30] and digested with BamHI and NcoI. Primers (5’-TAATACGACTCACTATAGGG-3’) and (5’-CCATGGATCCGCCCCCTCCCACTTTACAAGCTGTGAGAGACAC-3’) were used for the amplification. The pDsRed-Monomer (Clontech) sequence was digested by NcoI and NotI, and was ligated to the 3’ end of HLA-A*24:02 fragment to obtain the HLA-A24-DsRedm fragment. Then HLA-A24-DsRedm fragment was inserted into the multiple cloning site of pcDNA3.1/Hygro(+) vector (Invitrogen).
Mini-Nef and mini-Gag gene expression vectors containing two reporter genes, renilla luceferase (Rluc) and EGFP, together with hygromycin selection were constructed in a pTracer-CMV2 vector (Invitrogen) as follows. Wild-type and mutant mini-Nef genes (from amino acid position 123 to 153) were amplified by PCR using the plasmids containing HIV-1 SF-2 Nef gene with wild-type or mutant sequence as a template [4, 31, 32]. Primer sequences were (5’-GGTACCGCCGCCATGGATTGGCAGAATTACACA-3’) and (5’-GGATCCGCCCCCTCCTACCTTCTCTGGCTC-3’). A mini-Gag gene, extending from amino acid position 18 to 46 in p17, which includes the HLA-A*24:02-restricted CTL epitope Gag28-9 (KYRLKHIVW), was amplified using a plasmid containing the 5’ half of the HIV-1 SF2 strain [31, 32]. The primers used were 5’-GGTACCGCCGCCATGAAAATTCGGTTAAGG-3’ and 5’-GGATCCGCCCCCTCCGACTGCGAATCGTTC-3’. The Rluc and hygromycin genes were amplified from pGL4.77 hRlucP/Hygro (Promega) using primers 5’-GGATCCATGGCTTCCAAGGTGTAC-3’ and 5’-TCTAGAGTCGCGGCCTTAGACGTT-3’.
KpnI/BamHI fragments of mini-Nef or mini-Gag, BamHI/XbaI fragment of Rluc and hygromycin genes amplified from pGL4.77 hRlucP/Hygro were ligated to the KpnI/XbaI fragment of pTracer-CMV2 vector (Invitrogen) to create a mini-Nef or mini-Gag expression vector, pmNef(wt)-hRluc-EGFP and pmGag(wt)-hRluc-EGFP, respectively. In the final step the GFPz sequence was replaced by EGFP sequence (Clontech).
Synthetic peptides were purchased from Sigma-Genosys. The peptides used in the screening of immune response by ELISpot had a purity of 70% or more. All other peptides were more than 95% pure as determined by high-performance liquid chromatography and mass spectroscopy.
Cells and media
T2-A24, a kind gift from K. Kuzushima, was cultured in RPMI 1640 (Sigma) supplemented with 100 U of penicillin/ml, 100 U of streptomycin/ml, 10% heat-inactivated fetal calf serum (FCS) (Sigma), and 0.8 mg of G418 (Invitrogen)/ml . We established Nef126-10 and Nef134-10-specific CTL clones, I30-1 and H27-9, as previously described . CTL clones were cultured with RPMI 1640 supplemented with 50 U of interleukin-2/ml, 100U of penicillin/ml, 100U of streptomycin/ml, and 10% heat-inactivated FCS (R10/50), but the clones were cultured in the absence of interleukin-2 (R10) for two days before antigen presentation assays. pcDNA3.1-A24-DsRedm was introduced into 293FT cell line (Invitrogen) and the cells were treated by hygromycin for 2 weeks. After cloning by limiting dilution we obtained 293FT-A24DRm-CY0, and confirmed HLA-A*24:02 expression with FACS analysis by using anti-HLA-A9 serotype antibody (One Lambda, data not shown).
IFN-γ ELISpot assay
The gamma interferon enzyme-linked immunospot (IFN-γ ELISpot) assay was performed using patients’ PBMCs as previously described  with some modifications. In brief, 96-well plates (Millipore) were coated with anti-gamma-interferon (IFN-γ) MAb 1-D1k (Mabtech) overnight at 4°C. Peptides were added directly to the wells at a final concentration of 10-5 M. 5 ~ 10 × 104 cells were added to each well with a final volume of 100 μl of R10. For negative controls, PBMCs were incubated with R10 alone without peptides. After incubation at 37°C under 5% CO2 overnight (16 to 18 h), the plates were washed six times with phosphate-buffered saline containing 0.01% tween-20 (PBST). Biotinylated anti-IFN-γ MAb 7-B6-1 (Mabtech) was added, and was incubated for 2 hours at 37°C under 5% CO2. After washing with PBST, streptavidin-alkaline phosphatase conjugate (Mabtech) was added and the plates were kept at room temperature for 45 min. After washing with PBST, IFN-γ-producing cells were detected as dark spots after 10- to 20-min color reaction with 5-bromo-4-chloro-3-indolylphosphate and nitroblue tetrazolium by using AP Conjugate Substrate Kit (Bio-Rad). Spots were counted by KS ELISPOT compact (Carl Zeiss) and expressed as spot-forming units (SFU) per 106 PBMCs after subtracting the SFU of the negative control. Values with >50 SFU, >3 × mean SFU of negative control and > mean SFU of negative control + 3 SD per 106 input cells were considered as a positive response.
Since more cells were required for the immune response screening (Figure 1B) and functional avidity assays (Figures 4B and 6C), PBMCs were stimulated with anti-human CD3 antibody and the T cells were expanded for 2 to 3 weeks in R10/50 (BD Pharmingen). Culture media was changed to R10 two days prior to the assay date. For ex vivo IFN-γ ELISpot assay (Figures 5B and 6B), PBMCs were cultured for 6 hours in R10 media.
Peptide-HLA binding assay (Figure 2A)
Peptide binding to HLA-A*24:02 was assessed by using a T2-A24 stabilization assay as previously described [4, 33]. Briefly, after incubation for 16 hours at 26°C under 5% CO2, 2 × 105 T2-A24 cells were incubated with 10-4 to 10-9 M peptides for 1 h at 4°C. After keeping at 37°C under 5% CO2 for 3 hours, the cells were stained with biotinylated anti-human HLA-A9 monoclonal antibody (One Lambda), and streptavidin-APC conjugates (BD Pharmingen). The mean fluorescence intensity (MFI) was measured by FACSCalibur (Becton Dickinson). In each experiment, MFI of samples was normalized by the MFI of 10-4 M control peptide, Nef134-8(RYPLTFGW). Three independent experiments were performed.
CTL clones and Epitope recognition (Figure 2B)
Nef134-10- and Nef126-10-specific clones, H27-9 and I30-1, were established by Nef134-10(wt) and Nef126-10(wt) peptide stimulation respectively, and limiting dilution of PBMCs from HIV-1-infected patients harboring 133 T/135 F. For in-vitro peptide stimulation, 5 × 105 PBMCs were pulsed with 10 μM of each peptide for 1 h. The cells were washed twice with R10, then cultured with 1 × 106 autologous PBMCs and 4 × 106 irradiated (3300 rad) allogeneic PBMCs in R10. After 4 days, IL-2 was added to 50 U/ml and the cells were cultured for 2–3 weeks in R10/50. Peptide-specific CD8+ T cells were enriched by MACS separation (Miltenyi) using tetramers. The sorted cells were cloned by limiting dilution to 3 or 10 cells/well in 96-well round-bottom tissue culture plates, and cultured with 105 irradiated allogeneic PBMCs in R10/50 containing 5 μg/ml PHA-L.
CTL-recognition of the epitopes was assessed by serially diluted peptides. On day 0, 293FT-A24DRm-CY0 cells were seeded onto 96-well Flat-bottom transparent plate (BD Falcon) so that the cultures become confluent on day 2. On day 2, each peptide was pulsed with concentrations from 32 to 3-6 μM to the wells and incubated at 37°C under 5% CO2 for 1 h. Then CTLs (5,000 ~ 10,000 cells) were added and co-cultured at 37°C under 5% CO2 for 18 to 24 h. After the incubation, supernatants were harvested, and IFN-γ concentrations were quantified by Human IFN-γ ELISA Set (BD Bioscience). In each experiment, the IFN-γ value of samples was normalized to that of the highest wild type peptide concentration (9 μM). For example, each IFN-γ value of I30-1 CTL clone was divided by the value of the well pulsed with 9 μM Nef126-10(wt) peptide. Each assay was performed in duplicate and three independent experiments were conducted.
Antigen presentation assay (Figure 3)
Antigen presentation was assessed by measuring epitope-specific CTL responses to endogenously expressed antigen. First, intracellular expression of each antigen was extrapolated from the activity of reporter protein, Rluc. 293FT-A24DRm-CY0 cells were seeded in 96-well plate (Nunc) on day 0. On day 1, antigen expression plasmids, pmNef(wt)-hRluc-EGFP, pmNef(135F)-hRluc-EGFP, pmNef(133T135F)-hRluc-EGFP, or control vectors (pmGag(wt)-hRluc-EGFP) were transfected into the cells in each well using FuGENE HD (Promega). The cultures were incubated at 37°C under 5% CO2 for 18 to 24 h. On day 2, transfection efficiency was inspected roughly under a fluorescence microscope (KEYENCE BZ-9000), then Rluc activity in each transfected well was measured by using Dual-Glo Luciferase Assay System (Promega) and luminometer (Promega GloMax 96 Microplate Luminometer). The assay was performed in triplicate.
Second, CTL responses against endogenously expressed and processed epitopes were evaluated. 293FT-A24DRm-CY0 cells were seeded onto a 96-well Flat-bottom transparent plate (BD Falcon) on day 0. On day 1, expression plasmid was transfected to each well using FuGENE HD (Promega). CTLs (5,000 ~ 10,000 cells) were added to the transfected wells on day 2. After incubation for 18 to 24 h, the supernatant of each well was harvested, and IFN-γ secretion was quantified by Human IFN-γ ELISA Set (BD Bioscience). IFN-γ ELISA was performed in duplicate.
Experiments were performed in duplicate on three independent occasions. To normalize the values in each experiment, mean IFN-γ values of each sample were normalized by the mean value of reference wells in duplicate. In the reference wells, 9 μM of the wild type peptide was pulsed to the antigen presenting cells and then co-cultured with CTLs.
Protein expression in E. coli, refolding and purification
HLA-A*24:02 and β2m were expressed in E. coli and refolded from inclusion bodies as previously described with some modifications . HLA-A*24:02 (18 mg), β2m (6 mg) and peptide (4 mg) were mixed in 400 ml of refold buffer containing 100 mM Tris, pH 8.0, 400 mM L-arginine-HCl, 2 mM EDTA, 5 mM GSH, 0.5 mM GSSG, 0.2 mM PMSF. The refolded protein was purified by Superdex 75 column, followed by Mono Q column, and subsequently concentrated to 10 mg/ml in 20 mM Tris, pH 8.0, 50 mM NaCl for crystallization.
Crystallization, data collection and structure determination
The crystallization was done by the sitting drop vapor diffusion method at 20°C. Crystals of the A24/N126-10(8T10F) complex were obtained in 20% (w/v) PEG 3350, 200 mM sodium phosphate dibasic, and those of A24/N126-10(8I10F) were obtained in 20% (w/v) PEG 3350, 200 mM sodium nitrate. For cryoprotection, crystals were soaked briefly in reservoir solutions containing 20% ethylene glycol, and then frozen in liquid nitrogen before data collection. Data were collected at the beamline BL41XU in SPring 8 (Hyogo, Japan), and processed with HKL2000  and the CCP4 program suite .
The structure were determined by molecular replacement using Molrep . The search model was the coordinate file of PDB (Protein Data Bank) code 3I6L with omitted peptide for A24/N126-10(8T10F). Model building and refinement were carried out using Coot  and REFMAC5.6 implemented in CCP4, respectively. The structure of A24/N126-10(8I10F) was determined with the refined A24/N126-10(8T10F) as a search model and refined as described above. The stereochemistry of the refined models was assessed with RAMPAGE . All molecular graphic representations were created with the program PyMOL (DeLano Scientific; http://www.pymol.org). Data collection and refinement statistics are shown in Additional file 2: Table S1.
Functional avidity assay
Basically, the assay was performed using the same procedure with IFN-γ ELISpot assay as described above. PBMCs were cultured for 2 to 3 weeks in R10/50 after anti-human CD3 antibody (BD Pharmingen) stimulation, and culture media were changed from R10/50 to R10 two days before use. PBMCs were incubated with peptides at concentrations from 10-5 to 10-12 M, and SFU was calculated. The functional avidity to peptide dilutions was determined as a 50% of sigmoidal dose (SD50) SFU.
All data visualization and statistical analyses were performed using GraphPad Prism (GraphPad Software, La Jolla, CA). Student’s t-test and Mann–Whitney U-test were used to compare the antigen presentation and functional avidity between two groups, respectively. Spearman rank correlation was used to calculate the correlation between peptide-specific response and pVL. Dose at 50% response in sigmoidal dose–response curves (SD50) was calculated by drawing sigmoidal dose–response curves. Time to mutational escape, defined as the time elapsed between estimated date of HIV-1 infection and the first appearance of a full or partial amino acid change consistent with the specific escape mutation of interest, was calculated using Kaplan-Meier (survival) methods.