Inhibition of HIV-1 immune modulation by small molecules targeting viral Nef-host CD80 interface

HIV-1 causes diverse immunomodulatory responses in the host, including the down-regulation of co-stimulatory proteins CD80/86, mediated by HIV-1 protein Nef, blunting T-cell activation. Using a screening cascade of biochemical and cell-based assays, we identified potent small molecules representing three chemical scaffolds namely amino pyrimidine, phenoxy acetamide and bi-aryl heteroaryl carbamate which target the protein-protein interaction interface of CD80/86 and Nef with sub-micromolar potency. These molecules restore CD80/86 surface levels in HIV-1-Nef infected antigen presenting cells and T-cell activation. Nef-CD80 interface and small molecule binding sites were mapped by using computational docking and structural studies, followed by validation by mutational analysis. This analysis resulted in the identification of two key residues, K99 and R111, which were associated with down-modulation of CD80 surface levels by Nef and important for small molecule binding. Targeting these interacting residues disabled Nef-mediated down-modulation of CD80 surface levels, consequently restoring T-cell activation. Thus, we validate a new target, the Nef-CD80/86 protein-protein interaction interface, with a potential to develop new inhibitors to counteract the immunomodulatory consequences of HIV-1.


Introduction
Over the past decade, there has been tremendous effort in finding newer therapies 1 for HIV/AIDS. Increased use of anti-retroviral drugs has been accompanied by the 2 steady increase in HIV drug resistance and viremia that often result in 3 immunosuppression leading to morbidity 3 . Drug resistance is mainly transmitted at 4 the time of infection or acquired during previous treatment for instance in women 5 given anti-retroviral drugs to prevent mother-to-child transmission of HIV 4 . Other 6 treatment strategies like using anti-HIV-1 antibodies (bnAbs) are reported in 7 combination with ART which have the capacity to impact on HIV-1-specific T cell 8 immune responses in infected humans but whether it controls the virus remains to be 9 determined 5 . In addition, with the vaccine trials failing to elicit broad plasma 10 neutralization of primary virus isolates 6 , there is an urgent need to fast-track the 11 transition to newer antiviral drug regimens which needs to be administered in multi-12 drug combinations in order to combat the ever-evolving virus. Increasing drug 13 resistance also emphasizes the requirement of new molecules that can target the 14 host-viral interface with high specificity 7,8 . 15 To obtain targets for the next generation of HIV-1 therapy we explored the interaction 16 interface of the essential HIV-1 accessory protein, Nef 9 , with host proteins. Nef is a 17 27-35 kDa protein expressed during early phases of viral replication and helps 18 study their binding sites.
In the current study, we show that Nef directly binds to cytoplasmic tail peptides of 49 CD80 and CD86. We have identified small molecules, which can abrogate Nef 50 interactions with CD80 and/or CD86. The compounds mainly belong to 3 scaffolds 51 amino pyrimidine (AP), phenoxy acetamide (PA) and bi-aryl heteroaryl carbamate 52 (BC) having nanomolar to micromolar inhibition potencies in vitro. Representative 53 actives from these scaffolds were then validated in functional cell-based assays for 54 reverting the down regulation of cell surface Nef-mediated co-stimulatory protein 55 expression and re-establishing T-cell activation. These identified actives also 56 reversed similar Nef-mediated effects after viral infections. 57 To further improve the efficacy of these leads, we used an in-silico approach to 58 explore the binding mode of Nef with the co-stimulatory molecules. Full length Nef 59 protein was modeled and potential binding sites for CD80 were identified. Based on 60 these predictions a selected subset of CD80 interacting residues in Nef were 61 mutated and the corresponding mutants were examined in assay platforms to 62 confirm the model and its functional consequences. The understanding of the 63 druggable pocket opens the scope for future lead optimization work. Altogether, we 64 report a chemical strategy to inhibit Nef-mediated immunomodulatory functions, 65 which prevents immune evasion of HIV-infected cells. While these molecules may be 66 developed for future therapeutic intervention, at the current stage, they may be used 67 as chemical biology tools to understand the role of host-pathogen interface in the 68 form of Nef-CD80/86 interaction surface in HIV immune evasion. 69

levels; relevant controls included T-cells alone, T-cells and B-cells in the absence of 172
anti-CD3 antibody (Fig. 3b). 173 We tested two different modes of addition of compounds: in the first mode, we pre-174 treated APCs with 1, 10 and 100 µM of compounds (AP5, PA4 and BC5) for 24h 175 then exposed the cells to Nef-carrying viral particles and assayed for IL-2 release. 176 We observed a dose dependent response with all 3 compounds AP5, PA4 and BC5 177 showing a 4-fold increase (**p ≤ 0.01) in IL-2 release at 100 µM as compared to Nef 178 virus control. At lower concentrations, compounds AP5 and PA4 showed 2-fold 179 increase (*p ≤0.05) at 10 µM. PA4 did not show significant IL-2 release at 10 µM 180 ( Fig. 3c), thus, indicating restoration of T-cell activation in a dose dependent 181

manner. 182
In the second mode, we added compounds at 1, 10 and 100 µM concentrations post 183 viral exposure for 96h. AP5 showed significant increase in IL-2 levels (3-fold) at all 184 concentrations when compared to Nef virus control (**p ≤ 0.01), while PA4 showed 185 2-fold increase (*p ≤0.05) in IL-2 levels at 10 µM concentration and no changes in IL-186 2 levels were observed in BC5. Interestingly, there was a reduction in IL-2 release at 187 100 µM dose in comparison to 1 µM and 10 µM, perhaps due to the toxicity 188 associated with this dose alongside viral effects (Fig. 3d). Thus far, these data 189 shows that AP5 is a potent molecule both in vitro biochemical assays and in 190 restoring T-cell activation in an assay designed to assess the role of co-stimulation 191 dependent T-cell activation mediated by CD80/86. 192 193

Structural insights into Nef-CD80 interaction 194
To further understand the nature of inhibition of Nef-CD80 interaction by small pocket. This requires the characterization of the interaction surface between  CD80 and an analysis of the ligand binding pocket for CD80. Despite repeated 198 attempts we were unable to obtain crystals of full-length Nef that diffracted better 199 than 4 Å, wherein the structure could be fully resolved. In the absence of a high-200 resolution X-ray structure of the full-length Nef, we created a computational model of 201 Nef using a multi-template modeling approach. The major structural information was 202 acquired from the NMR structure PDB ID: 2NEF, as this structure has information for 203 the highly flexible loop region of the core domain (55-66) which contains important 204 interacting residues 30 , as well as from the crystal structure PDB ID: 3RBB which 205 contains structural information of C-terminal folded core (residues 79-206). After 206 modeling, the lowest energy state structure was obtained by energy minimization via 207 SYBYL (Version 7.1) (Tripos Associates Inc.) and validated using PROCHECK. 208 PROCHECK results for the model shows more than 95% of the residues are in 209 allowed regions (79.9% in the strictly allowed region and 17.2% in partially allowed 210 region of the Ramachandran plot) which is better than the template structure (62.3% 211 in the strictly allowed region and 34.2% in partially allowed region of the 212

Ramachandran plot). 213
The structure of full-length Nef can be divided into two parts: a flexible and 214 structurally diverse N-terminal region of about 70 residues followed by a well-215 conserved and folded core domain of about 120 amino acids. The core domain is the 216 only part of the Nef protein which has a stable tertiary structure. It forms an α-β 217 domain in which a central anti-parallel β-sheet of four strands (β1-β4) is flanked by 218 two long anti-parallel α helices (α4 and α5) and two short α helices (α1 and α5). 219 An independent verification of some aspects of the model was obtained from the 221 predicted Small Angle X-ray Scattering (SAXS) envelope of soluble Nef protein. 222 SAXS-patterns of full-length Nef were obtained at three different concentrations 1, 3 223 and 5 mg/ml (Sup. Fig. S7a). The Guinier plots at low angles appeared linear and 224 confirmed good data quality with no indication of protein aggregation (inset Sup. 225   Fig. S7b). The averaged solution shape calculated using the 1 233 mg/ml scattering data clearly indicated that Nef is monomeric in solution. This 234 solution model also revealed a two-domain architecture, a large domain that is well 235 overlaid with the available 3D structure of folded C-terminal core (PDB ID: 3RBB) 236 (Sup. Fig. S7c) 31,32 . The small domain corresponds to the N-terminal region 237 (residues 1-78) that contain a long flexible loop (residues 24-68). While structural 238 details of N-terminal region are available from NMR studies of a peptide regions from 239 residues 2-26 33 and 2-57 30 , information about the relative orientation with C-terminal 240 core domain is missing. The SAXS data envelope along with the computational 241 model (Fig. 4b) provides structural information about spatial arrangement of the C-242 terminal folded core and the flexible N-terminal region of full-length Nef in solution, 243 consistent with the computational predictions. 244

Identification of crucial residues involved in Nef-CD80 interaction surface 246
We next utilized the computational prediction of full-length Nef structure and 247 molecular docking studies with the cytoplasmic tail of CD80, to identify key residues 248 at the interaction surface. The putative binding sites of cytoplasmic CD80 to full 249 length Nef were mapped onto the template model utilizing SiteMap program for 250 binding site prediction. In characterizing binding sites, SiteMap provided quantitative 251 and graphical information in terms of site score and druggability score with properties 252 such as hydrogen bond donor, acceptor, hydrophobic and hydrophilic regions in the 253 predicted site. Docking studies of cytoplasmic tail region of CD80 with full length Nef 254 revealed that CD80 may interact with the interface between the flexible N-terminal 255 and C-terminal core domain. Potential binding sites were predicted with a good site 256 and druggability score (>0.5; Sup. Table 3). 257 The sites of interacting regions of Nef with other cellular proteins have been 258 previously characterized. A polyproline motif (68-78aa) present on the core domain 259 of Nef binds to the SH3 domain of Src kinases with high (nM to μM) affinity 34 . Other 260 than the polyproline motif within the core domain, a number of residues on the core 261 domain are involved in multiple interactions, such as FPD126-128 with human 262 thioesterase and W61 and L115 with CD4 19 . An acidic cluster (EEEE65) close to the 263 core domain is required for interaction with PACS1 and controls MHC-I down-264 regulation 35,36 . The unstructured regions of Nef also provide an extensive accessible 265 surface that could be used to connect to other molecules. Since there is no prior 266 information about the binding pattern of Nef with CD80 it was necessary to score 267 each pose based on energy calculations. From the top ranked docked poses, the 268 best complex with the lowest energy (-246.30 kcal/mol) was chosen as the model 269 complex for Nef and CD80 interaction. In this predicted pose, CD80 cytoplasmic region (indicated in cyan) interacts with the Site-1 and 2 residues in the core domain 271 of Nef (indicated in blue) (Fig. 4c). Based on this pose, the interacting residues 272 were mapped (left inset Fig. 4c). The side chains of site-1 residues W61, E68, K99 273 and R111 are in favourable position to interact with the N-terminus of the CD80 274 cytoplasmic tail. It should be noted that the residues K99 and R111 potentially make 275 polar contact with CD80 backbone carbonyl oxygen of F4 and side chain hydroxyl 276 group of Y2, respectively. In addition, the C-terminus of the CD80 cytoplasmic tail 277 potentially interact with E160 and D180 residues. The side of E160 potentially makes salt 278 bridge interaction with R16 and R23 of CD80 (right inset Fig. 4c,). Based on our in-279 silico predictions we chose four residues that include, 3 from Site-1 (W61, K99 and 280 R111) and one from Site-2 (E160) for further analysis. 281 282

Functional validation of predicted residues mediating Nef-CD80 interaction 283
Single site mutant of Nef such as Nef W61A , Nef K99A , Nef R111A and Nef E160A were 284 designed and purified (Sup. Fig. 8a & b). These mutants were tested for their affinity 285 for CD80 peptide in the ELISA assay (Fig. 5a). Two mutants Nef K99A and Nef R111A 286 showed a loss of affinity to the CD80 peptide, whereas, the Nef W61A and Nef E160A 287 exhibited an affinity comparable to full length Nef WT . These mutants were also 288 assessed for their ability to affect CD80 surface levels after delivery into APCs. 289 Indeed, the two mutants Nef K99A and Nef R111A did not show any reduction in CD80 290 levels, while Nef W61A showed slightly lesser reduction in CD80 receptors and Nef E160A 291 behaved similar to Nef WT (Fig. 5b). Consistent with a key role for the K99 and R111 in modulated CD80 significantly, it did not result in a loss of T-cell activation. Since T-296 cell activation assay requires a minimum of 4-5 h we reasoned that the Nef W61A 297 protein delivered into the APCs may be less stable than the other isoforms for the 5 298 h required for this assay. Indeed, western blot analysis of the protein at 2 h versus 5 299 h in cell lysates shows that the level of Nef W61A protein was drastically decreased 300 after 5 h, while the levels of the other Nef protein variants remained substantial (Sup 301 Fig. 8c). The other Nef E160A mutant exhibited similar reduction of IL-2 release as 302 Nef WT consistent with its ability to bind CD80 peptides as well as down modulate 303 CD80 at the APC surface ( Fig. 5b-d). 304 The two mutants Nef W61A and Nef R111A also did not down regulate MHC-I receptors 305 as much as Nef WT ; whereas Nef K99A and Nef E160A exhibited a similar reduction in 306 MHC-I levels as Nef WT , as assessed by surface MHC-I antibody staining (Fig. 5d). 307 W61 residue (W57 in Subtype B) has been previously reported to be important in CD4 308 down regulation and R111 residue (R106 in subtype B) is located in the oligomerization 309 domain of Nef. Many residues of Nef have been identified that promote interaction 310 with MHC-I, including W57 and R106 in subtype-B, NL4-3 strain, but their mutation did 311 not hinder MHC-I down regulation 37 . However, W61 and R111 residues in subtype C, 312 appear potentially important for Nef interaction with MHC-I, indicating subtle 313 differences in the modulation of host proteins by different Nef variants. Nevertheless, 314 these functional studies provide strong support to the predicted binding mode of 315 CD80 peptide with Nef via residues K99 and R111 in Site 1 (Fig. 4c). Mutation in these 316 residues leads to loss of binding capacity, resulting in the inability to down-modulate 317 CD80 thereby restoring T-cell activation function of the transduced APCs.
Considering the predictive potential of the computational model of Nef-CD80 321 interaction surface, the most potent binding inhibitor molecule, AP5 was docked with 322 Nef protein to capture its binding pattern and important residues involved in 323 interaction. AP5 fits nicely into the hydrophobic cavity formed by the residues W61, 324 V71, L115 and W118, which are part of the N-terminal loop region and the α4 helix of 325 core domain ( Fig. 6a and inset). The aromatic residues W61 and W118 mainly have π 326 -π stacking interaction with aromatic ring B in AP5. In addition to hydrophobic 327 interactions, the side chain of S50 and backbone of E68 form hydrogen bond 328 interaction with amine group of the ligand. The side chain of K99 from α3 helix of core 329 domain, interacts with the CF3 (triflouro methyl) group. Moreover, AP5 binding site 330 overlaps with Site-1 of Nef-CD80 binding pocket, and the docking results showed 331 that the Nef-CD80 and Nef-AP5 binding sites are overlapping with two important 332 common residues such as W61 and K99. 333 Consistent with these predictions, at 10 µM AP5 was neither able to inhibit the 334 interaction between CD80 peptide and Nef W61A mutant nor further reduce the residual 335 interaction of Nef K99A and Nef R111A mutants in vitro (Fig. 6b). However, AP5 was able 336 to displace both Nef WT and Nef E160A from CD80 peptides adsorbed on the ELISA 337 plate. Thus, W61 is an important residue for AP5 binding to Nef. Furthermore, in 338 agreement with the predictions, AP5 treatment did not result in any change in 339 surface levels of CD80 in APCs transduced with Nef W61A , Nef K99A and Nef R111A 340 mutant proteins (Fig. 6c). The levels of IL-2 release in all three mutants Nef W61A , 341 Nef K99A and Nef R111A also remained unchanged, with the Nef W61A mutant mimicking 342 the inhibition observed with wild type Nef. AP5 restored IL-2 release in Nef E160A 343 treated cells, comparable to Nef WT (Fig. 6d), consistent with the inability of Nef E160A 344 to affect neither CD80 nor AP5 binding, thereby serving as a negative control. These results predict and functionally validate the residues in the Nef protein that are 346 important for AP5 binding. 347 348 Hit refinement of AP series 349 We chose AP5 as our starting point, given its nanomolar and micromolar potencies 350 in biochemical and cell-based assays respectively as well as specificity to CD80. An 351 initial hit refinement of AP5 was conducted to understand the preliminary structure-352 activity relationships (SAR). New analogs were synthesized in two series by 353 modifying the rings B and C with un/substituted aryl (heteroaryl) moieties, and 354 without -CF3 group at the 6 th position (Fig. 7a). In series-1, four analogs AP(S1-S4) 355 were prepared (Sup. In series-2, six analogs AP(S5-S10) were prepared by modifying both 4 th and 5 th 358 positions (Sup. Fig. S2; Scheme-4). Apart from these, another analog AP-S11 359 (Sup. Fig. S2; Scheme-2) was also synthesized where -CF3 was maintained at the 360 6 th position and substituted aryl rings at 4 th and 5 th positions. 361 All the synthesized molecules were evaluated for their effect on Nef-CD80/CD86 362 inhibition by ELISA. The analogs AP(S1-S10) without -CF3 group at 6 th position of 363 ring A didn't show any activity (Fig. 7b), however, the analog AP-S11 with CF3 at 6 th 364 position of ring A showed the activity. These results disclosed that presence of CF3 365 group at 6 th position of ring A important for the activity (Fig. 7c). Moreover, docking 366 studies revealed that CF3-group showed hydrophobic interactions with non-polar 367 residues such as W61, L91, I114 and L115 while in vitro and in vivo experiments PA4 belonged to three diverse scaffolds. While PA4 and BC5 are able to inhibit 386 both Nef mediated CD80 as well as MHC1 down-regulation, possibly indicating 387 different interaction points in Nef or counteracting Nef at more than one protein-388 protein interface. PA4 and BC5 molecules are leads to explore for molecular 389 interaction promiscuity and investigating some of the multiple interactions of Nef. 390 Since AP5 selectively inhibits Nef-mediated CD80 down-regulation, we chose to 391 pursue its detailed characterization in this study, 392

393
To gain an insight into the interactions of Nef with CD80, computational approaches 394 followed by experimental validation were used to identify possible binding sites on Nef for both CD80 peptide and the small molecule inhibitor AP5. Due to the lack of 396 crystals with suitable diffraction properties to provide high-resolution structures, and 397 limited experimental information on full length Nef structure, possibly due to its 398 inherent flexibility we adopted a computational strategy. Full length Nef was modeled 399 by a multi-template computational approach, and their spatial conformation was 400 validated using the constraints obtained from SAXS experiments. The interaction site 401 for CD80 was obtained by docking of CD80 cytoplasmic tail with full length Nef 402 model, and a number of possible binding sites for its already known protein-protein 403 interaction sites were identified. These predictions provide key insights that could be 404 correlated with the experimental results, identifying key residues that are involved in 405 Nef-CD80 interaction. A caveat to be noted is that there is a limitation in finding the 406 best biologically relevant orientation of CD80 since the docking was restricted to only 407 the cytoplasmic tail peptide of CD80 which does not impose spatial conformation of 408 the full length CD80 embedded in the membrane. The free cytoplasmic CD80 region 409 fits in the energetically favorable orientation, given its steric constraints, providing 410 verifiable insights from mutation studies. In parallel the AP5 docking results confirms 411 that W61 and K99 residues of Nef contributes to the interaction interface with this lead 412 molecule. Based on the Nef-CD80 and Nef-AP5 docking results, hotspot residues 413 such as W61, K99 and R111 were mutated. This revealed that the K99 and R111 414 residues are crucial for CD80 binding and additionally, W61 plays an important role in 415

421
Our results also indicate that Nef interaction with the co-stimulatory receptors 422 CD80/86 cytoplasmic tails are distinct from reported Nef -MHC-I interactions. Nef 423 interacts with MHC-I cytoplasmic via E62-65 and P78 residues 41  Our study aims to develop small molecule inhibitors that disrupt the interaction 461 interface between HIV-1 viral protein Nef and host CD80 / CD86 in Antigen 462 Presenting Cells (APC). The disruption of this interaction will makes infected APCs 463 more visible to the immune system, increasing cytotoxic lymphocyte activity on these 464 HIV-infected cells, potentially leading to viral clearance from macrophage reservoirs. 465 Here we identify and structurally characterize small molecule inhibitors that indeed 466 disrupt the protein-protein interaction interface of Nef-CD80 and restore the T-cell percentage Inhibition of compounds; of Z-factor>0.5 analysis was used to qualify the NPI normal distribution curve for CD86 with Cutoff percentage for CD86 NPI>20% 522 was considered as hits (g) Scheme shows the hits belonging to 9 scaffolds that 523 were identified in the primary screen (h) Dose response curve of a hit compound 524 from "AP" scaffold; x axis= log concentration of compounds; y axis= Normalized 525 percentage Inhibition, (inset) Structure of AP5 compound and its molecular 526 properties 527    shows the important residues for the interaction between AP5 and Nef. The non-587 polar residues such as W61, L91, I109 and L115 contribute to hydrophobic interactions 588 with CF3. AP5 ligand docking studies shows that the binding interactions occurs 589 between the α4 and α5 helices along with few residues such as W61, E65 and R111 590 which are crucial for AP5-Nef interaction (b) Graph shows colorimetric signal of 591 immobilized CD80 cytosolic peptide upon binding to Nef WT or Nef mutants in the 592 presence /absence of 10 M AP5 as measured by ELISA at OD450 nm. (g) Graph 593 shows surface levels of CD80 receptors in RAJI cell line after the delivery of Nef WT or 594 levels with AP5 addition. (c) Graph shows cytokine (IL-2) release in supernatants 597 after the co-culture T-cell activation assay. RAJI cells were pre-treated with 10 M 598 AP5 for 1 h and then the cells were delivered with Nef mutants or wild type Nef 599 protein for 2 h before co-culture with Jurkat T-cells for 3 h. The IL-2 levels remain 600 unchanged with and without addition of AP5 compound in all three mutants Nef W61A , 601 Nef K99A and Nef R111 . Reduction in IL-2 seen with mutant Nef E160A comparable to 602 Nef WT . 603 The identified hits such as BC2, 6 and 8, and PA4 and PA9 were not synthesized 700 and were procured in somewhat large quantities from the original commercial 701 vendors because of the unavailability of the starting materials. These hits were 702 characterized by using NMR and Mass spectroscopy and then were taken up for 703 validation study. 704 Compound stock and storage-All compounds were dissolved 100% DMSO to 705 make a 10mM stock. Multiple aliquots were prepared from mother stock to avoid 706 multiple freeze-thaw cycles and were stored at -80°C 707

Expression and Purification of recombinant Nef 708
HIV-1 Nef gene sequence was cloned into pET28b expression vector with antibiotic 709 resistance to chloramphenicol (25µg/ml) and kanamycin (50µg/ml) and recombinant 710 for 15minand pellet stored at -80ºC. The protein was purified in 20mM Tris HCl, 713 150mM, NaCl, 3 mM DTT, 5% glycerol, 0.2% Tween-20 in a Ni-NTA column and gel 714 filtration chromatography on Sepharose-75pg (GE) in Akta FPLC purifier. Nef protein 715 with single mutants W61A, K99A, R111 and E160A were designed by site-directed 716 mutagenesis. The buffer conditions were same as full length WT-Nef. and Z score for active compounds, and Normalized percent inhibition to determine 757 the compound activity. 758

Robustness of screen 759
The plate controls showing a Z factor >0.5 was considered to qualify for further 760 analysis. The activity of compounds was determined with the Normalized Percent 761 Inhibition (NPI) and Z scores were used to assess the efficacy of the compounds in The compounds identified as 'hits' in primary screen were defined as those 768 displaying more than 30% and 20% inhibition for CD80 and CD86 respectively and 769 two of three repeats of a particular compound should have Z score between -2 and - The formula used to calculate the percentage toxicity: (Average (X) -Average (high control) / Average (low control) -Average (high control) )*100 784 Where, Average(X)= average OD of individual test; Average (high control) = average OD 785 of cells alone control; Average (low control) = average OD of cells with 1% triton X 786

Measurement of surface levels of CD80/CD86 receptor by flow cytometry 787
Nef protein was delivered into RAJI B cells using Chariot TM protein delivery reagent 788 according to manufacturer's protocol (Active Motif). In brief, 1.  The potential actives from the plates were selected based on the following metrics. 799 a) Z score: 800 Where Ix is the measurement of each triplicate, D: mean of population and D: 801 Standard deviation of population (excluding the positive and negative control). Z 802 score was used to select hits reversal of CD80/86 downregulation by Nef. 803 Compound repeats were qualified by Z score selection, with a Z score between 1.5 804 to -1.2. Each graph was calculated with student t-test using Graphpad prism 7.0 (*p ≤0.05; **p ≤ 0.01; ***p ≤ 0.001) was used to determine the significant difference 806 between the means of control group and treatment groups. 807

Calculation of surface CD80 and CD86 down regulation-was calculated by 808
normalizing raw fluorescent measurements relative to controls (Normalized Inhibition 809 of Down-regulation) 810

NID= Ix/I 811
Where Ix, is the raw measurement each triplicate and I is the mean of the 812 measurements on the positive control.

Modelling of full-length Nef 864
To build a full-length Nef protein, multi-template modelling approach was performed, 865 where more than one experimentally determined structure was utilized for building 866 the model. For the N-terminal part, NMR structure of Nef anchor domain(1QA5) and 867 for core domain, the NMR structure of HIV-1 Nef (2NEF: A) and X-ray structure HIV-868 1 NEF protein, in complex with engineered HCK SH3 domain (3RBB: A) are used as 869 templates. Among the three templates, major structural information is acquired from 870 2NEF structure which covers maximum region of core domain. The tool 871 The modeled full-length Nef protein is utilized for examining the potential-ligand 877 association site using the SiteMap c tool in Schrödinger software. SiteMap identifies 878 potential peptide/ligand binding sites considering van der Waals forces and hydrogen 879 donor ⁄ acceptor characteristics. SiteScore is the most important property generated GNOM, which provides the distance distribution function P(r ) of the maximum particle 928 dimension, Dmax as well as the radius of gyration, Rg, qualitative particle motion was 929 inferred by plotting the scattering patterns in the normalized Kratky plot 930 ((qRg)2(I(q)/I(0)) vs qRg). Ab initio low-resolution models of the proteins were built by 931 the program DAMMIF 48 considering low angle data (q<2nm -1 ). Ten independent ab 932 initio reconstructions were performed for each protein and then averaged using 933 DAMAVER 49 . Superimposition between ab initio reconstruction and atomic model was 934 performed using the software SUPCOMB 50 . The averaged ab initio model (surface representation) overlay with the crystal 978 structure of folded C-terminal core (cartoon representation). 979   x axis= peptide concentration (nM) and y axis= percentage normalized fluorescence (ΔFnorm); Plots represent the mean ± SD (error bars) from three independent experiments.(c) Similarly, CD86 peptide was titrated from 10 µM with a 2-fold serial dilution upto 16 points against a fixed Nef concentration (35 nM). A sigmoidal curve with kD=112 nM, showing higher affinity as compared to CD80. (d) Graph shows ELISA assay where change in OD is observed when the immobilized CD80 and CD86 cytosolic peptides binds to Nef. The OD measurement was done at 450nm. CD74, a negative peptide control shows minimal OD value (e) Graph shows Normalized Percentage Index (NPI) on a normal distribution curve for statistical significance of active compounds across qualified plates showing CD80 actives. X axis= the number of compounds screened in ELISA assay; y axis= normalized percentage Inhibition of compounds; of Z-factor>0.5 analysis was used to qualify the plates. Compounds with NPI>30% for CD80 was considered as hits (f) Similarly, NPI normal distribution curve for CD86 with Cutoff percentage for CD86 NPI>20% was considered as hits (g) Scheme shows the hits belonging to 9 scaffolds that were identified in the primary screen (h) Dose response curve of a hit compound from "AP" scaffold; x axis= log concentration of compounds; y axis= Normalized percentage Inhibition, (inset) Structure of AP5 compound and its molecular properties    cytosolic peptide upon binding to Nef WT or Nef mutants as measured by ELISA at OD450nm. Two mutants Nef K99A and Nef R111A showed reduced affinity to CD80 peptide (b) Graph shows FACS data of surface levels of CD80 receptors in RAJI cell line after delivery of Nef WT or Nef mutant protein delivery . No significant down regulation seen with mutants Nef K99A and Nef R111A (c) Graph shows the levels cytokine (IL-2) released in supernatants of cells in the co-culture functional T-cell activation assay after delivery of the Nef mutants as compared to the wild type Nef protein. (d) Graph shows FACS data of MHC-1 levels after delivery with Nef WT and mutants. Nef WT or mutants were delivered into RAJI cells using Chariot TM delivery reagent. MHC-I was detected by flow cytometry and shown as Ix/I plots. . The inset shows the important residues for the interaction between AP5 and Nef. The non-polar residues such as W 61, L 91, I 109 and L 115 contribute to hydrophobic interactions with CF 3 . AP5 ligand docking studies shows that the binding interactions occurs between the α4 and α5 helices along with few residues such as W 61 , E 65 and R 111 which are crucial for AP5-Nef interaction (b) Graph shows colorimetric signal of immobilized CD80 cytosolic peptide upon binding to Nef WT or Nef mutants in the presence /absence of 10 µM AP5 as measured by ELISA at OD450 nm. (c) Graph shows surface levels of CD80 receptors in RAJI cell line after the delivery of Nef WT or Nef mutant protein delivery as measured by FACS in the presence /absence of 10 µM AP5. Nef W61A , Nef K99A and Nef R111 did not show any further change in CD80 levels with AP5 addition. (d) Graph shows cytokine (IL-2) release in supernatants after the co-culture T-cell activation assay. RAJI cells were pre-treated with 10 µM AP5 for 1 h and then the cells were delivered with Nef mutants or wild type Nef protein for 2 h before co-culture with Jurkat T-cells for 3 h. The IL-2 levels remain unchanged with and without addition of AP5 compound in all three mutants Nef W61A , Nef K99A and Nef R111 . Reduction in IL-2 seen with mutant Nef E160A comparable to Nef WT .