A fluorescence polarization assay for high-throughput screening of inhibitors against HIV-1 Nef-mediated CD4 downregulation

Therapeutic inhibition of the viral protein Nef is an intriguing direction of antiretroviral drug discovery—it may revitalize immune mechanisms to target, and potentially clear, HIV-1–infected cells. Of the many cellular functions of Nef, the most conserved is the downregulation of surface CD4, which takes place through Nef hijacking the clathrin adaptor protein complex 2 (AP2)-dependent endocytosis. Our recent crystal structure has unraveled the molecular details of the CD4–Nef–AP2 interaction. Guided by the new structural knowledge, we have developed a fluorescence polarization–based assay for inhibitor screening against Nef’s activity on CD4. In our assay, AP2 is included along with Nef to facilitate the proper formation of the CD4-binding pocket and a fluorescently labeled CD4 cytoplasmic tail binds competently to the Nef–AP2 complex generating the desired polarization signal. The optimized assay has a good signal-to-noise ratio, excellent tolerance of dimethylsulfoxide and detergent, and the ability to detect competitive binding at the targeted Nef pocket, making it suitable for high-throughput screening.

Therapeutic inhibition of the viral protein Nef is an intriguing direction of antiretroviral drug discovery-it may revitalize immune mechanisms to target, and potentially clear, HIV-1-infected cells.Of the many cellular functions of Nef, the most conserved is the downregulation of surface CD4, which takes place through Nef hijacking the clathrin adaptor protein complex 2 (AP2)-dependent endocytosis.Our recent crystal structure has unraveled the molecular details of the CD4-Nef-AP2 interaction.Guided by the new structural knowledge, we have developed a fluorescence polarization-based assay for inhibitor screening against Nef's activity on CD4.In our assay, AP2 is included along with Nef to facilitate the proper formation of the CD4-binding pocket and a fluorescently labeled CD4 cytoplasmic tail binds competently to the Nef-AP2 complex generating the desired polarization signal.The optimized assay has a good signal-to-noise ratio, excellent tolerance of dimethylsulfoxide and detergent, and the ability to detect competitive binding at the targeted Nef pocket, making it suitable for high-throughput screening.
AIDS, caused by HIV infections, remains one of the deadliest diseases affecting human lives.As estimated by the World Health Organization, about 38.4 million people were living with HIV infections at the end of 2021, and a total of 650,000 people died from AIDS or other HIV-related causes in the year of 2021.Antiretroviral therapy (ART) has dramatically reduced the death rate, and thus transformed the treatment, of HIV infections.Current antiretrovirals used in ART target either viral enzymes (protease, integrase, or reverse transcriptase) or viral entry.Although they can effectively block viral transformation, they are not curative; latent viral reservoirs persist in the infected individuals, and, if ART is stopped, viral rebound occurs within weeks (1).Due to such limitations, treatment of HIV-1 infections needs to be continuous and lifelong.However, such prolonged treatment often results in drug resistance and/or severe side effects such as metabolic disorders and cardiovascular complications (2).Novel antiretrovirals that can better treat, or even eliminate, HIV infections are therefore highly desired.
An attractive direction of developing novel antiretrovirals is the therapeutic inhibition of the viral accessory protein Nef.Through downregulating cellular immune factors, Nef enables the infected cell to evade host immunity and thus plays an important role in viral pathogenesis.Defects in nef genes are associated with disease nonprogression-patients infected by such HIV-1 strains do not develop AIDS for decades in the absence of ART (3)(4)(5)(6).The promise of Nef inhibition is that it may revitalize immune surveillance mechanisms leading to killing of infected cells by the empowered immune cells.This may be particularly relevant in the proposed shock and kill strategy (7-9): following latency reversal, clearance of replication-competent cells by host immunity may benefit significantly from Nef inhibition.
Among the many cellular functions of Nef, CD4 downregulation is the most conserved (10).Although CD4 is the primary receptor for HIV infections, its cellular presence postentry negatively impacts viral replication and transmission.Association between CD4 and the viral Env protein would interfere with the processing and anterograde transport of Env as well as impede new virion release at the cell surface (11)(12)(13)(14).Furthermore, surface CD4 would cause superinfection, leading to premature cell death and thus limiting viral replication (15,16).Finally, binding between CD4 and Env at the cell surface would expose antigenic Env epitopes, sensitizing the infected cell to antibody-dependent cellular cytotoxicity (17,18).
Nef downregulates surface CD4 via hijacking the clathrin AP2-dependent endocytosis, which subsequently leads to degradation of CD4 in the lysosomes (19)(20)(21).A recent crystal structure solved by us has revealed the molecular details of the CD4-Nef-AP2 association (22).As shown by the structure, Nef uses a conserved pocket to recruit the CD4 cytoplasmic domain (CD4 CD ), and the overall shape of the CD4-binding pocket is dependent on the Nef-AP2 association (Fig. 1A).Intriguingly, it has been shown, by another structure of ours, that the same Nef pocket is used for binding and downregulating major histocompatibility complex class I (MHC-I), although the binding site there is shaped differently through the association between Nef and the hijacked clathrin AP1 complex (23).These structural findings strongly suggest that this common pocket on Nef should be the focus of inhibitor screening; they also indicate that the specific cellular factor hijacked by Nef should be incorporated in the assay when screening for inhibitors against the corresponding function of Nef.Furthermore, the CD4-binding pocket, which is constructed partially by Nef's N-terminal flexible arm (Fig. 1A), is of intermediate size: the interface between CD4 and Nef covers an area of only 887 Å 2 .This suggests that the pocket should be susceptible to inhibition by small molecules.Inspired and guided by these structure-derived insights, we have designed, developed, and optimized a fluorescence polarization (FP) assay for screening inhibitors against Nef-mediated CD4 downregulation.

Results
Construct design and initial FP assay using TMR-linear-CD4 CD We followed the three-step protocol described by Moerke to develop an in vitro FP assay for high-throughput screening (HTS) of inhibitors against Nef-mediated CD4 downregulation (24).We first designed and commercially synthesized the fluorescent probe: a linear CD4 CD peptide with a fluorescence tag, tetramethylrhodamine (TMR), attached at the N terminus.We then engineered a Nef-AP2 construct competent in binding CD4 CD and suitable for the FP assay.As shown by our recent structure (22), CD4 binding is facilitated by the Nef-AP2 association-the otherwise flexible N-terminal loop of Nef becomes well-ordered upon binding to AP2 and a portion of this loop then engages productively in binding CD4 (Fig. 1A).This suggests to us that the binding (and the FP assay) would be the most competent if Nef and AP2 are stably bound.There are two challenges in maintaining a stable complex of Nef and AP2 in solution.First, the tetrameric AP2 complex tends to stay in the closed conformation, which is incapable of binding Nef.This could be dealt with by using a truncated version of AP2 containing the N-terminal two-thirds (residues 1-398) of the a subunit and the s2 subunit.Such a hemicomplex has its substrate-binding site fully exposed and has been shown to bind Nef in vitro (25).Furthermore, according to our structure, Nef, when bound to this hemicomplex of AP2, should be fully competent in recruiting CD4 CD (22).The second challenge is that binding between the a/s2 hemicomplex and Nef is still of moderate affinity (25).To overcome this challenge and ensure a stable association between Nef and a/s2, we created a fusion protein by fusing Nef to the C terminus of a (1-398) via a 31-aa linker, which should be flexible enough to not interfere with Nef's binding with a/ s2.The a-Nef fusion was coexpressed with s2, and the complex of a-Nef/s2 was conveniently purified to homogeneity in high yield (Fig. 1B).The above constructs-TMR-linear-CD4 CD and the a-Nef/ s2 protein-indeed gave the expected signal in the FP assay (Fig. 1C).Here, we selected 200 nM as the working concentration of TMR-linear-CD4 CD for the FP assay because, at this concentration of the probe, intensity of the emitted total fluorescence is more than 10-fold higher than that of the background (buffer only).
Importantly, we found that a significant amount of noise exists in the observed FP signal.Such noise is best revealed by the usage of a TMR-mutant-CD4 CD , which carries three mutations-I410D, L413D, and L414D-that completely abolish CD4's ability to bind Nef (22).FP assay using this TMRmutant-CD4 CD showed that, at concentrations of a-Nef/s2 higher than 60 mM, FP signal above the baseline was observed despite that TMR-mutant-CD4 CD is incapable of binding (Fig. 1C).We believe that such noise should have come from light scattering by high concentrations of a-Nef/s2 (26).Indeed, the a-Nef/s2 protein, in the absence of the fluorescent probe, produced a positive FP signal in a concentrationdependent manner (data not shown).
While the undesired light scattering demands that we use low concentrations of a-Nef/s2 in the FP assay, the low binding affinity between a-Nef/s2 and TMR-linear-CD4 CD dictates, however, that binding only happens at high concentrations of a-Nef/s2.As a result, the range of workable a-Nef/ s2 concentrations is very narrow, and within it the FP signal window is at best modest (Fig. 1C).

Using TMR-cyclic-CD4 CD improves the competency of the FP assay
We then reasoned that, to improve the FP assay, we need a fluorescent probe that is capable of binding a-Nef/s2 with higher affinity.The structure, once again, helped us design such a probe.As shown by the structure, CD4 CD adopts a near-circular shape when bound in the hydrophobic pocket of Nef (Fig. 1D).We therefore hypothesized that a cyclized CD4 CD peptide should bind competently to a-Nef/s2 and should do so more efficiently than the linear CD4 CD , because binding with the former should be more favorable entropically.We thus designed and commercially synthesized a TMR-cyclic-CD4 CD probe.When tested in the FP assay, the TMR-cyclic-CD4 CD showed binding at lower concentrations of a-Nef/s2, than that of TMR-linear-CD4 CD , indicating higher affinity binding was indeed achieved through this design (Fig. 1C).After subtracting out the noise signal (see details in Experimental section), a K D of 9.8 mM was derived for the binding between the TMR-cyclic-CD4 CD probe and the a-Nef/s2 protein, while the K D for the TMR-linear-CD4 CD probe could not be accurately derived because the binding failed to reach saturation despite of the high concentrations of a-Nef/s2 used (Fig. 1E).Importantly, the usage of TMR-cyclic-CD4 CD significantly increases the signal window of the assay.To ensure that the assay is sufficiently sensitive to competition, the a-Nef/s2 concentration of 25 mM was selected, which gives a signal window of around 50 mP.

Assay optimization through incorporation of detergents
During our experiments, we noticed that CD4 CD is prone to aggregation: adding unlabeled CD4 CD to the TMR-cyclic-CD4 CD probe led to increased FP signals (Fig. 2A), which is most consistent with TMR-cyclic-CD4 CD binding nonspecifically (or aggregating) with unlabeled CD4 CD .This is presumably due to the presence of multiple hydrophobic residues within the short CD4 CD peptide.To mitigate this issue and improve the assay, we tried adding detergent-either Tween 20 or Triton X-100-at different concentrations.As shown in Figure 2A, among all tested, 0.01% (v/v) Triton X-100 worked the best in reducing (although not eliminating) the aggregation-related signal.We therefore included 0.01% Triton X-100 in the buffer for all subsequent experiments; no adverse effect was observed from it.

Assessment of the assay's tolerance toward DMSO
Since compounds in most small molecule libraries are dissolved in DMSO, dimethylsulfoxide (DMSO), we investigated whether, and to what extent, our FP assay is stable in DMSO.As shown in Figure 2B, DMSO concentrations of 3% or lower led to minimal, if at all, decrease of the FP signal.At DMSO concentrations of 6.3%, 12.5%, and 25.0%, the FP signal window (DmP) decreased by 10.1%, 11.2%, and 19.8%, respectively.Overall, our FP assay is very stable at up to 3% of DMSO and is thus suitable for library screening.

The optimized assay is responsive to competitive binding of inhibitors
To assess the FP assay's ability to detect competitive inhibition, we first tested whether unlabeled linear CD4 CD can displace TMR-cyclic-CD4 CD .As shown in Figure 3A, addition of unlabeled CD4 CD led to a dose-dependent decrease of the FP signal consistent with a competitive binding scenario; in contrast, the control using TMR-mutant-CD4 CD as the probe showed stable, baseline polarization signal throughout.These results suggest that the competitor is not damaging the fluorophore in any way and that the observed dose-dependent decrease of the FP signal should be due to competition between the unlabeled linear CD4 CD and the TMR-cyclic-CD4 CD probe.However, it appears that unlabeled CD4 CD , even at the highest concentration tested (3 mM), did not displace the bound TMR-cyclic-CD4 CD completely (Fig. 3A).As mentioned earlier, the short CD4 CD peptide contains several hydrophobic residues, and therefore is prone to aggregation.We suspect that some microscopic aggregates of unlabeled CD4 CD may be present, and as a result, the monodispersed, competitioncompetent CD4 CD may be just a fraction of the total unlabeled CD4 CD .
We then tried to make the competitor CD4 CD peptide more soluble and less aggregation prone.To do so, we attached a solubility-enhancing maltose-binding protein (MBP) to the N terminus of the CD4 CD peptide.When tested in the FP assay, this MBP-CD4 CD protein caused dose-dependent decrease of FP at apparent concentrations lower than that of unlabeled CD4 CD (Fig. 3B); this seems to agree with our hypothesis above that concentrations of competition-competent unlabeled CD4 CD may be lower than its apparent concentrations.Interestingly, competition by MBP-CD4 CD benefitted from incubation: for the first 4 h of incubation but not afterward, the FP signals progressively decreased (Fig. 3B).
As indicated by the TMR-mutant-CD4 CD curve (Fig. 3A, red curve), when the probe is completely dissociated from a-Nef/ s2, the FP signal should drop to around 105 mP.Yet, in the competition with MBP-CD4 CD , the FP signal decreased only to 127 mP (Fig. 3B), which indicates that the displacement of the probe also may not be complete here.In fact, after reaching the lowest FP values at the MBP-CD4 CD concentration of around 30 mM, the curves (especially those of 3-h incubation or longer) started to rise as the MBP-CD4 CD concentration further increased.We believe that this may be due to either nonspecific binding of TMR-cyclic-CD4 CD to MBP-CD4 CD aggregates (similar to what was observed between TMR-cyclic-CD4 CD and unlabeled CD4 CD in Fig. 2A) or light scattering by MBP-CD4 CD , both of which would occur more readily at higher concentrations of MBP-CD4 CD leading to increased FP signals.
To further prove the specificity of this assay and further investigate the nature of the noise, we performed another control experiments using MBP (without CD4 CD fused to it) as a mock competitor.Under optimized conditions with TMR-cyclic-CD4 CD used as the probe, no dose-dependent decrease was observed; instead, at concentrations of MBP higher than 20 mM, the FP signal experienced a gradual, modest increase (top portion of Fig. 3C).This indicates that MBP cannot compete with TMR-cyclic-CD4 CD for binding the a-Nef/s2 protein, which further confirms that the dose-dependent decrease of FP signal observed in Figure 3B should be due to competition between MBP-CD4 CD and the TMR-cyclic-CD4 CD probe.Thus, our optimized FP assay is indeed capable of detecting competitive inhibition at the targeted pocket of Nef.
Interestingly, when MBP was used to "compete" against the TMR-mutant-CD4 CD probe, which is unable to associate with a-Nef/s2, the trend observed was similar to that when MBP was used against the TMR-cyclic-CD4 CD probe (Fig. 3C).Given that TMR-mutant-CD4 CD lacks hydrophobic residues and is thus very soluble, we believe that the background noise revealed in this experiment should not come from aggregation of the fluorescent probes but should be caused by light scattering by MBP, a 42-kDa protein (Fig. 3C).Given that MBP-CD4 CD and MBP are similar in size, it is perhaps reasonable to believe that light scattering by high concentrations of MBP-CD4 CD is the major source of noise observed in Figure 3B.
Notably, data in Figure 3C also confirms that, in the optimized assay, binding between TMR-cyclic-CD4 CD and a-Nef/ s2 reaches equilibrium within 1 h of incubation-the FP signal here reached 155 mP after 1 h of incubation and remains stable afterward for at least another 4 h.It should also be noted that the superb stability of the signal exhibited here is in great contrast with the observed time effect on the FP signal when MBP-CD4 CD was used as a competitor (Fig. 3B).We postulate that the time effect observed during competition by MBP-CD4 CD was likely caused by a low off-rate (k off ) of the TMRcyclic-CD4 CD probe.Overall, the observed dose-dependent competitions indicate that our optimized FP assay is capable of detecting, and is sensitive to, competitive binding against TMR-cyclic-CD4 CD .Furthermore, the lack of complete displacement of TMR- cyclic-CD4 CD by either unlabeled CD4 CD or MBP-CD4 CD indicates that, to evaluate the assay's competency for HTS, neither of these two competitors can serve competently as the positive control.

The FP assay is robust and suitable for HTS
To assess whether our assay is competent for HTS, we calculated the Z 0 factor.Given the incomplete competition by either unlabeled CD4 CD or MBP-CD4 CD (Fig. 3, A and B), we instead decided to use a solution of TMR-mutant-CD4 CD mixed with a-Nef/s2 as the positive control.Such a control resembles the scenario where the fluorescence probe is completely displaced by an inhibitor and thus should help reveal the true signal window of our assay.From 66 positive controls (a-Nef/s2 with TMR-mutant-CD4 CD ) and 66 negative controls (a-Nef/s2 with TMR-cyclic-CD4 CD ), the Z 0 factor was calculated to be 0.56 (Fig. 4A).
Finally, to get a glimpse of whether addition of compounds might or might not lead to significant noise in our assay, we tested the addition of three tool compounds: maltose, chloramphenicol, and MHC-I CD peptide.None of these compounds is expected to inhibit the Nef-CD4 interaction.In particular, while Nef targets the cytoplasmic domain of MHC-I for its downregulation in cells, binding between Nef and MHC-I CD is not expected to occur in vitro in the absence of clathrin AP1 or its m1 subunit (23).As shown in Figure 4B, consistent with their lack of activity, addition of these compounds at 300 mM did not decrease the FP signal in any meaningful way.Importantly, no significant noise was observed in any case.Overall, the measured Z 0 factor and the validation with tool compounds together indicate that our assay is competent and suitable for HTS (24).

Discussion
We have developed a robust FP-based assay that is suitable for HTS of small molecule inhibitors against HIV-1 Nef-Figure 3. Dose-dependent competition of the fluorescence-labeled probe by unlabeled CD4 CD peptide or MBP-CD4 CD .A, unlabeled CD4 CD showed dose-dependent competition against TMR-cyclic-CD4 CD while the control using TMR-mutant-CD4 CD showed baseline signals without dose-dependent decrease.B, MBP-CD4 CD showed dose-dependent competition against TMR-cyclic-CD4 CD .Incubation up to 4 h improved the competition, but longer incubation did not improve it further.Dose effect was observed up to 30 mM of MBP-CD4 CD for all incubation time.For incubation longer than 2 h, when the concentration of MBP-CD4 CD increased past 30 mM, the FP signals increased.C, in contrast, adding MBP did not result in dose-dependent competition with TMR-cyclic-CD4 CD (top portion).Similar trend was observed when TMR-mutant-CD4 CD was used as the probe (bottom portion).In both cases, the FP signal modestly increased at high concentrations of MBP, and the signal was insensitive to incubation.CD4 CD , CD4 cytoplasmic domain; FP, fluorescence polarization; MBP, maltose-binding protein; TMR, tetramethylrhodamine. mediated CD4 downregulation.Our work here benefitted greatly from the high-resolution structure that we solved earlier (22); it not only guided our protein engineering efforts aimed at stabilizing Nef into the desired conformation but also inspired us to use a cyclized version of the CD4 CD peptide in our probe design, which helped us gain the needed signal window.Thus, our work here underscores the key role of highresolution structures in enabling and facilitating drug discovery efforts against challenging targets.
One issue we had in developing this FP assay is the noise signal, which was typically observed when the protein concentration used was high.With a control experiment using MBP as a mock competitor and a soluble, bindingincompetent TMR-mutant-CD4 CD as the probe (Fig. 3C), we showed that the noise observed in our competition should be mainly due to light scattering by the sizable MBP or MBP-CD4 CD .Importantly, since library compounds are small and therefore unlikely to cause significant light scattering, such noise may not be present during HTS.Of course, some small molecule may form aggregates, which may lead to light scattering and thus interfere with HTS; such an issue, if indeed occurs, might be mitigated by the presence of 0.01% Triton X-100 in our assay solution.
Parallel to this work, we have also developed, as reported in the companion publication (27), a similar FP assay for inhibitor screening against HIV-1 Nef-mediated MHC-I downregulation.The two FP assays share similar time windows (4 h) and comparable Z 0 factors.Intriguingly, however, the dose-dependent competition of the other FP assay was not impacted by incubation time, which is in great contrast with the FP assay reported here (Fig. 3B).This difference in the inhibition kinetics is intriguing and should be due to the drastically different ways that protein-protein interaction takes place in each case (22,23).Given the dynamic feature of the inhibition for the FP assay reported here, it would be necessary to include, on every assay plate, positive and negative controls so that the effect of screened compounds can be evaluated safely and reliably.
We plan to use both this FP assay, which targets Nef's activity on CD4, and the other FP assay, which targets Nef's activity on MHC-I (27), in parallel to screen against the same compound libraries, which should help filter out false positives and allow us to identify true Nef inhibitors of potentially unique properties (e.g., specific to one Nef function or dual functional).

Fusion protein design, expression, and purification
The a-Nef/s2 fusion was constructed by fusing HIV-1 Nef (26-206, NL4.3) to the C terminus of a (1-398) subunit of AP2 via a flexible linker of 31 amino acids.Genes encoding the above a-Nef fusion (with a N-terminal 6xHis tag) and the s2 subunit of AP2 were cloned into the two multiple cloning sites, respectively, of the pETDuet expression vector.Escherichia coli cells transformed with the plasmid were grown at 37 C till A 600 reached 0.8.Protein expression was then induced with 0.1 mM IPTG and continued at 16 C overnight.Cells were then lysed using sonication.The protein of interest was purified sequentially through a nickel-nitrilotriacetic acid affinity column, a HiTrap Q anion exchange column, and finally a Superdex 200 size-exclusion column.
For MBP-CD4 CD , the gene encoding 394 to 419 of CD4 was cloned into a pMAT9 vector.The MBP-CD4 CD protein was FP assay for HTS against Nef-mediated CD4 downregulation expressed overnight at 16 C in NiCo21 (DE3) cells in terrific broth after induction with IPTG.Expressed protein was purified sequentially by a MBP affinity column, a HiTrap Q anion exchange column, and finally a Superdex 200 size-exclusion column.
For purifying the MBP protein, MBP-CD4 CD , after being eluted from the MBP affinity column, was subjected to digestion by the Mpro protease, which cleaves at a site between MBP and CD4 CD .The freed MBP was then purified similarly as MBP-CD4 CD .

FP assay
Purified a-Nef/s2 was buffer exchanged into the assay buffer (50 mM Tris, 150 mM NaCl, 0.5 mM DTT, 0.01% Triton X-100, pH 8.0).A stock protein solution of 200 mM a-Nef/s2 was then prepared and was subsequently used to create different dilutions.Assays were carried out in Corning 384-well black microplates (3820).In each well, 200 nM TMR-labeled CD4 CD peptide was mixed with a-Nef/s2 at varied concentrations in a total volume of 15 ml.Incubation was done for 1 or 2 h at room temperature with minimal exposure to light.FP was then measured using the EnVision plate reader (Perki-nElmer) with excitation at 535 nm and emission at 595 nm.Experiments were done in triplicates, and data was plotted using nonlinear regression as a function of protein concentration in a logarithmic scale using OriginLab (www.originlab.com).
For calculating the dissociation constant (K D ), the noise, represented by the FP signal generated with TMR-mutant-CD4 CD being used as the probe, was first subtracted from the measured FP signal with TMR-cyclic-CD4 CD .The adjusted data was then fitted to nonlinear regression for one-site binding using the following equation: Y = B max *X/(K D + X) (B max : maximum specific binding in the same units as Y).

DMSO tolerance
Assay solutions were prepared containing 25 mM a-Nef/s2, 200 nM TMR-cyclic-CD4 CD , and different concentrations of DMSO (0-25%).After incubation at room temperature for 2 h, FP values were measured and recorded.All experiments were done in triplicates.
Competition of TMR-cyclic-CD4 CD by unlabeled CD4 CD peptide or MBP-CD4 CD For competition using unlabeled CD4 CD peptide, a stock solution of 3 mM unlabeled CD4 CD peptide was first prepared.The stock solution was then serial-diluted (2-fold each) ten times.For making the final assay solutions, a-Nef/s2 and TMR-cyclic-CD4 CD were first mixed and incubated at room temperature for 30 min.Then, unlabeled CD4 CD peptide was added (final concentrations: 25 mM a-Nef/s2, 200 nM of TMR-cyclic-CD4, and varied concentrations of unlabeled CD4 CD peptide).The plate was incubated at room temperature for 2 h.FP values were recorded.All experiments were done in triplicates.
For competition using MBP-CD4 CD (or, in the control experiment, using MBP), an initial MBP-CD4 CD (or MBP) stock solution of 120 mM was prepared, and procedure used for measurements was similar as the above.The plate was read hourly for a total of 6 h.DMSO.Samples of positive controls and negative controls (66 wells each) were prepared in a 384-well plate and incubated at room temperature for 2 h.FP values were recorded using the plate reader.The Z 0 factor was calculated using the following equation: where m N and m P are the averages of mP values of negative and positive controls, respectively.SD N and SD P are the standard deviations.

Figure 1 .
Figure 1.Structure-guided design and initial tests of the FP assay.A, the N terminus of Nef (cartoon) becomes ordered upon AP2 association and participates in CD4 binding (PDB ID: 6URI).B, SDS-PAGE of purified a-Nef/s2.C, Nef-bound CD4 CD (406-419) adopts a near-circular shape.D, FP assays assessing the binding between the fluorescent probes and a-Nef/s2.The TMR-mutant-CD4 CD curve represents both the baseline (at lower concentrations of a-Nef/s2) and the interference caused by light scattering of a-Nef/s2 at higher concentrations.FP signals from using TMR-linear-CD4 CD as the probe are above the baseline but modestly so.FP signals from using TMR-cyclic-CD4 CD as the probe are sufficiently higher than the baseline.E, FP signals after subtraction of the background noise.AP2, adaptor protein complex 2; CD4 CD , CD4 cytoplasmic domain; FP, fluorescence polarization; TMR, tetramethylrhodamine.

Figure 2 .
Figure 2. Effects of detergents and DMSO on the FP assay.A, 200 nM TMR-cyclic-CD4 CD was either alone in solution or mixed with 60 mM unlabeled CD4 CD .Buffers contain either no or different concentrations of detergent.FP signals were read after 2-h incubation.While the FP signals of TMR-cyclic-CD4 CD alone (the group of columns on the left) were slightly decreased by the presence of detergents, more substantial reductions in FP signals were observed in samples containing both TMR-cyclic-CD4 CD and 60 mM unlabeled CD4 CD (the group of columns on the right) with the greatest reduction seen in 0.01% Triton X-100.B, the DMSO tolerance of the assay.Left y-axis: FP signal (mP) of the assay normalized by substracting out the background (buffer only) FP signal; the obtained DmP values here should correlate directly with binding.Right y-axis: percentage of the FP signal correlated with binding (100% corresponds to the FP signal of the binding without DMSO added; 0% corresponds to the no-binding background).Analysis using one-way ANOVA indicates that these data are statistically significant (p < 0.0001).CD4 CD , CD4 cytoplasmic domain; DMSO, dimethylsulfoxide; FP, fluorescence polarization; TMR, tetramethylrhodamine.

Figure 4 .
Figure 4. Assessing the assay's suitability for HTS.A, measurement of the Z 0 factor.The mean values of positive and negative controls are indicated by the dashed lines, while solid lines indicate the range where data points were considered (one data point in the positive controls was considered as an outlier).The Z 0 factor was calculated to be 0.56.B, tool compounds did not cause noise in the assay.Chloramphenicol, maltose, and MHC-I CD peptide (319-330 of MHC-I) was each dissolved in dimethylsulfoxide and then added to the assay solution to a final concentration of 300 mM (final dimethylsulfoxide concentration = 3%).MHC-I, major histocompatibility complex class I.