Mechanism of Multivalent Nanoparticle Encounter with HIV-1 for Potency Enhancement of Peptide Triazole Virus Inactivation

Results: Gold nanoparticle conjugates of peptide triazole Env inhibitors demonstrated impressive picomolar potencies. Conclusion: Nanoparticle conjugates enhanced antiviral functions by multivalent attachment to Env spikes.


INTRODUCTION
There is an urgent need for new antiretroviral agents for the prevention and treatment of HIV-1 infection. Current inhibitors used in most combination antiretroviral therapies target the viral enzymes reverse transcriptase, integrase and protease. Although the use of combination antiretroviral therapy has resulted in reducing viremia and HIV-1 related morbidity and mortality (1,2), these drugs target stages of HIV-1 infection after cell entry and have significant challenges such as complex regimens and treatment-resistant variants of HIV-1. Currently, there are only two FDA approved antiretroviral drugs targeting the initial entry of the virus into host cells, namely T20 (enfuvirtide) and maraviroc (3)(4)(5)(6)(7)(8), but these are used mainly in salvage therapy due to such limitations as injection site reaction and cost of production for the former, and incomplete co-receptor breadth for the latter (7,(9)(10)(11). T20, a fusion inhibitor, targets transient conformations of the virus Env protein, gp41, and hence has a relatively short time window to act against the HIV-1 host cell membrane fusion process (12,13). Maraviroc, on the other hand, targets the host cell co-receptor CCR5 and can only block R5-tropic HIV-1 virions from infecting the cell (7,11,14). Therefore, prolonged maraviroc treatment can result in the development of X4tropic HIV-1 (7,11,14).
Host cell infection by HIV-1 is mediated by cell receptor interactions with trimeric envelope glycoprotein (Env) spikes that are exposed on the virus membrane surface (15,16). Env is the only virally encoded protein on the virion surface (16,17). The external gp120 component of Env is essential for cell receptor (CD4) and coreceptor (usually CCR5 or CXCR4) interactions and subsequent virus-cell fusion (17,18). Hence, Env gp120 presents an obvious target to attack the virus directly in order to block the cascade of integrated binding and conformational change steps that lead to host cell infection. Env-specific inhibitors that could inactivate the virus before receptor encounter would hold great promise in preventing AIDS transmission and progression. In spite of the potential of gp120 antagonists for prevention and intervention of infection, progress has been limited for such agents, due to such factors as high rate of mutation of gp120 protein leading to the threat of resistance to developing inhibitors (19)(20)(21), low potency of small molecule inhibitors (22)(23)(24), the high cost and potential toxicity of protein inhibitors (25)(26)(27), and the potential risk of infection enhancement with CD4-mimicking ligands (22,28).
Previously we showed that a family of entry inhibitors, peptide triazoles (29)(30)(31)(32)(33)(34) derived from the peptide 12p1 (35,36) bind specifically and with high affinity to HIV-1 Env gp120, antagonize the interactions of Env binding sites for both host cell receptors CD4 and CCR5/CXCR4 and cause gp120 shedding from the virus surface, leading to HIV-1 inactivation before host cell encounter (37,38). The most potent compositions of peptide triazole inactivators discovered so far are multivalent forms conjugated to monodisperse gold nanoparticles. The conjugates initially derived were found to have low nanomolar inhibitory IC50 values and to act as virucidal agents leading to the irreversible breakdown of the virus before the virus has a chance to engage the host cell (37,38).
The functional properties of virolytic gold nanoparticle-peptide triazole (AuNP-PT) conjugates make it feasible for long term use of this type of agent to inactivate HIV-1 before host cell encounter. In the current work, we sought to investigate the mechanism of cell-free HIV-1 inactivation using the AuNP-PT multivalent nanostructures. We evaluated the impact of [1] altering the properties of the AuNP-PT constructs, by increasing size and PT coverage density of the nanoparticles, and [2] varying the HIV-1 virion functional spike density.
Our investigations demonstrate the importance of multivalent contact between nanoparticle and metastable virus interaction partners. The results define Envtargeting NP designs that take advantage of HIV-1 Env metastability and cause specific and irreversible HIV-1 inactivation.
Overall, the investigation demonstrates that the multivalent physical contact approach can be used as a general strategy for antagonizing HIV-1 transmission and proliferation.
The HOS.T4.R5 cells were grown in DMEM supplemented with 10% FBS, 2.5% HEPES, 1% Penicillin-Streptomycin, 2% L-Glutamine and 1 mg of puromycin. 293T cells were obtained from American Type Culture Collection and grown in the same culture medium as the HOS.T4.R5 cells except without puromycin. The plasmids for HIV-1 BaL-01 gp160 and VSV-G Env DNA were obtained from the NIH AIDS Reagent Program. Replication competent HIV-1 BaL-01 virions were obtained from the University of Pennsylvania Center for AIDS Research (Penn CFAR). Mouse anti-p24 and rabbit anti-p24 antibodies, and p24 protein, were purchased from Abcam. Fully Infectious HIV-1 (BaL) was a gift from Dr. Michele Kutzler and obtained from the Penn CFAR. The gp120 monomer protein was produced using an already-established protocol (33), anti-gp120 D7324 was purchased from Alto chemicals. Gp41 protein, Enfurvirtide (T20) and anti-gp41 antibodies 4E10 and 2F5 were obtained from the NIH AIDS Reagent Program. Enhanced Chemiluminescence western blot detection system was obtained from Amersham. Ophenylenediamine (OPD) was from Sigma Aldrich. All other materials were purchased from Fisher Scientific.

Peptide triazole synthesis:
To conjugate peptide triazole on nanoparticles, the peptide triazole KR13 (RINNIXWSEAMMβAQβACamide, X=ferrocenyltriazolePro) was synthesized to contain the 12-residue N-terminal sequence of the HNG-156 parent peptide (RINNIXWSEAMMamide) (39) with a C-terminal extension containing a free thiol group. KR13 itself contained the essential peptide triazole pharmacophore, Iletriazole-Pro-Trp, and was functionally active (37). The free sulfhydryl group was incorporated to enable direct thiol conjugation to the gold surface of AuNP (38). The peptide KR13 binds to gp120 with a KD of 11.3 nM (37).

Gold nanoparticle synthesis:
The citrate reduction method developed by Frens (40) was modified in order to synthesize size-controlled, stable and monodisperse AuNPs. The citrate acid concentration was varied to obtain AuNP with various sizes ranging from 13 nm -123 nm. The citrate reaction solution, initially at 100˚C, was cooled to room temperature and BSPP (18 mg/ml for 20 nm AuNP) was added to the synthesized particles and stirred overnight at room temperature. The resulting particles were further washed with phosphate buffer at pH 7.2 and concentrated using a Millipore 100,000 kDa filter. The particle size was determined by dynamic light scattering (DLS) in the Zetasizer NS90 (Malvern Instruments), and the particle concentration was calculated using the absorbance at 450 nm and the peak absorbance of the UV spectrum (41). The AuNP particle morphology was characterized using transmission electron microscopy. Samples were prepared by adding a drop of the AuNP solution onto a lacey carbon grid film and then allowing evaporation. TEM images of AuNPs were taken with a JEM 2100 camera operated at 200 kV.

Gold
nanoparticle-peptide triazole conjugate production: Peptide conjugation to the AuNPs was achieved via covalent linkage between the thiol moiety (-SH) on C-terminal cysteine residue and the gold surface of AuNP (38). A flocculation assay (42) was conducted to determine the ratio of AuNP to KR13 to use for complete peptide conjugation without peptide aggregation. The BSPP-stabilized AuNP-KR13 mixtures in phosphate buffer were stirred vigorously for 30 minutes at 25ºC. Excess peptide was removed from the nanoparticle fraction using a 10,000 MW cutoff centrifugation filter. PT content of AuNP-KR13 preparation was determined using amino acid analysis after acid hydrolysis (Keck Institute, Yale University). To prepare AuNP-PT particles with varying density of KR13, we used 20 nm AuNPs. These were mixed with different molar ratios of KR13 and L-cysteine during the step of peptide conjugation to AuNP. Both molecules present a single reactive thiol for covalent attachment to the BSSP-stabilized nanoparticle. However, AuNP-Cys has no virolytic activity (Table 2), and only the KR13 component will bind to HIV Env on the virion surface.
Incorporation of KR13 was measured directly by amino acid analysis. We calculated the average areal peptide density (ρ) assuming spherical NP's, Where πd 2 is surface area of nanoparticle and CKR13/CAuNP is the molar ratio of peptides per nanoparticle. The conjugates were validated for peptide conjugation by testing their change in zeta potential, which was measured by the Zetasizer NS90 (Malvern Instruments) (data not shown). Furthermore the monodispersity was assessed using both DLS ( Figure 1B) and TEM ( Figure 1C).

HIV-1 pseudotyped virion production:
The recombinant virus consisted of the pro-viral envelope plasmid sequence corresponding to the CCR5 targeting HIV-1BaL strain or a VSV (Vesicular Stomatitis Virus) pseudotype, and the backbone sequence corresponded to an envelopedeficient pNL4-3 luc + , envprovirus developed by N. Landau (43,44). Four µg of envelope DNA and eight µg of backbone DNA were co-transfected into 293T cells using FuGene 6 as the transfection reagent (Promega).
Fourteen hours posttransfection, the medium was changed, and the VLP containing supernatants were collected at 48 hours post transfection. Soluble proteins (including Env gp120 and p24) from the viral supernatants were removed via fractionation on 6-20% iodixanol gradient for 2 hours at 210,700 x g (SW41 rotor, Beckman ultracentrifuge). The collected fractions were validated for p24 content using capture ELISA as well as gp120 content using the western blot detection. Virions purified on the 6-20% iodixanol gradient exhibited a characteristic distribution profile of p24 and gp120 content, enabling viral fractions (18)(19).2% iodixanol) and soluble protein fractions (6-8% iodixanol) to be isolated. The gradient-purified virus samples, which exhibited full or greater infectivity (against HOS.T4.R5 cells (38)) compared to the unfractionated control virions, were collected, aliquoted and stored at 80°C until further use.

Env spike presentation on the virus surface:
To make viruses with varying spike density, HEK293T cells were transfected with backbone vector, pNL4-3.Luc R-E-, and a mixture of active Env plasmid (HIV-1 BaL -WT) with an Env plasmid encoding inactive Env gp120 S375W BaL. The S375W mutation has been found previously to be fusion competent (45,46) but it does not bind significantly to KR13 and hence causes resistance to PT (36,47). Of note, varying density does not in itself eliminate the potential for local clustering, and indeed evidence has been obtained showing that HIV-1 Env spikes have the tendency to cluster (48, 49). Control virions included those with all-[BaL or S375W Env] (allactive or resistant for peptide triazole binding respectively). Protease digestion of the spike varying virion was conducted in order to eliminate non-functional envelopes. This digestion of pseudoviruses was carried out by treating the culture supernatants with a protease cocktail of 1 μg of trypsin, chymotrypsin, subtilisin, and/or proteinase K (Sigma) at 37°C as described by Crooks et al. (50). The treated virions were spun on a 6-20% iodixanol gradient as described above. Spike density was quantified using western blot analysis of gp120, viral infection and p24 content (ELISA) as explained above (data not shown). As expected, the S375W mutant was similar to the wild type BaL in infecting the HOS.T4.R5 cells (data not shown).
nanoparticle size on viral inhibition and virolytic activity, we synthesized AuNPs with diameters ranging from 10-200 nm as previously described and functionalized them with the KR13 peptide. We utilized assays for HIV cell infectivity and for virus contents, including p24 and gp120, in order to correlate nanoparticle diameter and surface area of the AuNP-KR13's with antiviral effects. Purified virus was treated with AuNP-KR13 constructions for thirty minutes at 37°C and spun on a 6-20% iodixanol gradient for 2 hours at 210,700 x g (ultracentrifugation as above). The collected virus fraction and the supernatant fraction were tested for infectivity, p24 and gp120 content from the virus post treatment. Virus treated with PBS for 30 minutes at 37°C was used as negative control for background signal, and virus incubated for 5 minutes with 2% Triton-X at 98°C was the positive control for p24 and gp120 content.

Modulating nanoparticle surface density of KR13 on AuNP:
KR13 coverage on 20 nm diameter AuNP was varied by mixing different molar ratios of KR13 and L-cysteine during the step of KR13 conjugation to AuNP. Both molecules present a single reactive thiol for covalent attachment to the BSSP-stabilized nanoparticle. However, only the KR13 was expected to bind to HIV envelope glycoproteins on the surface of the virion. AuNP and KR13 were synthesized and functionalized as described above. Cysteine/KR13 molar ratios were varied from 0:1000 to 1000:0 in increments of 50. We calculated the density of peptide coverage (ρ) as explained above. Purified virus was treated with AuNP-KR13 constructions for thirty minutes at 37°C and spun on a 6-20% iodixanol gradient for 2 hours at 210,700 x g (ultracentrifugation as above) (37). The collected virus fraction and the supernatant fraction were tested for infectivity and both p24 and g120 contents post treatment.

AuNP-KR13 effects on replication competent HIV-1: Replication competent HIV-1
BaL virions were obtained from the Penn CFAR, purified using 6-20% iodixanol gradient and p24 content measured using ELISA. The inhibitor concentration that gives 50% reduction of the viral titer in tissue culture (IC50) was determined from dose response data using nonlinear least-squares regression analysis with Origin V.8.1 (Origin Lab).
The working dilution of virus was pre-treated with a serial dilution of AuNP-KR13 (20 nm and 80 nm), starting from an effective KR13 concentration on the AuNP surface of 50 µM and 10 µM respectively, for 30 min at 37 °C. HOS.T4.R5 cells, seeded at 8,000 cells per well, were incubated for 24 hours followed by addition of pre-incubated inhibitor-virus mixtures. After 48 hours infection, with a growth medium wash step 24 hours post treatment, the p24 content of virions in the supernatant was measured in order to determine the IC50 value of inhibition of virus infection. For cellfree p24 release experiments, the peptide triazole and cell-free replication competent HIV-1 BaL virus were pre-incubated for 30 min at 37°C followed by spinning for 2 hours at 23,000 x g in a table top centrifuge (Eppendorf) at 4°C. Supernatant was separated and p24 content determined using the p24 capture ELISA as described above. The extents of viral inhibition of, and p24 release from, fully infectious HIV-1 BaL were also evaluated for the free control peptide KR13 (data from Bastian et al. 2013) (37).

Exposure of gp41 in viruses treated with AuNP-KR13 and KR13:
The HIV-1 BaL pseudotype was incubated with increasing concentration of either KR13 starting at 50µM or AuNP-KR13 starting at 1 µM (KR13 concentration) for 30 minutes at 37°C. The samples were spun at 23,000 x g (table top centrifuge) for 2 hours at 4°C, and the virion fraction for each sample was collected and fixed using a homemade fixative with 0.1% formaldehyde and 0.1 % gluteraldehyde in PBS. Glycine (0.1 M) was added to stop the crosslinking reaction by the fixative. The virions then were loaded onto an ELISA plate (high protein binding) and incubated overnight on a shaker at 4°C. The plate was blocked with 3% BSA, and ELISA was used to detect gp41 epitopes with human gp41 antibodies 2F5 and 4E10 followed by addition of anti-human IgG HRP secondary antibody.
A corresponding experiment was conducted to test whether T20 inhibited virolysis of pseudotyped HIV-1 BaL induced by AuNP-KR13 (20.2 nm diameter of AuNP). The AuNP-KR13 was kept at an IC80 concentration of 25 nM, and serial dilutions of T20 starting at 1 µM were co-incubated with pseudotyped HIV-1 BaL for 30 min at 37°C. Treated virion samples were fractionated on a 6-20% iodixanol gradient (above). Gradient fractions were quantified for p24 using ELISA, and relative p24 release was quantified and plotted using Origin V.8.1 (Origin Lab).

Time dependence of AuNP-KR13 induced HIV-1 breakdown:
The time dependence of gp120 shedding, p24 release and inhibition of pseudotyped HIV-1 BaL by an IC80 concentration of 20.2 nm AuNP-KR13 (25 nM) was measured. The AuNP-KR13 was incubated with the virus for different times ranging from 1 min to 24 hours at 37°C, the samples were spun on a 6-20% iodixanol gradient and the gradient fractions were tested for gp120 shedding, p24 release and inhibition of viral infection. For gp120 shedding, the fractions were tested using a western blot analysis with primary anti-gp120 antibody, D7324, and secondary antibody, anti-sheep HRP. For p24 release, the ELISA explained above was used. The western blots were analyzed using Image J western blot band quantification. For inhibition of viral infection, the luciferase reporter assay was used as explained above (37).

Anti-viral effects of AuNP-KR13 vs. KR13 on HIV-1 pseudovirions with modified Env spike presentation on the virus surface:
Comparisons were made of the lytic deformation of viruses by KR13 and AuNP-KR13 with viruses containing a varying proportion of KR13-resistant mutant vs. WT Env. In the western blot analysis, the amount of loaded virions was normalized using both p24 content from ELISA and total protein content measured using Bradford assay. The p24 contents of the different virions and their total protein contents were similar, within ± 1.5%. Thus, virus amounts were determined by p24 content (3000 molecules per virion (51,52)). Antiviral assays included inhibition of infection by KR13 compared to AuNP-KR13 on HOS.T4.R5 cells, and cell-free HIV-1 p24 release.

Morphological analysis of HIV-1 after AuNP-KR13 conjugate treatments
Transmission electron microscopy (TEM) was conducted to visualize the morphology of the virions treated with AuNP-KR13 (20.2 nm and 80.2 nm diameter AuNP). Purified pseudotyped HIV-1 BaL virus and AuNP-KR13 were pre-incubated from 5-1440 min at 37°C. Following incubation, samples were fixed with 2% glutaraldehyde for 30 minutes at room temperature, and then embedded in Spurr's low viscosity epoxy medium after acetone washes to dry the virions. Slices (100 nm thick) were prepared using an ultra-microtome (Leica EM UC6), loaded onto a holey carbon TEM 200 mesh grid (Electron Microscopy Science) and imaged using the JEM 2100 camera operated at 120 kV (JEOL, Japan). Sixteen images were taken per sample, and the sizes of observed particles were determined, using Image J software to derive average diameters of the virion particles from TEM images measured from 5 angles. In order to confirm identification of collapsed virions in samples imaged, we used EDAX (Energy Dispersive Spectroscopy) and related the element ratios on untreated virions to those of the collapsed particles.

Monodispersity of gold nanoparticle-peptide triazole conjugates
In this study, we investigated the mechanism underlying the antiviral functions of AuNP-KR13 conjugates against HIV-1, and in particular the ability of the nanoconjugates to cause specific virus lytic inactivation. To achieve this, the dependence of lysis on variations of surface characteristics, of both nanoparticle inhibitor and virion envelope protein, was measured.
The physical characteristics of gold nanoparticlepeptide triazole conjugates were validated by DLS, TEM and UV absorbance. Particle size was obtained by DLS using the Zetasizer NS90 (Malvern Instruments), while the particle concentration was calculated using the UV spectrum as described in the methods section (41). The total number of peptide molecules per AuNP was obtained by amino acid analysis. The DLS data are shown in Figure 1, along with representative TEM images of AuNP-KR13 conjugates of various AuNP sizes. Figure 1A also shows the comparison between the expected sizes of AuNP using varying concentrations of citric acid as well as the experimentally obtained diameters (DLS). The AuNP-KR13 conjugates showed good monodispersity as observed by the TEM images in Figure 1C. Further, the zeta potential values for AuNP-KR13 conjugates compared to the AuNP alone verified the successful peptide conjugation to AuNP (data not shown) (53).

Anti-viral effects of AuNP-KR13 conjugates
Relationship between AuNP-KR13 particle size and antiviral activity: AuNP-KR13 conjugates were tested for both inhibition of HIV-1 BaL pseudotype infection of HOS.T4.R5 cells as well as cell-free p24 release, in both cases after 30 min exposure of viruses to AuNP-KR13 conjugates. Figure 2 shows the dose responses of the effects, and IC50 values are summarized in Table  1. Strikingly, the anti-viral effects were greatly enhanced with increasing size of the nanoparticle. This effect is likely due to decreasing nanoparticle curvature and corresponding increasing contact area between the AuNP-KR13 nanoparticle and the HIV-1 virion. Table 1 also shows the number of peptides per nanoparticle calculated from amino acid analysis and the KR13 areal density on the AuNP surface.
To control for differences in peptide density on separate populations of AuNPs, we prepared AuNPs of 20.2 nm, 42.3 nm and 80.2 nm diameters with nearly equal peptide densities by empirically adjusting the peptide concentration used during conjugation. The antiviral effects observed upon size variation of AuNP at (nearly) fixed KR13 areal density in AuNP-KR13 conjugates are shown in Figure 3, and the IC50 values obtained from the dose response profiles are given in Table 2. Here, a clear correlation was observed between the effects of AuNP size on inhibition of viral infection and p24 release, with increasing size leading to increasing potency. This result suggests that the AuNP-KR13 induced p24 release from HIV-1 is related to the extent of contact of the HIV-1 virion surface with multiple KR13 monomers on the nanoparticle surface. In retrospect, the more limited increase in p24 release potency initially observed with nanoparticle size increase in initial experiments ( Figure 2) likely was due to the fact that the densities of KR13 per unit area on the AuNP surface decreased with increasing size (Table 1), while, in Figure 3B, increasing size of AuNP, while maintaining a similar density of KR13 on the AuNP surface, caused a drastic increase in the potency of p24 release from HIV-1.

Impact of density variation of peptide triazoles conjugated to AuNP on anti-viral functions:
AuNP-KR13 conjugates were prepared with varying KR13 densities on 20 nm nanoparticles and tested for both inhibition of HIV-1 BaL pseudotype cell infection with HOS.T4.R5 cells and cell-free p24 release after 30 min of exposure of virus with the conjugates. The values computed for KR13 coverage per square nanometer of the AuNP surface assumed that AuNP is a perfect sphere and that individual KR13's are randomly distributed on the NP surface. As shown in Figure 4 and Table 3, both viral infection inhibition and p24 release activities were greater with increasing areal density of KR13 on the AuNP surface. The AuNP particles without KR13 had 100% cysteine coverage and did not exhibit nonspecific anti-viral effects (Table 3 and Figure 4).

AuNP-KR13 potencies with replication competent HIV-1:
The anti-viral effects of 20 and 80 nm AuNP-KR13 particles were tested with replication-competent Bal-01 HIV-1, as shown in Figure 5. The results demonstrate that the AuNP-KR13 is effective and has a greater potency than KR13 alone (approximately 300 fold greater for the 20 nm AuNP-KR13 and 1400 fold greater for the 80 nm AuNP-KR13 compared to KR13 (37)).

Comparison of AuNP-KR13 and KR13 actions on pseudoviruses
We previously observed that KR13-induced lysis of HIV-1 embodies virus breakdown characteristics reminiscent of those occurring during virus-cell fusion. These include exposure of the membrane proximal external region (MPER) of gp41 and inhibition by fusion inhibitor T20. In the current work, we sought to evaluate these characteristics for AuNP-KR13 lysis. As shown in Figure 6A, AuNP-KR13 treatment did not induce a concentration-dependent exposure of MPER on the residual virion, while treatment with KR13 did. In addition, as seen in Figure 6B, the fusion inhibitor, T20, did not lead to the inhibition of AuNP-KR13 induced p24 release, in contrast to the inhibition seen with KR13. Figure 6C shows that, like KR13, AuNP-KR13 induced virolysis was accompanied by gp120 shedding. However, for AuNP-KR13, gp120 shedding, p24 release and viral infection inhibition occurred simultaneously, in all cases starting within 10 min of incubation with the virus. By comparison, in KR13 virolysis, p24 release occurred with a time lag compared to the gp120 shedding and virus infection inhibition as shown in Bastian et al. 2013 (37). Hence, from the data shown above, the AuNP-KR13 induced virolysis process appears to be controlled differently than that of KR13. In order to further compare AuNP-KR13 with KR13, we turned our attention to the surface properties of the virus.

Impact of Env trimer composition on the antiviral potencies of peptide-triazole inhibitors
The enhanced antiviral activity of AuNP-KR13 compared to KR13 suggests that multivalent interactions between the virion and nanoparticle are important determinants of AuNP-KR13 potency. Such interactions could involve multivalent attachment to single Env trimers and/or localized trimer cross-linking. To further probe the role of multivalency in inhibition, we compared the antiviral activities of AuNP-KR13 and KR13 against HIV-1 pseudotyped with both susceptible and non-susceptible BaL Env variants. The nonsusceptible Env contained the gp120 S375W mutation that ablates peptide-triazole binding, by steric blockade of the PT Trp indole side chain (36,47), but has negligible impact on glycoprotein expression, processing or fusogenic activity (data not shown). Viruses were produced from cells coexpressing WT (denoted W) and S375W (denoted M) Env at different relative levels. The resulting viral populations shared the same total Env content (W + M) but varied in the average surface density of WT Env. When co-expressed in virus producing cells, WT and S375W Env form four trimeric species (W3, W2M, WM2 and M3), the distribution among which can be approximated through binomial statistics assuming the protomers assort randomly: Here, αi is fraction of Env trimers in a viral population that contain i WT protomers (0 to 3) and f is the abundance of the WT protomer relative to the total protomer level. As the mutation does not alter Env expression level, f was equated with the fraction of WT gp160 DNA (relative to total gp160 DNA) used for transfection into virus producing cells. Values for α0, α1, α2 and α3 for each ratio of WT:S375W DNA tested are given in Figure 7. Figure 8A, 8B, 8C, and 8D, and the IC50 values are given in Table 4. As expected, both AuNP-KR13 and KR13 were most potent against the 100% WT Env HIV-1 population (4:0) and demonstrated no antiviral activity against the 100% S375W Env HIV-1 population (0:4). For inhibition of viral infection, there was a similar loss of both AuNP-KR13 and KR13 potency as the fraction of WT Env decreased. However, even at the lowest WT Env fraction tested (f = 0.125 for the 0.5:3.5 population), infection was almost fully blocked at high inhibitor concentrations.

Dose-response curves of infection inhibition and viral lysis for the different viral populations are shown in
In this viral population, the amount of Env trimers with two or three WT protomers (W3 and W2M) is small (~4%), and viruses lacking such Env trimers (i.e., containing only WM2 and M3 trimers) should form a nontrivial fraction of the population. The fact that all viruses of the 0.5:3.5 WT DNA:Mutant DNA population appear sensitive to AuNP-KR13 and KR13 suggests that these inhibitors can effectively block the function of Env trimers containing only one susceptible gp120.
The logarithm of the IC50 values for infection inhibition varied linearly with the fraction of WT protomer in the viral populations ( Figure 8E). This linearity suggests that inhibitory potency directly depends on the energy of inhibitor binding. Suppose the energy of dissociation for a single KR13 is G, then the energy of n such events is nG. The average value of this occupancy number n can be reasonably assumed to vary linearly with the average number of the available sites on HIV-1, which, in turn, is proportional to f. In this simplistic model, effective inhibitory concentrations can be treated as de facto equilibrium binding constants, leading to the observed proportionality: The linear relationships for AuNP-KR13 and KR13 share the same slope, as expected since single inhibitor-binding energy G is the same for both species. The result suggests that the two inhibitors utilize the same mechanism to block viral infection and that the enhanced inhibitory activity of AuNP-KR13 is a consequence of its ability to concentrate additional KR13 near the viral surface after the first peptide triazole is bound.
While similar dependence on f was observed for viral lysis induced by AuNP-KR13, the lytic activity of KR13 was markedly different ( Figures  8B and 8D). These KR13 titrations were largely independent of the WT protomer fraction until f dropped to 0.25. The KR13 titrations for f = 0.25 (1:3) and f = 0.125 (0.5:3.5) were significantly shallower than those for f > 0.25 and for all of the viral lysis titrations of AuNP-KR13.
The divergence in lytic activity between AuNP-KR13 and KR13 is reflected in the different slopes of the ln (EC50) versus f plots ( Figure 8F). The AuNP-KR13 data had the same slope as the plots for infection inhibition (Figure 8E), suggesting a similar dependence of lytic activity on KR13 binding energy. By contrast, the KR13 data for f > 0.25 had zero slope, implying a fundamental difference in how AuNP-KR13 and KR13 encounter the virus to cause lysis.

Morphological analysis of AuNP-KR13 conjugates effects on HIV-1
We sought to image the virus-nanoparticle interface and consequent virus lysis using transmission electron microscopy. Viruses were treated with the AuNP-KR13 (20 nm and 80 nm) conjugates for 10 minutes, and then fixed and loaded on a TEM grid. Uranyl acetate was used as a contrast agent to stain the lipids in order to visualize the virion. Representative images obtained are shown in Figure 9. In the images for both the 20 nm and 80 nm AuNP-KR13, a limited number of AuNP-KR13 particles (average 1-3) were found associated with individual virions, suggesting that a small number of NP's is sufficient to induce virus lysis. The 80 nm AuNP-KR13 appeared to cause a more drastic disruption of the virion compared to the 20 nm AuNP-KR13. As shown by the representative images in Figure 9, we observed that AuNP-KR13 led to a drastic disruption and collapse of the virus particle, with the residual virion partially engulfing the nanoparticle conjugate. In contrast, KR13 was shown previously (37) to lead to a more subtle virion collapse and at the same time greater virion shrinkage as observed using TEM.

DISCUSSION
The current study has demonstrated the ability to dramatically enhance the potency of anti-HIV-1 function of PT inhibitors by gold nanoparticle display, and at the same time has derived a conceptual model of nanoparticle-virus encounter that can hijack the intrinsic metastability of the virus spike protein complex (15,(54)(55)(56). Previously, we demonstrated that conjugating PT thiols on 20 nm AuNP improved their anti-viral functions (38). Here, we found that altering the AuNP properties of core size and PT density coverage led to dramatic enhancement in potency of both viral infection inhibition and cell-free virolysis (Figures 2, 3 and 4). More importantly, potency enhancement was also observed in replication competent HIV-1 ( Figure 5). The potent antiviral functions of nanoparticle-PT conjugates, in particular irreversible inactivation of viruses by specific interaction with the virus surface protein gp120, makes this a worthy type of inhibition to understand mechanistically.
Considering that both KR13 and AuNP-KR13 exhibit a similar pattern of antiviral activities though with vastly different potencies, we asked if there are fundamental mechanistic differences for these two inhibitor compositions. Observations made in this study ( Figure 6) demonstrated important differences. The MPER of gp41 on the HIV-1 Env is known to be exposed during the HIV-1 / host cell fusion process, but to be unexposed in the ground (unliganded) state of the virus (57). For KR13, we previously observed both concentration dependent and time dependent exposure of the MPER epitope on the HIV-1 virion when viruses were treated with KR13 (37). However, in the case of AuNP-KR13, we did not observe MPER epitope exposure ( Figure 6A). In addition, the data in Figure 6B show that T20, a fusion inhibitor that targets 6-helix bundle formation of gp41 during HIV-1 fusion with host cell, inhibits KR13-induced p24 release (as reported previously (37)) but does not inhibit such release by AuNP-KR13 ( Figure  6B). Further, the data of Figure 6C show that gp120 shedding, infection inhibition and virolysis tracked by p24-release of the HIV-1 BaL pseudotype all had a similar time-dependence for the AuNP-KR13 exposure, in contrast to the unequal kinetics for KR13 (p24 release lagged compared to gp120 shedding and infection inhibition (37)). We note that the difference in kinetics of virus breakdown could have an impact on the extent of T20 sensitivity detected with AuNP-KR13 treatment. However, the lack of exposure of 4E10 and 2F5 epitopes with AuNP-KR13, vs exposure with KR13, argues that kinetics alone cannot explain the differences seen between the modes of action of these two types of virus inactivators. Collectively, the characteristics of MPER exposure, T20 sensitivity and time-dependence of antiviral effects observed for KR13 all argue that KR13-induced lysis bears a relationship to the physiological events of virus-cell fusion. In contrast, the lack of these phenotypes with AuNP-KR13 suggests that the latter responds to different forces. Several other differences between the mechanism of action of KR13 and AuNP-KR13 have emerged during ongoing investigations of KR13-induced lysis and its relationship to HIV-1 biological functions (58). A gp120 antibody, 2G12, that binds to the V3 loop and inhibits KR13 induced virolysis, has been observed (58) not to inhibit AuNP-KR13 induced virolysis. In addition, DTNB (dithionitrobenzene), a thiol-blocking agent that inhibits KR13 induced virolysis, did not affect AuNP-KR13 mediated virolysis (unpublished results). These data clearly reinforce the need to consider mechanisms specific to gold nanoparticles in order to explain the virolytic activity of AuNP-KR13.
The responses of both p24 release and infection inhibition potency enhancements to nanoparticle and virus surface properties argue that multivalent contact between virus and nanoparticle surfaces is a major driving force for the enhanced antiviral potencies of AuNP-KR13. Three lines of evidence obtained in this work support the conclusion of multivalency. [1] Increasing the size of the AuNP particle leads to dramatic potency enhancement ( Figure 2, Table 1). Since size increase causes a decrease of nanoparticle curvature, the area of contact of virus with nanoparticle would be expected to increase, and more points of contact would ensue between the peptide triazole NP surface and multiple spikes on the virus surface. [2] Increased density of KR13 on the AuNP surface strongly enhances anti-viral potency (Figure 3 and Tables 2 and 3). Increased areal peptide density is expected to increase the extent of multipoint attachment between peptide triazoles and Env spikes on the nanoparticle and virus surfaces, respectively. [3] Decreasing the amount of wildtype (fully active) spikes on the HIV-1 virion surface by replacing the gp120 components with the KR13-insensitive mutant gp120, S375W, suppressed KR13-induced p24 release more drastically than AuNP-KR13 activity ( Figure 8C and 8D). The functional data ( Figure 8) can be correlated with the statistically predicted relative abundance of heterogeneous and homogenous spikes expected for different mutant-WT mixtures (Figure 7). We observed that, while KR13 induced p24 release required a statistical average of ≥1 fully active spike (containing 3 active WT gp120 monomers in individual spikes per virion), AuNP-KR13 induced up to 50% p24 release even with an average of close to zero fully active spikes (the case of 0.5-3.5 WT-mutant). The contrasting effects of KR13 vs AuNP-KR13 induced p24 release ( Figure  8F) suggest that AuNP-KR13 can cause lysis by engaging multiple spikes with an average of only 1 active gp120/spike, while KR13 likely requires a greater number of active gp120 monomers per spike trimer. It is worth noting that the ability of both KR13 and AuNP-KR13 to fully inhibit virus at spike populations containing a significant (though not complete) proportion of S375W mutant shows that mixed trimers can still exhibit PT sensitivity to the extent that they contain WT gp120 protomers. The difference in HIV-1 virion encounter by KR13 and AuNP-KR13 is shown schematically in Figure 10. The physical nature of encounter in the model is based on the previously known dimensions of the virus Env (17,55) and assumes that KR13 components on the AuNP are isotropic and equidistant from each other.
One may ask why multivalent contact should cause such a profound effect leading to irreversible inactivation. We speculate that multivalent contact as envisioned in Figure 10 causes an elevated stress on the HIV-1 membrane of the metastable HIV-1 virion, leading to irreversible collapse of the virion particle. As envisioned in the model in Figure 10, KR13 action requires an Env spike with full or close to full native Env protomer content to exert a lytic force on the virus. In contrast, AuNP-KR13 exerts a lytic force by crosslinking Env spikes and hence can carry out this function as long as multiple Env spikes each contain at least one active protomer. That AuNP-KR13 causes significant loss in virus integrity was visualized by TEM imaging (Figure 9). Here, the physical impact on the virion is much greater for AuNP-KR13 as compared to KR13 alone (37). Further investigation is needed to better understand the nature of the virus membrane transformation caused by AuNP-KR13 and how this type of effect compares to membrane transformations triggered by other types of viruslysing agents (59)(60)(61).
The work presented here reinforces the potential usefulness of inhibitor display on nanoparticles to improve potency for disease prevention and intervention. While nanocarriers have been used as vehicles for drug delivery, their use for forming therapeutic agents has been more limited (62)(63)(64)(65). Jiang et al. have shown that gold nanoparticles ranging from 2-100 nm in diameter conjugated to multiple trastuzumab antibodies enable targeting and cross-linking of human epidermal growth factor receptor (HER)-2 in human SK-BR-3 breast cancer cells, by causing reduced (HER)-2 receptor expression and inducing receptor mediated endocytosis (66). Jiang also showed that the anticancer potencies were altered by variations of the diameter of the gold nanoparticles, with the larger particles having more protein on their surface and increased avidity, while the smaller particles exhibited greater receptor mediated endocytosis (66). In the area of HIV-1, nanomaterial design for HIV-1 infection inhibition has been limited (67). Gold nanoparticle display was found to increase the antiviral potency of the HIV-1 inhibitor SDC-1721, a derivative of TAK-779, that targets the cellular coreceptor, CCR5 (68,69). However, such studies have used very small (2 nm) AuNP that could be taken up by cells non-specifically (70). Heterogeneity of the HIV-1 also has limited antibody and drug targeting (71). The current work argues that gold nanoparticle display could help overcome the heterogeneity of Env spikes by enabling multivalent attachment of neighboring partially defective spikes. The current work provides a generalizable approach for nanoparticle drug design not only for HIV-1 but also for other infectious enveloped viruses such as influenza and hepatitis B virus.
Several types of inhibitors that bind to the HIV-1 envelope protein gp120 have now been found to be lytic inactivators of HIV-1. These include DAVEI (dual acting virus entry inhibitor containing cyanovirin-N fused to MPER) (61), the dodecameric CD4 construct (72), KR13 and other PT thiols (37,38,58) and PT thiols on AuNP's. There are differences in the modes of action of these inactivators at the molecular level. DAVEI causes lysis but no observed gp120 shedding, and the requirement for the MPER domain for DAVEI lysis argues for unique interactions of this sequence with gp41 and/or membrane of the virus envelope (61). KR13 and AuNP-KR13 both cause gp120 shedding and membrane lysis but differ mechanistically as enumerated above. Nonetheless, in spite of these differences, the common feature of cell-independent virus lysis hints that there likely is a common mechanistic underpinning for all of these inhibitors. We posit that this lies in the intrinsic metastability that has been argued to be a property of HIV-1 Env (15,(54)(55)(56) and of the HIV-1 virion itself (73,74). Metastability is a feature common to all viruses, since these must protect their genetic contents and at the same time be able to open up to transfer their contents during cell entry to propagate more virus (75)(76)(77). For HIV-1, we propose that virus metastability appears to be a general property that can be hijacked for specific antiviral activities, as observed previously (37,38,61,72). At the same time, the actions of lytic inactivators can inform about fundamental properties of virus metastability, including the membrane transformations that underlie both virolysis by gp120-inhibitor interactions and virus-cell fusion by gp120-cell receptor interactions.   Table 1. Error bars represent the standard deviation of the mean, n = 3.  Table 3. Error bars represent the standard deviation of the mean, n = 3.

Figure 4:
Anti-viral effects of AuNP-KR13 (19.9 nm diameter) with increasing surface density coverage of KR13, with 100% being 69 peptides per AuNP as represented in Table 3. (A) Inhibition of cell infection measured using a single round infection assay as explained in methods. The IC50 values are reported in Table 2.  Table 2. Error bars represent the standard deviation of the mean, n = 3.     Table 4. (E) Linear relationship between the natural log of the IC50s of the virus infection inhibition (Table 4)   After treatment for 10 minutes at 37°C, samples were fixed as explained in the Methods. Collapsed virions of the type shown were validated to be virus-derived, as opposed to impurities, by detecting distinctive element signatures using the EDAX system attached to the TEM instrument. The samples were stained with 1% uranyl acetate, loaded onto a TEM copper grid and imaged using the JEM 2100 operated at 120 kV (JEOL, Japan).      Figure 8 using origin Pro. 8. The relative amounts of DNA of WT vs S375W gp160 used for transfection to produce virions with different ratios of WT and S375W gp120 are listed. N =3.