Growth Inhibition of <i>Staphylococcus aureus</i> by Mixed Monolayer Gold Nanoparticles

Abstract

A library of 120 nanoparticle conjugates is produced by simple one-pot thiol exchange reactions. The antibiotic activity of the conjugates toward Staphylococcus aureus is found to depend upon the combination of thiols assembled on the nanoparticles. Synthetic nanometer-scale systems have the potential to overcome many limitations of conventional small-molecule therapeutic agents.1 For instance, small molecules often have short blood circulation times (half-life, t1/2, of hours), rely on a single high-affinity contact to a disease target, and are typically incapable of disrupting protein–protein interactions that can drive disease pathogenesis. In contrast, nano­scale systems can provide long circulation half-lives (days to weeks), have tunable valency and aqueous solubility, and are adept at preventing protein–protein interactions.2, 3 However, a significant advantage of small-molecule drugs is the ease with which large chemical and structural diversity can be manufactured and screened for biological activity. Synthetic routes to the creation of diversity in nanoparticle composition space would serve to combine many of the advantages of both small-molecule and nanoscale therapeutics. A number of materials have been explored as scaffolds for the design of nanometer-scale therapeutics. Notable recent examples are the pyrimadine-coated gold nanoparticle antibiotics studied by the Zhang laboratory4 and the dextran-coated iron oxide nanoparticles synthesized in Weissleder's laboratory.5, 6 The former experiments suggest that nanoparticles may be able to withstand pathogen evolutionary resistance mechanisms that plague small-molecule drugs, while the latter highlighted the remarkable ability of multivalent binding to strengthen drug-target binding interactions. Our laboratories have focused on gold nanoparticles as a platform for the discovery of novel therapeutics for the treatment of infectious disease. We chose gold for a number of reasons, including the ability to access gold nanoparticles in a range of well-defined sizes from 1 to 10 nm7, 8 and the straightforward gold modification chemistry afforded via formation of gold–thiolate bonds.9-11 In addition, using thiol exchange reactions, combinations of two or more chemically distinct organothiol ligands can be attached to a single particle to create multivalent and multifunctional systems.12 The ability to assemble mixed thiol monolayers on a nanoscale platform provides a powerful tool that can be used to tune binding affinity to a biological target and control cellular internalization and subcellular localization.13, 14 The potential benefits of gold nanoparticle therapeutics were demonstrated recently in our research groups by transforming a weak CCR5 binding small molecule, which by itself was biologically inactive, into a multivalent gold conjugate that effectively inhibited HIV-1 fusion to peripheral blood mononuclear cells (PBMCs) in vitro.15 The biological activity of ligand-coated gold nanoparticles in the prevention of HIV-1 entry suggests that known, weak binding or perhaps even resistance-compromised small-molecule drugs may be transformed into potent therapeutics via conjugation to gold nanoparticles. We were also interested in determining whether completely new biologically active compounds could be discovered using ligand-coated gold nanoparticles. Specifically, could we identify nanoparticle formulations whose biological activity was dictated by a specific combination of ligands displayed on the surface of the particle? Indeed the answer appears to be yes, as we have found that gold nanoparticles with potent activity for Escherichia coli (E. coli) growth inhibition could be discovered from a library of mixed thiol-monolayer-coated gold nanoparticles.16 Here, we show that the gold nanoparticle library created to search for inhibitors of the Gram-negative E. coli could be used to discover inhibitors of the Gram-positive bacterium Staphylococcus aureus (S. aureus). The active nanoparticles that emerged from this screen consisted of a different subset of the library compared to those discovered in the previous search. This suggests that the display of ligand mixtures on gold nanoparticles could present new opportunities in the rapid identification of nanomaterials with biological activity toward a range of microbes. The library of nanoparticle conjugates was assembled by first synthesizing 2.0 nm diameter gold nanoparticles capped with p-mercaptobenzoic acid (pMBA).15, 16 These particles have a proposed empirical formula of [Au144(SC6H4COOH)60].17 The ten thiols shown in Figure 1 were chosen as a representative library of molecules containing H-bond donor/acceptor and hydrophilic/hydrophobic properties. These ligands were incubated with gold nanoparticles in combinations of three (initially at 1:1:1 molar ratios) to build a library of 120 nanoparticle conjugates. The conjugates were purified by salt and methanol precipitation to remove free thiols. A large subset of these formulations (62 combinations) displayed poor aqueous solubility under the equimolar ligand exchange concentrations used initially. This was rectified by simply adjusting the molar ratio of thiols added into the reaction mixture. For example, nanoparticle conjugates 28 and 50 were relatively insoluble in aqueous solution due to the low solubility of thiol 1 (see structure in Figure 1). The amount of thiol 1 in the exchange reaction was thus reduced to 67% of the original feed. With the exchange reaction optimized for solubility, a total of 95 nanoparticle conjugates could be screened for antibiotic activity (see Supporting Information, Table S1 for the composition of the entire library). An initial screen against methicillin- susceptible S. aureus (MSSA) revealed activity that depended upon the combination of thiols conjugated to pMBA-coated gold nanoparticles (Figure 2). The ten thiols chosen as a representative library of molecules containing H-bond donor/acceptor and hydrophilic/hydrophobic properties. Antibiotic activity of several mixed ligand-coated gold nanoparticles. In parentheses are the thiols combined to generate each nanoparticle. Nanoparticle concentrations were 25 μM. POS indicates the positive control, and CFU indicates the colony forming unit. Nanoparticle conjugates that showed >90% growth inhibition at nanoparticle concentrations of 25 μM in the initial screen were chosen for further analysis. Cultures of MSSA were incubated with varying concentrations (10–50 μM) of nanoparticles. Assays were conducted under standard broth dilution procedures followed by colony counting to assess bacterial viability after incubation with nanoparticles. The nanoparticles shown in Table 1 displayed the highest decrease in growth, with conjugate 6 yielding 99.9% growth inhibition at 10 μM. The nanoparticles listed in Table 1 were then tested for activity against methicillin-resistant S. aureus (MRSA) and the Gram-negative bacterium E. coli. This screen allowed us to assess whether nanoparticle formulations could be discovered that were not susceptible to current mechanisms of drug resistance and whether nanoparticle formulations were also active against Gram-negative bacteria. All conjugates were as active against MRSA as they were toward MSSA except conjugate 56, which showed only 99.0% growth inhibition at 50 μM. None of the conjugates were active toward the inhibition of E. coli. The inhibitory activities of the individual, unconjugated thiols were then determined. Thiols 2, 5, 6, 8, 9, and 10 showed little to no inhibition of MSSA growth at concentrations as high as 2 mM (<0.4 log decrease). Thiols 3 and 4 were potent growth inhibitors, which is not surprising since phenols and anilines are known antiseptics. It was also found that thiol 7 showed inhibitory activity of ca. 1 log at 2 mM. However, for this thiol alone to be responsible for the activity of nanoparticles 50 and 56 it would require that more than 100 of them were coordinated to the gold surface, an unlikely scenario given that these nanoparticles can only accommodate 60 ligands total. Thiol 1 could not be screened in solution due to poor solubility in the broth used. However, on agar containing 500 μM 1 and 10% dimethyl sulfoxide (DMSO), no inhibition was observed. Various combinations of the free thiol monomers were then incubated with MSSA. Surprisingly, binary mixtures of 50 μM pMBA and 50 μM thiol 1 or 150 μM pMBA and 150 μM thiol 2 showed 99.9% MSSA growth inhibition. The activity of conjugates containing pMBA and thiols 1 or 2 is thus independent of their attachment to the nanoparticle; however, conjugation of 1 to the nanoparticle has the advantage of converting it into a water-soluble conjugate. Once we had identified active nanoparticle formulations, we employed IR spectroscopy to confirm the presence of thiols on conjugates 6, 28, and 50 (Figure 3). Characteristic vibrations for thiols 1 and 8 were observed for conjugate 6, and thiols 1 and 5 for conjugates 28 and 50. Thiols 2, 8, and 7 were not detected in conjugates 6, 28, and 50, respectively, likely because they were not present in sufficient quantities to be detected by IR or did not have vibrations that could be assigned unambiguously given the other thiols present. IR spectra of active nanoparticle conjugates 6, 28, and 50 confirms the formation of mixed thiol monolayers. A–D) Spectra of nanoparticle conjugates 6, 50, 28, and pMBA-gold nanoparticles, respectively. The * indicates a representative band for pMBA–Au nanoparticle conjugates. The ^, §, and £ correspond to vibrations unique to 3-(nitrobenzyl)mercaptan, glutathione, and 3-mercapto-1-propane sulfonate, respectively, as determined from the spectra of the free thiols. Finally, the toxicity of conjugate 50A was assessed with a hemolysis assay, yielding an HC50 of 40 μM. This value corresponds to a hemolytic index (HC50/MIC99.9) of 2 where HC50 is the concentration capable of causing 50% red blood cell lysis and MIC99.9 is the minimal inhibitory concentration resulting in 99.9% growth inhibition of S. aureus. Nanoscale systems including DNA aptamers, antibodies, proteins, and inorganic nanoparticles such as the gold particles described herein are attractive as therapeutics in part because of their tunable valencies, blood circulation times, and biodistribution profiles. In addition, nanoscale therapeutics are often adept at disrupting protein–protein interactions that can drive disease pathogenesis. In contrast, small-molecule therapeutics typically rely on a single high-affinity contact to a disease target and have difficulty blocking protein–protein interactions. A significant advantage of small-molecule drugs, however, is the ease with which large chemical and structural diversity can be manufactured and screened for biological activity. It has thus been proposed that methods capable of blending the properties of nanoscale systems with the chemical diversity of small molecules will lead to the discovery of superior therapeutic agents.1 We have shown that a library of small-molecule ligand-coated gold nanoparticle conjugates may be generated rapidly via one-pot thiol exchange reactions. The nanoparticle conjugates are prepared at room temperature in aqueous solution and purified using a simple aqueous salt/methanol precipitation and resuspension procedure. Considering solely the number of commercially available thiols (>200), there is potential to access significant chemical and structural diversity with this approach. While the aqueous solubility of the resulting nanoparticle conjugates may in some cases be low (as experienced with many of the compounds in our initial 120-member library), this can be overcome by simply tuning the molar ratios of the ligands during the exchange reaction or by combining thiols with low aqueous solubility with highly water-soluble thiols. The library of 95 unique ligand-coated gold nanoparticles investigated here revealed differential activity toward the inhibition of bacterial growth, with one conjugate displaying 99.9% growth inhibition at 10 μM for both MSSA and MRSA. Whether the bacterial growth inhibition observed for these conjugates is due to efficient internalization, nanoparticle aggregation inside of the cells, or enhanced binding to a biomolecule target located in the cell membrane or inside of the cell is currently not known. As a comparison we note that the minimum inhibitory concentrations of vancomycin, ciprofloxacin, and cefixime against MSSA are ca. 0.7, 1.5, and 17 mM, respectively; thus nanoparticle formulations can be rapidly identified from simple thiol building blocks that are comparable to conventional antibiotics with respect to in vitro bacterial growth inhibition. Synthesis of 2.0 nm Gold Nanoparticles: Two-nanometer dia­meter [Au144(SC6H4COOH)60] gold nanoparticles were synthesized as previously described.1 In short, a solution of HAuCl4 (11.1 mM), pMBA (37.8 mM), and NaOH (180 mM) in aqueous methanol (55.6% (v/v)) was prepared and allowed to equilibrate for 24 h with constant stirring. Fifty milliliters of this solution were diluted with the addition of methanol (260 mL) and water (740 mL). The Au+ was reduced with the addition of aqueous NaBH4 (10 mL, 0.25 M). The final methanol concentration was adjusted to 25% with the addition of water (100 mL). The reduction of gold was allowed to proceed for 48 h at room temperature with constant stirring. Gold nanoparticles were precipitated with the addition of NaCl (70 mmol) and methanol (500 mL) (final methanol concentration of 47% v/v) followed by centrifugation (3200 rcf, 5 min). The precipitated nanoparticles were reconstituted in water. The concentration was measured by UV–vis spectroscopy, using the molar extinction coefficient at a wavelength of 510 nm, ϵ510 nm, of 409 440 M−1 cm−1. Place Exchange Reactions: One-pot place exchange reactions were conducted with the addition of thiol added in 1:1:1 molar ratio (740 μM total) to gold nanoparticles (7.4 μM) in sodium phosphate buffer, pH 9.5 (20 mM, 15 mL). These solutions were prepared from 20 mM stock solutions of the individual thiols. Thiols 5, 6, 7, 8, 9, and 10 stock solutions were prepared in H2O, while thiols 1, 3, and 4 were dissolved in DMSO, and thiol 2 was dissolved in 20% glycerol. It is important to note that stock solutions of thiol 1 had to be be stored at –80 °C to avoid conversion into a species that fails to place exchange properly onto pMBA-capped gold nanoparticles. Reactions were placed on a plate shaker and agitated for 24 h at room temperature. The exchange product was harvested through the addition of NaCl (11 mL of 4 M stock for a final concentration of 0.8 M) and a volume of methanol equal to that of the reaction volume plus added salt water (26 mL). Reactions were centrifuged (3200 rcf, 30–60 min). Precipitated nanoparticles were resuspended and precipitated with the addition of NaCl and methanol two times to wash out excess unreacted thiol. Particles were allowed to dry to completion overnight at room temperature and resuspended in water. Resuspended nanoparticles were washed with water over a 30 000 MWCO (mole­cular weight cut-off) centricon filter to remove excess salt and thiol followed by buffer exchange into Mueller–Hinton broth for assay. Optimizing the molar feed ratios resulted in decreasing the concentration of thiols 1, 2, 3, 4, 6, 9, and 10 by 67% of the original value, with the rest of the reaction remaining as described above. The initial screen suggested that thiol 5 (glutathione) was an important ligand in the preparation of gold nanoparticle conjugates with activity toward MSSA. It was thus of interest to determine whether activity could be enhanced by starting with glutathione-capped gold nanoparticles rather than pMBA-capped gold nanoparticles. Glutathione-capped gold nanoparticles were synthesized by mixing HAuCl4•xH2O (0.4 mmol) in methanol (20 mL) with glutathione (1.4 mmol) in H2O (15.4 mL) supplemented with NaOH (0.6 mL of 10 M stock solution). This solution was then divided into thirds. The following was then added to each reaction: methanol (62 mL) and H2O (178 mL), followed by aqueous NaBH4 (2.4 mL, 0.25 M), and finally water (24 mL). Purification of the particles was performed via precipitation of nanoparticles with the addition of NaCl (40 mmol) and methanol (250 mL). Centrifugation of samples allowed for the purification of particles from the solution. Particles were then dissolved in water and washed over a 10 000 MWCO centricon filter to remove excess glutathione and salts. Place exchange reactions were then conducted on these particles as described above. Nanoparticles prepared in this way were designated conjugate 50A. Generation of nanoparticle conjugate 50A occurred through the addition of pMBA (2.95 μmol) and (3-nitrobenzyl)mercaptan (2.95 μmol) to glutathione-capped gold nanoparticles (0.05 μmol) in water (4 mL). These particles did not show enhanced activity versus conjugates prepared with pMBA-capped gold nanoparticles, although batch-to-batch variability was observed to improve. Bacterial Growth Inhibition Assays: Inoculation of S. aureus into Mueller–Hinton broth (3 mL, BD) was carried out by touching the top of 4 well-isolated colonies of MSSA (ATCC 29213), MRSA (ATCC BAA-44), or E. coli (ATCC 25922) from a Mueller–Hinton agar (BD) plate with an inoculation loop. The culture was allowed to grow at 37 °C, 225 rpm until mid-log phase after which it was diluted to 1 × 106 CFU/mL in Mueller–Hinton broth. Equal volumes of diluted inoculum and nanoparticle sample (adjusted to the correct assay concentration in Mueller–Hinton broth) were mixed to make the final inoculum concentration 5 × 105 CFU/mL. Samples were incubated at 37 °C, 225 rpm for 18 h. End points were determined by colony counting on Mueller–Hinton agar after dilution of each sample in phosphate buffered saline (PBS) and incubation of the plates at 37 °C for 24 h. Initial assays were conducted in nutrient broth and nutrient agar with S. aureus ATCC 9144, with final growth inhibition data documented for Mueller–Hinton broth and agar with S. aureus ATCC 29213. Infrared Spectroscopy: Nanoparticle samples were reconstituted and washed of contaminants over a 30 K MWCO centricon filter with water. Samples were then spotted onto potassium bromide Real Crystal IR cards (International Crystal Laboratories) in their appropriate solvent and allowed to dry. IR analysis was carried out on a Thermo Nicolet Avatar 360 FT-IR spectrometer. Blood Hemolysis Assay: Hemolysis assays were performed on mechanically difibrinated sheep's blood (Hemostat Labs: DSB100). Briefly, blood (1.5 mL) was placed into a microcentrifuge tube and centrifuged (10 000 rpm, 10 min). Cells were resuspended and washed with PBS (1 mL). The final cell suspension was then diluted tenfold and nanoparticle compound was added in PBS. PBS alone was used as a zero hemolysis marker and a 1% Triton X sample was used as a 100% lysis marker. Samples were then incubated at 37 °C, 200 rpm for 1 h followed by centrifugation (10 000 rpm, 10 min). The resulting supernatant was diluted by a factor of 40 in distilled water. The absorbance of the supernatant was measured with a UV–vis spectrometer at a 540 nm wavelength. As gold nanoparticles absorb readily at a 540 nm, the nanoparticles of each test sample were precipitated out of solution with NaCl and methanol. The resulting pellet was resuspended and the absorbance was measured with a UV–vis spectrometer at a 540 nm wavelength. This A540 nm reading represents nanoparticles and was subtracted from the initial A540 nm reading of the supernatant to yield A540 nm of cell lysis only. Supporting Information is available from the Wiley Online Library or from the author. The authors wish to thank the Bill and Melinda Gates Foundation for funding. The article is dedicated to Professor Chad Mirkin, a pioneering scientist and devoted mentor. This Communication is part of the Special Issue dedicated to Chad Mirkin in celebration of 20 years of influential research at Northwestern University. Detailed facts of importance to specialist readers are published as ”Supporting Information”. Such documents are peer-reviewed, but not copy-edited or typeset. They are made available as submitted by the authors. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.