Water-Soluble Nitric Oxide-Releasing Gold Nanoparticles

Joseph S Fernandez-Moure , Michael Evangelopoulos , Kayla Colvill .... Ammar T. Qureshi , William T. Monroe , Vinod Dasa , Jeffrey M. Gimble , Daniel ...
1 downloads 0 Views 320KB Size
4938

Langmuir 2007, 23, 4938-4943

Water-Soluble Nitric Oxide-Releasing Gold Nanoparticles Mark A. Polizzi, Nathan A. Stasko, and Mark H. Schoenfisch* Department of Chemistry, UniVersity of North Carolina at Chapel Hill, Chapel Hill, North Carolina 27599 ReceiVed NoVember 20, 2006. In Final Form: January 24, 2007 The synthesis and characterization of water-soluble nitric oxide (NO)-releasing monolayer-protected gold clusters (MPCs) are reported. Tiopronin-protected MPCs (∼3 nm) were functionalized with amine ligands and subsequently exposed to 5 atm of NO to form diazeniumdiolate NO donors covalently bound to the gold MPC. Diazeniumdiolate formation conditions, NO-release, and nanoparticle stability were examined as a function of the structure of the protecting ligand, pH, and storage time. Despite their aqueous solubility, proton-initiated decomposition of the diazeniumdiolate-modified Tio-MPCs resulted in only modest NO-release (10 mol % of the tiopronin protecting ligands with DETA resulted in aggregation and permanent loss of water solubility, likely the result of DETA reacting with adjacent MPCs. At 10 mol % DETA modification, the size of the gold core remained constant, indicating such coupling did not compromise ligand/ nanoparticle stability (Figure 1B). The resulting DETA-modified Tio-MPCs were again purified by dialysis for 2 d before characterization. As expected, the sharp multiplets for the methylene protons of DETA at 2.7 and 2.8 ppm were absent in the 1H NMR spectrum of 10 mol % DETA-modified Tio-MPCs. Rather, broad peaks consistent with MPC coupling34 appeared downfield at ∼3.0 and 3.1 ppm (Supporting Information), supporting DETA immobilization. A Kaiser test was employed to further confirm the covalent attachment of DETA and the resulting amine-functionalized exterior. As expected, an absorption maximum at 570 nm, characteristic of the reaction of ninhydrin with primary amines, was observed for amine-modified Tio-MPCs but not for unmodified Tio-MPC controls. Elemental analysis was also used to confirm Tio-MPC synthesis. The measured wt % correlated well with the predicted values for MPCs with core diameters of ∼2.8 nm (CHNtheor: %C 8.24, %H 1.23, %N 1.92. CHNobs: %C 8.30, %H 1.42, %N 1.89).34 As expected, an increase in organic content was observed upon DETA functionalization (CHNexp: %C 9.45, %H 1.87, %N 3.63). Previous studies have indicated that diazeniumdiolate formation is favored in the presence of a strong base that promotes deprotonation of the backbone secondary amines during NO addition.35 For example, Zhang et al. employed sodium methoxide base to increase diazeniumdiolate formation on fumed silica particle-based NO donor systems.28 Similarly, Rothrock et al. observed enhanced NO-release from amine-modified alkanethiolMPCs suspended in 0.5 M NaOMe/MeOH during NO exposure.27 Initially, a similar suspension was used for the DETA-modified Tio-MPCs. Unfortunately, the Tio-MPCs formed an aggregated suspension in NaOMe/MeOH, thus precluding the use of methoxide base during diazeniumdiolate formation. As shown in Figure 1C, the aggregation yielded larger, irregularly shaped gold clusters. Elemental analysis confirmed that the aggregation was a result of ligand etching by the strong base resulting in deprotection of the gold cores (decreases from 9.45 to 4.25, 1.87 to 0.98, and 3.64 to 1.23 wt % for C, H, and N, respectively). The real-time and total NO-release profiles for control and DETA-modified Tio-MPCs are shown in Figure 2. After exposure to conditions necessary to synthesize diazeniumdiolates, control Tio-MPCs still released minute levels of NO over the first 5 min of solution immersion (Figure 2A), indicative of adsorbed or physically entrapped NO. In contrast, the release of NO from DETA-modified Tio-MPCs was distinct with an NO-release halflife and duration of 16 min and ∼2.5 h, respectively. The conversion efficiency of amines to diazeniumdiolates was calculated to be ∼1%, a value similar to that reported for waterinsoluble alkanethiol-MPCs.27 The low conversion efficiency is attributed to ligand etching and MPC aggregation (confirmed via TEM) during the diazeniumdiolate formation reaction, which inherently limits the number of available amine precursors for reaction with NO. (34) Hostetler, M. J.; Wingate, J. E.; Zhong, C.-J.; Harris, J. E.; Vachet, R. W.; Clark, M. R.; Londono, J. D.; Green, S. J.; Stokes, J. J.; Wignall, G. D.; Glish, G. L.; Porter, M. D.; Evans, N. D.; Murray, R. W. Langmuir 1998, 14, 17-30. (35) Drago, R. S.; Ragsdale, R. O.; Eyman, D. P. J. Am. Chem. Soc. 1961, 83, 4337.

4942 Langmuir, Vol. 23, No. 9, 2007

Polizzi et al.

Table 1. Nitric Oxide Release Characteristics as a Function of the Water-Soluble MPC System

a

MPC type

[NO]m (ppb/mg)

total NO (µmol/mg)

t1/2 (min)

release longevity (h)

conversion efficiency (%)

DETA Tio-MPCsa TEPA Tio-MPCsa PEHA Tio-MPCsa DETA-MPCsb TEPA-MPCsb PEHA MPCsb

155 143 344 289 1009 3522

0.007 0.013 0.023 0.121 0.297 0.386

8 9 12 97 119 55

1.5 1.5 1.5 5 >16 >16

2 1.4 1.9 33.4 17.5 15.6

0.1 M NaOH/H2O used for NO exposure. b 0.5 M NaOMe/MeOH used for NO exposure. Table 2. Nitric Oxide Release Characteristics of PEHA-Protected MPCs as a Function of pH pH

[NO]m (ppb/mg)

t1/2 (min)

3 7.4 10

17 527 3533 170

2.4 56 >960

To circumvent the aggregation observed during NO exposure in methoxide/MeOH, diazeniumdiolate formation was carried out under aqueous conditions (0.1 M NaOH/H2O). The TioMPCs were fully soluble under these conditions, and the core sizes were unchanged following NO exposure as evidenced via TEM micrographs (Figure 1D). Since water may serve as a proton source for diazeniumdiolate decomposition, sodium hydroxide was employed to maintain solution basicity and circumvent premature breakdown of the NO donor functionality. A concentration of 0.1 M hydroxide proved optimal for preventing aggregation. The level of NO released from diazeniumdiolatemodified DETA-Tio-MPCs was slightly enhanced in the absence of aggregation (Figure 3A). Modification of Tio-MPCs with larger polyamines including tetraethylenepentamine (TEPA) and pentaethylenehexamine (PEHA) was also evaluated in an attempt to increase NO storage capacity by introducing a greater number of potential diazeniumdiolate formation sites. Regardless of the length of the polyamine employed, modifications were limited to 10 mol % to maintain particle solubility. The number of secondary amines was systematically altered from one (DETA) to three (TEPA) to four (PEHA). As expected, the total NO-release for the three polyamine MPCs scaled with increasing number of secondary amines per ligand (Figure 3). Despite the higher initial levels of NO-release ([NO]m), the calculated conversion efficiency remained low for each polyamine (2.0%, 1.4%, 1.9% for DETA-, TEPA-, and PEHA-modified Tio-MPCs, respectively), with NOrelease expiring at ∼1 h for all Tio-MPC systems. Because MPC aggregation was not observed, the lower level of NO released was likely the result of ineffective diazeniumdiolate formation in the weaker hydroxide base solution. Synthesis of Nitric Oxide-Releasing Polyamine-Protected MPCs. To create a water-soluble MPC system with greater stability in both methoxide and MeOH, polyamine protecting ligands (i.e., DETA, TEPA, and PEHA) were used to stabilize the gold clusters directly. Because the polyamine-protected MPCs were formed using only amine-protecting ligands, the potential NO storage was enhanced by avoiding further ligand modification. Amine-protected MPCs have been previously synthesized with size and stability comparable to that of thiol-protected analogues.36,37 Diethylenetriamine (DETA)-, tetraethylenepentamine (TEPA)-, and pentaethylenehexamine (PEHA)-protected MPCs were (36) Brown, L. O.; Hutchinson, J. E. J. Am. Chem. Soc. 1999, 121, 882-883. (37) Aslam, M.; Fu, L.; Su, M.; Vijayamhanan, K.; Dravid, V. P. J. Mater. Chem. 2004, 14, 1795-1797.

synthesized following a method outlined by Schulz-Dobrick et al.19 to produce nanoparticles with a multitude of NO-donor precursors. This method is based on the initial reduction of a gold salt, forming gold nanoparticle clusters that are weakly protected by diglyme solvent interactions. Addition of appropriate amine-terminated ligands permanently pacifies the gold core, thereby allowing direct attachment of NO-donor precursors in one step, facilitating the synthesis of MPCs with a greater number of polyamine ligands than the Tio-MPC system. In contrast to Tio-MPCs, the polyamine-protected MPCs precipitated from solution following their synthesis and could thus be more easily purified by centrifugation and washing with methylene chloride. In this respect, the lengthy dialysis purification steps were avoided, significantly reducing the overall preparation time from days to hours. Regardless of the polyamine protecting ligand, the diameter of the gold cores was found to be 5.6 ( 1.6 nm. In contrast to the Tio-MPCs, the polyamineMPCs did not aggregate during subsequent synthesis of diazeniumdiolate NO donors in 0.5 M NaOMe/MeOH (Figure 4). Thermogravimetric analysis data also supported this claim as the organic content of the MPCs did not change following diazeniumdiolate formation for the three polyamine systems. In fact, the organic content of the DETA-, TEPA-, and PEHAprotected MPCs (1.8, 5.4, and 6.6 wt %, respectively) was lower than that predicted for similarly sized alkanethiol-protected MPCs (4.8%, 8.4%, and 10.1% for DETA-, TEPA-, and PEHA-MPCs, respectively), assuming 41% ligand surface coverage on a 5.6 nm core consisting of 6266 core gold atoms and 601 protecting ligands.34 These results are not surprising, however, because not all of the amine protecting ligands would extend outward, perpendicularly from the gold core. Rather, some ligands likely bind to the core at both ends (i.e., via both amines), sequestering two gold binding sites and preventing uniform ligand packing. Increased steric hindrance and/or charge repulsion of protonated amines at pH 7.4 also may limit ligand packing density at the gold surface. To examine the possibility of polyamine ligands binding via both primary amine termini, N-propylethylenediamine-protected MPCs (PEDA-MPCs) were synthesized. The structure of PEDA is identical to DETA except that one of the primary amines is replaced with a terminal methyl group, eliminating potential double binding. Indeed, TGA revealed that the organic content for PEDA-MPCs was 7.3 wt % as compared to 1.8 wt % for the DETA-MPCs, supporting our hypothesis that a greater number of PEDA ligands were attached to the gold core than when DETA was used. The stability of the polyamine-protected MPCs in methoxide/ MeOH, coupled with the increased concentration of amines (diazeniumdiolate precursors) over those attainable using aminemodified alkanethiol- and Tio-MPCs, resulted in substantially elevated levels of NO-release (Figure 5). Both the initial maximum level of NO-release ([NO]m, ppb/mg) and the total amount of NO released (t[NO], µmol/mg) for amine-protected MPCs were significantly greater than those observed for the amine-modified

Water-Soluble NO-Releasing Gold Nanoparticles

Tio-MPCs (Table 1). The duration of NO-release was also extended to >16 h for TEPA- and PEHA-MPCs as expected on the basis of the exceptional stability of small molecule polyamine NO adducts.9 Indeed, polyamine diazeniumdiolates have been shown to have markedly different NO-release kinetics governed by the presence of remote amines offering alternative sites for protonation and thus slowing decomposition.10 For example, the biphasic character of the TEPA-MPCs (Figure 5B) may be a result of two distinct diazeniumdiolates on the polyamine structure that have independent NO-release kinetics and correspondingly different [NO]m. The total NO-release from the polyamine MPCs corresponds to diazeniumdiolate conversion efficiencies of 33.4%, 17.5%, and 15.6% for DETA-, TEPA-, and PEHA-MPCs, respectively. The lower conversion efficiencies for TEPA and PEHA (i.e., the larger polyamines) are attributed to charge repulsion effects that hinder diazeniumdiolate formation on neighboring amines of the same ligand. Because diazeniumdiolate decomposition in solution is proton initiated,10 the effect of pH on NO-release from PEHA-MPCs was also investigated. Previous studies have demonstrated that the kinetics of NO-release from small molecule diazeniumdiolates increases significantly at acidic pH.10 Polyamine-MPC samples were immersed in 0.1 M citric acid buffer (pH 3.0), phosphate buffered saline (pH 7.4), and 0.1 M sodium borate buffer (pH 10.0). As expected, the [NO]m and rates of NO-release increased with decreasing pH (Table 2).

Conclusions The synthesis of 3-5 nm water-soluble NO-releasing gold nanoparticles represents an important step toward developing NO-release scaffolds for biological applications. As demonstrated herein, polyamine-protected MPCs facilitate the storage and

Langmuir, Vol. 23, No. 9, 2007 4943

release of NO at levels equivalent to larger (200 µm), waterinsoluble scaffolds (e.g., silica and polyethyleneimine microspheres).28,38 Selection of the protecting ligand (amine) used during nanoparticle synthesis allows for tunable NO-release properties varying from 0.007 to 0.386 µmol/mg. Addtionally, the polyamine-protected MPCs present a multivalent surface for use in subsequent delivery efforts. Future studies aim to modify the cationic primary amine functional groups with charge neutral groups to eliminate the potential toxicity associated with the amines and/or harmful nitrosamine formation following NOrelease. Functionalization of the nanoparticle exteriors with ligands for tissue-specific targeting or molecular probes for imaging may further facilitate the use of NO-releasing polyamineprotected MPCs for a wide range of pharmacological applications. During preliminary experiments, we have labeled PEHA-MPCs with fluorescein isothiocyanate (FITC). Such MPCs were modified to release NO and reconstituted in water for fluorescence measurements. Studies to quantify the quantum yield of fluorescence in the presence of NO-release are currently underway. Acknowledgment. This research was supported by the National Institutes of Health (NIH EB 000708). Supporting Information Available: 1H NMR, UV-vis, TGA, and NO-release characterization of the diazeniumdiolate-modified MPCs. This material is available free of charge via the Internet at http:// pubs.acs.org. LA0633841 (38) Pulfer, S. K.; Ott, D.; Smith, D. J. J. Biomed. Mater. Res. 1996, 37, 182-189.