A noncovalent, fluoroalkyl coating monomer for phosphonate-covered nanoparticles

Abstract

Gadolinium-containing phosphonate-coated gold nanoparticles were prepared and then non-covalently coated with an amphiphilic fluorous monomer. The monomer spontaneously self-assembles into a non-covalent monolayer shell around the particle. The binding of the shell utilizes a guanidinium-phosphonate interaction analogous to the one exploited by the Wender molecular transporter system. Particle-shell binding was characterized by a 27% decrease in ¹⁹F T₁ of the fluorous shell upon exposure to the paramagnetic gadolinium in the particle and a corresponding increase in hydrodynamic diameter from 3 nm to 4 nm. Interestingly, a much smaller modulation of ¹⁹F T₁ is observed when the shell monomer is treated with a phosphonate-free particle. By contrast, the phosphonate-free particle is a much more relaxive ¹H T₁ agent for water. Together, these observations show that the fluoroalkylguanidinium shell binds selectively to the phosphonate-covered particle. The system’s relaxivity and selectivity give it potential for use in ¹⁹F based nanotheranostic agents.

1. Introduction

The development of “smart” nanoparticulate medicinal entities is an emerging area at the interface of chemical and clinical sciences. For example, novel drug delivery systems and imaging agents based on drug-loaded polymers or nanoparticles have appeared recently,¹ and are making way into the clinic.² Nanotechnology further offers opportunities to design combined therapeutic plus diagnostic agents, or “theranostic” systems; examples of such include paramagnetic nanoparticles for MRI imaging,^{3 64}Cu-loaded liposomes for PET imaging,⁴ and others.⁵

One key factor to controlling the in vivo behavior of these systems is the ability to control their surface properties.⁶ Along these lines, we are pursuing the development of monomeric organic compounds that engage in predictable self-assembly around nanoparticulate molecular probes and that have the ability to control the bioactivity of the former by modifying their size, surface, and access to the external aqueous milieu.⁷

We present here a general strategy for applying a poly)ethyleneglycol (PEG)-terminated fluoroalkyl coating to phosphonate-covered nanoparticle agent and illustrate the use of bioorthogonal ¹⁹F magnetic resonance relaxivity to track the behavior of the coating monomer as it associates with the particle. The coating monomer (1, Fig. 1) exploits a guanidinium head group, which can interact with particle surface phosphonate groups through a double hydrogen bonding system analogous to the one used by the Wender molecular transporter system for its initial adhesion to cell surfaces.⁸ Particles utilizing such coreshell interactions have various clinical applications, such as vectors for gene delivery, as investigated in the Wender lab⁹. Monomer 1 further features a fluoroalkane region that enables the self-assembly of a fluorous, Teflon-like layer that can limit the particle’s exposure to its aqueous surroundings.

Fig. 1.

Structure of monomer 1.

We have previously characterized the binding of monomer 1 with small-molecule phosphonates,¹⁰ and we hypothesize that a particle with a phosphonate surface can utilize the same interaction(s) to bind species 1, thus enabling the self-assembly of a monolayer coating around a particle (Fig. 2). Incorporating a paramagnetic species into such a particle can thus enable the use of monomer 1 as a ¹⁹F MRI contrast system for an appropriately filtered, T₁-weighted image plane. Although ¹⁹F MRI-based nanotheranostic technology is not as well developed as the analogous ¹H MRI based systems, ¹⁹F MRI based systems are beginning to emerge,^11,12 and offer important bioorthogonality unavailable to other imaging modalities.¹³ We show here potential utility of monomer 1 as a flexible tool for the installation of ¹⁹F groups on the surface of a phosphonatecovered nanoparticle and illustrate the corresponding modulation in ¹⁹F T₁ upon particle-1 binding. We further illustrate 1’s selectivity for a phosphonate surface over a phosphonate free particle and illustrate that the coverage provided by 1 corresponds to a monolayer in aqueous solution.

Fig. 2.

Hypothetical interactions between phosphonate core and guanidine shell to yield shell coated phosphonate particle 2

2. Results and discussion

2.1. Design and synthesis of shell monomer 1

Shell monomer 1 features a guanidinium head group, which is designed to participate in hydrogen bonding with phosphonates on the periphery of a particle. This is attached to a fluorous phase group¹⁴ that prevents intercalation of the shell in to the interior of the particle and limits the particle’s exposure to its surroundings. Appropriate fluorous starting materials for this design are readily available side products from teflon synthesis.¹⁵ A hydrophilic PEG chain is then appended as a solubilizing group. Scheme 1 summarizes the synthesis of monomer 1, which we have previously reported.¹⁰ This route relies on the intermediacy of a bench top-stable alkyl triflate (4), which is a versatile starting material for fluoroalkyl amines.

Scheme 1.

Synthesis of monomer 1.

2.2. NMR properties of monomer 1

Shell monomer 1 exhibits 3 different peaks in the ¹⁹F NMR spectrum, which can be rigorously assigned by a combination of ¹⁹F-¹³C HMBC and ¹³C-¹H HSQC NMR spectra (see supporting information). The most downfield signal (−117.9 ppm, peak A) corresponds to the fluorine CF₂ group nearest the guanidine of 1. An upfield (−123.5 ppm, peak C) multiplet corresponds to the two central CF₂ groups, and the middle peak of the ¹⁹F spectrum (−119.8 ppm, peak B) corresponds to the fluorines closest to the PEG tail: see Supporting Information regarding the ¹⁹F signal assignments. The T₁ relaxation times were measured in 25 mM pH = 7.6 TRIS-HCl buffer and were found to be 457(8) ms, 436(7) ms, and 497(7) ms for peaks A, B, and C respectively. These data are summarized in Fig. 3

Fig. 3.

NMR Properties of 1.

2.3. Design and synthesis of paramagnetic nanoparticles

Our design for a template nanoparticle is based on a very simple gold-centered structure, which is decorated with a periphery of thiol-terminated phosphonic acid surfactants (Fig. 4). The selection of this construction is based on its popularity in several scaffolds currently being developed for drug delivery and clinical imaging applications.¹⁶ It enables concise size control and flexibility of surface functionalization.^{17, 18} In this section we present synthetic routes for paramagnetic particles functionalized with both [Gd(DOTA)] and phosphonate peripheries.

Fig. 4.

Composition of nanoparticles.

A 1.5 nm phosphine-stabilized gold core (10) was first synthesized based on previously reported conditions.¹⁹ Thiols 6 and 7 were assembled onto the gold core via interfacial ligand exchange in a biphasic water/dichloromethane system (Scheme 2A).²⁰ The resulting particles form visible aggregates at neutral pH, but are stable in a pH = 7.6 TRIS-HCl buffer. Monomer 1 spontaneously self assembles onto particles 8 immediately upon introduction to an aqueous solution of 8 to yield hybrid gadolinium-phosphonate particles 2.

Scheme 2.

Synthesis of nanparticles.

Phosphonate-free particles 9 are apparently unstable to aggregation when prepared according to the procedure above, but in-situ deprotection of trityl-protected thiol 11 followed by treatment with HAuCl₄ can be used to generate these particles (Scheme 2B).²¹ This method yields particles that are of 7 nm in hydrodynamic diameter as measured by dynamic light scattering (DLS). ¹⁹F longitudinal relaxivity time constants (¹⁹F T₁) provide evidence for the binding of particle and monomer 1. Upon binding of 1 to the particle core, the fluorine nuclei are brought into close proximity of the particle’s paramagnetic gadolinium centers, where the paramagnetic gadolinium center causes a decrease in the T₁ relaxation times of the ¹⁹F nuclei.²²

2.4. Particle-shell binding

Upon addition of an aqueous solution of particle 8 to 1, the ¹⁹F T₁ relaxation times for 1 decreased to 332(14) ms, 348(4) ms and 385(4) ms respectively (Table 1). Significant decrease of the observed T₁ values upon the addition of paramagnetic particles is consistent with an intimate interaction between 1 and the particle core. To assess the role of the phosphonate functionality on particle 8, phosphonate free particles 9 were also treated with 1. In this experiment, the T₁ relaxation times were 434(6) ms, 431(7) ms, 481(5) ms. These data show that 9 exhibits lower changes in T₁ than those observed for particle 8. This suggests a more intimate interaction between 1 and 8 versus 1 and 9. Furthermore, exposure of the shell (1) to 1.0 mM GdCl₃ (0.25 molar equivalents relative to 1, 158 ppm wt/wt [Gd]) also resulted in a less significant change in ¹⁹F T₁ than our phosphonate coated particles (8, Table 1), which shows that high [Gd] alone can not account for the observed decrease in ¹⁹F T₁ observed in the presence of 8.

Table 1.

Decrease of ¹⁹F T₁ when particle is added to 1.

	Peak A −117.9 ppm	Peak B −119.8 ppm	Peak C −123.5 ppm
1	457(8) ms	436(7) ms	497(7) ms
1 + 8	332(14) ms	348(4) ms	385(4) ms
1 + 9	434(6) ms	431(7) ms	481(5) ms
1 + GdCl3	429(13) ms	396(8) ms	466(6) ms

¹H T₁ time constants of H₂O for aqueous solutions of particles 8 and 9 reveal that, although 8 is a more effective ¹⁹F T₁ contrast agent, 9 is a more efficient ¹H T₁ contrast agent. We measured the H₂O ¹H T₁ relaxation of the same particle solutions used for acquisition of the foregoing ¹⁹F T₁ data. Particle 8 showed a ¹H T₁ of 1.18(5) s, and particle 9 exhibited a T₁ of 0.495(3) s. The significantly lower H₂O ¹H T₁ of 9 shows that it is a more relaxive ¹H contrast agent, whereas particle 8 is a more relaxive ¹⁹F T₁ contrast agent. This remarkable contrast between ¹H and ¹⁹F highlights the importance of a guanidinium-phosphate interaction in binding 1 to the particles: whereas phosphonaterich 8 holds 1 in the proximity of the paramagnetic gadolinium, phosphonate-free 9 does so less well, relative to a water standard. Moreover, absolute gadolinium concentration measurements by ICP (ppm wt/wt) indicate [Gd] = 52 ppm in the experimental solution of 8 and [Gd] = 37 ppm in the experimental solution of 9. Although these values are similar, they are inconsistent with the relatively short ¹H T₁ of the 9 solution, which can be attributed to slower tumbling of the larger particle.

2.5. Particle sizes

The foregoing relaxivity-based shell binding data are supported by dynamic light scattering (DLS) measurements. DLS data for particle 8 indicate monodisperse particles of hydrodynamic diameter of 3.0 nm (Fig. 5A). Addition of shell to the particle increased the size of the particle to 4.0 nm (Fig. 5B). The increase in diameter of 1 nm upon association of 1 is appropriate for the addition of two equivalents of 1, along with associated water, along the diameter. Further, these data also show that the particles are not aggregating upon changing their surface from phosphonate to PEG.

Fig. 5.

(A) DLS of particle 8 (average hydrodynamic diameter = 3.0 nm) (B) DLS of particle 2 (after the addition of monomer 1, hydrodynamic diameter = 4.0 nm).

3. Conclusion

In conclusion we have characterized the non-covalent coating of a gadolinium-containing gold nanoparticle with a fluorous shell. The coating utilizes a phosphonate-guanidine interaction and is characterized by both a decrease in ¹⁹F T₁ and an appropriate increase in the particle’s hydrodynamic radius. We further use ¹H and ¹⁹F NMR relaxivity data to show that phosphonate groups on the surface of a particle increase the particle’s affinity for fluoroguanidine 1 relative to water. Because these observations describe a general procedure for the selective self-assembly for fluoroaklyl groups around the surface of a phosphonate-covered material, we believe that they have potential applications in biomaterials and ¹⁹F MRI-based nanotheranostic systems.

4. Experimental section

General procedures

Deuterated NMR solvents were purchased from Cambridge Isotopes Labs. All NMR spectra were obtained on a Varian 500 MHz spectrometer at 25 °C. DLS were acquired on a Wyatt Dynapro Titan instrument. UV-Vis spectra were acquired on a Shimadzu UV-1800 spectrometer. Phosphine-stabilized gold nanoparticles (9) with diameter 1.5 nm were synthesized according to previously reported conditions,²³ and characterized by a match of their UV-Vis spectrum. TRIS buffer was purchased from Angus Chemicals. ICP-MS data were acquired by the University of Illinois at Urbana-Champaign School of Chemical Sciences Microanalysis Laboratory.

4.1. Synthesis of gadolinium-thiol 6

Gadolinium thiol 6 was synthesized based on previously reported conditions.²³ Mass data for 6 is consistent with a known sample. MALDI for C₁₉H₃₂GdN₅O₇S: Calculated 632.13 g/mol; found 631.94 g/mol. Measured isotopic distribution matches calculated prediction.

4.2. Synthesis of particle 8

Phosphine stabilized gold nanoparticles 10 (1.0 mg) were dissolved in dichloromethane (1 mL), and TRIS-HCl buffer (1 mL of 25 mM pH = 7.6) was added to form a biphasic mixture. While stirring at room temperature, gadolinium thiol 6 (40 µL of a 10 mM aqueous solution) was added, and the solution was stirred for 4 hours. Thiol phosphonic acid 7²⁴ was then added and the reaction was stirred until the brown color partitioned completely into the aqueous layer. The aqueous layer was diluted with TRIS-HCl buffer (1.0 mL of a 25 mM solution at pH = 7.6) and extracted with dichloromethane²⁵ (5 mL, 5 times), then filtered through a 0.45 µm syringe filter to yield an aqueous solution of particle 8.

4.3. Synthesis of particle 9

Trityl-protected gadolinium complex 11 (33.4 mg, 38 µmol) was suspended in tetrahydrofuran (0.4 mL). Chloroauric acid²⁶ (HAuCl₄, 13.0 mg, 38 µmol) was added to the solution at room temperature and the reaction was stirred for 1 hour. Triethyl silane²⁷ (Et₃SiH, 4.2 mg, 38 µmol) was then added, and the solution stirred for 18 hours at room temperature. The solution was filtered and ethanol²⁸ (10 mL) was added to initiate precipitatation. The precipitate was washed with ethanol (5 times, 10 mL), and the filter paper was sonicated in H₂O (10 mL) to dissolve precipitate. The resulting aqueous solution was then filtered through a 0.45 mm syringe filter to yield particles 9 as an aqueous solution. DLS measurements on the resulting solution showed a monodisperse nanoparticulate material with hydrodynamic diameter of 7 nm.

4.4. Measurement of ¹⁹F T₁ of 1-coated particles

A portion of 1 (3 µL of a 100 mM solution) was lyophilized to an oil in a 1/2 dram glass vial. A 125 µL aliquot of aqueous particle solution was delivered to the vial containing lypholized 1, stirred, then placed in a 3 mm diameter coaxial NMR tube insert and ¹⁹F T₁ time constants were acquired using the parameters previously stated.