Tuning Cationic Block Copolymer Micelle Size by pH and Ionic

Aug 3, 2016 - The formation, morphology, and pH and ionic strength responses of cationic block copolymer micelles in aqueous solutions have been ...
0 downloads 0 Views 5MB Size
Download PDF
Article pubs.acs.org/Biomac

Tuning Cationic Block Copolymer Micelle Size by pH and Ionic Strength Dustin Sprouse,† Yaming Jiang,‡ Jennifer E. Laaser,† Timothy P. Lodge,*,†,‡ and Theresa M. Reineke*,† †

Department of Chemistry, and ‡Department of Chemical Engineering & Materials Science, University of Minnesota, 207 Pleasant Street SE, Minneapolis, Minnesota 55455, United States

Downloaded via UNIV OF SUSSEX on July 27, 2018 at 07:03:37 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

S Supporting Information *

ABSTRACT: The formation, morphology, and pH and ionic strength responses of cationic block copolymer micelles in aqueous solutions have been examined in detail to provide insight into the future development of cationic micelles for complexation with polyanions such as DNA. Diblock polymers composed of a hydrophilic/cationic block of N,N-dimethylaminoethyl methacrylate (DMAEMA) and a hydrophobic/nonionic block of n-butyl methacrylate (BMA) were synthesized [denoted as DMAEMA-bBMA (X-Y), where X = DMAEMA molecular weight and Y = molecular weight of BMA in kDa]. Four variants were created with block molecular weights of 14−13, 14−23, 27−14, 27−29 kDa and low dispersities less than 1.10. The amphiphilic polymers self-assembled in aqueous conditions into core−shell micelles that ranged in size from 25− 80 nm. These cationic micelles were extensively characterized in terms of size and net charge in different buffers over a wide range of ionic strength (0.02−1 M) and pH (5−10) conditions. The micelle core is kinetically trapped, and the corona contracts with increasing pH and ionic strength, consistent with previous work on micelles with glassy polystyrene cores, indicating that the corona properties are independent of the dynamics of the micelle core. The contraction and extension of the corona scales with solution ionic strength and charge fraction of the amine groups. The aggregation numbers of the micelles were obtained by static light scattering, and the Rg/Rh ratios are close to that of a hard sphere. The zeta potentials of the micelles were positive up to two pH units above the corona pKa, suggesting that applications relying on micelle charge for stability should be viable over a wide range of solution conditions.

■

INTRODUCTION Amphiphilic block copolymers are being widely studied for applications ranging from drug delivery, cosmetics, and personal care formulations to dispersants, emulsifiers, water purification, and nanosensors.1−3 Through creative polymer design and meticulous synthesis, amphiphilic polymers can be tailored to achieve specific assembly characteristics. It has been shown that increasing the volume fraction of the hydrophobic block (relative to the hydrophilic block) can alter type of assembly (micelle or vesicle) or shape (spherical or cylindrical).4 In addition, the incorporation of polyelectrolyte character to the water-soluble component allows the overall charge (anionic, neutral, cationic), pKa, and charge density to be tuned,5−7 especially when coupled with a rubbery micelle core.8−12 Therefore, polymer composition can be readily adjusted to customize the assembly behavior for various applications,4,7 including stimuli responsive polymers for biological detection and delivery.13 For example, control of the size, shape, and assembly/disassembly behavior could aid the custom design of monodisperse and reproducible drug and nucleic acid delivery vehicles for a wide-range of biomedical applications.14,15 Our group and others have investigated the assembly of polycations with nucleic acids (polyanions) to form interpolyelectrolyte complexes termed polyplexes, which are known © 2016 American Chemical Society

to improve stability and cellular delivery efficiency of a variety of nucleic acid types.16−19 While packaging polynucleotides into polyplexes has advantages, their assembly behavior is complicated by many factors. For example, the lack of control in polycation over-charging, size, pH- and serum-mediated aggregation, and unpackaging, all present difficulties in reproducing defined complex formations and biological behavior. To this end, preassembly of polycations into micelles offers advantages over polyplexes in that assemblies can be reproduced and well characterized prior to DNA complexation.17,20−23 Utilizing preassembled micelles as vehicles also provides an alternative approach to carefully control, characterize, and understand the complexation of micelle-based complexes (micelleplexes) under physiological conditions. Micelleplexes also have the potential to incorporate multiple biologically active modalities such as drugs, genes, proteins, and/or imaging agents in a controlled manner.3,24,25 The polymer molecular weight, ratio of the hydrophilic to hydrophobic block lengths, and choice of monomer chemistry all impact micelle formation, shape, size, dispersity, ability to complex with other molecules, and biological activity.16 Received: May 8, 2016 Revised: July 16, 2016 Published: August 3, 2016 2849

DOI: 10.1021/acs.biomac.6b00654 Biomacromolecules 2016, 17, 2849−2859

Article

Biomacromolecules

Scheme 1. Synthetic Scheme for the Poly(DMAEMA) macroCTA and the Poly(DMAEMA-b-BMA) Block Copolymersa

a

Naming of the polymers reflects the Mn of each block as obtained by SEC and 1H NMR spectroscopy.

Table 1. Summary of Polymer Characteristics polymer DMAEMA (14) DMAEMA-BMA DMAEMA-BMA DMAEMA (27) DMAEMA-BMA DMAEMA-BMA a

(14−13) (14−23) (27−14) (27−29)

DMAEMA Mna (kDa)

BMA Mnb (kDa)

total Mnc (kDa)

mass % of DMAEMA

Đc

14 14 14 27 27 27

− 13 23 − 14 29

13 23 34 24 38 50

100 52 38 100 66 48

1.07 1.08 1.10 1.09 1.10 1.06

From SEC. bFrom 1H NMR; cFrom MALDI-TOF-MS

micelle core affect the corona behavior. Poly(DMAEMA) is a weak polyelectrolyte29 and when incorporated into a micelle acts as a polyelectrolyte brush.30−32 The poly(DMAEMA) chains are extended outward due to the osmotic pressure of counterions within the micelle corona.33,34 Above the critical micelle concentration (CMC), poly(BMA)-based amphiphilic block copolymers are well-known to assemble into micelles, driven by poly(BMA) burial into the micelle core (expelling water molecules).35 Indeed, previous studies investigating statistical copolymers of BMA and DMAEMA as well as poly(BMA-co-DMAEMA)-b-PEG polymers have shown promise as pH and temperature sensitive drug and gene delivery vehicles, tunable hydrogels, and antimicrobial agents.23,29,36−39 More recently, a self-assembled micelleplex made from poly(ethylene glycol-b-butyl acrylate-b-DMAEMA) complexed with siRNA was found to have improved gene silencing in vitro, and had increased tumor uptake relative to PEG-b-PDMAEMA polyplexes.21 These studies suggest that hydrophobic groups in polymers aid the uptake of delivery vehicles and the polymer microstructure (co- vs block-polymers) is vital to the stability and toxicity of the delivery vehicle. Herein, we detail the synthesis of four well-defined poly(DMAEMA-b-BMA) polymer variants that differ in the DMAEMA and BMA block molecular weights (Scheme 1, Table 1). In addition, we seek to understand the responsive behavior of these four micelle variants to various pH and ionic strength environments, to correlate the shape, size, and charge of the micelle corona within buffers ranging from pH 5 to 10 with ionic strengths from 20 mM to 1 M, and assess whether the rubbery core affects the observed corona behavior.

Previously, we studied poly(N,N-dimethylaminoethyl methacrylate)-block-poly(styrene) micelles in monoprotic and polyprotic buffers and their pKa in micelle, polymer, and monomer forms.26 These uniformly sized spherical micelles have a glassy poly(styrene) core, but it has been reported that cationic micelles with rubbery cores have a higher plasmid cellular delivery (transfection) efficiency than those with glassy cores.22 Related studies also show that block copolymers and copolyesters with lower glass transition temperatures are more susceptible to enzymatic degradation.27,28 Here, block copolymers with poly(N,N-dimethylaminoethyl methacrylate) (DMAEMA) as the corona block and poly(n-butyl methacrylate) (BMA) as the hydrophobic core are explored for their ability to form stable micelles in physiological conditions. Cationic micelles offer utility to condense and complex nucleic acids into micelleplexes, which offer an alternative motif for efficient transfection. Poly(BMA) has a glass transition temperature (Tg) of approximately 20 °C, and is more dynamic near room temperature than the high-Tg poly(styrene) cores previously studied.26 The longer-term goal of this work is to examine these self-assembled micelles for transporting and delivering biologically active macromolecules. The cationic rubbery micelles presented here have been created as a model system specifically tailored with the potential to achieve high gene delivery efficiency as lipid-like amphiphilic molecules are known to facilitate membrane permeability. In the current study, we sought to design and examine a system of cationic micelles with rubbery cores and cationic hydrophilic coronas and understand their fundamental behavior at differing pH and buffer strengths and to investigate whether the dynamics of the 2850

DOI: 10.1021/acs.biomac.6b00654 Biomacromolecules 2016, 17, 2849−2859

Article

Biomacromolecules

NMR measurements were performed with a temperaturecontrolled Bruker 500 MHz Avance III spectrometer (Bruker, Billerica, MA) equipped with Prodigy TCI CryoProbe operating at 500 MHz for 1H. Samples were prepared in MeOD, THF, and various combinations thereof. Spectra were recorded for each polymer at 30 °C. Block copolymer compositions were determined by comparing resonances of the DMAEMA block with those associated with the BMA block. The known Mn from SEC of the first block was then used to compare the DMAEMA/BMA ratio to calculate the final polymer molecular weights (Mn). For MALDI-TOF-MS characterization, an AB SCIEX TOF/TOF 5800 matrix-assisted laser desorption/ionization-time-of-flight mass spectrometer (Framingham, MA) equipped with a 1 kHz solid-state laser at 355 nm was used. Polymer samples were dissolved in methanol and THF and spotted on the MALDI sample plate with a 50:50 mixture of dihydroxybenzoic acid (DHB) and αcyano-4-hydroxycinnamic acid (CHCA) as the matrix. In some occasions a doping agent [either a cationic resin (DOWEX 50w X2 hydrogen form) or trifluoroacetic acid (TFA)] was added after the original spot had air-dried to enhance the resolution or assist with higher molecular weight polymers. The MALDITOF-MS was calibrated with fresh myoglobin before each experiment and the laser power was adjusted for each spot and polymer/matrix type. Micelle Formation. Dry polymer powder was dissolved in ultrapure water at a concentration of 1 mg/mL. This concentration was selected for its utility in a number of analyses including dn/dc determination, DLS, SLS, cryo-TEM, zeta potential, and for future biological assays such as complexation with DNA. The solution was vortexed for 2−3 min to aid dissolution. Other micelle preparation methods were explored, including thin film, sonication, heating, and cosolvent strategies; however, for this system direct dissolution was the most consistent way to generate micelles with low size dispersities (see Supporting Information Figure S7). The micelle stock solutions were then pipetted into dialysis bags (8000 g/mol MWCO) and dialyzed against buffers of desired ionic strengths and pH for 4 d, with media changed every 12 h. Final micelle concentrations were determined with UV−vis spectroscopy (see Supporting Information Figures S9 and S10). Buffer Preparation. Buffers of different pH and ionic strengths were prepared to test the effects of pH and ionic strength. It should be noted that, unless stated otherwise, all micelles were prepared in water first, and then dialyzed into each buffer. The concentration of the buffers was held constant at 20 mM; monoprotic buffers were chosen to yield a range of pH, and sodium chloride was added to adjust the total ionic strength. The buffers chosen were as follows: pH 5 - acetate; pH 6 - 2-(N-morpholino)ethanesulfonic acid (MES); pH 7 - 3(N-morpholino) propanesulfonic acid (MOPS); pH 7.5 - 4-(2hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES); pH 8 - tri(hydroxymethyl)aminomethane (TRIS); pH 9 - TRIS; pH 10 - N-Cyclohexyl-2-aminoethanesulfonic acid (CHES). OptiMEM was also used as a model of biological media; the measured pH of Opti-MEM was 7.2, and the ionic strength was estimated to be 140 mM according to the specifications provided by the manufacturer. Also, the role of buffer ionic strength was examined, and the ionic strengths were adjusted to 20, 50, 100, 200, 500, and 1000 mM solutions with added NaCl. Dynamic Light Scattering (DLS). Dynamic light scattering (DLS) measurements were performed on a single-angle

Understanding the impact of polymer block length and solution environment on micelle morphology and charge will inform future development of well-defined, stable, and stimuliresponsive micelleplexes with polyanions such as nucleic acids that are of interest for a myriad of applications.

■

EXPERIMENTAL SECTION Materials. All solvents were purchased from Thermo Fisher Scientific and used as received, unless mentioned otherwise. Milli-Q water or ultrapure distilled water was used for all studies, unless mentioned otherwise. Opti-MEM culture media was purchased from Life Technologies (Grand Island, NY). 2(dimethyamino) ethyl methacrylate (DMAEMA), n-butyl methacrylate (BMA), 2,2′-azobis(2-methylpropionitrile) (AIBN), 4,4′-azobis(4-cyanovaleric acid) (V-501), and 4cyano-4-(dodecylsulfanyl thiocarbonyl)sulfanyl pentanoic acid (CDT) were purchased from Sigma-Aldrich (St. Louis, MO) and were either recrystallized or distilled before use. Methods. Polymer Synthesis. Two DMAEMA blocks were first polymerized and from these polymers two different lengths of the BMA block were grown to yield four different block copolymers (Scheme 1). A representative procedure is as follows: A cationic macro-chain transfer agent (macroCTA) was synthesized by adding DMAEMA (15.0 g, 95.4 mM), V501 (13.4 mg, 0.048 mM), and CDT (193 mg, 0.478 mM) to a 100 mL round-bottom flask with a magnetic stir bar, and dissolving in 63 mL of a mixture of DMF, water, and THF (5:3:2 by volume). The solution was then acidified to pH 5 with 6 M HCl. The round-bottom flask was sealed with a septum and purged with nitrogen gas for 1 h before being heated to 70 °C in an oil bath and stirred for 3 h. The reaction was stopped by exposure to the atmosphere and the product was purified via dialysis in a 3500 g/mol MWCO (molecular weight cut off) regenerated cellulose membrane against ultrahigh purity deionized water for 4 d and then lyophilized. The dried polymer was characterized by size exclusion chromatography (SEC), 1H NMR spectroscopy, and MALDITOF-MS (see Supporting Information Figures S1−S6). For the second block, BMA (2.39 g, 16.5 mM), macroCTA (1.19 g, 0.0828 mM), and AIBN (1.36 mg, 0.00828 mM) were added to a 50 mL round-bottom flask and dissolved in 28 mL of a mixture of DMF, THF, and water (20:5:3 by volume). The round-bottom flask was equipped with a stir bar, sealed with a rubber septum, and purged with N2 gas for 1 h before being heated in a 60 °C oil bath for various amounts of time. The resulting polymer solution was purified via dialysis against ultrahigh purity deionized water with a 8000 g/mol MWCO membrane for 4 d and lyophilized to dry polymer powder and then characterized via NMR and MALDI-TOF-MS (see Supporting Information Figures S2, S3, and S6). Polymer Characterization. SEC (Agilent, Santa Clara CA) was used to determine the molecular weight [number average (Mn) and weight average (Mw)] and dispersity (Đ) for the macroCTAs using a 1.0 wt % acetic acid/0.1 M Na2SO4 aqueous eluent. A flow rate of 0.4 mL/min, Eprogen (Downers Grove, IL) columns [CATSEC1000 (7 μm, 50 × 4.6), CATSEC100 (5 μm, 250 × 4.6), CATSEC300 (5 μm, 250 × 4.6), and CATSEC1000 (7 μm, 250 × 4.6)], a Wyatt HELEOS II light scattering detector (λ = 662 nm), and an Optilab rEX refractometer (λ = 658 nm, Wyatt Technologies, Santa Barbara, CA) were used for the analysis. Astra V (version 5.3.4.18, Wyatt Technologies, Santa Barbara, CA) was utilized for the determination of Mn, Đ, and dn/dc of the polymers. 1H 2851

DOI: 10.1021/acs.biomac.6b00654 Biomacromolecules 2016, 17, 2849−2859

Article

Biomacromolecules

block lengths and low dispersity. The DMAEMA block was synthesized first followed by chain extension with BMA such that the trithiocarbonate and its aliphatic dodecyl end group would be buried in the hydrophobic core of the micelles. It should be noted that after chain extension the four amphiphilic diblock polymers were not able to be characterized via SECSLS due to their low critical micelle concentration (CMC); all polymers formed at least some micelles in all mobile phases tested. This was particularly noticeable in the 1H NMR spectra of the diblock polymers taken in deuterated methanol at 50 °C. Under these conditions, the BMA moiety was undetectable (Figure S1). However, upon addition of deuterated tetrahydrofuran (2:1 MeOD:THF vol) the BMA block could be observed (Figures S2 and S3). Broadening of the BMA resonances was a result of the micelle cores behaving as one large particle sequestered from the solvent, as has previously been reported with poly[DMAEMA-b-(BMA-co-DMAEMA-coPAA)] micelles.37 In aqueous conditions, the BMA blocks in the micelle core are kinetically trapped due to the strong amphiphilic nature of the polymer43 and the long core block,44 and thus the overall core size and aggregation number is fixed.1 The NMR integration of the peaks from the DMAEMA and the BMA blocks allowed calculation of the final polymer composition as seen in Table 1. MALDI-TOF-MS was used to determine the dispersity of the four diblocks, all of which were less than 1.10 (Table 1). The MALDI results closely resemble those from NMR but systematically underestimate Mn by as much as 10%. This can be attributed to the bias toward volatilizing lower molecular weight fractions in MALDI (details available in the Supporting Information).45−49 Micelle Formation and Characterization. The assembly method used to promote micelle formation with strongly segregated core blocks can greatly affect their final size and morphology. Due to the ease of formation (and resulting low micelle size dispersity) with the direct dissolution technique, this method was used for preparation of all micelle formulations in this study. The sizes of the micelles formed by direct dissolution in pure water, determined by DLS, are displayed in Figure 1. It was observed that, as the total polymer molecular weight increases, the mean hydrodynamic radius (Rh) increases, while the size dispersities (normalized second cumulant) of the micelles remained very low (see Figure 1). CryoTEM was used to visualize the shape and size of each micelle type (Figure 2). This technique allowed visualization primarily of the core, due to its higher contrast. All the micelles adopted a spherical morphology, with the exception of DMAEMA-BMA (27−29), where about 10% of the micelles appear to exhibit a worm-like morphology. The presence of these wormlike micelles presumably reflects the increasing relative size of the hydrophobic block; whether they represent the equilibrium morphology is not yet clear. This small population of worm-like micelles explains why the size dispersity (as measured with DLS) is broader than the other samples (Figure 1). Here it should be noted that in some images, a small fraction of worms extend up to 500 nm in length. After this sample was sonicated, the size dispersity decreased (Supporting Information Figure S12) from 0.10 to 0.05 (1 min sonication), and 0.04 (10 min sonication), indicating the disruption of the wormlike micelles and formation of spherical micelle samples with low dispersity (Figure 2e). The core sizes of the spherical (Rcore) micelles increased with the length of the hydrophobic BMA block. We attempted to visualize the corona in the cryo-TEM image by

Malvern Zetasizer Nano ZS (λ = 637 nm) (Worcestershire, UK) or a multiangle Brookhaven Instruments BI-200SM light scattering system (λ = 617 nm). The average size of each micelle was determined at 23 or 25 °C with dynamic light scattering. Multiple measurements were taken at either five angles (60°, 75°, 90°, 105°, 120°) or at 173°, for the multiangle and single-angle instruments, respectively. Measurements are typically repeatable to within 2%; however, due to various systematic errors we conservatively estimate 5% uncertainty for the DLS radii. Static Light Scattering (SLS). For all samples used in SLS, micelle solutions were prepared by dialyzing 1 mg/mL micelle solutions in Milli-Q water against pH 7, 100 mM MOPS buffer for 4 d. The concentration of dialyzed micelles was determined with UV−vis spectroscopy [Spectronic Genesys 5 spectrophotometer (Houston, Texas)] over a wavelength range of 200− 600 nm using quartz cuvettes of 10 mm path length, and dilutions were then performed to obtain desired concentrations. The dn/dc value of each micelle solution was measured using the refractometer described for SEC analysis. For each micelle solution, at least four different concentrations were used to measure scattering intensity at 10 angles (between 45° and 135°) for each concentration. Berry plots were used to calculate the molar mass Mw, radius of gyration Rg, and second virial coefficient A2 for each micelle. Cryogenic Transmission Electron Microscopy (Cryo-TEM). The morphologies of the poly(DMAEMA-BMA) micelles in water were visualized by cryo-TEM with a FEI Tecnai G2 Spirit BioTWIN (Hillsboro, Oregon) equipped with an Eagle 4 megapixel CCD camera. Image analysis was performed using ImageJ. For each specimen, 3−4 μL of micelle solution was loaded onto a carbon-coated and lacey film-supported copper TEM grid in the climate chamber of a FEI Vitrobot Mark III vitrification robot. The climate chamber was kept at 26 °C with saturated water vapor. The loaded grid was blotted and then plunged into liquid ethane that was cooled by liquid N2. Vitrified samples were kept under liquid N2 before being imaged. Images were taken at an under focus for adequate phase contrast. Zeta Potential. The dialyzed micelles were put in a folded capillary cell equipped with gold plated electrodes, and the mobility was measured with a Malvern Zetasizer Nano ZS (Malvern Instrument Ltd., Worcestershire, United Kingdom). The zeta potential was calculated from the measured electrophoretic mobility using the Smoluchowski equation. All measurements were taken at 25 °C, and the viscosity was adjusted according to the amount of NaCl added to the buffered solution.

■

RESULTS AND DISCUSSION Polymer Synthesis. The amphiphilic block copolymers were created with two monomers, DMAEMA and BMA, both having relatively low glass transition temperatures (Tg ≈ 20 °C), promoting facile self-assembly40 and providing a rubbery core and a shell whose charge density can be tuned by changing the solution pH and ionic strength. While previous reports have indicated that acrylate monomers may be lower in toxicity than methacrylates,41,42 we chose methacrylates as they are much less susceptible to hydrolysis in aqueous media, thus aiding in characterization of these systems under a variety of pH and ionic strength conditions over long periods of time.4 Reversible addition−fragmentation chain transfer (RAFT) polymerization was used to synthesize the desired polymers with controlled 2852

DOI: 10.1021/acs.biomac.6b00654 Biomacromolecules 2016, 17, 2849−2859

Article

Biomacromolecules

close to 100 individual micelles, of which the core sizes of the stained samples were measured to be smaller than those for the unstained samples. This might be due to the fact that, in the unstained samples, part of the corona was visualized as a micelle core and therefore the core sizes obtained from the unstained samples were overestimated (Figure 3).

Figure 1. Hydrodynamic radius and distribution of micelles formed by direct dissolution in water at 1 mg/mL, by dynamic light scattering. Blue circles are DMAEMA-BMA (14−13) 27 nm − 0.03 μ2/Γ2, red diamonds DMAEMA-BMA (14−23) 37 nm − 0.04 μ2/Γ2, green triangles are DMAEMA-BMA (27−14) 42 nm − 0.05 μ2/Γ2, and purple squares are DMAEMA-BMA (27−29) 79 nm − 0.10 μ2/Γ2. Figure 3. A schematic illustration of the micelle structure with an extended corona (low pH) and a contracted corona (high pH), depicting the change in Rh, corona density, and the corona area per chain.

The micelles were examined via static light scattering at pH 7 and 100 mM ionic strength to estimate their molar mass (Mw), radius of gyration Rg, and aggregation number (Ng). The raw data are shown in the Supporting Information (Figures S14− S17), and the results are collected in Table 2. One important goal is to calculate the number of chains in each micelle and therefore the number of nitrogen atoms (molar amine content), which is necessary to calculate number binding ratios for future complexation studies with polyanions such as DNA. From SLS, the radius of gyration can also be calculated. The Rg/Rh values ascertained by SLS are close to that of a hard sphere (0.775 ± 0.15). These values are likely within the combined accuracy of the Rg and Rh values (estimated to be 20%). The error reflected in Table 2 defines the precision of the measurement and therefore our values are not statistically significantly different from 0.775. It should be noted that it has previously been shown that some core−shell micelles that have Rg/Rh values ≤ 0.775 are designated as hard spheres (further discussion is available in the Supporting Information).50−52 From the combination of cryo-TEM data and SLS aggregation values (Ng), we calculated the areal density of chains at the core−corona interface and the core density (Table 3). As expected, the aggregation number increases with the length of the hydrophobic block and decreases with the length of the hydrophilic block,53,54 and the packing density is independent of polymer length and comparable for all four micelles. The density of the micelle core, calculated using micelle radii measured from unstained cryo-TEM images, is much lower than the reported density of poly(BMA) at 1.07 g/ cm3, thus indicating that part of the corona is being visualized in the image.55 This discrepancy has been reported previously with core−corona micelles containing DMAEMA in weakly buffered solutions, and was attributed to the low contrast difference between the core and corona.56 In Table 3 we have

Figure 2. Cryo-TEM images of micelles formed in deionized water: (a) DMAEMA-BMA (14−13), (b) DMAEMA-BMA (14−23), (c) DMAEMA-BMA (27−14), (d,e) DMAEMA-BMA (27−29). Top section is unsonicated (d) while the bottom half (e) was sonicated for 5 min. Scale bar represents 200 nm. With the use of imaging software, the core radius of the micelles was estimated to be 12 ± 2 nm, 20 ± 2 nm and 12 ± 2 nm, for the micelles formed with the DMAEMA-bBMA (14−13), (14−23), and (27−14), respectively, and for only the spherical (27−29) micelles the size was 25 ± 2 nm (see Supporting Information Figure S11).

increasing its contrast with the staining reagent, sodium phosphotungstate hydrate, (1 wt % in final solution). This caused the micelles to aggregate but allowed much greater resolution of the corona and core (Supporting Information Figure S13). From some of the cryo-TEM images we identified 2853

DOI: 10.1021/acs.biomac.6b00654 Biomacromolecules 2016, 17, 2849−2859

Article

Biomacromolecules Table 2. Micelle Characteristics from SLSa polymer

Mw (×106 Da)

(14−13) (14−23) (27−14) (27−29)

4.4 ± 0.5 8.7 ± 1.0 3.8 ± 0.4 47 ± 5

A2 (×10−5 cm3 mol/g2) −7.1 −6.6 −6 −0.9

± ± ± ±

0.5 0.8 2 0.4

Ng 161 ± 19 236 ± 27 92 ± 11 843 ± 97

Rg (nm) 15 27 28 43

± ± ± ±

4 1 1 2

Rg/Rh 0.62 0.83 0.84 0.75

± ± ± ±

0.03 0.04 0.04 0.03

Micelle molar mass and second virial coefficient obtained by SLS in pH 7, 100 mM buffer. The dn/dc value for all four micelles is 0.18 ± 0.02 mL/ mg. The Ng is the corresponding micellar aggregation number calculated from micelle molar mass. a

physiological conditions. These results are more consistent with star-shaped poly(DMAEMA); however, the LCST can vary with molecular weight, degree of branching, pH, ionic strength, and morphology, and is therefore system dependent. Micelle Behavior In Buffers. As DMAEMA is a weak polyelectrolyte, and thus the fraction of charged monomers in the corona is strongly affected by the local proton and ion concentrations (but the core size should not be affected), pH and ionic strength can significantly affect the size of the micelle corona.33 The micelles were prepared via direct dissolution in ultra pure water, then dialyzed, filtered, and the size (Rh) and zeta potential (ζ) of the micelles were measured. As seen in Figure 4, the micelle size decreases with increasing pH. When

Table 3. Micelle Core Dimensions from Cryo-TEM and SLS polymer DMAEMABMA (kDa)

Rcorea (nm)

Rcoreb (nm)

core densitya (g/cm3)

packing areab (nm2/chain)

(14−13) (14−23) (27−14) (27−29)

12 20 12 25

9 13 8 17

0.507 0.307 0.204 0.253

7 8 9 10

a

Determined by cryo-TEM. bDetermined from SLS assuming the core density to be 1.07 g/cm3.

calculated the corrected radius of the core from micelle molar mass using eq 1 [assuming the literature density (ρ) of poly(BMA)]. R core =

3NgM w 3

4πρNAv

(1)

By comparison, the Rcore values obtained from SLS are slightly lower than those seen in unstained cryo-TEM, and the scaling of the core radius with the core and corona block lengths is consistent with previous reports57 for spherical micelles. This yields an average area per polymer chain at the core/corona interface of 8 ± 2 nm2/chain for all four micelle systems (Table 3). The ratio of the hydrophilic and hydrophobic block lengths has a direct impact on the size, shape, and stability of the micelles.15 To assess the stability of the micelles in various buffers, we measured micelle size as a function of time. Over the period of one month the micelles did not change size, indicating that these cationic polymer micelles are stable in water and buffers at 1 g/L when formed by direct dissolution (Supporting Information Figure S18). To evaluate micelle stability at higher pH, the diblock polymers were dialyzed in pH 9, 100 mM buffer at 1 g/L. The micelles were also consistently the same size over time (Supporting Information Figure S18). Even at high pH, close to the isoelectric point, and in a range where hydrolysis can occur, the sizes remained constant. Polymer integrity was also examined by forming micelles in D2O with sodium deuteroxide (40 wt %); no hydrolysis was observed by 1H NMR over at least 3 days. Here it should be noted that poly(DMAEMA) does have a reported lower critical solution temperatures (LCST) of 38 °C at pH 9, 45 °C at pH 7, and 69 °C at pH 4,58 whereas starshaped poly(DMAEMA) has a reported LCST of 77 °C at pH 7.59 Previously, Gil et al. noted that polymers with similar structures to those studied here might be used as temperature responsive micelles for gene delivery.36 While LCSTs can be useful for temperature responsive gene delivery vehicles,31 insolubility at elevated temperatures can also pose limitations. To test this, we conducted turbidity measurements over a range from 23 to 60 °C. Over this range no cloud points were observed under the conditions tested with the micelles in water at 1 g/L, thus these micelles are thermally stable including at

Figure 4. Size (Rh) of the four micelles in different buffers (pH) prepared at an ionic strength of 100 mM. Samples were 1 g/L. Blue circles are DMAEMA-b-BMA (14−13), red diamonds are DMAEMAb-BMA (14−23), green triangles are DMAEMA-b-BMA (27−14), and purple squares are DMAEMA-b-BMA (27−29). The pKa of the micelles in 100 mM buffer is 7.2. Error bars are the standard deviation of all the data collected, a minimum of three replicates.

pH ≪ pKa, the DMAEMA moieties are fully protonated, and the condensation of small counterions in the polymer brush leads to osmotic swelling and extension of the corona chains. As the pH of the solution increases, the tertiary amines are deprotonated and the osmotic pressure due to counterions in the corona decreases, thus allowing the DMAEMA chains to contract. This result is consistent with our previous work on poly(DMAEMA-b-PS) micelles, indicating that the rubbery nature of the BMA core in the present system does not significantly affect the corona physics.26 Similar results have been found with homopolymers of poly(DMAEMA), although the mechanism for chain collapse is somewhat different; at lower pH values the polymer is highly water soluble, while, at high pH, deprotonation occurs, and the DMAEMA methyl groups have a hydrophobic character and promote a compact 2854

DOI: 10.1021/acs.biomac.6b00654 Biomacromolecules 2016, 17, 2849−2859

Article

Biomacromolecules Table 4. Summary of the Micelle Corona Parameters from SLS and DLS measurement

buffer

(14−13)

(14−23)

(27−14)

(27−29)

units

conc. of DMAEMA in corona

water pH 5 pH 10 pH 5 pH 10 pH 5 pH 10/pH 5

27 25 19 410 1040 46 56

37 35 28 240 500 57 48

42 34 23 145 480 177 69

79 57 41 130 350 119 63

nm nm nm mM mM nm2/chain %

Rh

surface area/chain volume change in coronaa a

Change in corona volume was determined by Δ[4/3π(Rh3 − Rcore3)].

Figure 5. (a) Corona length of polymer DMAEMA-b-BMA (27−14) as a function of buffer ionic strength. Black squares represent samples at pH 5, dark green triangles are micelles at pKa, and light green circles are micelles at pKa + 1.5 (3% protonated). The error bars represent a very modest estimation of 5% error in DLS measurements. Hydrodynamic radii were measured at five different angles for good accuracy. (b) The corona length of the four micelle formulations at their pKa (50% protonated) at different ionic strengths ranging from 20 mM to 1 M. Blue circles are DMAEMA-bBMA polymers (14−13), red diamonds (14−23), green triangles (27−14), and purple squares (27−29). Error bars are the standard deviation of three sample replicates. The corona length (Lcorona = Rh − Rcore) scales as csn, where the scaling factor (cs) is the total ionic strength.

conformation.36,60 While DMAEMA hydrophobicity affects the degree of swelling, it is unlikely to be the primary driving force for the corona contraction in the micelle system. Lastly, it can be seen in Figure 4 that the chain contraction plateaus at its minimum value approximately 1 pH unit above the pKa, when the protonation of the corona chains drops below 10%. Knowing the aggregation number (Ng), Mw, and Rh, we calculated the average monomer concentration in the corona at pH = 5, 100 mM ionic strength and pH = 10, 100 mM ionic strength (see Table 4). The collapse of the corona is expected when the ionic strength in the osmotic brush approaches that of the bulk solution, which in this system occurs about 0.3 units above the pKa. This transition is higher than expected for some of our osmotic brushes, which could arise from the variation of the monomer concentration throughout the corona. The corona size increases with degree of protonation from 0 to 33% protonation (pKa + 0.3) in 100 mM salted buffer (Supporting Information Figure S21); however, above 33% protonation, the size plateaus as the protonation further increases. This behavior is qualitatively consistent with the predicted α1/2 scaling for annealed polyelectrolyte stars (where α is the charge density of the star arms or corona chains),34 though quantitative comparison is difficult since the DMAEMA-BMA micelles have a nonzero core radius and excluded volume interactions in the corona may play a significant role. It is also interesting to note that the two micelles of identical DMAEMA block length [DMAEMA-BMA

(27-14) and DMAEMA-BMA (27-29)] do not have the same change in corona size between pH 5 and 10. The micelle with the larger core block DMAEMA-BMA (27−29) appears to be able to contract its corona more due to the lower initial concentration of monomers in the corona. To examine the effect of ionic strength on corona size, the DMAEMA-b-BMA (27−14) micelles were dialyzed into buffers at pH = 5, pH = pKa,26 and pH = pKa + 1.5 with different ionic strengths ranging from 20−1000 mM (see Figure 5a and Supporting Information Figures S22 and S23). For these experiments, the concentration of the buffer salt was 20 mM and the total ionic strength was adjusted by the addition of NaCl. It should be noted that the pKa of poly(DMAEMA) increases with increasing ionic strength (from approximately 6.5 at 10 mM ionic strength to 7.9 at 1 M ionic strength), but is always lower than the that of the DMAEMA monomer (which has a pKa of approximately 8.3).26,60−63 The corona extension in polyelectrolyte micelles is primarily governed by the osmotic pressure induced by counterions condensed within the brush-like corona.64 As shown in Figure 5a, the decrease in corona length with ionic strength is welldescribed by a power-law relationship, as is predicted by scaling theory and has been observed in other polyelectrolyte brush and micelle systems.26,34,61,65,66 The scaling of the corona length (Lcorona = Rh − Rcore) is cs−0.10±0.01. The slopes labeled on the log−log plots are the scaling exponent n.67 The fitted scaling exponent of −0.10 is lower than that predicted by 2855

DOI: 10.1021/acs.biomac.6b00654 Biomacromolecules 2016, 17, 2849−2859

Article

Biomacromolecules theory (−1/5 for star-like micelles34 and −1/3 for planar brushes67), but is in good agreement with our previously reported data with poly(DMAEMA-b-PS) micelles, in which the corona extension scaled as cs−0.11. The weaker-thanexpected scaling with salt concentration has been reported previously, and has been attributed to excluded volume effects and the corona not being fully in the salt dominated regime.9,26,66,68,69 This quantitative agreement with the poly(DMAEMA-b-PS) case confirms the previous qualitative conclusion that the corona behavior does not depend on the nature of the micelle core. Interestingly, in Figure 5a the scaling of the corona in DMAEMA-b-BMA (27−14) is almost identical regardless of protonation state. At pH 5 the DMAEMA is almost fully protonated, the micelles are highly charged, and an increase in salt decreases the osmotic pressure difference between the corona and the bulk, leading to contraction of the cationic corona at higher ionic strength. Micelles at the pKa (50% charged) have fewer condensed counterions than those at pH 5, and while the scaling of the corona is almost identical, the change in corona length is less pronounced (11 vs 9 nm). The micelles dialyzed at pKa + 1.5 are almost completely neutralized, however they still exhibit a response to changes in the bulk ionic strength, with a change in corona length of 7 nm from 20 mM to 1 M. At low degrees of ionization, the micelle conformation is controlled by a balance between the short-range chain−chain repulsion and the conformational entropy of stretched out chains.33 Even at high pH and salt concentrations, these micelles still have enough charge to remain colloidally stable in solution, as no aggregation or precipitation was noted. In Figure 5b, the four polymer micelle types were dialyzed into buffers at their individual pKa and the ionic strength was systematically increased. The pKa of the micelles varies from 6.7 (20 mM), 7.0 (50 mM), 7.2 (100 mM), 7.4 (200 mM), 7.68 (500 mM), to 7.88 (1 M) depending on ionic strength.26 A similar trend to the previous data was observed, that micelle size decreased with increasing ionic strength. The DMAEMA-bBMA (14−23) and (27−14) polymers have similar length, and the micelle sizes are almost identical when the chains are extended (at 20 mM) (see Supporting Information Figure S23b). The DMAEMA-b-BMA (27−14) and (27−29) polymers showed interesting behavior. In Figure 5b there is no change in the scaling of the corona vs bulk ionic strength (cs−0.08±0.01 and cs−0.09±0.01, respectively); however the change in corona length (Lcorona) from 20 mM to 1 M for (27−14) was 6 nm whereas (27−29) had a change in length of 12 nm. Considering these systems have the same DMAEMA block length in the micelle, the larger size decrease for the (27−29) system is likely due to the fact that a larger micelle core has a lower concentration of DMAEMA in the corona, which allows for a higher compaction of polymer chains. The zeta potentials of the micelles were characterized at these different pH and ionic strengths (using monoprotic buffers and monovalent salts) to understand how zeta potential may affect future complexation and stability experiments. Here it should be noted that the zeta potential is the potential difference at the slip plane and is therefore essentially independent of micelle molecular weight or aggregation number, due to the fact that the zeta potential reflects only the effective surface charge on the micelles rather than the total charge. Our results were calculated from the measured electrophoretic mobility using Henry’s equation and the

Smoluchowski approximation.70 The environment of the sample (pH, ionic strength, additives, concentration, etc.) greatly affects the zeta potential.71,72 The point at which the zeta potential is zero (isoelectric point) is significant, as the micelles are the least stable against colloidal aggregation.70,71,73 According to our results (Figure 6), the zeta potential decreases

Figure 6. Measured zeta potential of the four polymer micelles at 100 mM ionic strength from pH 5 to pH 10. Error bars represent the standard deviation of triplicate measurements. Blue circles are polymers DMAEMA-b-BMA (14−23), red diamonds (14−23), green triangles (27−14), and purple squares (27−29). The error bars are the standard deviation from all the data collected, a minimum of three sample replicates.

with an increase in pH and was positive at all pH values except at pH 10. The isoelectric point falls between pH 8 and 10 at 100 mM ionic strength, close to the point where poly(DMAEMA) is almost neutral (