Article pubs.acs.org/Langmuir
Complex Coacervate Core Micelles for the Dispersion and Stabilization of Organophosphate Hydrolase in Organic Solvents Carolyn E. Mills, Allie Obermeyer, Xuehui Dong, Jeremy Walker, and Bradley D. Olsen* Department of Chemical Engineering, Massachusetts Institute of Technology, 77 Massachusetts Avenue, Cambridge, Massachusetts02139, United States S Supporting Information *
ABSTRACT: Organophosphate (OP) nerve agents are a class of chemical warfare agents (CWAs) that exist as bulk stocks in current and past war zones. Thus, a technology that can perform on-site decontamination in a safe and timely fashion is desirable. Here, complex coacervate core micelles (C3Ms) were used to encapsulate organophosphate hydrolase (OPH) and chemostabilize it to maintain activity after exposure to organophosphate simulants ethanol and dimethyl methylphosphonate (DMMP). C3Ms were formed by two polymerspoly(acrylic acid) (PAA) and poly(oligo(ethylene glycol) methacrylate)-b-poly(4-vinyl N-methylpyridyl iodide), (POEGMA-b-qP4VP). Complexes of the coacervate micelles with the enzyme OPH were investigated by small angle neutron scattering (SANS), dynamic light scattering (DLS), and transmission electron microscopy (TEM), demonstrating the formation of micellar structures in solution. The activity of OPH against methyl paraoxon in these C3Ms under aqueous conditions was assayed after heat treatment for 3 days at 37 °C. The OPH in C3Ms retained 88 ± 7% of its initial activity, as compared to the 48 ± 3% activity retained by OPH alone, indicating that the C3Ms were able to stabilize the enzyme to heat treatment. C3Ms transferred into the two organic solvents formed larger structures than inverse micelles formed by the block copolymer alone. The addition of OPH to the C3Ms in organic solvents did not significantly change their structure. The activity of OPH (again, against methyl paraoxon) after 24 h of incubation at 4 °C was measured and compared to that of OPH in C3Ms. While OPH alone retained less than 5% of its activity after this incubation in both solvents, OPH in C3Ms retained 35 ± 3% of its activity in DMMP and 26 ± 1% of its activity in ethanol.
■
INTRODUCTION Many organophosphate (OP) compounds are highly neurotoxic chemical warfare agents.1 Bulk stocks of these OP compounds are extremely dangerous, and current Center for Disease Control (CDC) protocols require off-site shipment to destroy the compounds. This is costly, hazardous,2 and may be impractical in many situations where these OP compounds are found. Therefore, there is great interest in developing a strategy for on-site decontamination of these bulk stocks. Organophosphate hydrolase (OPH), also known as phosphotriesterase (PTE), is an enzyme isolated from bacteria that catalyzes the hydrolysis of the phosphate ester bond, detoxifying a broad range of OP compounds,3−8 including tabun (GA), sarin (GB), soman (GD), cyclosarin (GF), and VX. Previous work with OPH has focused on mutating OPH to improve its activity against different substrates,9−13 conjugating OPH to polymers and other molecules to enhance its stability and immobilize it, or encapsulating OPH in polymer films, foams, and hydrogels to formulate specific sensing and decontamination products.14−23 Despite these efforts, OPH is still limited by its stability and activity under many desired use conditions.9 These limits are common to many enzymes, where the ability to catalyze reactions in organic media or other extreme conditions © XXXX American Chemical Society
could significantly expand the scope of biocatalytic technologies. For bulk OP decontamination, the OPH must remain stable in a relatively hydrophobic organic medium and must be dispersed, along with the water necessary for hydrolysis, in the hydrophobic agent. Inverse micelles have been successfully used to disperse OPH in hexanes; however, formation of these micelles required the addition of an inhibitory cosolvent, isopropanol.24,25 Although polar organic solvents competitively inhibit OPH activity, complexation with a positively charged polyelectrolyte can decrease these inhibitory effects and improve OPH stability in solvents such as ethanol.26,27 OPH has also been covalently immobilized in polyurethane foam using a prepolymer synthesis; this immobilization imparted improved stability to OPH in DMSO.16 The formation of complex coacervate core micelles provides a potentially attractive technology to simultaneously disperse OPH and water within an organic matrix, enabling improved bulk decontamination performance. Complex coacervation is a Received: July 26, 2016 Revised: October 24, 2016
A
DOI: 10.1021/acs.langmuir.6b02350 Langmuir XXXX, XXX, XXX−XXX
Article
Langmuir phase-separation phenomenon that occurs when a solution of two polyelectrolytes of opposite charge phase separates into polyelectrolyte-rich and polyelectrolyte-poor phases.28 This phenomenon has attracted interest for underwater adhesives,29,30 forming polyelectrolyte complex plastics,31,32 and a variety of applications in the food industry.33 More recently, complex coacervation has emerged as a tool to drive selfassembly and structuring at the nanoscale.34,35 Micelles with a complex coacervate at the core can be formed by a block copolymer with a neutral block and a charged block interacting with an oppositely charged polyelectrolyte homopolymer, block copolymer, or charged biomolecule such as DNA, RNA, or protein that carries sufficient charge to coacervate.36−38 Complex coacervate core micelles can be used for a broad variety of applications, including gene delivery,39,40 the growth of metal nanoparticles used in catalysis applications,41 and formulating MRI constrast agents.38 In the case of protein coacervation, the presence of amino acids with different charges adds an additional layer of complexity. Charge regulation of these residues can dramatically change the net charge of the protein,42 leading to the formation of complex coacervates between protein and polyelectrolyte even if the net formal charge of the protein is close to zero at a given pH.43 If a protein does not have sufficient charge to form a coacervate phase, protein can be added to coacervate micelles formed by a block copolymer and a synthetic polyelectrolyte; however, the degree of inclusion in the coacervate core is typically significantly lower when the protein is not one of the two coacervating polyelectrolyte components.44,45 Herein, complex coacervate core micelles (C3Ms) are shown to disperse and stabilize enzyme and water in organic solvents, enabling use of enzymes in organic solvents. C3Ms are formed between poly(oligo(ethylene glycol)methacrylate)-b-poly(4vinyl N-methylpyridyl iodide) (POEGMA-b-qP4VP) and poly(acrylic acid) (PAA) in aqueous buffer and are then transferred to solvents such as ethanol and dimethyl methylphosphonate (DMMP), simulants for the physical properties of organophosphate toxins. After suitable conditions for micelle formation are identified, OPH is incorporated into these coacervate micelle systems to show how they improve the stability of the enzyme in the organic medium.
■
Figure 1. (a) Structures of components in complex coacervate core micelle formulation. (b) Schematic of approach to using complex coacervate core micelles to encapsulate OPH and perform decontamination of chemical warfare agents. and 5 mM imidazole. This solution was thawed and lysed by mixing with ∼0.5 mg/mL lysozyme followed by sonication. Lysate was clarified by centrifuging at 14 000g for 40 min at 4 °C, and His6-tagged OPH was purified using cobalt affinity columns (product no. 89965, Thermo Scientific, USA) followed by fast protein liquid chromatography (FPLC) using a HiTrap Q HP column on an AKTA FPLC. The buffer exchange and enzyme concentration were achieved using spin concentration at 4000g at 4 °C using a centrifugal ultrafilter with a 10 kDa molecular weight cutoff. Polymerization and Quaternization. Block copolymers were synthesized in a two-step reversible addition−fragmentation chain transfer (RAFT) polymerization, followed by quaternization. 4-Cyano4-(phenylcarbonothioylthio)pentanoic acid (CPP) was used as a chain-transfer agent, and AIBN was used as an initiator. First, CPP (27.9 mg, 0.1 mmol), AIBN (3.3 mg, 0.02 mmol), and OEGMA (5.6 g) were dissolved in 11.2 g of 1,4-dioxane in a reaction flask equipped with a magnetic stirrer. The mixture was degassed by three freeze− pump−thaw cycles, followed by polymerization at 65 °C for a prescribed time. Polymerization was terminated by the removal of heat and exposure to ambient air. The POEGMA homopolymer was obtained as a dark-red oily sample after precipitation three times in excess hexanes. Second, POEGMA (2.4 g, Mn = 24 000 g/mol, 0.1 mmol), AIBN (3.3 mg, 0.02 mmol), and 4VP (2.4 g) were dissolved in a mixture of DMF and 1,4-dioxane (v/v = 1:1) in a reaction flask. After three freeze−pump−thaw cycles, the flask was placed in an oil bath at 70 °C for a prescribed time. The polymerization was terminated by the removal of heat and exposure to ambient air. Block copolymer POEGMA-b-P4VP was obtained by precipitation three times in excess cold ether and dried under vacuum overnight. Both the POEGMA homopolymer and POEGMA-b-P4VP were characterized using gel permeation chromatography (GPC), which gave a number-average molecular weight (Mn) of 24 000 g/mol for the POEGMA homopolymer and showed a low dispersity of the block copolymer (Figure S2). The molecular weight of the P4VP block was determined to be 12 000 g/mol by computing ratios of peak areas in 1 H NMR of the block copolymer (Figure S3a). For quaternization, POEGMA-b-P4VP (3.0 g, Mn = 36 000 g/mol) was dissolved in 10 mL of DMF. A 3−5-fold molar excess of methyl iodide was added. The mixture was stirred at room temperature overnight. Quaternized block copolymer was obtained by precipitating in excess cold ether three times. Finally, 100% quaternization of the P4VP was confirmed by 1H NMR performed before and after quaternization (Figures S3 and S4). Sample Preparation. Aqueous micelle samples were mixed on a rocker at 4 °C overnight prior to measurement. Samples containing
MATERIALS AND METHODS
Materials. Poly(acrylic acid) sodium salt, α-chymotrypsinogen, and methyl paraoxon were purchased from Sigma-Aldrich (product nos. 447013, C4879, and 46192). Poly(acrylic acid) sodium salt was characterized using gel permeation chromatography (GPC), and it was found to have a bimodal molecular weight distribution with peaks at 12.9 kDa, dispersity 1.48 and at 196.8 kDa, dispersity 2.04 (Figure S1). The following chemicals were used as received: methyl iodide (CH3I, Aldrich, >99%), 4-cyano-4-(phenylcarbonothioylthio)pentanoic acid (CPP, Aldrich, >97%), hexanes (ACS grade, VWR), N,N-dimethylformamide (DMF, Aldrich, 99%), and 1,4-dioxane (anhydrous, 99.8%). 2,2′-Azobis(isobutyronitrile) (AIBN, Aldrich, 98%) was purified by recrystallization from ethanol. Oligo(ethylene glycol) methyl ether methacrylate (OEGMA, average Mn = 300 g/mol, Aldrich) and 4-vinylpyridine (4VP, Aldrich, > 95%) were purified over basic alumina prior to use. OPH Expression and Purification. OPH was expressed and purified as described in detail elsewhere18 with minor modifications. OPH expression in LB medium with 1 mM ampicillin was induced using IPTG at OD600 ≈ 0.4 and was carried out for 18−22 h. CoCl2 (1 mM) was introduced for the last 3 h of expression. Cells were harvested by centrifugation at 4000g for 10 min and were subsequently frozen in pH 8 buffer containing 50 mM NaH2PO4, 300 mM NaCl, B
DOI: 10.1021/acs.langmuir.6b02350 Langmuir XXXX, XXX, XXX−XXX
Article
Langmuir only polymers (and no protein) were warmed to room temperature prior to measurement. Aqueous micelle samples containing protein were prepared by first mixing PAA and the protein of interest for at least 30 min on a rocker at 4 °C. Block copolymer was then added, and the sample was allowed to mix on a rocker overnight at 4 °C. Proteins and polymers not incorporated into coacervate core micelles were not removed for any of the samples. All samples containing protein were kept at 4 °C until just before measurements were performed. Stock solutions of coacervate micelles were also prepared at a total polymer concentration of 100 mg/mL (25×, to yield a final concentration of 4 mg/mL, consistent with the concentration used in the aqueous experiments) in pH 8 HEPES buffer of specified concentration and allowed to equilibrate overnight at this concentration. To disperse coacervates in organic solvent, the organic solvent was added to the concentrated aqueous solution stepwise in composition increments not exceeding 10%. Samples were vortex mixed briefly after each addition of organic solvent. The final total polymer (PAA + POEGMA-b-qP4VP) concentration in organic solvent was 4 mg/mL unless otherwise specified, and the final OPH concentration in samples containing protein was 0.55 mg/mL. After reaching the final composition, organic solvent solutions were mixed on a rocker at 4 °C overnight before measurement. Polymer-only samples were warmed to room temperature prior to measurement, and proteincontaining samples were kept at 4 °C until immediately before measurement. Turbidimetry. A Tecan Infinite M200 Pro microplate reader was used to measure the absorbance of 150 μL samples at a wavelength of 750 nm in a 96-well microplate. The total protein/polymer concentration was held constant at 4 mg/mL in all samples. Separate polymer/protein solutions were prepared at 4 mg/mL in buffer at 4 °C and were combined in the plate wells just prior to measurement. Measurements were performed within 10 min of preparation to prevent the settling of the coacervate phase. Each sample was prepared in triplicate. Protein Partitioning into the Coacervate Phase by Bicinchoninic Acid (BCA) Assay. As a commercially available model for OPH with similar charge density and molar mass, αchymotrypsinogen was purchased from Sigma-Aldrich and dissolved at 10 mg/mL in water. To eliminate any excess salt from the lyophilized powder, the protein was dialyzed into 50 mM pH 8 HEPES buffer. Two calibration curves were made with the protein (with protein concentrations ranging from 0.025 to 1.5 mg/mL) for the experimentone in 50 mM pH 8 HEPES buffer and one in 50 mM pH 8 HEPES buffer with 500 mM sodium chloride. Additional dialyzed αchymotrypsinogen was combined with a PAA solution to yield a solution containing 2 mg/mL PAA and 2 mg/mL α-chymotrypsinogen to be used in the preparation of complex coacervates. Samples were prepared at positive charge fractions ( f+) of 0.2, 0.3, 0.4, and 0.5 by mixing the 2 mg/mL PAA, 2 mg/mL αchymotrypsinogen solution with a 2 mg/mL qP4VP solution in various ratios to yield a total volume of 200 μL for each sample. To control for background due to the polymers, additional coacervate solutions (again, at f+ = 0.2, 0.3, 0.4, 0.5) were prepared without protein by mixing a 2 mg/mL PAA solution with a 2 mg/mL qP4VP solution in various ratios, again to yield a total volume of 200 μL. Finally, to measure the total protein content in each sample, solutions were prepared by mixing the same volume of 2 mg/mL PAA, 2 mg/ mL α-chymotrypsinogen solution used at each charge fraction with a volume of buffer equal to the volume of 2 mg/mL qP4VP used at each charge fraction. Samples were vortex mixed to ensure complete mixing. The two sets of samples containing coacervates were centrifuged in a benchtop microcentrifuge at 14 800 rpm for 20 min to ensure the full separation of the coacervate phase from the dilute phase. The dilute phase was extracted off of each sample, leaving just the coacervate phase. The coacervate phase was then resuspended in 200 μL of 50 mM pH 8 HEPES with 500 mM sodium chloride. A Pierce BCA protein assay (Thermo Scientific product 23225) was then run on all samples (using two calibration curves described above to account for differences between dilute and coacervate phases due to high sodium chloride concentration in the resuspended coacervate), and the
concentrations in each phase were determined by comparing to the relevant calibration curve. These concentrations were used to calculate the total quantity of protein in each phase by multiplying the concentration by the volume of each phase, as described previously.43 Partition coefficients were calculated by dividing the concentration in the coacervate phase by the concentration in the dilute phase (Table S1). Control curves were also run on the polymer alone and showed that the polymer did not contribute to the observed signal. Dynamic Light Scattering. All samples were prepared at a total polymer concentration of 4 mg/mL unless otherwise specified. All aqueous polymer and protein samples were filtered through 0.1 μm Whatman Anotop 10 syringe filters prior to mixing to prevent dust from affecting the integrity of light scattering data. All DMMP and ethanol solvents used were filtered through 0.1 μm Whatman Anotop 10 syringe filters prior to addition to aqueous samples. Dynamic light scattering (DLS) measurements of coacervate micelles in buffer and ethanol were performed on a Wyatt Möbiuζ with a 532 nm laser in a 45 μL quartz cuvette. Measurements of aqueous systems averaged 10 acquisitions of 10 s each, and measurements of ethanol systems averaged 5 acquisitions of 5 s each. Autocorrelation curves were fitted using the cumulant fitting algorithm in the Wyatt DYNAMICS 7.3.1.15 software. DLS measurements of aqueous samples with protein and all DMMP samples were performed on a Wyatt DynaPro Plate Reader with a 850 nm laser on 50 μL samples in a 384-well plate. All measurements averaged five to seven acquisitions of 5 s each. Autocorrelation curves were fitted using the cumulant fitting algorithm in Wyatt DYNAMICS 7.1.8. Small-Angle Neutron Scattering. Samples for small-angle neutron scattering (SANS) were prepared in a fashion analogous to that used to prepare DLS samples, but at a total polymer (PAA + POEGMA-b-qP4VP) concentration of 20 mg/mL. Samples containing OPH had a final protein concentration of 2.5 mg/mL. Deuterated water and deuterated ethanol (Cambridge Isotopes item numbers DLM-6-PK and DLM-31-10X1, respectively) were used in sample preparation, and samples were prepared within a week of measurement. Samples were measured at the NGB 30 m beamline at the NIST Center for Neutron Research (NCNR) at the National Institute of Standards and Technology (NIST).46 All samples were measured at 25 °C in a demountable titanium cell (25 mm diameter, 1 mm path length) that held the sample between two quartz windows sealed with Viton o-rings. Samples were measured at a wavelength of 6 Å at detector distances of 1.3, 4, and 13 m as well as at a wavelength of 8.4 Å at a detector distance of 13 m. Data was reduced using macros in Igor Pro.47 Transmission Electron Microscopy. Transmission electron microscopy (TEM) images were recorded on a FEI Tecnai G2 Spirit TWIN instrument. The accelerating voltage was 120 kV. Images were captured using an ORCA camera in a fixed bottom mount configuration. Samples were prepared by depositing a drop (ca. 3 μL) of each micelle solution onto copper grids. After 1 min, the excess solution was wicked away by a piece of filter paper. The grids were then allowed to dry under ambient conditions. Activity Assays. The specific activity of all solutions was measured using methyl paraoxon hydrolysis. Paraoxon degradation was monitored via tracking the formation of the hydrolysis product, pnitrophenol. In these assays, the p-nitrophenol concentration was monitored by measuring the absorbance at 405 nm. Serial dilutions of OPH solutions were assayed in triplicate in a Tecan plate reader over 5 min, and the highest concentration dilution that resulted in a linear absorbance versus time curve was used to calculate the specific activity. The specific activity was calculated using the following equation
⎛ μmol ⎞ slope × 106 × dilution factor specific activity⎜ ⎟= 1000 × 17 100 × PC ⎝ min ·mg ⎠ (1) where the dilution factor refers to how much the OPH solution was diluted (i.e., if a 100-fold dilution was required to obtain a linear slope, then the dilution factor would be 100), C is the concentration of OPH C
DOI: 10.1021/acs.langmuir.6b02350 Langmuir XXXX, XXX, XXX−XXX
Article
Langmuir
Figure 2. Percent intensity of micelles in aqueous solution as measured by DLS. (a) Percent intensity of micelles as a function of HEPES concentration at f+ = 0.4. The solution is completely composed of micelles up until 60 mM HEPES, at which point smaller scatterers are observed in solution. (b) Percent intensity of micelles as a function of the positive charge fraction in 50 mM pH 8 HEPES. The solution is made up completely of micelles over the range of f+ = 0.2−0.4.
using a model protein α-chymotrypsinogen in the BCA assay based on previous work43 (Table S1). These results indicate that the partition coefficient of α-chymotrypsinogen, a more widely available model for OPH, in the coacervate system ranges between 0.31 ± 0.03 and 0.9 ± 0.2 depending on the mixing ratio of the polymers. Although partitioning into the coacervate phase is relatively poor in an aqueous medium, the protein’s preference for an aqueous environment in the predominantly hydrophobic environments of interest in this article should increase protein incorporation into the core of the micelle. As predicted by turbidimetry measurements, PAA and POEGMA-b-qP4VP were found to form stable C3Ms across a range of buffer concentrations and charge fractions. Because the coacervating components in the C3M system were the two polymers PAA and POEGMA-b-qP4VP, conditions for C3M formation were first optimized using only these components, without the inclusion of protein. Figure 2a shows the percent intensity of micelles in solution at various buffer concentrations. Salt can dissolve a complex coacervate phase, preventing the formation of C3Ms; 49 however, because higher buffer concentrations stabilize the pH during OP hydrolysis, the highest possible buffer concentration that would not result in the dissolution of micelles was desired. These criteria are met for an f+ = 0.4 mixture in 50 mM HEPES buffer at pH 8. Figure 2b shows the percent intensity of micelles, as measured by DLS, in 50 mM HEPES buffer at pH 8 as a function of positive charge fraction, f+. For f+ = 0.2−0.4, the solution is composed entirely of micelles. Outside of this range, smaller structures begin to appear. On the basis of the literature,38 these smaller structures are likely a mixture of free polymer and small complexes formed between polymers with no well-defined structure, with excess positive or negative charge, depending on the location on the positive charge fraction axis. A positive charge fraction of f+ = 0.4 was selected for all future studies because this condition yields a solution composed entirely of micelles. After selecting the mixing ratio of PAA to POEGMA-bqP4VP on the basis of the experiments above, OPH was added to the system, and the formation of micelles in this threecomponent system was confirmed using DLS, TEM, and SANS. The micelles formed in the three-component system (OPH/PAA/POEGMA-b-qP4VP) were similar in all respects to those composed only of the two polymers (PAA/POEGMA-
in the original solution (mg/L), P is the path length of the light (cm), and 17 100 is the molar extinction coefficient of p-nitrophenol at 405 nm (M−1 cm−1). For assays performed after incubation in organic solvents, solutions of 4 mg/mL total polymer and ∼0.5 mg/mL OPH were incubated in 96 vol % of each organic solvent and 4 vol % 50 mM pH 8 HEPES buffer with 100 μM CoCl2 for 24 h prior to assaying the activity. Assays themselves were performed by diluting the incubated solutions 10-fold into 50 mM pH 8 HEPES buffer with 100 μM CoCl2 containing methyl paraoxon. The activity was then tracked by monitoring the absorbance at 340 nm for ethanol and at 345 nm for DMMP. These absorbances were converted to p-nitrophenol concentrations using a calibration curve prepared with p-nitrophenol in solutions made up of the same amount of buffer and organic solvent as in the assay. Cobalt(II) chloride (100 μM) was included in all OPH solutions because it is a cofactor for the enzyme’s active site and is required for enzymatic activity.
■
RESULTS AND DISCUSSION Complex Coacervate Core Micelles in Aqueous Solution. Because OPH does not coacervate with polycation qP4VP or polyanion PAA (Figures S5 and S6), complex coacervate phases were formed from two oppositely charged polyelectrolytes, qP4VP and PAA. Previous work has shown that these two polymers coacervate,48 and turbidimetry experiments confirmed that these polymers coacervate across a broad range of charge fractions ( f+ = 0.3−0.8) (Figure S5). Here, the positive charge fraction, f+, is defined as the fraction of positive charges in the system (on a molar basis) divided by the total charges in the system, where the number of charges per polymer is calculated assuming full ionization of each polymer. This yields a charge fraction of f+ =
N+ N + N− +
(2)
where N+ is the moles of positive charges and N− is the moles of negative charges. These results indicate that PAA and qP4VP coacervate strongly, and they were thus used as the coacervating components in the core of the C3Ms studied in this article. Measurement of the partitioning of the protein between the dilute and coacervate phases revealed that protein does not have a significant preference for the coacervate phase over the dilute phase. A study of protein partitioning between dilute and coacervate phases in the PAA/qP4VP system was performed D
DOI: 10.1021/acs.langmuir.6b02350 Langmuir XXXX, XXX, XXX−XXX
Article
Langmuir
OPH. Both data sets were fit to a fuzzy sphere model50,51 described by the following equation for intensity
b-qP4VP), indicating that the protein did not alter micelle formation. Table 1 shows the results of DLS measurements on Table 1. Dominant Scatterer Size of OPH Protein and Polymer Combinations in 50 mM pH 8 HEPES Buffer with 100 μM CoCl2 as Determined by DLS system
dominant scatterer size (hydrodynamic radius, nm)
OPH only polymers only ( f+ = 0.4) OPH in PAA
5.6 ± 0.6 21.0 ± 0.3
OPH in POEGMA-bqP4VP OPH in micelles ( f+ = 0.4)
5.7 ± 0.2
7.0 ± 0.9
25.5 ± 0.1
I(q) =
Ilor (Δρ)2 ϕ Pfuzzy Schulz(q) + +B ⟨Vparticle⟩ 1 + ξ 2q2
(3)
where φ is the volume fraction of the spheres, < Vparticle> is the average sphere volume, given by
classification free protein micelle
⟨Vparticle⟩ =
4π 3
∫0
∞
f (R )R3 dR =
σ 4 ⎞⎟ 4π ⎛⎜ R avg 3 + 3R avgσ 2 + 2 ⎜ 3 ⎝ R avg ⎟⎠
(4)
free protein/polymer complexes free protein/polymer complexes
Δρ is the difference in scattering length density between the solvent and the sphere, Ilor is the Lorentz scale, ξ is the fluctuation correlation length, B is the background, Ravg is the average particle radius, σ is the standard deviation of the particle size distribution, and PFuzzy Schulz(q) is the form factor, given by
micelle
PFuzzy Schulz(q) =
several different aqueous systems with OPH. OPH alone has a hydrodynamic radius of 5.6 ± 0.6 nm, consistent with the size of the OPH dimer, and the polymers alone (at charge fraction f+ = 0.4) form a micelle that has a hydrodynamic radius of 21.0 ± 0.3 nm. When OPH is added to the micelle system, the hydrodynamic radius increases to 25.5 ± 0.1 nm, which indicates a slight increase in micelle size as the OPH is included in the core of the micelle. DLS measurements on PAA/OPH and POEGMA-b-qP4VP/OPH systems confirmed that no micelle formation occurs between OPH and either of the two individual polymers. Figure 3a,b shows small-angle neutron scattering (SANS) data taken on the C3M solutions at f+ = 0.4, with and without
∫0
∞
9[sin(qR ) − qR cos(qR )]2 6
(qR )
2
e−(σsurf q) f (R ) dR
(5)
σsurf is the width of the smeared particle surface, and f(R) is the particle size distribution, here assumed to be Schulz distributed about an average particle radius: ⎛ R avg ⎞ R avg f (R ) = ⎜ 2 ⎟ ⎝ σ ⎠
2
/σ 2
1 ⎛ R avg ⎞ Γ⎜ 2 ⎟ ⎝ σ ⎠ 2
RR avg
2
/σ 2 − 1 −R avgR /σ 2
e
(6)
Data was fit using a nonlinear least-squares algorithm. Tabulated values for all fit parameters are included in Table S2.
Figure 3. SANS data for C3M systems at f+ = 0.4 (a) with POEGMA-b-qP4VP and PAA only and (b) with OPH included. Both curves are fit by the fuzzy spheres model, where R is the sphere radius, Đ is the sphere dispersity, and σsurf is the interface thickness of the fuzzy spheres. TEM images of (c) POEGMA-b-qP4VP and PAA only (f+ = 0.4) and (d) system c + OPH showing the formation of spherical micelles. E
DOI: 10.1021/acs.langmuir.6b02350 Langmuir XXXX, XXX, XXX−XXX
Article
Langmuir
Complex Coacervate Core Micelles in Organic Solvents. By applying the same polymer mixing ratio ( f+ = 0.4) as in the aqueous system, C3Ms (containing only POEGMA-b-qP4VP and PAA) were transferred into organic solvents. This work was performed in two organic solvents ethanol, which is completely miscible with water, and DMMP, which is marginally miscible with water (∼10%) and has physical properties, including density, water miscibility, molecular weight, and structure, that are similar to those of chemical warfare agents of interest for future applications. DLS was used to confirm the formation of C3Ms in each organic solvent and distinguish between the structures formed with the C3M system and the inverse micelles formed by the block copolymer only. Table 2 shows the results of DLS measure-
Although both systems have a similar interface thickness (σsurf = 4.1−4.7 nm), the mean radius, Ravg, of the C3M system without OPH (14.8 ± 0.2 nm) is slightly smaller than the radius of the C3M system with OPH (15.3 ± 0.1 nm), indicating that the protein inclusion into the core of the micelle increases the structure’s radius slightly. The sum of the interface thickness and average radius yields a radius very similar to that measured by DLS. These measured radii (∼20 nm) are reasonable considering that the block copolymer contour length (assuming a carbon−carbon bond length of 0.154 nm and a carbon− carbon bond angle of 109.5°) is 68 nm. TEM of the C3M systems confirms that the micelles are spherical (Figure 3c,d); however, the micellar size via TEM is significantly larger than the radii measured by SANS and DLS. This is likely due to sample preparation artifacts in which the micelles, which have fluid coacervate cores, flatten onto the surface of the TEM grid as water evaporates. Activity tests against methyl-paraoxon in aqueous solution show that the C3M-OPH system (composed of PAA/ POEGMA-b-qP4VP/OPH) is more stable to incubation at 37 °C as compared to free OPH in solution. The activity of OPH in and out of micelles (as well as with both polymers separately) is shown in Figure 4. Immediately after preparation
Table 2. Sizes of Micelle Scatterers in 96% Ethanol and DMMP Systems complex coacervate core micelle ( f+ = 0.4)
block copolymer only organic solvent ethanol DMMP
hydrodynamic radius (nm) 18.3 190 12 40 1
± ± ± ± ±
0.4 20 3 10 1
percent intensity
hydrodynamic radius (nm)
percent intensity
± ± ± ± ±
25 ± 1
99.9 ± 0.1
33 ± 1
100
59.2 40.6 50 10. 40
0.9 0.7 20 10. 20
ments comparing two polymer-only systemsPOEGMA-bqP4VP alone and POEGMA-b-qP4VP/PAA at f+ = 0.4 in 96 vol % organic solvent, 4 vol % 50 mM pH 8 HEPES buffer. Because the qP4VP block is highly charged, it is insoluble in both organic solvents, so the block copolymer alone forms a mixture of inverse micelles and other structures, including large aggregates, in both cases. In ethanol, DLS of the block copolymer shows the presence of only two structures with 18.3 ± 0.4 and 190 ± 20 nm hydrodynamic radii, which is likely a mixture of spherical micelles and some other larger structures. The presence of these larger structures was observed by DLS regardless of the sample preparation, so they are likely not kinetic; however, because they are quite large but make up less than half the intensity, these structures are only a small fraction of what is actually present in solution. In DMMP, DLS of the block copolymer alone shows the formation of structures with a hydrodynamic radius of 12 ± 3 nm, which were also assumed to be inverse micelles. When switching from the block copolymer only to the POEGMA-b-qP4VP/PAA ( f+ = 0.4) mixture, DLS of both DMMP and ethanol systems reveals that they form uniform micellar structures different in size from those formed by the block copolymer alone. In ethanol, the micelle radius increases to 25 ± 1 nm and larger structures no longer form. In DMMP, the radius of the micelles observed in the binary system doubles as compared to the block copolymer only, with a hydrodynamic radius of 33 ± 1 nm. This increase in radius is presumably due to the inclusion of PAA in the core of the micelle. Both micelles are much smaller than the block copolymer contour length of 68 nm. We hypothesize that this is due to a large degree of polymer overlap in the core of the micelle as well as overlap of the POEGMA corona block into the core of the micelle. SANS measurements of C3Ms in ethanol show that they have a bimodal size distribution. Figure 5 shows SANS data of the polymers only (f+ = 0.4) as well as polymers with OPH
Figure 4. Methyl paraoxon hydrolysis activity of OPH polymer blends. Activities were measured over time after incubating samples at 37 °C in buffer. All micelle preparations were carried out at f+ = 0.4. Results of ANOVA analysis on the data is available in Tables S5 and S6.
and storage at 4 °C, all solutions containing OPH have similar activities. After incubation at 37 °C for several days, the activity of OPH alone as well as OPH with PAA drops off significantly. This is consistent with previous studies that show that prolonged exposure to high temperatures precipitates and deactivates the enzyme.52 However, when the enzyme is incubated at 37 °C in the C3Ms (prepared at f+ = 0.4) or with the POEGMA-b-qP4VP alone, the initial activity is preserved even after incubation for several days. Interestingly, the presence of POEGMA-b-qP4VP both enhances and protects the activity of the OPH enzyme. This phenomenon was previously observed in OPH/Pluronic blends, and it is possible that POEGMA-b-qP4VP affects the OPH activity in a similar fashion to the Pluronic.18 Similar increases in activity with the addition of poly(ethylene oxide)-co-poly-2-methylvinylpyridinium to lipase have also been reported.53 Activity tests on resuspended lyophilized samples show that no significant loss of activity is observed in any systems due to lyophilization (Figure S8). F
DOI: 10.1021/acs.langmuir.6b02350 Langmuir XXXX, XXX, XXX−XXX
Article
Langmuir
Figure 5. SANS data of C3Ms in ethanol at f+ = 0.4 for (a) POEGMA-b-qP4VP and PAA only and (b) system a with OPH included and a fit to a bimodal fuzzy Schulz spheres model.
(again, f+ = 0.4) in ethanol. This data was fit to two different bimodal Schulz-distributed sphere form factorsone without and one with a fuzzy interface.50,51 (Fully tabulated fit parameters are available in Table S3.) For the fit with a fuzzy interface, the model for the scattering intensity is I(q) = +
(Δρ1)2 ϕ1 ⟨Vparticle, 1⟩ Ilor
1 + ξ 2q2
Pfuzzy
Schulz,1(q)
+
(Δρ2 )2 ϕ2 ⟨Vparticle,2⟩
Pfuzzy
the SANS measurements were made on samples with a total polymer concentration of 20 mg/mL, the formation of these smaller structures could also be a concentration-dependent phenomenon. A fit performed on SANS measurements of the block copolymer alone under the same solvent conditions (using the bimodal fuzzy Schulz-distributed spheres model) shows the presence of 7.2 ± 0.2 nm spheres with interfacial thickness 5.8 ± 0.2 nm (Figure S11, Table S2), confirming that the smaller species observed in the bimodal fit are the same size as structures formed by the block copolymer alone. A combination of SANS and DLS measurements on the POEGMA-b-qP4VP/PAA/OPH and POEGMA-b-qP4VP/ OPH systems shows that stable C3Ms are formed in both solvents with all three components. Table 3 shows the results of DLS measurements on organic solvent systems with OPH. In both organic solvents, OPH and OPH with PAA form mostly large aggregates, indicating that neither formulation is capable of effective dispersion in organic solvent. However, POEGMA-
Schulz,2(q)
+B
(7)
where Pfuzzy spheres and ⟨Vparticle⟩ are given by eqs 5 and 4, with different Ravg, σ, and σsurf for each subscript 1 and 2 above. For the fit without a fuzzy interface, the intensity equation is given below I(q) =
(Δρ1)2 ϕ1 ⟨Vparticle,1⟩
PSchulz,1(q) +
(Δρ2 )2 ϕ2 ⟨Vparticle,2⟩
PSchulz,2(q) +
Ilor 1 + ξ 2q2
+B
(8)
where the form factor PSchulz is given by the following equation PSchulz(q) =
∫0
∞
9[sin(qR ) − qR cos(qR )]2 (qR )6
Table 3. Dominant Scatterer Size of Various Systems with and without OPH in 96% Organic Solvent, 4% 50 mM pH 8 HEPES Buffer with 100 μM CoCl2, as Determined by DLS Measurements
f (R ) d R (9)
with a different Ravg and σ for each subscript 1 and 2. Although the model with the fuzzy interface most accurately captures all features of the observed data, the large number of parameters and covariance between parameters leads to large uncertainties in some of the key estimated parameters. The fit to bimodal Schulz spheres with no fuzzy component (Figure S12, Table S4), while unable to capture the intensity in the 0.3−0.7 nm−1 q range as accurately, yields similar predictions of radii for both species, with overall excellent fit quality. The high quality of fit for both bimodal spheres models supports the hypothesis that there exist populations of two different sizes in these samples. A fit of the polymers-only (f+ = 0.4) data to the bimodal fuzzy Schulz spheres model (Figure 5a) predicts average radii, R, of 6 ± 3 and 20.0 ± 0.2 nm for the two sphere sizes, with interface thicknesses, σsurf, of 3 ± 2 and 5 ± 5 nm, respectively. The fit to the SANS data predicts that the smaller spheres make up 30 ± 1 vol % of the objects in solution (Table S4). The 20.0 nm radius is consistent with the hydrodynamic radius given by DLS. The smaller radius could indicate the presence of smaller inverse micelles formed by the block copolymer alone. Because
organic solvent ethanol
DMMP
system OPH only
3 ± 4 μm
polymers only ( f+ = 0.4) OPH in PAA
27.5 ± 0.9 nm
OPH in POEGMA-bqP4VP
19.3 ± 0.5 nm 2 ± 2 μm
OPH in C3M ( f+ = 0.4) OPH only
25.5 ± 0.1 nm
polymers only ( f+ = 0.4) OPH in PAA
33 ± 1 nm
OPH in POEGMA-bqP4VP OPH in C3M ( f+ = 0.4) G
dominant scatterer size (hydrodynamic radius)
3 ± 4 μm
2 ± 2 μm
2.7 ± 0.8 μm 2 ± 1 μm 45 ± 2 nm
classification large aggregate micelle large aggregate micelle large aggregate micelle large aggregate micelle large aggregate large aggregate micelle
DOI: 10.1021/acs.langmuir.6b02350 Langmuir XXXX, XXX, XXX−XXX
Article
Langmuir
Figure 6. TEM images of (a) POEGMA-b-qP4VP/PAA (f+ = 0.4), and (b) POEGMA-b-qP4VP/PAA/OPH ( f+ = 0.4) in 96% DMMP, 4% pH 8 50 mM HEPES buffer.
Figure 7. Percent activity retained by OPH against methyl paraoxon in 90% aqueous solution after 24 h of incubation in organic solvents at 4 °C. The full ANOVA analysis is included in Table S7.
b-qP4VP/PAA with OPH (both prepared at f+ = 0.4) forms micelles. Here, the block copolymer is able to stabilize the other components because of the entropic repulsive interactions between the highly soluble POEGMA coronas of different micelles. SANS data of the C3M/OPH system in ethanol in Figure 5b supports the DLS data and is similar to SANS measurements on the POEGMA-b-qP4VP/PAA only system in ethanol. A fit to a bimodal fuzzy Schulz-distributed spheres model predicts average radii, R, of 8 ± 9 and 18 ± 1 nm, with interface thicknesses (σsurf) of 3 ± 3 and 3 ± 7 nm, respectively. The fit to SANS data predicts that 50 ± 6 vol % of the objects are smaller spheres (Table S4). TEM images of the C3M only and OPH/C3M systems in DMMP confirm the formation of uniformly sized micelles (Figure 6). Interestingly, the block copolymer alone with OPH formed exclusively aggregates in DMMP but formed some micelles (35 ± 3% intensity) in ethanol. The C3M system effectively stabilizes OPH against denaturation after incubation in both ethanol and DMMP. Figure 7 shows percent OPH activities retained against methylparaoxon in all systems (with polymers only as a control) after 24 h of incubation in each of the organic solvents at 4 °C. Both OPH and OPH with PAA lose most of their activity after treatment in both solvents. The block copolymer, POEGMA-bqP4VP, is able to chemostabilize the OPH against ethanol, retaining 37 ± 2% of OPH activity (as compared to OPH activity when stored in buffer), presumably because of the formation of inverse micelles observed in the block copolymer−OPH system in ethanol using DLS. Importantly, only the C3M system (at f+ = 0.4) is able to chemostabilize the OPH against both ethanol and DMMP. 35 ± 3% OPH activity is preserved in the C3M system in DMMP, and 26 ± 1% OPH activity is preserved in ethanol, which is significant (p < 0.01) compared to the 0−4% activity retained by OPH alone after
incubation in these solvents. The results of assays against bulk methyl paraoxon and in varying contents of DMMP indicate that the enzyme loses activity above ∼50 vol % DMMP or simulant (Figures S13 and S14). This indicates that whereas the enzyme is preserved by the coacervate micelles, solvent conditions prevent effective catalysis until the water content is increased. The ability of the C3M system to protect enzyme activity in the more hydrophobic DMMP is hypothesized to be a result of two factors. First, previous work on OPH stabilization has shown that complexation with a polyelectrolyte can decrease the inhibitory effects of organic solvents on OPH.27 It is possible that the complex coacervate core provides similar protection to the OPH enzyme in both ethanol and DMMP. Furthermore, the complex coacervate phase contains water and a charge-rich environment, which could provide bonds that stabilize the protein and help maintain its folded, active structure.
■
CONCLUSIONS OPH was stabilized for use in polar organic media by incorporating OPH into complex coacervate core micelles (C3Ms) formed between POEGMA-b-qP4VP and PAA. OPH stability to heat treatment at 37 °C improved upon incorporation into C3Ms. When C3Ms were transferred to organic solvents, they formed structures larger than those formed by the block copolymer alone (which are assumed to be inverse micelles). The OPH enzyme was then added to the system, and activity assays revealed that C3Ms were able to chemostabilize the enzyme against both solvents, whereas the block copolymer alone could not. DLS revealed that the POEGMA-b-qP4VP/OPH system formed micelles of radius 19.3 ± 0.5 nm only in ethanol and not in DMMP, whereas the C3M/OPH system formed micelles in both solvents (25.5 ± 0.1 nm radius in ethanol and 45 ± 2 nm radius in DMMP). H
DOI: 10.1021/acs.langmuir.6b02350 Langmuir XXXX, XXX, XXX−XXX
Article
Langmuir
(3) Raushel, F. M. Bacterial Detoxification of Organophosphate Nerve Agents. Curr. Opin. Microbiol. 2002, 5 (3), 288−295. (4) Richins, R. D.; Kaneva, I.; Mulchandani, A.; Chen, W. Biodegradation of Organophosphorus Pesticides by Surface-Expressed Organophosphorus Hydrolase. Nat. Biotechnol. 1997, 15 (10), 984− 987. (5) Russell, A. J.; Berberich, J. A.; Drevon, G. F.; Koepsel, R. R. Biomaterials for Mediation of Chemical and Biological Warfare Agents. Annu. Rev. Biomed. Eng. 2003, 5 (1), 1−27. (6) Hill, C. M.; Li, W.-S.; Thoden, J. B.; Holden, H. M.; Raushel, F. M. Enhanced Degradation of Chemical Warfare Agents through Molecular Engineering of the Phosphotriesterase Active Site. J. Am. Chem. Soc. 2003, 125 (30), 8990−8991. (7) Tsai, P.-C.; Fox, N.; Bigley, A. N.; Harvey, S. P.; Barondeau, D. P.; Raushel, F. M. Enzymes for the Homeland Defense: Optimizing Phosphotriesterase for the Hydrolysis of Organophosphate Nerve Agents. Biochemistry 2012, 51 (32), 6463−6475. (8) Singh, B. K.; Walker, A. Microbial Degradation of Organophosphorus Compounds. FEMS Microbiol. Rev. 2006, 30 (3), 428− 471. (9) Watkins, L. M.; Mahoney, H. J.; McCulloch, J. K.; Raushel, F. M. Augmented Hydrolysis of Diisopropyl Fluorophosphate in Engineered Mutants of Phosphotriesterase. J. Biol. Chem. 1997, 272 (41), 25596− 25601. (10) Cho, C. M.-H.; Mulchandani, A.; Chen, W. Bacterial Cell Surface Display of Organophosphorus Hydrolase for Selective Screening of Improved Hydrolysis of Organophosphate Nerve Agents. Appl. Environ. Microbiol. 2002, 68 (4), 2026−2030. (11) Lai, K.; Dave, K. I.; Wild, J. R. Bimetallic Binding Motifs in Organophosphorus Hydrolase Are Important for Catalysis and Structural Organization. J. Biol. Chem. 1994, 269 (24), 16579−16584. (12) Di Sioudi, B. D.; Miller, C. E.; Lai, K.; Grimsley, J. K.; Wild, J. R. Rational Design of Organophosphorus Hydrolase for Altered Substrate Specificities. Chem.-Biol. Interact. 1999, 119−120, 211−223. (13) Gopal, S.; Rastogi, V.; Ashman, W.; Mulbry, W. Mutagenesis of Organophosphorus Hydrolase to Enhance Hydrolysis of the Nerve Agent VX. Biochem. Biophys. Res. Commun. 2000, 279 (2), 516−519. (14) Lu, H. D.; Wheeldon, I. R.; Banta, S. Catalytic Biomaterials: Engineering Organophosphate Hydrolase to Form Self-Assembling Enzymatic Hydrogels. Protein Eng., Des. Sel. 2010, 23 (7), 559−566. (15) Mansee, A. H.; Chen, W.; Mulchandani, A. Detoxification of the Organophosphate Nerve Agent Coumaphos Using Organophosphorus Hydrolase Immobilized on Cellulose Materials. J. Ind. Microbiol. Biotechnol. 2005, 32 (11−12), 554−560. (16) LeJeune, K. E.; Mesiano, A. J.; Bower, S. B.; Grimsley, J. K.; Wild, J. R.; Russell, A. J. Dramatically Stabilized Phosphotriesterase polymers for Nerve Agent Degradation. Biotechnol. Bioeng. 1997, 54 (2), 105−114. (17) Pedrosa, V. A.; Paliwal, S.; Balasubramanian, S.; Nepal, D.; Davis, V.; Wild, J.; Ramanculov, E.; Simonian, A. Enhanced Stability of Enzyme Organophosphate Hydrolase Interfaced on the Carbon Nanotubes. Colloids Surf., B 2010, 77 (1), 69−74. (18) Kim, M.; Gkikas, M.; Huang, A.; Kang, J. W.; Suthiwangcharoen, N.; Nagarajan, R.; Olsen, B. D. Enhanced Activity and Stability of Organophosphorus Hydrolase via Interaction with an Amphiphilic Polymer. Chem. Commun. 2014, 50 (40), 5345−5348. (19) Frančič, N.; Košak, A.; Lyagin, I.; Efremenko, E. N.; Lobnik, A. His6-OPH Enzyme-Based Bio-Hybrid Material for Organophosphate Detection. Anal. Bioanal. Chem. 2011, 401 (8), 2631−2638. (20) LeJeune, K. E.; Wild, J. R.; Russell, A. J. Nerve Agents Degraded by Enzymatic Foams. Nature 1998, 395 (6697), 27−28. (21) Dennis, P. B.; Walker, A. Y.; Dickerson, M. B.; Kaplan, D. L.; Naik, R. R. Stabilization of Organophosphorus Hydrolase by Entrapment in Silk Fibroin: Formation of a Robust Enzymatic Material Suitable for Surface Coatings. Biomacromolecules 2012, 13 (7), 2037−2045. (22) Shimazu, M.; Mulchandani, A.; Chen, W. Thermally Triggered Purification and Immobilization of elastin−OPH Fusions. Biotechnol. Bioeng. 2003, 81 (1), 74−79.
Activity tests against methyl paraoxon were run on OPH samples that had been incubated at 4 °C with the two organic solvents for 24 h, and these tests agreed with what was observed structurally: OPH with POEGMA-b-qP4VP retained activity only in ethanol (37 ± 2%, as compared to OPH incubated at 4 °C in buffer) and not DMMP; however, the OPH in C3Ms retained activity in both organic solvents: 35 ± 3% in DMMP and 26 ± 1% in ethanol. Importantly, OPH alone as well as OPH mixed with PAA homopolymer retained
