Letter pubs.acs.org/NanoLett
Aiding Nature’s Organelles: Artificial Peroxisomes Play Their Role Pascal Tanner, Vimalkumar Balasubramanian, and Cornelia G. Palivan* Department of Chemistry, University of Basel, Klingelbergstrasse 80, CH-4056 Basel, Switzerland S Supporting Information *
ABSTRACT: A major goal in medical research is to develop artificial organelles that can implant in cells to treat pathological conditions or to support the design of artificial cells. Several attempts have been made to encapsulate or entrap enzymes, proteins, or mimics in polymer compartments, but only few of these nanoreactors were active in cells, and none was proven to mimic a specific natural organelle. Here, we show the necessary steps for the development of an artificial organelle mimicking a natural organelle, the peroxisome. The system, based on two enzymes that work in tandem in polymer vesicles, with a membrane rendered permeable by inserted channel proteins was optimized in terms of natural peroxisome properties and function. The uptake, absence of toxicity, and in situ activity in cells exposed to oxidative stress demonstrated that the artificial peroxisomes detoxify superoxide radicals and H2O2 after endosomal escape. Our artificial peroxisome combats oxidative stress in cells, a factor in various pathologies (e.g., arthritis, Parkinson’s, cancer, AIDS), and offers a versatile strategy to develop other “cell implants” for cell dysfunction. KEYWORDS: Polymer nanoreactor, artificial organelle, antioxidant enzymes, reactive oxygen species, peroxisome
C
synthetic membranes10 or by the insertion of channel proteins12 depending on the chemical nature of the copolymer. Even though a variety of nanoreactors that support specific reactions has been introduced,6,11−13 only recently have these been tested for uptake in cells. In a very few studies, nanoreactor activity inside cells pointed to behavior such as could be expected from a simple, artificial organelle.14−17 However, none has been proposed to mimic one of the existing cell organelles. To design mimics of cell organelles, complex requirements need to be fulfilled:18 (i) polymer compartments of appropriate size for cell-uptake, (ii) activity of encapsulated enzyme(s) maintained in situ, (iii) a polymer membrane, which is stable and permeable to small molecules, (iv) integrity and activity preserved in cells (i.e., an artificial organelle), and (v) biocompatibility and biodegradability of polymer compartments. This complexity could explain why only very few examples of nanoreactors have been reported to be integrated and active in cells. We recently introduced the concept of an antioxidant nanoreactor in which enzymes/enzyme mimics encapsulated in polymer vesicles successfully detoxified superoxide radicals14,19 and peroxynitrites,20 well-known reactive oxygen species (ROS) involved in oxidative stress. Beyond that, in THP-1 cells exposed to oxidative stress, two types of enzymes
reating artificial organelles that aid their natural counterparts in cells will have a dramatic impact on medicine in the treatment of disorders and in the design of artificial cells. Because organelles are key components in cells and comprise compartments loaded with molecules essential to life, their inadequate functioning can contribute to numerous pathological conditions.1 Scientists have already anticipated that the design of molecular factories, such as artificial organelles, will provide a new frontier in medicine.2 To this end, nanoscience provides a necessary tool, due to its specificity in the development of structures and assemblies that are compatible with the size range of natural organelles. In particular, amphiphilic copolymers, which self-assemble into supramolecular structures3 such as micelles, tubes, and vesicles, represent ideal candidates to form organelle-like compartments that can contain combinations of biomolecules, because their chemistry allows the adjustment of properties such as stability, flexibility, and functionality.4 For example, polymer vesicles5 have been reported as nanocarriers for active compounds6 ranging from small molecular weight drugs to proteins or DNA, with improved stability as compared to lipid vesicles.2 They can be functionalized at the outer surface for targeting approaches7 and possess stimuli-responsiveness8 for release of encapsulated compounds “on demand”. A step beyond this was realized by the design of nanoreactorspolymer compartments that encapsulate active compounds (enzymes, proteins, enzyme mimics) with the dual role of protecting the compounds from proteolytic attack and serving as a reaction space.9 The necessary exchange of substrates/products with the environment to support in situ reactions occurs through porous, © XXXX American Chemical Society
Received: April 4, 2013 Revised: May 2, 2013
A
dx.doi.org/10.1021/nl401215n | Nano Lett. XXXX, XXX, XXX−XXX
Nano Letters
Letter
cascade reaction and increasing their encapsulation efficiency serves to control the overall activity of an AP. We evaluated the processes by which an AP is taken-up and distributed, such as internalization kinetics, intracellular trafficking, and translocation, to determine functional aspects of an artificial peroxisome. The design of artificial organelles, such as our AP, will aid in the production of “cell implants” that can contribute to the treatment of various cell disorders and will ultimately support the design of artificial cells. In addition, the strategy we introduce here for the development of APs from the molecular level onward would provide the opportunity to tailor different combinations of active compounds (biomolecules, mimics, and drugs), thereby providing solutions for personalized medicine. Results and Discussion. Design of Artificial Peroxisomes. To produce an artificial peroxisome (AP) that shares the functioning and properties of a natural peroxisome, we developed and optimized the gated polymer vesicle in which a cascade reaction can be used to combat reactive oxygen species (ROS)16 (Figure 1D). To enhance the overall cascade reaction occurring as tandem activity of Cu/Zn-superoxide dismutase (SOD), the first enzyme, and lactoperoxidase (LPO), the second enzyme, we increased the enzymes encapsulation efficiency and changed LPO with catalase (CAT). In nature, SOD detoxifies superoxide radicals28 to H2O2 via a two-step reaction, while the second enzyme, LPO29 or CAT,30 serves to degrade H2O2 to harmless products, molecular oxygen and water. We selected the poly(2methyloxazoline)-b-poly(dimethylsiloxane)-b-poly(2-methyloxazoline), (PMOXA-PDMS-PMOXA) triblock copolymer to produce AP compartments. It has been shown that this type of amphiphilic polymer self-assembles in dilute solution and generates vesicles of high mechanical stability,31 sizes in the nanometer range, and features none of the organizational defects responsible for the known instability of liposomes. Vesicle membranes of PMOXA-PDMS-PMOXA copolymer are permeable to superoxide radicals and oxygen,19 while they are impermeable to other molecules of higher molecular weight, such as glucose, glycerol, and urea.32 We chose to incorporate the outer membrane protein F (OmpF), which allows passive transport of molecules up to 600 Da,33 to serve as a gate for the substrates/products of our cascade reaction, while the enzymes cannot escape from the vesicle cavity. Channel protein insertion in synthetic membranes is an elegant approach to the development of mimics of a biomembrane. We used a direct dissolution method for AP preparation based on the chemical nature of PMOXA-PDMS-PMOXA copolymers and to avoid organic solvents (Supporting Information). Such mild encapsulation conditions have already been reported to preserve the integrity and activity of a variety of enzymes encapsulated in nanoreactors.16,19,20 Variation of the hydrophilic-to-hydrophobic ratio of PMOXA-PDMSPMOXA copolymers served to optimize the supramolecular structures generated by self-assembly in terms of architecture (vesicles as major population of assemblies) and size. Homogeneous, empty PMOXA12-PDMS55-PMOXA12 polymer vesicles at a diameter of 200 ± 40 nm, which is comparable to the size range of a peroxisome from 200 nm to 1.1 μm,34 were generated in buffer by the self-assembly process, as established by dynamic and static light scattering (Supporting Information, Figure S1). The formation of vesicles was confirmed by transmission electron microscpy (TEM; Figure 2A). The larger value of the diameter of vesicles based on LS
that acted in tandem within polymer vesicles degraded superoxide radicals and related H2O2.16 Here, we go one step further, providing evidence that an enzyme-containing polymer vesicle mimics a natural cell organelle (Figure 1A), the peroxisome (Figure 1B), upon
Figure 1. (A) Schematic view of a cell showing various organelles. (B) Peroxisomes are spherical, cell organelles of sizes in the nanometer range that play a significant role in the regulation of ROS. (C) An artificial peroxisome (AP) is based on the simultaneous encapsulation of a set of antioxidant enzymes in a polymer vesicle, with a membrane equipped with channel proteins. The APs serve for ROS detoxification inside cells. (D) Schematic enzymatic cascade reaction occurring inside the AP and serving to detoxify superoxide radicals and related H2O2.
optimization in terms of properties and function. A peroxisome is a multipurpose, cell organelle involved in fatty acid αoxidation, the catabolism of purines, the biosynthesis of glycerolipids, and the regulation of reactive oxygen species (ROS) at the molecular level.21 Peroxisomes are nanometerscale, round, vesicle-like organelles that contain various antioxidant enzymes such as catalase, glutathione peroxidase, and Cu/Zn- or Mn-superoxide dismutase. Cellular antioxidant defense mechanisms, to which peroxisomes contribute, can be overwhelmed by an imbalance in ROS associated with oxidative stress. Increased levels of ROS, which contribute to the pathogenesis of various diseases (e.g., arthritis, Parkinson’s disease, cancer, and AIDS),22−25 and are responsible for inorganic nanoparticle toxicity,26 represent a serious concern in medicine. Thus, the design of artificial peroxisomes is expected to represent an excellent solution in this regard. Our artificial peroxisome (AP) is based on the tuned coencapsulation of two antioxidant enzyme types that support a specific cascade reaction in the cavity of a polymer vesicle. The vesicle membrane is provided with channel proteins that serve as gates for substrates and products (Figure 1C and D). The advantage of using the bottom-up approach of combining polymer compartments with active molecules serve for the optimization of an AP with respect to: (i) compartment properties, (ii) enzyme co-encapsulation efficiency, and (iii) enhancement of the cascade reaction. Optimization of compartment properties is necessary to improve cell integration of an AP.27 Substituting the types of enzymes involved in a B
dx.doi.org/10.1021/nl401215n | Nano Lett. XXXX, XXX, XXX−XXX
Nano Letters
Letter
Figure 2. (A) TEM micrograph of empty PMOXA-PDMS-PMOXA vesicles. (B) TEM micrograph of LPO-encapsulated vesicles. (C) TEM micrograph of SOD-LPO-containing APs. Scale bars = 200 nm. (D) APs solution used for cell-uptake studies. (E) FCS autocorrelation curves and fits of: (a) AlexaFluor-633, (b) LPO-AlexaFluor-633, and (c) LPO-AlexaFluor-633-containing vesicles. (F) FCS and FCCS of APs: (a) autocorrelation curve of vesicles containing SOD-AlexaFluor-488 together with its fit; (b) autocorrelation curve of vesicles containing LPOAlexaFluor-633 together with its fit, and (c) cross-correlation curve of SOD-AlexaFluor-488 and LPO-AlexaFluor-633 co-encapsulated in APs together with its fit. Dotted lines are experimental curves.
experiments compared to TEM is expected, as the DH from DLS experiments represents the sum of the vesicle and its surrounding hydration sphere. To obtain a membrane with high permeability while keeping the vesicles intact, we gradually increased the OmpF concentration initially used for insertion.35 A final concentration of OmpF of 0.05 μg/μL proved optimal for membrane permeability and did not affect membrane stability. The design of an AP first requires characterization and optimization of the encapsulation of each enzyme type in a polymer vesicle to provide the best conditions prior to enzyme co-encapsulation. An increase in the initial concentration of each enzyme to achieve higher encapsulation efficiency had to be balanced with respect to vesicle aggregation or disturbing the self-assembly process of vesicle formation. The initial enzyme concentration used for enzyme encapsulation did not affect the morphology or size of the nanoreactors, as established by a combination of light scattering and TEM for: SOD (Supporting Information, Figure S2), LPO (Figure 2B, Supporting Information, Figure S3), and CAT (Supporting Information, Figure S4). Each enzyme was labeled with a different fluorescent dye (AlexaFluor-488 for SOD, AlexaFluor-633 or dylight-633 for LPO, and dylight-488 for CAT), and its encapsulation efficiency was established by a combination of fluorescence correlation spectroscopy and brightness measurements (Supporting Information). A labeling efficiency of 3 AlexaFluor-488 molecules per SOD, 1 AlexaFluor-633 molecule, or 2.7 dylight633 molecules per LPO, and 1 dylight-488 molecule per CAT was obtained. The number of enzymes per vesicle was determined by dividing the value of the molecular brightness
of enzyme-containing vesicles, expressed as counts per molecule (CPM), by the CPM of freely diffusing, labeled enzymes. We calculated the encapsulation efficiency in vesicles (EEv) by dividing the number of enzyme molecules per vesicle by the maximum number of enzyme molecules possible to be encapsulated inside a vesicle via a statistical, self-assembly process of encapsulation, based on the initial enzyme concentration (Supporting Information, eqs S1−S4).36 The change in the diffusion time, from the value characteristic for free SOD-AlexaFluor-488 (τD = 109 μs) to a value of 3.5 ± 1 ms, indicates that the enzyme was encapsulated in vesicles (Supporting Information, Figure S5).19 Based on various initial amounts of SOD-AlexaFluor-488 (0.1−0.5 mg/ mL), an optimum EEv of 25% ± 10% was obtained using an initial enzyme concentration of 0.25 mg/mL (Supporting Information, Figure S5). To exclude unspecific binding by SOD-AlexaFluor-488 or AlexaFluor-488, which would lead to greater uncertainty in the number of encapsulated enzymes, empty polymer vesicles were incubated with dye-labeled enzyme or a dye molecule. In both cases, no unspecific binding was detected. A further increase in the initial SOD concentration used for encapsulation of up to 0.9 mg/mL was not possible, due to the intrinsically poor solubility of SOD. However, because SOD has very high reaction kinetics (kcat ≈ 10−9 M−1 s−1),37 only a few enzymes per vesicle were sufficient to support the in situ cascade reaction.16 Similar to SOD, the successful encapsulation of LPOAlexaFluor-633 in polymer vesicles was indicated by a change in the diffusion time from the value corresponding to the free, labeled-LPO (τD = 320 μs) to a value of 8.5 ± 1 ms, which corresponds to the enzyme-containing vesicles. Various initial C
dx.doi.org/10.1021/nl401215n | Nano Lett. XXXX, XXX, XXX−XXX
Nano Letters
Letter
Figure 3. (A) Viability in HeLa cells after incubation with different concentrations of APs (MTS assay). Blue bars: incubation time 24 h, red bars: incubation time 48 h. (B) Confocal laser scanning microscopy (CLSM) images of HeLa cells incubated with APs containing SOD-AlexaFluor-488 and LPO-AlexaFluor-633 for 24 h (scale bar = 20 μm). (C) CLSM images of HeLa cells incubated with APs containing SOD and LPO-AlexaFluor633 for different incubation times (scale bar = 20 μm). (D) Flow cytometry analysis of APs in HeLa cells for different incubation times (0, 2, 4, 8, 18, 24 h).
cases where no aggregates were present and the change of the dye did not improve the EEv of LPO, we used only LPOAlexaFluor-633 for AP characterization. The bottom-up approach to build an AP allows an exchange of enzyme types to optimize the cascade reaction in situ. We tested CAT instead of LPO as the second enzyme, to determine whether we would obtain an increased overall activity of the APs. The advantage of CAT is that it allows simplified conditions for the assessment of the overall activity of APs, because if detoxifies H2O2 in a catalytic reaction, without the necessity of additional cosubstrates.30 However, as CAT has a higher molecular weight (250 kDa),30 it creates problems for encapsulation efficiency, as shown before with lipid vesicles.38 The change in the diffusion time from the value corresponding to the free CAT-dylight-488 (τD = 340 μs) to a value of around 4 ± 1 ms, indicates that CAT-dylight-488 was successfully encapsulated in polymer vesicles. No aggregates were present across the entire range of initial enzyme concentrations (Supporting Information, Figure S9). An EEv of 16% ± 9%, significantly lower compared to that of LPO, was
amounts of LPO-AlexaFluor-633 (from 0.1 mg/mL to 0.5 mg/ mL) have been tested to optimize enzyme encapsulation efficiency. An optimum EEv of 66% ± 33% was obtained using an initial LPO-AlexaFluor-633 concentration of 0.25 mg/mL (Figure 2E, Supporting Information, Figure S6). The high EEv of LPO leads to the final activity of the APs. At higher initial concentrations of LPO-AlexaFluor-633 used for the encapsulation, for example 0.83 mg/mL, aggregates with a diffusion time of >30 ms were determined by FCS, and confirmed by TEM (Supporting Information, Figure S7). A diffusion time of around 7 ms was obtained when empty vesicles were measured in the presence of the free dye (50 nM). This unspecific binding of AlexaFluor-633 to the PMOXA 12 -PDMS 55 PMOXA12 vesicles explains the presence of large aggregates formed at the high, initial LPO-AlexaFluor-633 concentrations used for encapsulation. As expected, the labeling to LPO with a different dye, dylight-633, prevented the formation of aggregates when high initial concentrations of LPO were used for the encapsulation procedure (Supporting Information, Figure S8). However, because the optimum EEv of LPO was obtained at a significantly lower initial enzyme concentration, in D
dx.doi.org/10.1021/nl401215n | Nano Lett. XXXX, XXX, XXX−XXX
Nano Letters
Letter
Figure 4. (A) CLSM image of HeLa cells incubated for 24 h with APs (blue: nucleus stained with Hoechst, green: endosome/lysosome stained with pH-rodo, red: artificial peroxisomes) (scale bar = 5 μm). (B) TEM images of the peri-nuclear region of HeLa cells incubated with APs for 12 h (scale bar = 200 nm). (C) TEM images of endosome-like compartments in HeLa cells incubated with APs for 12 h (Scale bar = 200 nm). (D) Protective effect of APs against paraquat by flow cytometry analysis: viability of untreated cells considered as 100%, (cells), viability of cells after treatment with paraquat (PQ) and viability of cells pretreated with APs for 24 h, and sequentially treated with paraquat for a further 24 h (PQ-APs). (E) Real time ROS detoxification kinetics of APs in: (a) cells treated with pyocyanin, and (b) cells pretreated with APs (8 h) followed by treatment with pyocyanin.
calculated for CAT encapsulation based on brightness measurements correlated with FCS. We compared the enzymatic kinetics of LPO-containing vesicles to that of CAT-containing vesicles (Supporting Information, Scheme S1). For similar concentrations of LPO and CAT under free conditions and adjusted to the concentration range expected for APs, CAT competed with LPO with comparable kinetics (Supporting Information, Figure S10A). The equivalence of kinetics between peroxidase and catalase under physiological conditions at low substrate concentrations has already been reported.39 However, because of the lower encapsulation efficiency of CAT compared to LPO, H2O2 degradation by CAT-containing vesicles was less effective (Supporting Information, Figure S10B). This indicates that, under the given conditions of the APs, the key determinant of overall activity is encapsulation efficiency, and not the specificity of the enzyme. We chose to continue the development of APs by including the combination of SOD and LPO.
Fast, efficient ROS detoxification by APs requires successful co-encapsulation of SOD and LPO in polymer vesicles with OmpF-inserted membranes. The co-encapsulation process, leading to a turbid solution of APs (Figure 2D), influenced neither the morphology nor the size of vesicles, according to TEM (Figure 2C) and light scattering (Supporting Information, Figure S11). Fluorescence cross-correlation spectroscopy (FCCS) was used to establish the degree of simultaneous coencapsulation of fluorescently labeled SOD and LPO via the statistical process of self-assembly (Figure 2F). The positive cross-correlation curve (Figure 2F-c) was compared with the autocorrelation curves for each enzyme (Figure 2F-a,b) and proved the successful co-encapsulation of SOD-AlexaFluor-488 and LPO-AlexaFluor-633 in vesicles. Fluorescence brightness measurements performed in each detection channel showed that the ratio of SOD:LPO of 1:2 inside vesicles permitted the cascade reaction (Figure 1D), as in the case of single-enzyme encapsulation conditions. Cross-talk in the FCCS experiment was negligible (
