Protein Encapsulation via Porous CaCO3 ... - ACS Publications

Aug 19, 2004 - Lien Lybaert , Keun Ah Ryu , Lutz Nuhn , Riet De Rycke , Olivier De Wever , Alfred ...... M. Germain , S. Grube , V. Carriere , H. Rich...
1 downloads 0 Views 489KB Size
Biomacromolecules 2004, 5, 1962-1972

1962

Protein Encapsulation via Porous CaCO3 Microparticles Templating Dmitry V. Volodkin,*,†,‡ Natalia I. Larionova,‡ and Gleb B. Sukhorukov† Max-Planck Institute of Colloids and Interfaces, Golm/Potsdam, 14476, Germany, and Department of Chemistry, Moscow State University, Moscow, 119992, Russia Received June 7, 2004

Porous microparticles of calcium carbonate with an average diameter of 4.75 µm were prepared and used for protein encapsulation in polymer-filled microcapsules by means of electrostatic layer-by-layer assembly (ELbL). Loading of macromolecules in porous CaCO3 particles is affected by their molecular weight due to diffusion-limited permeation inside the particles and also by the affinity to the carbonate surface. Adsorption of various proteins and dextran was examined as a function of pH and was found to be dependent both on the charge of the microparticles and macromolecules. The electrostatic effect was shown to govern this interaction. This paper discusses the factors which can influence the adsorption capacity of proteins. A new way of protein encapsulation in polyelectrolyte microcapsules is proposed exploiting the porous, biocompatible, and decomposable microparticles from CaCO3. It consists of protein adsorption in the pores of the microparticles followed by ELbL of oppositely charged polyelectrolytes and further core dissolution. This resulted in formation of polyelectrolyte-filled capsules with protein incorporated in interpenetrating polyelectrolyte network. The properties of CaCO3 microparticles and capsules prepared were characterized by scanning electron microscopy, microelectrophoresis, and confocal laser scanning microscopy. Lactalbumin was encapsulated by means of the proposed technique yielding a content of 0.6 pg protein per microcapsule. Horseradish peroxidase saves 37% of activity after encapsulation. However, the termostability of the enzyme was improved by encapsulation. The results demonstrate that porous CaCO3 microparticles can be applied as microtemplates for encapsulation of proteins into polyelectrolyte capsules at neutral pH as an optimal medium for a variety of bioactive material, which can also be encapsulated by the proposed method. Microcapsules filled with encapsulated material may find applications in the field of biotechnology, biochemistry, and medicine. Introduction Encapsulation technologies have received considerable attention due to an increased interest in biotechnology, medicine, pharmaceutics, catalysis, ecology, nutrition, etc. Encapsulation can concentrate, shield, and protect biomolecules in a defined volume and creates a single compartment representing a microenvironment separated from the outer environment. Microcarriers (capsules, polymer spheres, microemulsions) filled with bioactive material can be applied for controlled release of drugs and other substances.1 Encapsulation technologies have found industrial applications based on chemical and physical methods: solvent evaporation, coacervation, interfacial polymerization, matrix (gel) entrapment, etc.2,3 The technology of ELbL polyelectrolyte adsorption, established to fabricate multilayer films on flat macroscopic substrates,4 explores the electrostatic interaction between oppositely charged macromolecules. Originally, the ELbL technology was not introduced as an encapsulation technique * To whom correspondence should be addressed. Tel.: +49-331-5679235. Fax: +49-331-567-9202. E-mail: dmitry.volodkin@mpikg-golm. mpg.de; [email protected]. † Max-Planck Institute of Colloids and Interfaces. ‡ Moscow State University.

but was developed as a surface coating, modification, and engineering tool. The ELbL approach was subsequently extended to utilize colloidal micro- and nanoparticles as the templates involving a wide range of substances to construct multilayers.5-7 The greatest advantage of the ELbL protocol is the striking simplicity with which the shell thickness can be tuned to nanometric precision by controlling the number of adsorbed molecular layers. Shell thickness and properties of polyelectrolytes allow the permeability of the capsule shell to be manipulated. Thus, the encapsulation of macromolecules, proteins, and other bioactive materials into such microcapsules is of great interest for pharmaceutics and biotechnology due to the possibility to apply such systems as microcontainers for catalysis, drug delivery, and controlled release. Applicability of the polyelectrolyte capsules, made by the ELbL technology, for microencapsulation of bio-substances with various physicochemical properties and biological functionalities (proteins, enzymes, DNA, and drugs), and subsequent release from them have been intensively studied in recent years. Enzymes can be involved in the multilayer shell of the capsules via alternative adsorption.8 However, this approach allows small amounts of protein to be incorporated exclusively in the capsule wall. Another method

10.1021/bm049669e CCC: $27.50 © 2004 American Chemical Society Published on Web 08/19/2004

Protein Encapsulation via Templating

of macromolecule encapsulation inside microcapsules based on different approaches: covering precipitated biomacromolecules (protein, DNA) or drugs,9-15 the controlled precipitation technique followed by colloidal support decomposition,16 and pH-controlled capsule shell permeation.17,18 Inclusion of proteins into the matrix of the capsule consisting of partially dissolved melamine formaldehyde resin was demonstrated.19,20 However, many of these methods are not suited to perform encapsulation of bioactive compounds because of several restrictions. Weakly cross-linked melamine formaldehyde particles were originally employed and most intensively studied templates for manufacturing of polyelectrolyte capsules. At the same time, incomplete elimination of melamine formaldehyde oligomers (formed during dissolution) and their biological incompatibility have strongly limited the use of these cores.21 The procedure of utilizing a precipitation protocol must be done at certain conditions when biomolecules are not soluble (organic solvents, high ionic strength, etc.). When used as sacrificial supports, the cores could be eliminated by dissolution applying extremely low or high pH, oxidizing agents, hydrofluoric acid, or organic solvents that depend on the type of microsupport used. Undoubtedly, such conditions used during encapsulation can influence the biological activity causing denaturation of sensitive biomolecules. Therefore, a very important task is to use mild conditions for encapsulation of bioactive materials and a biocompatible template. Crystallization of inorganic salts, such as calcium carbonate, from supersaturated solutions has been the subject of many studies due to its importance in geo-, bio-, and material sciences, as well as due to its wide applications in industry, technology, medicine, and many other fields.22-29 Great practical needs in CaCO3 preparation consisting of uniform, homogeneous size and nonaggregated microparticles stimulated many studies on controlling the crystallization process. The quality of the resultant microparticles was found to be strongly dependent on the experimental conditions such as the type of the salts used, their concentration, pH, temperature, rate of mixing the solutions, and the intensity of agitation of the reaction mixture, parameters affecting the rate of the nucleation process.22-32 Inclusion to the reaction mixture of such different additives as divalent cations, organic solvents and macromolecules (synthetic or natural) was shown to exert a profound effect on the morphology of the CaCO3 microparticles formed.33-36 We have standardized a simple and reproducible procedure for preparation of spherical nonaggregated and highly porous CaCO3 microparticles with a narrow size distribution ranging from 4 to 6 µm. These templates are biocompatible, inexpensive, and easy to produce and decompose by complexation with EDTA at neutral pH. In addition, CaCO3 microparticles offer a large surface area for adsorption of substances of interest due to their porous structure. Adsorption of proteins occurs on a variety of solid-liquid interfaces. For this propose, the use of porous colloidal particles with interfaces, characterized by a large surfaceto-volume ratio, is particularly interesting.

Biomacromolecules, Vol. 5, No. 5, 2004 1963

In summary, a template that can impart the multiple functionalities listed above is desirable, and we have found that the CaCO3 microparticles fulfill many of the desirable properties. The aim of the present work is to explore the porous microparticles from CaCO3 for encapsulation of proteins in polyelectrolyte capsules prepared using carbonate microparticles as sacrificial templates. We developed a new approach based on results proposed recently37 when calcium carbonate microparticles were used as a support for the microcapsules fabrication. The present method consists of physical adsorption of macromolecules in CaCO3 colloidal particles followed by coating of particles by means of ELbL of oppositely charged polyelectrolytes. Subsequent core dissolution by complexation with EDTA resulted in formation of protein-containing polyelectrolyte microcapsules. The adsorption capacity of CaCO3 microparticles for various proteins and dextran was examined as a function of their concentration and pH. Lactalbumin and horseradish peroxidase were encapsulated as model proteins. Materials and Methods The sources of the chemicals are as follows: sodium poly(styrene sulfonate) (PSS, Mw ∼ 70 kDa), poly(allylamine hydrochloride) (PAH, Mw ∼ 70 kDa) (Aldrich, USA); fluorescein isothiocynate (FITC, Sigma, USA), rhodamin B isothiocyanate, mixture of isomers (Rh, Aldrich), ethylenediaminetetraacetic acid, (EDTA, Sigma, USA), calcium chloride dihydrate (CaCl2‚2H2O, Sigma), Na2CO3 (Merck, Germany), CaCO3 microcrystals (10µm, Aldrich). Proteins: bovine milk R-lactalbumin, type III (Lact); chicken’s egg lysosyme (Lys); horseradish peroxidase, type I (Per); R-chymotrypsin from bovine pancreas (Cht) were purchased from Sigma. Dextran-FITC (Mw ) 4 kDa and 2000 kDa) was from Aldrich. All materials were used without further purification. Proteins and PAH were labeled with FITC or Rh by mixing their solutions in 50 mM borate buffer (pH 9.5) at a molar ratio of 1:1 for proteins and 1:100 for PAH (amino group PAH:dye). After stirring for 2 h at room temperature, the solutions were dialyzed against deionized water in a cellulose dialysis sack with a cutoff of 12-14 kDa overnight and used for further study. The water used in all experiments was prepared in a three stage Millipore Milli-Q Plus 185 purification system and had a resistivity higher than 18.2 MΩ‚cm. 1. Preparation and Characterization of CaCO3 Microparticles. Uniform, nearly spherical microparticles of CaCO3 with narrow size distribution were prepared by colloidal crystallization from supersaturated (relative to CaCO3) solution. The process was initiated by rapid mixing of equal volumes of CaCl2 and Na2CO3 solutions. In a typical experiment, 0.33 M Na2CO3 solution was rapidly poured into an equal volume of 0.33 M solution of CaCl2 at room temperature, and after intense agitation on a magnetic stirrer, the precipitate was filtered off, thoroughly washed with pure water, and dried in air. The course of the reaction was observed under a light microscope. The procedure results in highly homogeneous, spherical CaCO3 microparticles with an average diameter ranging from 4 to 6 µm.

1964

Biomacromolecules, Vol. 5, No. 5, 2004

Volodkin et al.

The Brunauer-Emmett-Teller (BET) method of nitrogen adsorption/desorption was used in order to determine the surface area of CaCO3 microparticles (mean diameter 4.75 µm determined by optical microscopy) and an effective pore size distribution. The data was collected with a Micromeritics TriStar system. 2. Adsorption Experiments. To demonstrate permeation of CaCO3 microparticles for macromolecules with different molecular weights, carbonate particles were suspended in water solutions of dextran-FITC (4 and 2000 kDa) with the concentration of 2 mg mL-1. After 40 min of incubation, the microparticles were washed 3 times (each for 15 min) with clean water to remove nonadsorbed dextrans. In the following experiments, the effects of concentration of macromolecules (proteins and dextrans) as well as medium pH on their adsorption capacity were studied. We used lactalbumine, lysozyme, peroxidase, and chymotrypsin as model proteins. The adsorption experiments were conducted at room temperature. To keep the system as simple as possible, the use of buffers was avoided. To investigate the amount of proteins and dextrans adsorbed in CaCO3 microparticles at different pH of the medium, the following procedure was used. Calcium carbonate microparticles were dispersed in water (6-10%, w/w), and the pH electrode was put into the suspension determining the pH value of about 10.2. The pH in the suspension was adjusted to the range 10.0-7.0 with a step of one unit by adding 10 mM HCl under intensive stirring. The concentration of microparticles at various pH values was equilibrated by centrifugation and removal of a part of the supernatant. A portion (500 µL) of the particle suspension with appropriate pH was added to an equal volume of the protein or dextran (1.0-2.0 mg mL-1) solutions with the same pH. The suspension was shaken for 1 h in an Eppendorf tube and then centrifuged (10000g, 5 min). The supernatant was removed later. 300 µL of supernatant was poured together with 2.7 mL of 0.1M borate buffer solution (pH 8.5) in a plastic cuvette, and the fluorescence was measured. The protein adsorption capacity was calculated from eq 1 Q ) P(A0 - A)/(A0N)

(1)

In eq 1, Q is the amount of protein (dextran) adsorbed per one CaCO3 microparticle (pg), A is the measured fluorescence, A0 is the fluorescence determined as described above using supernatant solutions of CaCO3 microparticles (6-10%, w/w, pH 10-7) instead of microparticle suspension, P is the common protein content in the particle suspension (pg), and N is the number of microparticles in the suspension. The concentration of particles was calculated with a bright-line hematocytometer (cell counting chamber, Sigma). The adsorption isotherms were determined using the same procedure as described above by adding the microparticle suspension (pH 10.0) to protein (dextran) solutions with the same pH but various concentrations (0.2-2.0 mg mL-1). To investigate the amount of Per adsorbed on CaCO3 microcrystals at different medium pH, the same procedure was applied as that for porous microparticles.

To investigate the release of protein adsorbed in CaCO3 microparticles at different medium pH, microparticles with Lact adsorbed at pH 7.0, as described above, were washed two times (2 mL, each for 30 min) with the water solution of pH 7.0 and 10.0 (adjusted with NaOH). After washing a fluorescence of the supernatant solutions was measured according to the procedure described above. 3. Fabrication of Protein-Filled Microcapsules. To encapsulate proteins in polyelectrolyte microcapsules, the protein was adsorbed in CaCO3 microparticles at pH 7.0, as described above in the adsorption experiments. After protein adsorption, the particles were centrifuged (250 g, 5 min) and used for microcapsule fabrication by consecutive polyelectrolyte adsorption of PAH/PSS and carbonate core removal with EDTA at pH 7.5 following the procedure described in ref 37. This procedure was used to fabricate microcapsules with fluorescent-labeled PAH-FITC and Lact-Rh for visualization of their localization. Microcapsules containing encapsulated protein were dispersed in water containing 0.1% NaN3 and stored at 4 °C. The amount of protein released on each step upon capsules preparation was calculated following the measurements of fluorescence of supernatant solutions according to the described above procedure. 4. Activity Measurements of Per. The specific activity of horseradish peroxidase was measured at room temperature. An aliquot of diluted peroxidase or peroxidase in capsules with the same protein content was added to mixture of 6.6 mmol L-1 H2O2 and 15 µmol ABTS in 0.1M acetate buffer, pH 5.1. An absorbance was monitored at 414 nm using a Shimadzu UV-265 FW (Japan) spectrophotometer. The amount of encapsulated Per was determined as a difference in activity of the protein adsorbed in CaCO3 particles and released upon encapsulation procedure. Termostability measurements were performed in a similar way after 1-h incubation at 50 °C in water of a native enzyme and capsules containing the same amount of Per. 5. Characterization. 5.1. Scanning Electron Microscopy (SEM). For SEM analysis, samples were prepared by applying a drop of the particle suspension to a glass slide and then drying overnight. After that, the samples were sputtered with gold. Measurements were conducted using a Gemini Leo 1550 instrument at an operation voltage of 3 keV. 5.2. Confocal Laser-Scanning Fluorescence Microscopy (CLSM). Confocal micrographs were taken with a Leica confocal scanning system mounted to a Leica Aristoplan and equipped with a 100X oil immersion objective with the numerical aperture of 1.4. The excitation wavelength was 488 nm for FITC-labeled compounds and 552 nm for the Rh-labeled ones. 5.3. Fluorescence Spectroscopy. Fluorescence spectra were recorded using a Fluorolog, Spex. The excitation wavelength was chosen to be 488 nm according to the FITC-labeled substrates and 560 nm for the Rh-labeled ones. Fluorescence was measured at 519 and 585 nm, respectively. 5.4. Electrophoretic Mobility. Electrophoretic mobilities of CaCO3 microparticles were measured using a Malvern Zetasizer 4. The pH value in the particle suspension was

Protein Encapsulation via Templating

Biomacromolecules, Vol. 5, No. 5, 2004 1965

Figure 1. Scheme of CaCO3 microparticle formation.

Figure 2. SEM images of CaCO3 micrparticles. A, in overwiev; B, brocken particle; C, recrystallization of particles; and D, calcium carbonate microcrystals (Aldrich).

adjusted to the values 10, 9, 8 and 7 before the measurements by adding HCl under intensive stirring. Results and Discussion 1. Synthesis and Characteristics of Porous CaCO3 Microparticles. Porous calcium carbonate microparticles were synthesized and used for capsule preparation in recent works.37,38 The scheme of microparticle formation is depicted in Figure 1. The amorphous nanoprecipitates instantly formed upon mixing of CaCl2 and Na2CO3 were found to transform into microparticles with spherical morphology and porous internal structure as presented on the scheme. SEM images of these microparticles are presented in Figure 2A,B. The surface is very rough and consists of a great number of carbonate nanoparticles combined to form the specific morphology (Figure 2A). Figure 2B shows the cross section of a broken microparticle illustrating the internal structure, which has specific channel-like structure formed during aggregation of primary nanoprecipitates of CaCO3. BET analysis revealed a high surface area of 8.8 ( 0.3 m2 g-1 and a pore size distribution from 20 to 60 nm (Figure 3). Because of recrystallization, CaCO3 microparticles must be stored in dry state (powder) but not as a water suspension. Recrystallization leads to transformation of spherical particles into rhombohedral calcite microcrystals as shown in Figure 2C (microparticles were analyzed by SEM when recrystal-

Figure 3. Pore size distribution of microparticles.

lization started). The largest fraction of CaCO3 particles (>80%) was found to be recrystallized after storage overnight in water at room temperature. Recrystallization can be easily observed by optical microscopy because of the shape change of the microparticles. The morphology of microcrystals formed after recrystallization was found similar to that of CaCO3 microcrystals (Aldrich) presented in Figure 2D. However, a considerable stabilization of microparticles to recrystallization was observed when microparticles were dispersed in solution of polyelectrolytes (PSS, PAH) or proteins used in this study at concentrations of 0.2-0.5 mg mL-1. After overnight storage, less than 10-20% of themicroparticles were transformed to crystals. Consecutive treatment with polyelectrolytes (PAH/PSS)3 resulted in polyelec-

1966

Biomacromolecules, Vol. 5, No. 5, 2004

Volodkin et al.

Table 1. Some Physical Properties of the Proteins and CaCO3 Microparticles charge at different pH substrate Lys Lact Per Cht Dextran

molecular weight, kDa

molecular dimensions,

15 14 44 24 4 2000

4.5 × 3.0 × 3.0 4.3 × 3.2 × 2.8 6.0 × 3.5 × 3.0 7.1 × 4.0 × 4.0 1.7b 28b

CaCO3 micoparticle b

nm3

,a

Adsmax pg

IEP

10

9

8

7

1.4 1.5 2.7 1.5

11.1 4.5 8.9 8.1

+ -

+ ( -

+ + (

+ + +

8.5

-

-

+

+

a Amount of the proteins adsorbed per CaCO microparticle was calculated based on the molecular dimensions for a closed-packed monolayer. 3 Molecular radii.39

trolyte covering and no crystallization was found in one week of storage at room temperature. Apparently polyelectrolytes adsorbing on calcium carbonate surface prevented the leakage of Ca2+ and CO32-, which would have resulted in recrystallization. 2. Protein Loading in Porous CaCO3 Microparticles. 2.1. Permeability Studies and Adsorption Isotherms. The high surface area and the presence of nanometer-sized pores and channels in CaCO3 particles offer a unique opportunity to capture biomacromolecules such as proteins via physical adsorption/pore diffusion, thus enabling very high substrate loading. To investigate the adsorption of macromolecules, we used dextrans and proteins (Lact, Lys, Cht, and Per) whose physicochemical characteristics are summarized in Table 1. Different permeability of CaCO3 microparticles for macromolecules with different sizes was demonstrated using confocal scanning fluorescence microscopy. The microparticles were incubated in substrate solutions and then washed with water to remove nonbound macromolecules, thus allowing to observe the distribution of the adsorbed material inside the microparticles (Figure 4). Low molecular weight dextran (4 kDa) adsorbed inside the internal volume is distributed uniformly (Figure 4A), whereas the adsorption of high molecular weight dextran (2000 kDa) predominates on the external surface of the particles (Figure 4B). Apparently, small macromolecules of dextran with a hydrodynamic diameter of about 3 nm (Table 1) can penetrate through the porous structure into the CaCO3 microparticle allowing adsorption on internal surface. The size of high-molecularweight dextran, 56 nm, is approximately double the hydrodynamic radius (Table 1). This size is more than the average pore size of CaCO3 microparticles, which is 35 nm (Figure 3). Therefore strong diffusion limitations occur for highmolecular-weight dextran to permeate into particles leading to adsorption mostly on the surface of the particles. Thus, porous CaCO3 microparticles have a different permeability for macromolecules with different molecular weight due to diffusion limitations. In this study, we used proteins whose molecular weight is not high within the limits of 44 kDa for Per (Table 1). These molecules have average size of about 3-4 nm and hence can easily permeate into the internal volume of CaCO3 microparticles through the pores of 20-60 nm. Figure 5 shows a CLSM image of CaCO3 microparticles dispersed in water in the presence of Cht-Rh. This image is typical for the proteins used in this study. One can see that

Figure 4. Confocal images of CaCO3 microcores with adsorbed dextran-FITC, 4 kDa (A) and 2000 kDa (B). Incubation time 40 min.

fluorescence is observed not only on the surface of microparticles but also inside. Thus, protein molecules can penetrate inside the porous microparticle architecture, and their adsorption capacity is determined by their amount adsorbed on the external surface of the particles as well as on the surface in internal pores. The effect of initial concentrations of dextrans and proteins with different molecular weights on adsorption was investigated. Figure 6 shows the amount of adsorbed macromolecules per CaCO3 microparticle, plotted against the concentration of proteins and dextrans in the bulk solution after adsorption at pH 10. The amount of adsorbed macromolecules was increased with the initial concentration, and adsorption isotherms showed well-defined saturation values. The adsorbed amount of proteins and dextran (4 kDa) used was not changed within 1 and 2 h of incubation at the concentration 1 mg mL-1 of macromolecules in suspension.

Protein Encapsulation via Templating

Biomacromolecules, Vol. 5, No. 5, 2004 1967

Figure 7. Change of ξ-potential of CaCO3 microparticles as a function of pH. Figure 5. CLSM image of CaCO3 microparticles in the presence of Cht-TRITC. Incubation time 40 min.

Figure 6. Adsorption isotherms for dextran with molecular weight 4 kDa (A), Lact (B), Cht (C), and dextran 2000 kDa (D) in CaCO3 microparticles. Incubation time 1 h.

It indicates a steady-state has occurred after 1 h of incubation. The protein capacity relative to the amount of protein in solution is a measure of the protein affinity to the surface. High-affinity isotherms are characterized by a strong increase in protein capacity when the amount of adsorbed protein reaches a constant value at low concentration in solution. The affinity of both dextrans and proteins (Cht, Lact) to the surface of CaCO3 microparticles is low, reflected in the initial slope of the observed isotherms. Such type of the isotherms can probably be attributed to diffusion of macromolecules through the pores of the microparticles. Adsorption isotherms were measured also for Lys and Per, and the saturation of the particle surface was found at a concentration less than 0.7 mg mL-1 (data not shown). From the comparison of the plateau values in Figure 6, a remarkable dependence of the adsorption capacity for dextrans on their molecular weight can be noticed. The adsorbed amount is strongly decreased with increasing the molecular weight (Figure 6, parts A and D). It is caused by restriction in size for dextran diffusion inside particles, which results in a decrease of the particle surface available for dextran adsorption.

On the other hand, one can clearly see that the adsorption of Cht is 2 times less than that of Lact (Figure 6B,C). The molecular weight of Cht is 24 kDa and 14 kDa for Lact; hence, they have small molecular dimensions of several nanometers (Table 1) and, in principle, can easily penetrate through the pores. It is evident that the reason for different adsorption amount of the proteins consists of different interaction with the CaCO3 surface. When we consider macromolecule loading in such types of particles, molecular interactions between the substrate and the surface of CaCO3 microparticles play an important role. Therefore, protein loading in such types of particles can be reached not only by manipulating molecular dimensions of macromolecules but also exploiting the interaction of proteins to the surface of CaCO3 microparticles. The pH of the medium would be a decisive factor when we are dealing with a charged substrate or a macromolecule bearing ionizable groups (e.g., protein). 2.2. Effect of pH. The effect of medium pH on the amount of various proteins and dextran adsorbed is investigated to understand the interaction of the proteins with the carbonate surface of the microparticles. In general, adsorption of globular proteins from aqueous solution on a solid surface is the net result of several sub-processes:40,41 (a) electrostatic interaction due to overlap of the electrical double layer around the charged protein molecule and the charged sorbent surface; (b) steric interactions due to the polymeric components at the sorbent surface that extend into the surrounding aqueous phase; (c) changes in the state of hydration; and (d) rearrangements in the protein structure. It is known from the literature that, as a rule, protein adsorption is enhanced in the vicinity of the isoelectric point (IEP)40-44 because protein molecules have no net charge at the IEP and the contribution of lateral intermolecular repulsion is minimum. In such a way, the number of protein-sorbent contacts is determined by a balance between intermolecular and intramolecular forces. It is necessary to consider the charge of a solid support on which proteins adsorb. The ξ-potential determined by an electrophoretic mobility measurement was changed at pH between 8.0 and 9.0 from negative to positive values (Figure 7) that can be explained by an excess of Ca2+. The carbonate ion reacts with additional H+, whereas Ca2+ remains in excess within the solution adsorbing on the particle surface. The surface charge of the particles is negative at pH 9.0 and

1968

Biomacromolecules, Vol. 5, No. 5, 2004

Figure 8. (A) Amount of proteins (dextran) adsorbed per porous CaCO3 microparticle as a function of pH. (B) Amount of peroxidase adsorbed per pg of CaCO3 porous microparticles (PM) and microcrystals (MC) as a function of pH. Incubation time 1h.

10.0 and positive at pH 7.0 and 8.0, as presented in Table 1, together with the protein charge in the range of pH 7.010.0. One can assume that H+ easily permeates through the pores of CaCO3 microparticles and reaches the surface of the internal volume. Therefore, the surface charge of the microparticles corresponds to the charge of the internal surface in the pores leading to the same interaction of protein molecules both with the external surface of the particles and with the surface in the internal volume. Several proteins with various IEP (Table 1) and dextran were used to analyze their adsorption at different pH leading to the results of Figure 8A,B. The amount of protein adsorbed in a CaCO3 microparticle was determined to be more than 1 pg per particle and was varied for each protein used. The adsorption capacity for Lys is strongly increased in the pH range from 7.0 to 10.0 in contrast to Lact for which it decreased toward the alkaline pH (Figure 8A). The proteins have opposite charge at pH 7-10, and their IEPs are far from the pH range used (Table 1). Usually proteins adsorb preferably at a pH equal to the IEP showing a bell-shaped curve and the adsorption capacity decreases at higher or lower pH due to lateral repulsion. In ref 43, a change in the

Volodkin et al.

adsorbed amount of less than 20% for both Lys and Lact adsorption on charged poly(styrenesulfonate) latices was demonstrated in the pH range 7-10. In the referred paper, the protein adsorption reached the estimated maximum, pointing to a close-packed monolayer. This indicates that a contribution of lateral repulsion of Lact and Lys in a compact monolayer structure influences the pH-dependent adsorption but does not induce a large change. It was found that a pronounced increase in the adsorption amount for Lys from 0.02 to 0.19 pg per microparticle occurs when the pH is changed from 7.0 to 10.0, respectively. We observe a doubled extent of adsorption for Lact compared to their alkaline counterparts (pH 9.0 and 10.0) at which the surface charge of the CaCO3 particles is reversed to negative. One can assume that electrostatics plays a remarkable role in adsorption of proteins. This is caused by a different charge of the CaCO3 solid support and proteins in the pH range used. According to Table 1, the charge of the microparticles is positive at pH 7.0 and 8.0. This results in the higher adsorbed amount of Lact, which is negatively charged at these pH values, and a very small amount of Lys due to electro-repulsion with the protein having identical charge. At pH 9.0 and 10.0 the charge of the particles is reversed which reduces the amount of negatively charged adsorbed Lact, but the adsorption capacity of Lys (positively charged) is strongly increased. In the same figure, the amount of adsorbed dextran is presented versus pH. It is found to be independent of adsorption pH that can be explained by the absence of a charge on the polysaccharide. Similar behavior of pH-profile was found for Per and Cht which have the same sign of the charge as microparticles in the pH range 7-10 (Table 1). A little increase in the adsorbed amount of Cht was observed at pH 9.0 (Figure 8A). The amount of dextran adsorbed was found to be higher than that of the proteins (Lys, Per, and Cht) although dextran has no charge that allows expectation of low adsorption capacity. Apparently, other factors play a role in the adsorption. One can conclude that the proteins adsorb in accordance with their net electrostatic interaction with the surface of CaCO3 microparticles. This interaction, caused by overlap of the electrical double layers at the solvated surface and the protein surface, results in electrostatic attraction if the protein macromolecule and the particle surface have opposite charge or in repulsion if their electrokinetic charges are of the same sign. It is important to stress that the adsorption amount of Lact in the pH range 7.0-10.0 is always higher than that of Lys (Figure 8A). These two proteins have similar molecular weight, shape, and size that should give rise to comparable adsorption. However, they have a different IEP (Table 1) and structural stability. One can assume that several factors influence the higher adsorption amount of lactalbumin. Lact has 17 positive and 26 negative residues instead of respectively 19 and 14 for Lys. A greater number of negative residues of Lact results in a huge amount of protein adsorbed, especially at the favorable conditions pH 7.0 and 8.0 (Lact and the particles have opposite charge). Proteins with lower native-state stability possess a stronger driving force for

Protein Encapsulation via Templating

adsorption because this process is driven by conformational rearrangements and an increase in entropy. ∆Gnative-denat. ) 0.7 and 4.1 J g-1 for Lact and the more stable Lys, respectively.45,46 This can be another reason for the higher adsorption capacity of Lact. The surface of Lys is also more hydrophobic than for Lact47 which can lead to an increase in adsorption amount of Lact on the hydrophilic CaCO3 surface. A significant difference in adsorption capacity for Per is observed when porous CaCO3 microparticles prepared in this study and CaCO3 microcrystals were utilized at the same pH (Figure 8B, (PM), (MC)). In the case of porous microparticles, the adsorbed amount is approximately 5-8 times higher than for microcrystals; this can be explained by the larger surface area of porous particles. The SEM images of Figure 2, parts A, B, and D, of two types of microparticles show their structure and morphology. Based on the molecular dimensions and molecular weights of the proteins used to examine the adsorption capacity (Table 1), the surface area of CaCO3 microparticles (8.8 ( 0.3 m2 g-1), and the average weight of one microparticle (88 pg),37 one can determine the amount of protein adsorbed as a close-packed monolayer per one microparticle. Calculated maximum occupancies and the experimentally determined amounts of adsorbed proteins are presented in Table 1 and Figure 8, respectively. One can realize that the loading achieved is lower than theoretically expected for a closepacked monolayer. In the case of Per, this difference is 1 order of magnitude. It points to a loosely packed structure of adsorbed molecules. At the same time, one can assume that not all internal surfaces of CaCO3 microparticles determined by BET are available for protein molecules to be in contact. Molecules of N2 (BET) can easily reach all internal surfaces in even small pores and the surface of irregular and rough internal channels of the porous CaCO3 architecture. Because of steric difficulties of penetrating into small pores, only a part of the internal surface area is available for adsorption of the proteins that can lead to a smaller amount of adsorbed protein than needed to form a compact monolayer. Unfolding and spreading upon adsorption of a protein on the sorbent surface could not lead to a significant change in size.48 To answer the question if protein adsorption in calcium carbonate microparticles is an equilibrium or a nonequilibrium process with respect to dilution and pH change, we tested the desorption of Lact which was adsorbed at pH 7.0 in a medium with the same pH and pH 10.0. Two times washing of microparticles with adsorbed Lact at pH equal to the adsorption pH value resulted in a small protein release (about 8% of the adsorbed amount). This indicates that Lact adsorption in CaCO3 particles is likely an irreversible process, as is the case for many protein adsorption systems. However, after washing with medium of pH 10.0, 47% of adsorbed protein was detected in the supernatant which means that the change of the conditions leads to desorption of protein when the particles have lower adsorption capacity. We suppose that protein binding caused by electrostatics is a reversible equilibrium process. This should be expected due

Biomacromolecules, Vol. 5, No. 5, 2004 1969

to a change of ionization degree of protein surface groups while the pH shifts. In this study we observe a relatively large adsorption of proteins (up to 0.4 pg per microparticle) on the surface of calcium carbonate particles even when both particles and proteins possess the same sign of charge. The same is true when an uncharged substrate like dextran is exposed to CaCO3. Although global electrostatic forces undoubtedly affect adsorption, they are not the only prevailing forces and hence do not dominate this process. The pH of the medium alters the degree of influence of the electrostatic part of the particle-substrate interaction (especially when the substrate is charged or chargeable) providing a tool to manipulate adsorption-desorption kinetics, hence an opportunity to control the amount of charged macromolecules incorporated. 3. Protein Encapsulation in Microcapsules. In the previous study,37 we showed that the porous nature of the CaCO3 microparticles, which could be used as sacrificial templates to obtain polyelectrolyte capsules, offers yet another interesting feature: the possibility to obtain a matrix existing of an interpenetrating network of polyelectrolytes (PSS and PAH) within the capsule interior with similar chemistry as the shell. The scheme illustrating fabrication of microcapsules is depicted in Figure 9, gray dashed arrows. According to the scheme, the CaCO3 microparticle was exposed by consecutive treatment of oppositely charged polyelectrolytes PSS and PAH (Figure 9A-C) resulting in adsorption of the polyelectrolytes not only on the surface of the microcore but also within the interior of the porous support due to diffusion of polyelectrolytes inside. Then the core is dissolved, extraction of calcium by complexation with EDTA (C-D) leads to formation of a microcapsule. Proteins or dextran can be spontaneously loaded by direct mixing with capsules (DH).37 Based on the proposed approach, in this study we demonstrate the possibility of protein encapsulation by addition of the preliminary stage (A-E) of protein adsorption within CaCO3 microsupports before coating with polyelectrolytes (Figure 9, black arrows). According to the scheme, a procedure of consecutive treatment of oppositely charged polyelectrolytes (E-G) followed by core dissolution (GH) results in formulation of microcapsules with incorporated protein. The size of the microcapsule is determined by the size of the template meaning that the capsule size can be predicted before the encapsulation takes place. Figure 10 shows CLSM images of microcapsules fabricated using the proposed approach and exploiting PAHFITC and Lact-Rh. Images A and B represent distribution of protein and PAH in the microcapsule, respectively. Fluorescent profiles of the images show the localization of labeled material in the capsule. Adsorption of polyelectrolytes in CaCO3 microparticles with the ELbL technique occurs easier near the external surface rather than close to the center due to diffusion through the porous material. It leads to formation of the capsule shell. Therefore, the fluorescence intensity of PAH-FITC is higher on the edges of the capsule than in the interior. Gellike matrix formation inside the capsule is more or less uniform as demonstrated in Figure

1970

Biomacromolecules, Vol. 5, No. 5, 2004

Volodkin et al.

Figure 9. Scheme of microcapsule fabrication and encapsulation of macromolecules into capsules.

10, but the presence of clusters of polyelectrolyte complex inside the capsule was also observed. A similar fluorescence profile as for PAH was found for Lact-Rh because the protein binds to the interpenetrating polyelectrolyte network. Otherwise Lact molecules could be released because the porous structure of the microcapsules offered no difficulties for macromolecules to permeate into.37 The amount of Lact released on each step of the microcapsule fabrication is presented in Figure 11. PSS started ELbL due to positive charge of CaCO3 particles. According to this figure, 1.2 pg of protein was adsorbed in a CaCO3 microparticle but less than 0.6 pg was found to be encapsulated per microcapsule (about 5% of microcapsule weight).37 The protein loss of about 50% can be explained by a release of Lact molecules adsorbed into the CaCO3 microparticle due to replacement for the period of permeation of the polyelectrolytes inside the particle during the coating. The major part of the released protein was determined on the first step of adsorption of PSS. It should be noticed that a small amount of Lact was released on the core dissolution step. This means that Lact was bound strongly to the polyelectrolyte complex formed in the CaCO3 microparticles during the ELbL coating process and after core removal the protein molecules remain attached to the complex. After 1 month of storage, only a slight amount of Lact was determined in the supernatant of the capsules (4% of protein in capsules) indicating that capsules can be stored as a suspension preventing the protein release. We have tested the activity of Per encapsulated by the proposed technique and found it to be 37% compared to native Per. At the same time, the revealed activities of encapsulated Per and native enzyme after incubation at 50 °C were found to be 82% and 39%, respectively, in comparison with the activity before heating. It indicates that termostability of Per is improved by encapsulation. Apparently, the protein lost its activity as a result of interaction with polyelectrolyte complex (PSS-PAH) from which the capsules consist of. Together with inhibition by polyelectrolytes one can assume limitations of substrate diffusion through porous polyelectrolyte interpenetrating network that

Figure 10. CLSM images of microcapsules with encapsulated LactRh in TRITC (A) and FITC (B) mode supported with fluorescent profile (C). PAH-FITC was used for capsule fabrication.

Protein Encapsulation via Templating

Biomacromolecules, Vol. 5, No. 5, 2004 1971

biocompatibility but also possess multiple functionalities with respect to biological applications. Formulation of hollow microcapsules exploiting nonpenetrable high-molecularweight polyelectrolytes as well as utilizing biocompatible polymers to produce completely biocompatible capsules is the subject of our current49 and further communications. The proposed method of encapsulation is a promising way to implement the encapsulation of biomaterials for a large number of applications.

Figure 11. Amount of Lact adsorbed per one CaCO3 microparticle and encapsulated in one microcapsule and amount of released protein during encapsulation procedure.

leads to activity decrease. Finding the reason for the loss of the activity will be the subject of our further research. Conclusion In the present investigation, we have applied porous CaCO3 microparticles as sacrificial templates for protein encapsulation. The polyelectrolyte complex was built by the ELbL protocol around the CaCO3 microparticles with adsorbed protein and the core was extracted with a chelating agent. The resultant assembly was not just the hollow polyelectrolyte capsule but a capsule with defined matrixtype interior that embedded protein. The microcapsule size is determined by the size of the template (CaCO3 microparticle), meaning that the capsule size can be predicted before the encapsulation takes place. Biocompatible and easily decomposable (at neutral pH) microtemplates from CaCO3 make the technique suitable for biomolecule immobilization where the sensitivity of the biocompound is often a serious problem. Moreover, a high surface area of 8.8 m2 g-1 and pore size of 20-60 nm provides the opportunity to deliver macromolecules, such as proteins, inside CaCO3 microparticles. They also allow to adsorb molecules in the internal volume and enable immobilization of a relatively high amount. Protein loading in porous CaCO3 microparticles is influenced by diffusion into pores and depends on molecular weight as well as the interaction between proteins and carbonate surface. Electrostatics has a significant contribution in adsorption of proteins. The termoactivity of Per in capsules was improved following encapsulation; but the enzyme partially lost its bioactivity. This can be attributed to inhibition with polyelectrolytes as well as diffusion limitations for a substrate. The use of high molecular weight polymers might improve the encapsulation yield because it adsorbs only on the surface of porous microparticles and does not get in contact with a material previously adsorbed in pores. Thus, the proposed method could be used for encapsulation of biological material such as proteins, enzymes as well as a combination of several compounds or various types of bioactive macromolecules. Encapsulation of reactive polymers or enzymes promotes the concept of an open type microreactor where small solutes can penetrate the shell wall, while the macromolecules stay in the interior. CaCO3 microcores comply not only with the requirement of

Acknowledgment. This work was supported by the Sofja Kovalevskaja Program of the Alexander von Humboldt Foundation and the German Ministry of Education and Research. D.V.V. thanks the DAAD for support (referat 325, number A/03/01495). The authors are grateful to Prof. Dr. H. Mo¨hwald for reading the manuscript and stimulating discussions. We thank A. I. Petrov for help with the preparation of CaCO3 microparticles, Dr. A. Skirtach for reading the manuscript and corrections, and also D. Shchukin and M. Prevot for SEM measurements. References and Notes (1) Kreuter, J. Colloidal Drug DeliVery Systems; Marcel Dekker: New York, 1994. (2) Arshady R. Manufacturing methodology of microcapsules. In Microspheres, microcapsules and liposomes Vol. I: Preparation and chemical applications; Arshady, R., Ed.; Citus books: London, 1999; p 279. (3) Vorlop, K. D., Muscat, A.; Beyersdorf, J. Biotechnol. Tech. 1992, 6, 483. (4) Decher, G.; Hong, J. D. Macromol. Chem., Macromol. Symp. 1991, 46, 321. (5) Sukhorukov, G. B.; Donath, E.; Lichtenfeld, H.; Knippel, E.; Knippel, M.; Budde, A.; Mohwald, H. Colloid Surf A-Physicochem Eng Asp. 1998, 137, 253. (6) Sukhorukov, G. B. In Studies in Interface Science; Mo¨bius, D., Miller, R., Eds.; Elsevier: Amsterdam, 2001; p 383. (7) Schlenoff, J. Multilayer thin films: sequential assembly of nanocomposite materials; G. Decher, Joseph B., Eds., Weinheim, WileyVCH: 2003; 524 p. (8) Schu¨ler, C.; Caruso, F. Macromol. Rapid. Com. 2000, 21, 750. (9) Trubetskoy, V. S.; Loomis, A.; Hagstrom, J. E.; Budker, V. G.; Wolff, J. A. Nucl. Acids Res. 1999, 27, 3090. (10) Caruso, F., Trau D., Mohwald, H.; Renneberg R. Lanqmuir 2000, 16, 1485. (11) Bobreshova, M. E.; Sukhorukov, G. B.; Saburova, E. A.; Elfimova, L. I.; Sukhorukov, B. I.; Shabarchina, L. I. Biophysics 1999, 44, 813. (12) Volodkin, D. V.; Balabushevitch, N. G.; Sukhorukov, G. B.; Larionova, N. I. STP Pharm. Sci. 2003, 13 (3), 163. (13) Volodkin, D. V.; Balabushevitch, N. G.; Sukhorukov, G. B.; Larionova, N. I. Biochemistry (Moscow) 2003, 68 (2), 283. (14) Qiu, X. P.; Donath E.; Mohwald, H. Macromol. Mater. Eng. 2001, 286, 591. (15) Ai, H.; Jones, S.; Villiers, M. M.; Lvov, Y. M. J. Controlled Release 2003, 86, 59. (16) Radchenko, I. L.; Sukhorukov, G. B.; Mo¨hwald, H. Colloids Surf. A 2000, 202, 127. (17) Tiourina, O. P.; Antipov, A. A.; Sukhorukov, G. B.; Larionova, N. I.; Lvov, Y.; Mo¨hwald, H. Macromol. Biosci. 2001, 1, 209. (18) Lvov, Y.; Antipov, A. A.; Mamedov, A.; Mo¨hwald, H.; Sukhorukov, G. B. Nano Lett. 2001, 1 (3), 125. (19) Gao, C. Y.; Donath, E.; Mo¨hwald, H.; Shen, J. Angew. Chem., Int. Ed. 2002, 41(20), 3789. (20) Balabushevich, N. G.; Tiourina, O. P.; Volodkin, D. V.; Larionova, N. I.; Sukhorukov, G. B. Biomacromolecules 2003, 4 (5), 1191. (21) Gao, C. Y.; Moya, S.; Lichtenfeld, H.; Casoli, A.; Fiedler, H.; Donath, E.; Mohwald, H. Macromol. Mater. Eng. 2001, 286 (6), 355. (22) Koga, N.; Nakagoe, Y.; Tanaka, H. Thermochim. Acta 1998, 318, 239. (23) Tracy, L. S.; Franc¸ ois, C. J. P.; Jennings, H. M. J. Cryst. Growth 1998, 193, 374. (24) Tracy, S. L.; Williams, D. A.; Jennings, H. M. J. Cryst. Growth 1998, 193, 382.

1972

Biomacromolecules, Vol. 5, No. 5, 2004

(25) Horn, D.; Rieger, J. Angew. Chem., Int. Ed. Engl. 2001, 40, 4330. (26) Ogino, T.; Suzuki, T.; Sawada, K. Geochim. Cosmochim. Acta 1987, 51, 2757. (27) Spanos, N.; Koutsoukos, P. G. J. Cryst. Growth 1998, 191, 783. (28) Kitamura, M. J. Colloid. Interface Sci. 2001, 236, 318. (29) Kitamura, M.; Konno, H.; Yasui, A.; Masuoka, H. J. Cryst. Growth 2002, 236, 323. (30) Wada, N.; Yamashita, K.; Umegaki, T. J. Cryst. Growth 1995, 148, 297. (31) Wang, L.; Sondi, I.; Matijevic´, E. J. Colloid. Interface Sci. 1999, 218, 545. (32) Sondi, I.; Matijevic´, E. J. Colloid. Interface Sci. 2001, 238, 208. (33) Co¨lfen, H.; Antonietti, M. Langmuir 1998, 14, 582. (34) Kato, T.; Susuki, T.; Amamiya, T.; Irie, T.; Komiyama, M.; Yui, H. Supramol. Sci. 1998, 5, 411. (35) Manoli, F.; Dalas, E. J. Cryst. Growth 2000, 218, 359. (36) Co¨lfen, H.; Qi, L. Chem. Eur. J. 2001, 7, 106. (37) Volodkin, D. V.; Petrov, A. I.; Prevot, M.; Sukhorukov, G. B. Langmuir 2004, 20(8), 3398.

Volodkin et al. (38) Antipov, A. A.; Shshukin, D.; Fedutik, Y.; Petrov, A. I.; Sukhorukov, G. B.; Mo¨hwald, H. Colloid. Surf.: Physicochem. Eng. Aspects 2003, 224, 175. (39) Dong, L. C.; Hoffman, A. S.; Yan, Q. J. Biomater. Sci., Polym. Ed. 1994, 5, 473. (40) Norde, W.; Macritchie. F.; Nowicka, G.; Lyklema, J. J Colloid Interface Sci. 1986, 112, 447. (41) Haynes, C. A.; Norde, W. Colloids Surf. B Biointerfaces 1994, 2, 517. (42) Malmsten, M. J Colloid Interface Sci. 1998, 207, 186. (43) Galisteo, F.; Norde, W. Colloids Surf. B: Biointerfaces 1995, 4, 375. (44) Oliva, F. Y.; Avalle, L. B.; Camara, O. R.; Pauli, C. P. J Colloid Interface Sci. 2003, 261, 299. (45) Privalov, P. L. AdV. Protein Chem. 1979, 33, 167. (46) Kronman, M. J. Crit. ReV. Biochem. Mol. Biol. 1989, 24, 565. (47) Eisenberg, D.; MacLachlan, A. D. Nature 1986, 319, 199. (48) Norde, W. AdV. Colloid Interface Sci. 1986, 25, 267. (49) Sukhorukov, G. B.; Volodkin, D. V.; Gu¨nther, A.; Petrov, A. I.; Shenoy, D. B.; Mo¨hwald, H. J. Mater. Chem. 2004, 14, 2073.

BM049669E