1212
Biomacromolecules 2009, 10, 1212–1216
Glucose-Sensitive Microcapsules from Glutaraldehyde Cross-Linked Hemoglobin and Glucose Oxidase Wei Qi,† Xuehai Yan,†,‡ Li Duan,†,§ Yue Cui,†,‡ Yang Yang,| and Junbai Li*,† Beijing National Laboratory for Molecular Science, International Joint Laboratory, CAS Key Laboratory of Colloid and Interface Science, Institute of Chemistry, Chinese Academy of Sciences, Beijing 100190, Max Planck Institute of Colloids and Interfaces, Golm/Potsdam, D-14476, Germany, Northwest Institute of Nuclear Technology, Xi’an 710024, China, and National Center for Nanoscience and Technology, Beijing 100190, China Received December 26, 2008; Revised Manuscript Received February 19, 2009
We have fabricated the glucose-sensitive protein hollow microcapsules from hemoglobin (Hb) and glucose oxidase (GOx) via layer-by-layer technique with glutaraldehyde (GA) as a cross-linking agent. We demonstrated the coupled enzymatic reactions between Hb and GOx within capsules by means of CLSM. The real-time monitoring of the consumption of glucose may be useful for the design of a glucose sensor. Furthermore, to further study the effect of glucose on the capsules, we investigated the wall permeability of (Hb/GOx)5 capsules. It was found that the glucose consumption could enhance the permeability of (Hb/GOx)5 capsules. This is significant for the development of advanced drug delivery devices. Therefore, the fabricated Hb/GOx capsules may be a potential candidate for the creation of multifunctional devices.
Introduction Biocompatible stimuli-sensitive delivery systems are receiving broad scientific attention because of the controlled release of materials upon triggering by external stimuli.1,2 For instance, in recent years, stimuli-responsive microcapsules for the controlled drug delivery have been fabricated by the layer-by-layer (LbL) approach.3-7 The properties and structure of LbL multilayer capsules can be influenced by a variety of physical and chemical conditions of the surrounding media.8 Different stimuli, such as temperature, pH, ionic strength, and so on, have been reported to tune the LbL multilayer properties and structure.9-12 The investigation of their permeation properties remain essential for further progress in the application of microcapsules as drug delivery vehicles and the development of new smart capsules.8 The LbL technique as a sequential combination of linear polycations and polyanions was originally proposed by Decher et al.13 And later polyion-protein assemblies were constructed through alternate electrostatic adsorption and their successful application in catalysis has been demonstrated.14-16 However, now the LbL method is no longer restricted to charged materials.1 For instance, in recent publications, we report on the development of a simple LbL approach to fabricate protein microcapsules by using glutaraldehyde (GA) as a cross-linker.17-19 Proteins were immobilized by means of the Schiff base reaction between the aldehyde groups of GA and the free amino sites of the proteins. In this work, we report on protein multilayer capsules that may be usable as glucose sensors and potential drug delivery systems. They are fabricated by the alternate assembly of hemoglobin (Hb) and glucose oxidase (GOx) with GA crosslinking. Scheme 1A shows the coupled enzymatic reactions * To whom correspondence should be addressed. E-mail:
[email protected]. † Chinese Academy of Sciences. ‡ Max Planck Institute of Colloids and Interfaces. § Northwest Institute of Nuclear Technology. | National Center for Nanoscience and Technology.
Scheme 1. (A) Coupled Enzymatic Process Based on Glucose Oxidase (GOx) and Hemoglobin (Hb) Coimmobilized as LbL Microcapsules Components; (B) Glucose-Stimulated Enhancement of the Hb/GOx Capsules Wall Permeabilitya
a The consumption of glucose enhanced the penetration of FITCdextran (green dots) into the capsules.
involving GOx and Hb. GOx catalyzes the oxidation and hydrolysis of β-D-glucose into gluconic acid and hydrogen peroxide. Hb can catalyze the reduction of H2O2 due to its peroxidase activity.20,21 This reduction can be monitored by nonfluorescent Amplex red, which is oxidized by H2O2 into resorufin, a fluorescence dye.22,23 Such a system offers several advantages. The fluorescence of resorufin is a convenient measure of the reaction process through its sensitivity for H2O2. This may be used to develop new fluorescence sensors for glucose. On the other hand, it was observed that the consumption
10.1021/bm801502r CCC: $40.75 2009 American Chemical Society Published on Web 03/26/2009
Microcapsules of Hemoglobin and Glucose Oxidase
Biomacromolecules, Vol. 10, No. 5, 2009
1213
Figure 1. TEM image of (Hb/GOx)5 capsule at dried state.
of glucose enhanced the permeability of the capsule walls (Scheme 1B). This is highly attractive for the fabrication of glucose-responsive release systems.
Materials and Methods Materials. Glucose oxidase from Aspergilus niger (GOx, Mw ∼ 186000), hemoglobin from bovine blood (Hb, Mw ∼ 64500), and FITCdextran (2000 kDa) were purchased from Sigma-Aldrich. β-D-glucose and Na2EDTA were obtained from Beijing Chemical Reagents Company. Amplex red reagent (Molecular Probes) was used for enzymatic activity experiments. Ultrapure water with a resistivity greater than 18.2 MΩ · cm was used in all experiments. Fabrication of Capsules. MnCO3 particles approximately 6 µm in diameter serve as templates for hollow capsule formation. MnCO3 microparticles were synthesized by mixing MnSO4 and NH4HCO3 solutions according to the previously reported method.24,25 MnCO3 particles were first dispersed into 1 mg/mL poly(ethylenimine) (PEI, Mw 50 ∼ 100 kDa) in 0.5 M NaCl solution for 20 min, followed by three times centrifugation and washing with 0.5 M NaCl solution. PEI adsorbs to the particles and then the amino end of PEI is reacted with the aldehyde groups GA by mixing the particles with 0.025% glutaraldehyde (GA) in pH 7.2 phosphate buffer solution (PBS) for 12 h. After three washings with PBS, the coated particles were dispersed into 4 mg/mL Hb solutions in PBS and stayed for 12 h, followed by washing with PBS. The GA and Hb were alternately adsorbed until five layers of GA/Hb were constructed on MnCO3 particles. With the same method, another five layers of GA/GOx were assembled on the (GA/Hb)5 multilayer. After the assembly of (GA/Hb)5/(GA/GOx)5, the coated particles were incubated in 0.1 M Na2EDTA solution (pH ) 7.0) to remove the MnCO3 templates. The obtained hollow protein capsules cross-linked by GA ((Hb/GOx)5) were stored at 4 °C. Characterization of Capsules. TEM measurements were performed on a Tecnai 20 microscope (Philips FEI, American) at 120 kV. A drop of sample solution was placed on the copper grid with carbon film and uranyl acetate (UAc) solution was dropped onto the grid for TEM observation. UV-Visible Spectra Measurement. The amount of proteins assembled at each layer was determined comparing the absorbance at 595 nm before and after assembly using the Bradford protein assay with a U-3010 spectrophotometer (HITACHI, Japan).26 Resorufin has absorption/emission maxima of 570/585 nm. Production of resorufin from the coupled reaction of GOx and Hb was investigated to follow the color change of mixture via the spectrophotometer. The time scan of the coupled enzymatic reaction was recorded
Figure 2. (A) Absorption of resorufin at 570 nm change as a function of time in two reaction systems: (1) protein solution (red), (2) (Hb/ GOx)5 capsules (black) upon the addition of glucose solution; (B) Comparison of enzymatic activity of protein solution (red) and protein capsules (black).
Figure 3. Storage stability of protein capsules incubated at different temperatures: 4 °C (black) and 40 °C (red).
at 570 nm. Hb solution and GOx solution were equally mixed in a cuvette and the addition of glucose began the reaction, immediately followed by time scan. The same experiment was carried out for (Hb/ GOx)5 capsules suspension, the only exception being that the immobilized Hb was used in the coupled reaction rather than Hb in solution. To study the long-term stability of the capsule, the (Hb/GOx)5 capsules suspension was stored and incubated at 4 and 40 °C, respectively. During half a month, remaining activities of the immobilized enzymes at two temperatures were determined periodically with the above method. Confocal Laser Scanning Microscopy (CLSM) Measurement. Realtime imaging of glucose consumption was performed with FV1000 confocal system (Olympus, Japan) using 559 laser excitation. The image of capsules was taken with the red fluorescent resorufin, the product of the coupled enzymatic catalysis reaction of GOx and Hb. The experiment was carried out by placing capsules/Amplex red suspension (10 µL) on cover glass, followed by addition of glucose (10 mM, 1 µL). The image acquisition was performed immediately. The images
1214
Biomacromolecules, Vol. 10, No. 5, 2009
Qi et al.
Figure 4. (A) Real-time monitoring of glucose catalysis within (Hb/GOx)5 capsules to analyze glucose consumption with respect to time with CLSM. The top is the CLSM image of (Hb/GOx)5 capsules and the bottom is the fluorescence intensity change of the selected region with time. (B) Fluorescence emission spectra of the selected region in CLSM image of (Hb/GOx)5 capsules at the dried state to prove the red fluorescence of resorufin. The spectra showed a peak at 585 nm, which further proved the production of resorufin.
fonic acid (pyranine) encapsulated inside the microcapsule, as a fluorescent pH-sensitive probe. Excitation spectra of pyranine at different reaction times were scanned from 390 to 490 nm at an emission wavelength of 513 nm immediately after the addition of glucose into (Hb/GOx)5 capsules suspension.
Results and Discussion
Figure 5. Fluorescence excitation spectra of pyranine encapsulated in the capsules at different reaction times: 0, 5, 10, 20, 30, 60, 90, and 180 min from top to bottom. The intensity at 460 nm decreased and the intensity at 406 nm increased gradually, indicating that the medium is becoming acidic with glucose consumption.
of dried capsules were taken when the reaction finished for 560-600 nm emission collection. Meanwhile, CLSM analysis was also used to show permeability change of capsules before adding and after adding glucose into the capsules suspension for 3 h. The FITC-dextran (Mw: 2000 kDa) was selected as the fluorescent probe. Fluorescence Spectra Measurement. The catalysis of glucose by Hb/GOx microcapsules was monitored with an F-4500 fluorescence spectrophotometer (Hitachi, Japan). The generation of gluconic acid during catalysis was monitored using an 8-hydroxyprene-1,3,6-trisul-
The protein microcapsules were prepared in the following way: MnCO3 particles were selected as capsule templates because of their mild dissolution conditions. Onto these template cores, poly(ethylenimine) (PEI) was adsorbed first. This produces a surface with amino groups. An imine linkage between these amino groups and some of its aldehyde groups bound GA to this surface. The amino groups of the following protein reacted with the residual free aldehyde groups of GA. Subsequently, alternating protein layers of Hb or GOx were deposited onto coated particles and immobilized by cross-linking with GA. After the assembly of the desired number of Hb/GOx multilayers, the MnCO3 templates were dissolved with Na2EDTA solution (pH 7.0). Thus, hollow protein capsules are obtained as can be proven by the folds of dried (Hb/GOx)5 capsules observed in TEM images (Figure 1). Here, it needs to be mentioned that on the cytotoxicity of glutaraldehyde, some studies have reported that it largely depends on the initial glutaraldehyde concentration used for cross-linking.27,28 According to the reports, the concentration of glutaraldehyde (0.025% now) for cross-linking is relative safe for cells even if the system is to be used in vivo.
Microcapsules of Hemoglobin and Glucose Oxidase
Figure 6. CLSM image of (Hb/GOx)5 microcapsules mixed with FITCdextran (2000 kDa): (A) without glucose, (B) with adding glucose for 3 h.
The protein immobilized by this technique retains its electroactivity. This has been proven by cyclic voltamometry in our previous reports.17,18 Here we are concerned with the enzymatic activity of both immobilized proteins (Hb and GOx) and with the coupled enzymatic reaction between two proteins. To examine the effect of immobilization procedure on enzymatic activity, we performed control experiments where the same amount of enzyme that was immobilized was added to solution.16 As described above, with active proteins, in the presence of GOx, glucose is catalyzed into H2O2. Subsequently, in the presence of this H2O2, nonfluorescent Amplex red is catalyzed into fluorescent resorufin. The positive features of Amplex red reagent are its very low background which changes little with time, the stability of the fluorescent product once formed, the high fluorescent yield of the oxidized Amplex red, all resulting in increased sensitivity for the detection of H2O2.29,30 Furthermore, because resorufin has excitation/emission maxima of ∼570/585 nm, there is much less interference from autofluorescence in most biological samples. The production of resorufin from the coupled reaction of GOx and Hb was investigated at 570 nm and the time scan was recorded to follow the color change of the reaction system. For activity analysis, the raw activity data were linearly fitted so that the enzymatic activity could be quantified.23 The results were shown in Figure 2 and we found that the coupled enzyme system immobilized in the capsule multilayer was indeed still active. The analysis showed that the immobilized enzyme in multilayer exhibited 30% of
Biomacromolecules, Vol. 10, No. 5, 2009
1215
the activity of free protein. According to the previous report, the 70% loss of activity mainly originates from physical blocking of the substrate binding site and also from the diffusion limitations of glucose within multilayer.16 To further evaluate the long-term stability of the protein capsule, we investigated the storage stability of the capsules at different temperatures during a half-month. The (Hb/GOx)5 capsules were stored and incubated at 4 and 40 °C, respectively. And the enzymatic activity of capsules was then evaluated periodically and compared. In Figure 3, for the capsules stored at 4 °C, almost no loss of activity was observed for half a month. By comparison, only 25% of the activity remained when incubated at 40 °C for two weeks. The results indicated that the present approach is suitable for constructing such protein films in which the bioactivity could be retained for a long period at lower temperature. Therefore, it is feasible to apply such protein capsules for probing glucose. Next, to demonstrate the coupled enzymatic reactions directly and obviously, confocal laser scanning microscopy (CLSM) was applied. CLSM analysis was used to monitor glucose consumption by the production of resorufin with respect to time following the addition of glucose. Figure 4A shows the fluorescence imaging (excitation at 559 nm) of microcapsules upon glucose addition. The graph presents the increase of the fluorescence intensity with time from a selected region in the image. From the production of the resorufin in the capsule wall regions it is apparent that the glucose catalysis occurs on the walls of the microcapsules. Via the increasing fluorescence intensity, the glucose consumption can be observed directly. So it is anticipated to apply this system in developing a fluorescence sensor to detect glucose qualitatively in vitro. In addition, the emission of 585 nm in the fluorescence spectra of the selected capsule further supports the production of the resorufin in the capsule wall regions (Figure 4B). As mentioned above, the coupled enzymatic reactions of Hb and GOx for glucose will produce gluconic acid, which may lead to the decrease of pH in the microenvironment. To detect pH variation during the process of glucose transformation we introduced a pH-sensitive probe inside the capsules.31 Pyranine is a highly water-soluble compound from the group of chemicals known as arylsulfonates. It has a ground state pKa ) 7.2, but on excitation, pKa* ) 0.5. Fluorescence quantum yields are almost 100% in both alkaline and acidic solution when excited with light of λ higher than 400 nm. Fluorescence decay times of it are invariably 5.2 ns in the pH 2-13 range so that deprotonation occurs over a wide pH range.32,33 Pyranine was loaded into capsules by mixing (Hb/GOx)5 capsules dispersion with pyranine solution (in PBS buffer), gently stirring for several minutes followed by centrifuging and washing.18,19 Upon the addition of glucose into the microcapsules suspension, fluorescence wavelength scan was immediately done at the emission of 513 nm at different time point. The excitation spectra of pyranine in Figure 5 showed that the intensity at 460 nm decreased and the intensity at 406 nm increased gradually with increasing time, indicating that the microenvironment of capsules was becoming more and more acidic. These results demonstrate that the produced gluconic acid is dropping the pH value of the microenvironment. According to the above two facts, we investigated the effects of glucose on the permeability of Hb/GOx microcapsules, shown in Figure 6. For the fabricated (Hb/GOx)5 microcapsules in PBS, the confocal images show the wall is impermeable for FITCdextran (2000 kDa). In contrast, the microcapsules were dispersed into glucose solution (100 mM) for 3 h. Then the
1216
Biomacromolecules, Vol. 10, No. 5, 2009
capsules were separated by centrifugation and resuspended in FITC-dextran solution. Being observed under confocal microscopy, the interior of microcapsules is found to be fluorescent equally to the exterior part. It indicates that the fluorescent molecules have penetrated into capsules through the wall. Thus, it is obvious that the capsule wall permeability can be changed by the glucose oxidation and hydrolysis. The capsules are constructed via Schiff base reaction between GA and protein, while the C)N bond becomes weak and even broken partly in acidic medium. So we believe that the change in the wall permeability may be caused by the production of gluconic acid. That decreases the pH in the microenvironment, thus loosening the layer structure and enhancing the wall permeability. Still the mechanism of the glucose-induced permeability change is subject to further studies.
Conclusions In conclusion, we successfully fabricated glucose-sensitive protein microcapsules from hemoglobin and glucose oxidase by the layer-by-layer method. This is achieved by cross-linking the protein layers with glutaraldehyde. The proteins were still enzymatically active after their immobilization in the multilayer. GOx oxidizes the glucose under the concomitant production of H2O2. This H2O2 in turn oxidizes nonfluorescent Amplex red to fluorescent resorufin. The production of this fluorescence dye provides a convenient method to real-time monitor the consumption of glucose. This may be used for microscale glucose sensor applications. In addition, we demonstrate the possibility of changing the Hb/GOx capsule wall permeability by glucose. It is found that the FITC-dextran (2000 kDa) macromolecules, excluded from the capsules formerly, have penetrated into the capsules after the glucose reaction. This may be attributed to a decrease in the local pH due to the production of gluconic acid and thus a loosening of the multilayer structure. We hope that the observed glucose-stimulated enhancement of the wall permeability could play an important role in the development of drug delivery vehicles and may find widespread application in drug loading and release in the coming future. Acknowledgment. This work was financially supported by the National Nature Science Foundation of China (Project No. 20833010) and National Basic Research Program of China (973 Program) 2009CB930100.
References and Notes (1) De Geest, B. G.; Sanders, N. N.; Sukhorukov, G. B.; Demeester, J.; Smedt, C. D. Chem. Soc. ReV. 2007, 36, 636–649. (2) Tang, Z. Y.; Wang, Y.; Podsiadlo, P.; Kotov, N. A. AdV. Mater. 2007, 19, 906–928.
Qi et al. (3) Kreft, O.; Skirtach, A. G.; Sukhorukov, G. B.; Mo¨hwald, H. AdV. Mater. 2007, 19, 3142–3145. (4) An, Z. H.; Lu, G.; Mo¨hwald, H.; Li, J. B. Chem.sEur. J. 2004, 10, 5848–5852. (5) Wang, K. W.; He, Q.; Yan, X. H.; Cui, Y.; Qi, W.; Duan, L.; Li, J. B. J. Mater. Chem. 2007, 17, 4018–4021. (6) De Geest, B. G.; Jonas, A. M.; Demeester, J. S.; Smedt, C. D. Langmuir 2006, 22, 5070–5074. (7) Chu, L. Y.; Liang, Y. J.; Chen, W. W.; Ju, X. J.; Wang, H. D. Colloid Surf., B 2004, 37, 9–14. (8) Johnston, A. P. R.; Cortez, C.; Angelatos, A. S.; Caruso, F. Curr. Opin. Colloid Interface Sci. 2006, 11, 203–209. (9) Ibarz, G.; Da¨hne, L.; Donath, E.; Mo¨hwald, H. AdV. Mater. 2001, 13, 1324–1327. (10) Sukhorukov, G. B.; Antipov, A. A.; Voigt, A.; Donath, E.; Mo¨hwald, H. Macromol. Rapid Commun. 2001, 22, 44–46. (11) Antipov, A. A.; Sukhorukov, G. B.; Mo¨hwald, H. Langmuir 2003, 19, 2444–2448. (12) Radt, B.; Smith, T. A.; Caruso, F. AdV. Mater. 2004, 16, 2184–2189. (13) Decher, G. Science 1997, 277, 1232–1237. (14) Onda, M.; Lvov, Y.; Ariga, K.; Kunitake, T. Biotechnol. Bioeng. 1996, 51, 163–167. (15) Onda, M.; Ariga, K.; Kunitake, T. J. Biosci. Bioeng. 1999, 87, 69– 75. (16) Caruso, F.; Schu¨ler, C. Langmuir 2000, 16, 9595–9603. (17) Duan, L.; He, Q.; Yan, X. H.; Cui, Y.; Wang, K. W.; Li, J. B. Biochem. Biophys. Res. Commun. 2007, 354, 357–362. (18) Duan, L.; Wei, Q.; Yan, X. H.; He, Q.; Cui, Y.; Wang, K. W.; Li, D. X.; Li, J. B. J. Phys. Chem. B 2009, 113, 395–399. (19) Qi, W.; Duan, L.; Wang, K. W.; Yan, X. H.; Cui, Y.; He, Q.; Li, J. B. AdV. Mater. 2008, 20, 601–605. (20) Zhang, Y. Y.; Hu, X.; Tang, K.; Zou, G. L. Process Biochem. 2006, 41, 2410–2416. (21) Mieyal, J. J.; Ackerman, R. S.; Blumer, J. L.; Freeman, L. S. J. Biol. Chem. 1976, 251, 3436–3441. (22) Kreft, O.; Prevot, M.; Mo¨hwald, H.; Sukhorukov, G. B. Angew. Chem., Int. Ed. 2007, 46, 5605–5608. (23) Stein, E. W.; Volodkin, D. V.; Mcshane, M. J.; Sukhorkov, G. B. Biomacromolecules 2006, 7, 710–719. (24) Hamada, S.; Kudo, Y.; Okada, J.; Kano, H. J. Colloid Interface Sci. 1987, 118, 356–368. (25) Sukhorukov, G. B.; Shchukin, D. G.; Dong, W. F.; Mo¨hwald, H.; Lulevich, V. V.; Vinogradova, O. L. Macromol. Chem. Phys. 2004, 205, 530–535. (26) Bradford, M. M. Anal. Biochem. 1976, 72, 248–252. (27) Oliver, R. F.; Grant, R. A.; Cox, R. W.; Hulme, M. J.; Mudie, A. Clin. Orthop. Rel. Res. 1976, 115, 291–302. (28) Mcpherson, J. M.; Sawamura, S.; Armstrong, R. J. Biomed. Mater. Res. 1986, 20, 93–107. (29) Mohanty, J. G.; Jaffe, J. S.; Schulman, E. D.; Raible, D. G. J. Immunol. Methods 1997, 202, 133–141. (30) Zhou, M.; Diwu, Z.; Panchuk-Voloshina, N.; Haugland, R. P. Anal. Biochem. 1997, 253, 162–168. (31) Overly, C. C.; Lee, K. D.; Berthiaume, E.; Hollenbeck, P. J. Proc. Natl. Acad. Sci. U.S.A. 1995, 92, 3156–3160. (32) Wolfbeis, O. S.; Fu¨rlinger, E.; Kroneis, H.; Marsoner, H.; Fresenius, Z. Anal. Chem. 1983, 314, 119–124. (33) Politi, M. J.; Brandt, O.; Fender, J. H. J. Phys. Chem. 1985, 89, 2345–2354.
BM801502R
