Microfluidic Self-Assembly of Insulin Monomers

According to our analysis using optical and fluorescence microscopy, insulin amyloid preferentially ... and Parkinson's.1 The current method for study...
4 downloads 0 Views 1MB Size
Download PDF
7068

Langmuir 2008, 24, 7068-7071

Microfluidic Self-Assembly of Insulin Monomers into Amyloid Fibrils on a Solid Surface Joon Seok Lee,† Eujin Um,‡ Je-Kyun Park,‡ and Chan Beum Park*,† Department of Materials Science and Engineering, and Department of Bio and Brain Engineering, KAIST, 335 Gwahangno, Daejeon 305-701, Republic of Korea ReceiVed March 24, 2008. ReVised Manuscript ReceiVed May 23, 2008 We report the self-assembly of insulin monomers into amyloid fibrils within microchannels. To demonstrate the microfluidic amyloid formation and fibril growth on a solid surface, we seeded the internal surfaces of the microchannels with insulin monomers via N-hydroxysuccinimide ester activation and continuously flushed a fresh insulin solution through the microchannels. According to our analysis using optical and fluorescence microscopy, insulin amyloid preferentially formed in the center of the microchannels and, after reaching a certain density, spread to the side walls of the microchannels. By using ex situ atomic force microscopy, we observed the growth of amyloid fibrils inside the microchannels, which occurred at a much higher rate than that in bulk systems. After 12 h of incubation, insulin formed amyloid spherulites having “Maltese cross” extinction patterns within the microchannels according to the polarized microscopic analysis. Microfluidic amyloid formation enabled low consumption of reagents, reduction of incubation time, and simultaneous observation of amyloid formation under different conditions. This work will contribute to the rapid analysis of amyloid formation associated with many protein misfolding diseases.

Introduction The self-assembly of peptides into amyloid aggregates is a hallmark of many protein misfolding diseases such as Alzheimer’s and Parkinson’s.1 The current method for studying amyloid formation, however, is often time-consuming, expensive, and labor intensive, lasting from months to years, making the analytical process very slow. Thus, a critical need exists for an analytical system that would enable the rapid investigation of amyloid formation with a very small amount of amyloidogenic peptides and reagents. In our present work, we first report the microfluidic self-assembly of amyloid aggregates on a solid surface. Microfluidics can manipulate a laminar flow of fluids in 10-200 µm width channels, typically fabricated by embossing of polymers using soft lithography.2 Microfabrication involving microfluidics has become a key technology for various biological applications such as drug delivery, 3D cell culture, cell sorting, and biosensors.3,4 The vital advantages of microfluidic technology include the requirement of minimal amounts of reagents, separations and detections with high resolution/sensitivity, low cost, and short analytical time.5 In this study, we used insulin as a model amyloidogenic peptide for amyloid self-assembly and fibril growth within microchannels on a solid surface. Insulin is a globular peptide having a mostly R-helical secondary structure in the native state, but then it transforms into a cross-β-sheet dominant structure under denaturing conditions, resulting in the formation of long amyloid fibrils through hydrogen bonding between monomers by an ordered stacking.6 Studying amyloid deposition over a “solid surface” rather than in a solution is physiologically more relevant, * To whom correspondence should be addressed. Phone: +82-42-8693340. Fax: +82-42-869-3310. E-mail: [email protected]. † Department of Materials Science and Engineering. ‡ Department of Bio and Brain Engineering.

(1) Murphy, R. M. Annu. ReV. Biomed. Eng. 2002, 4, 155–174. (2) Weibel, D. B.; Whitesides, G. M. Curr. Opin. Chem. Biol. 2006, 10, 584– 591. (3) El-Ali, J.; Sorger, P. K.; Jensen, K. F. Nature 2006, 442, 403–411. (4) Sato, K.; Yamanaka, M.; Takahashi, H.; Tokeshi, M.; Kimura, H.; Kitamori, T. Electrophoresis 2002, 23, 734–739. (5) Whitesides, G. M. Nature 2006, 442, 368–373.

as under the pathological conditions for different neurodegenerative disorders plaque deposition is associated with surfaces.7 Previously, we demonstrated the aggregate formation and deposition of insulin amyloid8 as well as Alzheimer’s β-amyloid9 on a “bulk” solid surface. For the high-throughput analysis of amyloid formation, however, minimizing the consumption of a large amount of reagents and accelerating the rate of amyloid formation were critical. Here, we have introduced a miniaturized microfluidic system for the rapid formation and deposition of insulin amyloids, which enabled the simultaneous investigation of their changes against environmental conditions.

Experimental Section Materials. Bovine insulin, bovine serum albumin (BSA), thioflavin S (ThS), 3-amino-propyltriethoxysilane (APTS), and N,N′-disuccinimidyl carbonate (DSC) were purchased from Sigma-Aldrich (St. Louis, MO) and used without further purification. Fabrication of Microchannels. Microchannels were designed with AutoCAD (Autodesk, San Rafael, CA). The dimensions of each microchannel were 100 µm (width) × 100 µm (height) × 15 000 µm (length). A silicon wafer was spin-coated with a negative photoresist (PR), and SU-8 (MicroChem, Newton, MA) was then exposed to UV light, followed by development with SU-8 developer to make a mold. The microchannels were then obtained by poly(dimethylsiloxane) (PDMS) (Dow Corning, Midland, MI) molding on the silicon wafer. PDMS was chosen as the channel material because of its inertness with the reagents and transparency for easy observation of amyloid formation. For the optical detection of amyloid fibrils, a glass slide was irreversibly bonded with the PDMS microchannels by plasma oxygen treatment. To observe the morphology of amyloid fibrils on the surface of the channel with ex situ atomic force microscopy (AFM), the upper PDMS part had to be removed after the whole process of fibril growth. Instead of plasma bonding, the glass slide was physically pressed against the PDMS microchannels with aluminum slabs and fixed with screws (6) Nielsen, L.; Khurana, R.; Coats, A.; Frokjaer, S.; Brange, J.; Vyas, S.; Uversky, V. N.; Fink, A. L. Biochemistry 2001, 40, 6036–6046. (7) Esler, W. P.; Stimson, E. R.; Jennings, J. M.; Vinters, H. V.; Ghilardi, J. R.; Lee, J. P.; Mantyh, P. W.; Maggio, J. E. Biochemistry 2000, 39, 6288–6295. (8) Ha, C.; Park, C. B. Biotechnol. Bioeng. 2005, 90, 848–855. (9) Ha, C.; Park, C. B. Langmuir 2006, 22, 6977–6985.

10.1021/la800907c CCC: $40.75  2008 American Chemical Society Published on Web 06/13/2008

Letters

Langmuir, Vol. 24, No. 14, 2008 7069

Figure 1. Schematic illustration of amyloid formation inside microchannels. (A) Fresh insulin seeds were immobilized on N-hydroxysuccinimide ester activated glass covered with PDMS microchannels, and remaining functional sites were blocked with BSA. (B) Deposition and growth of amyloid fibrils for the incubation occurred inside the microchannels by a continuous flow of fresh insulin solution.

Figure 2. Effect of incubation time on amyloid formation inside a single-channel. (A) Optical images of amyloid aggregates formed after 3, 6, 9, and 12 h of incubation. (B) Fluorescence images of ThS-induced amyloid aggregates. (C) Fluorescence intensity across the microchannel. (D) Ex situ AFM image of insulin amyloid fibrils formed after 6 h of incubation on the glass surface area within the microchannel.

for easy separation. All glass slides were cleaned in a piranha solution of 70% H2SO4/30% H2O2 (v/v) for 15 min at 60 °C before bonding. N-Hydroxysuccinimide (NHS) Ester Activation of Glass Slides and PDMS Microchannels. The glass slides and PDMS walls within each microchannel were chemically activated for the covalent attachment of insulin monomer seeds. First, a 3% solution of APTS in ethanol/water (95:5 v/v) was injected at a flow rate of 5 µL/h into the microchannel for 1 h for aminopropylation. The microchannel was then washed with 100% ethanol and cured at 100 °C. Finally, the microchannel was activated with NHS by injection of a 20 mM DSC solution in a sodium bicarbonate buffer (50 mM, pH 8.5) for 3 h at room temperature. After the reaction was completed, the microchannel was washed with deionized water and dried. Formation of Insulin Amyloid inside a Microchannel. Fresh insulin (1 mg/mL) was dissolved in 40 mM HCl solution (pH 2.0)

for dissociation of hexamers into monomers and was injected at a flow rate of 1-10 µL/h into the microchannel for 10 min to covalently immobilize insulin monomers, and the microchannel was rinsed with deionized water. A solution of 0.1% BSA in a 50 mM phosphate buffer (pH 7.5) was then introduced for 45 min to block the remaining functional sites of the NHS activated microchannel. The microchannel was washed with phosphate buffer (pH 7.5) and deionized water for 30 min to remove physically adsorbed molecules. The growth of amyloid fibrils was initiated with the introduction of insulin solution (1 mg/mL) in a 40 mM HCl solution at 50 °C. After incubation, the microchannel was washed with HCl solution (40 mM) and deionized water and then dried with N2 gas for further analysis. Microscopic Observation of Amyloid Aggregation. The resulting amyloid aggregates and spherulites formed inside the microchannel were observed with an optical microscope (Eclipse TS100;

7070 Langmuir, Vol. 24, No. 14, 2008

Letters

Figure 3. Formation of amyloid aggregates inside multichannels after incubation for 0, 5, 7, and 9 h. (A) Optical image of the microchannels. (B) Fluorescence image of the microchannels and fluorescence intensity profiles.

Figure 4. Insulin amyloid spherulites formed inside a microchannel after 12 h of incubation. Observation with (A) optical, (B) polarized, and (C) confocal microscopy. Arrows indicate amyloid spherulites having “Maltese cross” patterns of light extinction according to the polarized microscopic analysis.

Nikon, Japan) and a polarized microscope (Eclipse E600; Nikon, Japan). We used ThS-induced fluorescence for the detection of amyloid in the microchannel. The solution of 0.015% ThS in ethanol/ water (50:50, v/v) was introduced through the microchannels for 3 min, washed with deionized water, and dried. Fluorescence occurring in the microchannels was observed with a fluorescence microscope (Eclipse 80i; Nikon, Japan) and a confocal microscope (LSM 510; Carl Zeiss, Germany). The filter had an excitation wavelength of 380-420 nm and an emission wavelength of 450 nm. The relative intensity of florescence from ThS binding was analyzed with the Image J program.10 Ex Situ Atomic Force Microscopy (AFM). Amyloid fibrils grown on the solid surface were detected using AFM analysis. After 5 h of incubation, the glass slide, reversibly bonded with PDMS channels, was separated from the aluminum slab and subsequently rinsed twice with deionized water and dried with N2. AFM images were acquired on a Nanoscope III multimode atomic force microscope using an “E”-type scanner (Digital Instruments Inc., Santa Barbara, CA).

Results and Discussion Amyloid Fibrillation within a Single-Channel Microfluidic Network. The concept of amyloid self-assembly in microchannels is illustratively described in Figure 1. The internal surfaces of each microchannel were seeded with insulin monomers by covalent

bonding after NHS activation of the surfaces. A continuous flow of fresh insulin solution (1 mg/mL) passed through the microchannels at a rate of 5 µL/h. According to our observation using optical microscopy, the amount of insulin aggregates increased in the microchannels with the duration of incubation (Figure 2A). For confirmation of whether the aggregates had the properties of amyloids, we treated the microchannels with ThS, a fluorescent dye that specifically binds only to amyloid aggregates.11 Figure 2B shows the increase of the fluorescence intensity within the microchannels with respect to incubation time. Interestingly, the fluorescence patterns changed with incubation time according to our quantitative analysis (Figure 2C). The intensity was highest at the center of the microchannel incubated for 6 h, whereas, in the case of microchannels incubated for 9 or 12 h, the intensity was higher at the periphery. This indicates that initially, during the injection of fresh insulin solution, the insulin monomers were focused at the center of the microchannel, resulting in higher deposition of aggregates. After reaching a certain density, amyloid formation started on the side walls of the microchannel, and therefore, the deposition spreading near the walls. (10) http://rsb.info.nih.gov/ij/.

Letters

Ex Situ AFM Analysis of Amyloid Aggregates Formed within a Microchannel. In order to observe the morphology of insulin aggregates formed on the surface of the glass substrate within the microchannel using ex situ AFM, we detached the PDMS channel from the glass slide, which was previously fixed with aluminum slabs and screws. As shown in Figure 2D, we found numerous fibrils that had grown on the solid surface within the microchannel incubated for 6 h. Hill et al. suggested that the continuous shear flow influences the orientation of amyloid fibrils growing on a bulk surface.12 Based on the report, we suppose that fibril orientation within the microchannels may also be affected by the microfluidic flow, but it was almost impossible to confirm experimentally due to the difficulties of applying in situ AFM analysis to microchannels. The density of fibrils grown inside the microchannel was much higher than that of fibrils formed on a non-microfabricated bulk surface (Figure S1, Supporting Information). The accelerated amyloid fibrillation in the microchannel should be caused by the continuous flow allowing fresh insulin monomers to be available at the amyloid nucleation sites within the microchannel. According to our previous study, the rate of amyloid fibrillation increased almost linearly with the increase of insulin concentration in solution.8 In addition, all the inner surfaces within the microchannel could serve as a nucleation site for amyloid growth when the growth of amyloid takes place inside the narrow microchannel, which makes microfluidic amyloid formation a three-dimensional process. Furthermore, the ratio of surface area to volume (40 000:1) within the microchannel was much larger than that on a bulk surface, which should accelerate amyloid growth even with such a tiny amount of insulin of less than 25 µL. Formation of Amyloid Fibrils within Multichannels. In order to investigate the growth of amyloid fibrils under different conditions simultaneously, we prepared four microchannels having dimensions of 100 µm (width) × 50 µm (height) in one device with a distance of 70 µm between each microchannel. Amyloid formation in the multichannels was conducted by using the same procedures as that in the single-channel case, and the insulin flow time was differentiated as 0, 5, 7, and 9 h for the microchannels (Figure 3A). In Figure 3B, the fluorescence image of each microchannel is shown with the intensity profile across the microchannels. The increment pattern of fluorescence intensity with the duration of incubation was similar to the tendency of the single-channel case; amyloid fibrils preferentially grew on the center of the channel and then spread to the side walls of the PDMS channels. We also investigated the effect of insulin flow rate on amyloid formation in the microchannel. The flow rate of insulin solution was set at 1, 3, 5, and 10 µL/h for the microchannels. When we analyzed amyloid formation using an optical microscope after 6 and 9 h of incubation, we found that the rate of amyloid formation increased with the elevated flow rate up to 5 µL/h (Figure S2, Supporting Information). This acceleration of fibrillation should be caused by the increased availability of the fresh monomers at the nucleation site. At a flow rate of 10 µL/h, however, amyloid aggregates formed on a narrower region of the glass surface, as compared to the flow rate of 5 µL/h. We speculate that the higher flow rate may make the incoming insulin monomers more focused at the center of the microchannel and cause a higher shear rate at the walls, preventing further amyloid formation near the walls. (11) Urbanc, B.; Cruz, L.; Le, R.; Sanders, J.; Hsiao Ashe, K.; Duff, K.; Stanley, H. E.; Irizarry, M. C.; Hyman, B. T. Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 13990–13995. (12) Hill, E. K.; Krebs, B.; Goodall, D. G.; Howlett, G. J.; Dunstan, D. E. Biomacromolecules 2006, 7, 10–13.

Langmuir, Vol. 24, No. 14, 2008 7071

Spherulite Formation within Microchannels. We further extended our study to observe the morphology of amyloid deposits in microchannels after a longer incubation time. After 12 h of incubation, we found that insulin formed spherulite-like particles within the microchannels (Figure 4A). Amyloid spherulites are radial, spherical bodies consisting of a large quantity of amyloid fibrils and exhibit a “Maltese cross” pattern of light extinction.13–16 These spherical aggregates were observed in the brains of patients with Alzheimer’s disease or Down’s syndrome. Rogers et al. described the formation of insulin amyloid spherulites attached to bulk surfaces.15 The dome-shaped spherulites that formed on the surfaces grew to diameters of ∼100 µm. According to Krebs et al., amyloid spherulites formed in various sizes within 24 h by the concurrent occurrence of fibril formation and random aggregation.16 In our present investigation, most of the insulin spherulites formed inside the microchannel had a uniform diameter of about 20 µm after 15 h of incubation (Figure S3, Supporting Information). These amyloid spherulites clearly showed “Maltese cross” patterns of light extinction according to the polarized microscopic analysis (Figure 4B). We also observed amyloid spherulites by ThS fluorescence using confocal scanning microscopy and confirmed that amyloid spherulites consisted of amyloid fibrils (Figure 4C). These results show that microfluidic amyloid formation can facilitate the growth of amyloid spherulites and can be further used for the aggregation study of other amyloidogenic peptides.

Conclusion We report the first documented example of microfluidic amyloid formation, demonstrating the possible utility of microfluidic technology for amyloid deposition study within microchannels. In this work, we investigated the effect of incubation time and insulin flow rate on the formation of insulin fibrils by using multiple tools, such as ex situ AFM, optical microscopy, fluorescence microscopy, confocal scanning microscopy, and polarized microscopy. The microfluidic system enabled the simultaneous detection of amyloid formation under different environmental conditions using small volumes of reagents. We expect that our microfluidic system for insulin amyloid deposition can be extrapolated to the study of other amyloidogenic proteins such as β-amyloid of Alzheimer’s disease, R-synuclein of Parkinson’s disease, and prions of Mad Cow disease as well. Acknowledgment. We thank Dr. Koyeli Girigoswami for valuable comments. This research was supported by grants from the BioGreen 21 Program (20070301034038), the Korea Research Foundation (KRF-2006-D00078), and the Korea Science and Engineering Foundation (KOSEF) NRL Program (R0A-2008000-20041-0, R0A-2008-000-20109-0). Supporting Information Available: AFM images of insulin amyloid fibrils formed in a bulk solution and on the glass surface area within the microchannel; optical microscope images of amyloid aggregates formed at different flow rates; and optical and polarized optical microscope images of insulin spherulites formed on the glass bottom and the PDMS top. This material is available free of charge via the Internet at http://pubs.acs.org. LA800907C (13) Jin, L.-W.; Claborn, K. A.; Kurimoto, M.; Geday, M. A.; Maezawa, I.; Sohraby, F.; Estrada, M.; Kaminksy, W.; Kahr, B. Proc. Natl. Acad. Sci. U.S.A. 2003, 100, 15294–15298. (14) Krebs, M. R. H.; MacPhee, C. E.; Miller, A. F.; Dunlop, L. E.; Dobson, C. M.; Donald, A. M. Proc. Natl. Acad. Sci. U.S.A. 2004, 101, 14420–14424. (15) Rogers, S. S.; Krebs, M. R. H.; Bromley, E. H. C.; Linden, E.; Donald, A. M. Biophys. J. 2006, 90, 1043–1054. (16) Krebs, M. R. H.; Bromley, E. H. C.; Rogers, S. S.; Donald, A. M. Biophys. J. 2005, 88, 2013–2021.