Anal. Chem. 2007, 79, 4887-4891
Analytical Method for β-Amyloid Fibrils Using CE-Laser Induced Fluorescence and Its Application to Screening for Inhibitors of β-Amyloid Protein Aggregation Masaru Kato,*,†,‡,§ Hiroyuki Kinoshita,‡ Mitsue Enokita,‡ Yukiko Hori,| Tadafumi Hashimoto,| Takeshi Iwatsubo,| and Toshimasa Toyo’oka‡
Center for NanoBio Integration and Department of Applied Chemistry, Graduate School of Engineering, The University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-8656, Japan, Department of Analytical Chemistry, School of Pharmaceutical Sciences and COE Program in the 21st Century, University of Shizuoka, 52-1 Yada Suruga-ku Shizuoka, Shizuoka 422-8526, Japan, and Department of Neuropathology and Neuroscience, Graduate School of Pharmaceutical Sciences, The University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-0033, Japan
More than 20 million people are suffering from Alzheimer’s disease, and the number of patients will dramatically increase with the arrival of an aging society unless preventive or curative medications are discovered. A fast and sensitive analytical method for β-amyloid (Aβ) aggregates was developed by the combination of CE-laser induced fluorescence and the fluorescence reagent, thioflavine T. The developed method separates two different fibrils within 5 min. The first peak, which migrated at ∼4 min, was supposed to be derived from a precursor of a fibril that migrated at ∼3.5 min. The developed method was also applicable to the high-throughput screening of the Aβ aggregation inhibitors, which was expected to be an effective therapeutic agent candidate of Alzheimer’s disease. Three compounds (daunomycin, 3-indolepropionic acid (3-IPA), melatonin) were used for the assay. The order of the antiaggregation activity of these compounds was daunomycin > 3-IPA > melatonin, which was the same as that of the reported one. These results suggest that this analytical method may be used to analyze the Aβ fibrils and identify potential therapeutic agents for the treatment of Alzheimer’s disease. It is well-known that protein folding abnormalities are recognized to be responsible for diseases like type II diabetes, transmissible spongiform encephalopathies, and neurodegenerative disorders associated with the accumulation of fibrillar proteins, that is, Parkinson’s and Alzheimer’s diseases.1-4 Alzheimer’s * Corresponding author. E-mail:
[email protected]. Fax: +81-3-5841-1841. Tel: +81-3-5841-1841. † Center for NanoBio Integration, The University of Tokyo. ‡ University of Shizuoka. § Department of Applied Chemistry, The University of Tokyo. | Department of Neuropathology and Neuroscience, The University of Tokyo. (1) Dobson, C. M. Trends Biochem. Sci. 1999, 24, 329-332. (2) Merlini, G.; Bellotti, V. N. Engl. J. Med. 2003, 349, 583-596. (3) Binder, W. H.; Smrzka, O. W. Angew. Chem., Int. Ed. 2006, 45, 73247328. (4) Caughey, B.; Lansbury, P. T., Jr. Annu. Rev. Neurosci. 2003, 26, 27-198. 10.1021/ac0701482 CCC: $37.00 Published on Web 05/31/2007
© 2007 American Chemical Society
disease is a progressive neurodegenerative disease of the brain that is the most common form of age-related dementia. Many people are suffering from Alzheimer’s disease, and the number of patients will dramatically increase by the arrival of an aging society unless preventive or curative medications are discovered.5-7 Therefore, finding treatments that can delay, prevent, or reverse Alzheimer’s disease would have a substantial impact. The primary cause of the inflammation, cytotoxicity, and neurotoxicity associated with the progression of Alzheimer’s disease is probably related to the formation of β-amyloid (Aβ) aggregates and senile plaques.8 Aβ is produced by endoproteolysis of the amyloid precursor protein (APP) and can form noncovalent fibrillar aggregates in the brain that have been related to the neurotoxicity of Alzheimer’s disease. The aggregation occurs in a stochastic manner and involves nucleation steps9 based on the misfolding of Aβ.10 Short peptide sequences have been proposed as nucleation sites in this process.11 Above a certain protein concentration, there is a lag phase in the fibril formation, which can be shortened if a preformed nucleus, such as the seed, is present in the supersaturated solution. During the aggregation process, small beadlike structures (prefibllar aggregates) are first generated, and these subsequently assemble into more distinct morphologies (protofibrils, protofilaments), which is a potent neurotoxin (Figure 1).12-14 Finally, mature fibrils are formed.10 (5) Hebert, L. E.; Scherr, P. A.; Bienias, J. L.; Bennett, D. A.; Evans, D. A. Arch. Neurol. 2003, 60, 1119-1122. (6) Hardy, J.; Selkoe, D. J. Science 2002, 297, 353-356. (7) Selkoe, D. J.; Schenk, D. Annu. Rev. Pharmacol. Toxicol. 2003, 43, 545584. (8) De Felice, F. G.; Ferreira, S. T. Cell. Mol. Neurobiol. 2002, 22, 545-563. (9) Inouye, H.; Krischner, D. A. J. Struct. Biol. 2000, 130, 123-129. (10) Jahn, T. R.; Radford, S. E. FEBS J. 2005, 272, 5962-5970. (11) Inouye, H.; Sharma, D.; Goux, W. J.; Kirschner, D. A. Biophys. J. 2006, 90, 1774-1789. (12) Hartley, D. M.; Walsh, D. M.; Ye, C. P.; Diehl, T.; Vasquez, S.; Vassilev, P. M.; Teplow, D. B.; Selkoe, D. J. J. Neurosci. 1999, 19, 8876-8884. (13) Walsh, D. M.; Hartley, D. M.; Kusumoto, Y.; Fezoui, Y.; Condron, M. M.; Lomakin, A.; Benedek, G. B.; Selkoe, D. J.; Teplow, D. B. J. Biol. Chem. 1999, 274, 25945-25952. (14) Hori, Y.; Hashimoto, T.; Wakutani, Y.; Urakami, K.; Nakashima, K.; Condron, M. M.; Tsubuki, S.; Saido, T. C.; Teplow, D. B.; Iwatsubo, T. J. Biol. Chem. 2007, 282, 4916-4923.
Analytical Chemistry, Vol. 79, No. 13, July 1, 2007 4887
Figure 1. Sequence of Aβ and aggregation scheme of the aggregation process. (* showed the labeled aggregates by ThT.)
However,the details of the aggregation process are still unknown,15,16 because there have been no report that could separate aggregated fibrils with a high efficiency. A microplate reader is used for the conventional assay method for monitoring the aggregation reaction,17 but only the total amount of aggregate compounds was measured in this method. To decrease the production of Aβ, inhibit amyloid aggregation, or facilitate the clearance of amyloid deposits decrease the risk of Alzheimer’s disease. Currently, there are mainly two targets for the therapeutic agents of Alzheimer’s disease that have attracted attention. One is inhibition of the Aβ production, and the other is inhibition of the Aβ aggregation. Aβ is produced through proteolytic cleavage of the APP through sequential cleavages by two proteases, that is, the β- and γ-secretases.18,19 The proteolytic product, the Aβ peptides, is strongly related to the development of Alzheimer’s disease. Therefore, one of the strategies currently being investigated for preventing Aβ formation and Alzheimer’s disease consists of inhibiting the γ- and β-secretases.20,21 Another approach to the treatment or prevention of Alzheimer’s disease is the inhibition of the Aβ aggregation. Several assay methods have been developed to help identify compounds that inhibit the β-amyloid protein aggregation. For example, surface plasmon resonance,22 cell-based assays,23 capillary electrophoresis (15) Choo-Smith, L.-P.; Garzon-Rodriguez, W.; Glabe, C. G.; Surewicz, W. K. J. Biol. Chem. 1997, 272, 22987-22990. (16) Frost, D.; Gorman, P. M.; Yip, C. M.; Chakrabartty, A. Eur. J. Biochem. 2003, 270, 654-663. (17) LeVine, H., II. Methods Enzymol. 1999, 309, 274-284. (18) Masters, C. L.; Simms, G.; Weinman, N. A.; Multhaup, G.; McDonald, B. L.; Beyreuther, K. Proc. Natl. Acad. Sci. U.S.A. 1985, 82, 4245-4249. (19) Kang, J.; Lemaire, H.-G.; Unterbook, A.; Salbaum, J. M.; Masters, C. L.; Gzeschik, K.-H.; Multhaup, G.; Beyreuther, K.; Mu ¨ ller-Hill, B. Nature 1987, 325, 733-736. (20) Kan, T.; Tominari, Y.; Rikimaru, K.; Morohashi, Y.; Natsugari, H.; Tomita, T.; Iwatsubo, T.; Fukuyama, T. Bioorg. Med. Chem. Lett. 2004, 14, 19831985. (21) Wolfe, M. S. Nat. Rev. Drug Discovery 2002, 1, 859-866. (22) Cairo, C. W.; Strzelec, A.; Murphy, R. M.; Kiessling, L. L. Biochemistry 2002, 41, 8620-8629. (23) Apostol, B. L.; Kazantsev, A.; Raffioni, S.; Illes, K.; Pallos, J.; Bodai, L.; Slepko, N.; Bear, J. E.,; Gertler, F. B.; Hersch, S.; Housman, D. E.; Marsh, J. L.; Thompson, L. M. Proc. Natl. Acad. Sci. U.S.A. 2003, 100, 5950-5955.
4888
Analytical Chemistry, Vol. 79, No. 13, July 1, 2007
(CE),24 and LC-MS25 methods were used for screening the potential therapeutic agents for the treatment of Alzheimer’s disease. However, these assay method could not separate or detect the fibrils, because the fibrils were not a stable formation and had various kinds of formations. CE is a high-efficiency separation technique with a short time, and we have reported efficient analysis methods of biomolecules using CE.26-28 Although Sabella et al. have reported an analysis of Aβ aggregation reaction using CE, they used filtrate samples for the analysis, and a diode array detector was used for the detection.24 Therefore, only high-concentration monomer and small aggregates were analyzed in their method. In this study, we developed a fast analytical method for Aβ fibrils using CElaser induced fluorescence (LIF). The method has a highthroughput, is sensitive, and overcomes the limitation of previous analytical methods. Although several Aβ peptides have been reported including Aβ1-40, Aβ1-41, and Aβ1-42, Aβ1-42 was selected for use in this study due to its higher aggregation activity.31 EXPERIMENTAL SECTION Materials and Chemicals. Acetonitrile, 1,1,1,3,3,3-hexafluoro2-propanol, acetic acid, glycine, sodium hydroxide, disodium hydrogen phosphate 12-water, and trifluoroacetic acid (TFA) were purchased from Kanto Chemical Co., Inc. (Tokyo, Japan). Aβ1-42(trifluoroacetate form) was from Peptide Institute, Inc. (Osaka, Japan). Aβ1-42(HCl form), melatonin, daunomycin, 3-indolepropionic acid, and sodium dihydrogen phosphate dihydrate (24) Sabella, S.; Quaglia, M.; Lanni, C.; Racchi, M.; Govoni, S.; Caccialanza, G.; Calligaro, A.; Bellotti, V.; De Lorenzi, E. Electrophoresis 2004, 25, 31863194. (25) Cheng, X.; van Breemen, R. B. Anal. Chem. 2005, 77, 7012-7015. (26) Kato, M.; Sakai-Kato, K.; Matsumoto, N.; Toyo’oka, T. Anal. Chem. 2002, 74, 1915-1921. (27) Sakai-Kato, K.; Kato, M.; Toyo’oka, T. Anal. Chem. 2002, 74, 2943-2949. (28) Kato, M.; Sakai-Kato, K.; Jin, H.-M.; Kubota, K.; Miyano, H.; Toyo’oka, T.; Dulay, M. T.; Zare, R. N. Anal. Chem. 2004, 76, 1896-1902. (29) Bitan, G.; Kirkitadze, M. D.; Lomakin, A.; Vollers, S. S.; Benedek, G. B.; Teplow, D. B. Proc. Natl. Acad. Sci. U.S.A. 2003, 100, 330-334. (30) Krebs, M. R. H.; Bromley, E. H. C.; Donald, A. M. J. Struct. Biol. 2005, 149, 30-37. (31) Naiki, H.; Higuchi, K.; Hosokawa, M.; Takeda, T. Anal. Biochem. 1989, 177, 244-249.
were provided by Wako Pure Chemical Industries, Ltd. (Osaka, Japan). Thioflavin T (ThT) and L-arginine were supplied by SigmaAldrich Japan K.K. (Tokyo, Japan). 3-Chlorocarbonyl-6,7-dimethoxy1-methyl-2(1H)-quinoxalinone (DMEQ-COCl) was from Dojindo Laboratories (Kumamoto, Japan). Water was purified by Milli-Q apparatus (Millipore, Bedford, MA). Apparatus. All CE experiments were performed on a P/ACE System 5000 equipped with a LIF detector (Beckman, Fullerton, CA). A He-Cd laser (Kimmon Electric Co., Ltd., Tokyo, Japan) was used for the light source for excitation. The uncoated fusedsilica capillary (50-µm i.d., 30-cm total length) was from Polymicro Technologies Inc. (Phoenix, AZ). Unless otherwise stated, 10 µM ThT in 0.2 M Gly-NaOH buffer (pH 9.5) was used as electrolyte. The injection of the sample was carried out by applying a pressure of 20 psi for 5 s. The separations were carried out at 10 kV with the anode at the sample injection end. The LC system was used for the purification of the DMEQArg. The LC system consisted of an SCL-10A system controller, LC-10AD pump, DGU-12A degasser, SIL-10AXL autoinjector, CTO10A column oven, and RF-10AXL fluorescence detector (Shimadzu, Kyoto, Japan). Analysis was performed by using an Inertsil ODS80Ab column (4.6 mm × 150 mm) from GL Sciences Inc. (Tokyo, Japan) with the mobile phase consisting of water and acetonitrile (3:7, v/v) containing 0.1% TFA at a flow rate of 1.0 mL/min. The column temperature was 40 °C. 1. Preparation of the Internal Standard (I.S.). A mixture solution of 100 µL of 5 mM DMEQ-COCl in acetonitrile, 100 µL of 50 mM Arg in water, and 200 µL of 100 mM potassium carbonate was reacted at room temperature for 1 min. The reaction solution was injected into the HPLC system, and the DMEQ-Arg peak was eluted at ∼3.7 min. The fractions were then lyophilized, redissolved in water, and stored at -80 °C until used. 2. Preparation of the Seed. The seed was prepared by incubation of the 20 mM phosphate buffer (pH 7.4) containing Aβ1-42(HCl form) at room temperature for one week. The prepared seed solution was added to the sample solution for acceleration of the aggregation reaction. 3. Preparation of Analysis Sample. A mixture solution of 33 µL of 20 mM phosphate buffer (pH 7.4) containing 20 µg of Aβ1-42(TFA form), 1 µL of the DMEQ-Arg stocked solution, and 6 µL of the seed solution was incubated at 37 °C. In the case of the inhibition assay, we added 5 µL of 20 mM inhibitor solution to the mixture solution before the adding the I.S. and the seed solution. The incubated solution was sequentially injected into the CE system. RESULTS AND DISCUSSION Optimization of ThT Concentration. ThT has been extensively used for characterizing the presence of amyloid fibrils, because this reagent binds to the β-sheet structure of the fibrils. 24 In the absence of amyloid fibrils, ThT faintly fluorescenced at the excitation and emission maximums of 350 and 450 nm, respectively. In the presence of the amyloid fibrils, ThT brightly fluoresenced at the excitation and emission maximums of 450 and 482 nm, respectively.31 Although ThT is a highly sensitive and selective reagent for the fibrils, the reagents showed an inhibitory
Figure 2. Effect of ThT concentration on the peak ratio of fibril and internal standard.
activity for fibril formation.32 When the aggregation reaction was examined in the presence of ThT, the fibril formation was affected and changed by ThT, and it was thought that the aggregation reaction did not reflect the physiological condition. To examine the true aggregation reaction, we developed an assay method for fibrils such that the incubation was performed without ThT and then the separation and detection were performed with ThT. Other measurement difficulties of the amyloid fibrils included the aggregation speed that was very fast and the fibrils changed their structure very quickly during the analysis procedures.33 The length of the fibrils reached more than a few micrometers, and it was expected that these large molecules could not penetrate the highperformance separation column that was packed with small particles.34 Therefore, we developed an analysis method of fibrils by CE-LIF that used ThT only in the separation solution. We first optimized the ThT concentration in the separation solution. To achieve the fast separation and highly sensitive detection, the fibrils were derivatized with ThT immediately after injection into the separation capillary. DMEQ-Arg was used as the internal standard, and comparison of the ratio of the sample peak and internal standard peak was performed to determine the optimum concentration. Figure 2 shows the relationship between the ThT concentration and peak ratio of the fibril to I.S. The highest signal intensity ratio was obtained when the ThT concentration was 10 µM. A higher concentration of ThT did not improve the detection of the fibrils, and it caused a decrease in the signal ratio. This result showed that the derivatization reaction occurred immediately after the injection and 10 µM ThT was sufficient for the reaction. The migration time of I.S. and the sharp fibril peak were around 2 and 4 min, respectively, and their times were not affected by the ThT concentration (Figure 3). We chose 10 µM ThT as the optimum concentration and used it as the separation solution for further analysis. Aggregation Reaction Analysis Using the Seed. We tried to analyze the aggregation reaction of Aβ using the developed method. It was reported that the fibril formation critically depends on the seed, which was a piece of the aggregated Aβ fiber.24 The aggregation reaction was accelerated by the addition of the seed, and the fibril was formed within 1 h. Figure 4 depicts the (32) Klunk, W. E.; Wang, Y.; Huang, G.-f.; Debnath, M. L.; Holt, D. P.; Mathis, C. A. Life Sci. 2001, 69, 1471-1484. (33) Ban, T.; Hoshino, M.; Takahashi, S.; Hamada, D.; Hasegawa, K.; Naiki, H.; Goto, Y. J. Mol. Biol. 2004, 344, 757-767. (34) Kato, M.; Dulay, M. T.; Bennett, B.; Chen, J.-R.; Zare, R. N. Electrophoresis 2000, 21, 3145-3151.
Analytical Chemistry, Vol. 79, No. 13, July 1, 2007
4889
Figure 6. Change in peak heights depending on the incubation time. Figure 3. Electropherogram of Aβ fibril using 10 µM ThT.
Figure 4. Electropherograms of fibrils (a, b) with and (c, d) without seed for different incubation times. (a) and (c) 5 min, and (b) and (d) 24 h.
Figure 5. Effect of the seed on the Aβ aggregation reaction.
comparison of the fibril formation. Figure 4a and b shows the electropherograms of the Aβ solution with the seed. Figure 4a is the electropherogram of the sample 5 min after adding the seed, Figure 4b is that 24 h later. Figure 4c and Figure 4d are electropherograms of the samples without the seed and incubation times of 5 min and 24 h, respectively. Figure 5 depicts the relationship between the incubation time and the ratio of the sharp peak and I.S. peak (sharp peak/I.S. peak). A sharp peak, around 3.5 min, was observed in the 5-min incubated sample with the seed (Figure 4a). The peak height and area rapidly increased with the incubation time, and the migration time of the peak gradually became longer. The peak height of the sharp peak in the 24-h incubated sample became six times higher that in the 5-min incubated sample. On the contrary, no sharp peak was observed in the sample without the seed. This result coincided with the reported ones that the aggregation reaction was accelerated by the presence of the seed.35 Furthermore, this result indicated the possibility that this developed method was applicable for monitoring the aggregation reaction. (35) Hasegawa, K.; Yamaguchi, I.; Omata, S.; Gejyo, F.; Naiki, H. Biochemistry 1999, 38, 15514-15521.
4890 Analytical Chemistry, Vol. 79, No. 13, July 1, 2007
Aggregation Reaction without the Seed. Next, the aggregation reaction was observed using the developed method. At this time, we did not add the seed to the incubated solution for the examination of the aggregation reaction under physiological conditions. Although the sharp peak was not observed within 24 h in the sample without the seed (Figure 4d), the broad peak, which migrated ∼4 min, was observed in the sample after the 5-min incubation (Figure 4c). Furthermore, a sharp peak also appeared in the sample with the 50-h incubation. Figure 6 indicates the changes in the sharp and broad peaks depending on the incubation time. The broad peak became the largest and gradually decreased after the 50-h incubation. In contrast, a sharp peak appeared around 50 h and the peak rapidly increased. We supposed that the sharp peak was derived from the same fibril that was observed in Figure 4b, because the migration behavior (shape and time) of these peaks was similar. Under this analytical condition, only the fibrils were selectively detected by the combination of ThT and LIF detection. As an interesting point, the peak shapes of these two peaks are totally different from each other. These results showed that the aggregation reaction divides into more than two phases. At first, an aggregate of the first peak became large, and an aggregate of the sharp peak sequentially grew. At this time, a decrease in the broad peak was observed. Therefore, it was assumed that the broad peak was derived from a precursor of the sharp peak compound, and it is postulated to be an intermediate compound of the aggregation reaction. Further experiments are ongoing to clarify the structure of these fibrils. High-Throughput Screening of the Inhibitor. The developed method was used for the analysis of both the fibrils and the aggregation reaction. Aβ can form noncovalent fibrillar aggregates in the brain that have been related to Alzheimer’s disease neurotoxicity. Therefore, a compound that prevents, delays, or reverses the fibril formation is a promising candidate as a therapeutic agent of Alzheimer’s disease. We applied the developed method for the screening of the inhibitor and evaluated the effectiveness for a high-throughput screening. Melatonin,36 daunomycin,37 and 3-indolepropionic acid (3-IPA)38 were used to validate our assay since they have been reported to inhibit the formation of the Aβ aggregation. The fibrils in the mixed solution of the Aβ monomer, inhibitor, and the seed were measured using the developed method. For fast screening, the seed was added for the acceleration of the aggregation reaction. Figure 7 shows the (36) Pappolla, M. A.; Chyan, Y.-J.; Poeggeler, B.; Frangione, B.; Wilson, G.; Ghiso, J.; Reiter, R. J. J. Neural Transm. 2000, 107, 203-231. (37) Howlett, D. R.; George, A. R.; Owen, D. E.; Ward, R. V.; Markwell, R. E. Biochem. J. 1999, 343, 419-423. (38) Bendheim, P. E.; Poeggeler, B.; Neria, E.; Ziv, V.; Pappolla, M. A.; Chain, D. G. J. Mol. Neurosci. 2002, 19, 213-217.
LIF increased the sensitivity of the detection, and (5) the inhibitor did not change the detection limit, because only the fibrils were detected in this system. Therefore, it is thought that this method is extremely effective for screening of the therapeutic agent candidate for Alzheimer’s disease.
Figure 7. Inhibitory activity assay using the three inhibtors.
relationship between the incubation time and the peak height ratio of the sharp peak/I.S. peak. Although the peak rapidly increased in the sample without the inhibitor, this increase was fairly restrained in the sample with inhibitors. In particular, no sharp peak was observed within 24 h in the sample with daunomycin. This result indicated that daunomycin showed the strongest inhibitory activity among the three. The order of the antiaggregation activities for the three test compounds was daunomycin > 3-IPA > melatonin. This order was the same as the reported antiaggregation activities of these inhibitors.19 In the case of melatonin, the peak decreased about one-sixth compared to that without an inhibitor. The peak of 3-IPA decreased about half compared to that of melatonin. This analysis system, which combined CE-LIF and ThT, has many advantages, such as the following: (1) the fast analytical method of fibrils was accomplished by CE, (2) the required sample volume was very small, (3) ThT improved the selectivity for the aggregated fibrils, (4)
CONCLUSIONS A fast analytical method for Aβ fibrils was developed by the combination of CE-LIF and ThT. Two different peaks were mainly detected by the analytical method, that is broad and sharp peaks. From the sequential analysis of the incubation samples, it was postulated that the broad peak was derived from a precursor compound of the fibril, which was detected as the sharp peak. The developed method may be used to discover compounds that inhibit Aβ aggregation for the potential treatment of Alzheimer’s disease. We hope that a significant treatment for Alzheimer’s disease will be discovered using this method. ACKNOWLEDGMENT This work was supported by grants from the Ministry of Education, Culture, Sports, Science and Technology of Japan and Takeda Science Foundation.
Received for review January 25, 2007. Accepted April 17, 2007. AC0701482
Analytical Chemistry, Vol. 79, No. 13, July 1, 2007
4891