Anal. Chem. 2010, 82, 3067–3072
Screening Kinase Inhibitors with a MicroarrayBased Fluorescent and Resonance Light Scattering Assay Tao Li,†,‡ Dianjun Liu,† and Zhenxin Wang*,† State Key Laboratory of Electroanalytical Chemistry, Changchun Institute of Applied Chemistry, Chinese Academy of Sciences, Changchun, 130022, China, and Graduate School of the Chinese Academy of Sciences, Beijing 100039, P. R. China In this technical note, a microarray-based spectroscopic assay with two readout principles, fluorescence and resonance light scattering (RLS), for screening kinase inhibitors has been reported. In this assay, the phosphorylation and inhibition events are marked by biotinylated antiphosphoserinen/antiphosphotyrosine antibodies, and gold nanoparticles are attached to the antibodies by standard avidin-biotin chemistry followed by silver deposition for RLS signal enhancement. The avidin conjugated fluorescein is used as a fluorescent probe. Assays for both serine kinase, the r-catalytic subunit of cyclic adenosine 5′-monophosphate (cAMP) dependent protein kinase (PKA), and tyrosine kinase, leukocyte-specific protein tyrosine kinase (LCK), have been developed. The utility of this assay to high-throughput screening was demonstrated with a commercial inhibitor library, a collection of 80 kinase inhibitors, and satisfactory results were obtained. In addition, quantitative determination of binding strength and the inhibiting type (type I) of these inhibitors are also demonstrated by the adenosine 5′triphosphate (ATP) competing assays. More than 500 kinases are encoded in the human genome and play particularly important roles in metabolic pathways by phosphorylation of substrate proteins. Rearrangement of kinase activity normally results in loss of cell-cycle control and the development of severe diseases, e.g., cancers, inflammations, and diabetes.1-6 Therefore, it is not surprising that the identification of potential inhibitors of kinases is necessary in drug discovery.7-10 Current kinase inhibition assays can be divided into three classes by the * Corresponding author. E-mail:
[email protected]; Fax: (+86) 431-85262243. † Changchun Institute of Applied Chemistry. ‡ Graduate School of the Chinese Academy of Sciences. (1) Uttamchandani, M.; Wang, J.; Yao, S. Q. Mol. BioSyst. 2006, 2, 58–68. (2) Hutti, J. E.; Jarrell, E. T.; Chang, J. D.; Abbott, D. W.; Storz, P.; Toker, A.; Cantley, L. C.; Turk, B. E. Nat. Methods 2004, 1, 27–29. (3) Houseman, B. T.; Huh, J. H.; Kron, S. J.; Mrksich, M. Nat. Biotechnol. 2002, 20, 270–274. (4) Hunter, T. Cell 2000, 100, 113–127. (5) Manning, G.; Whyte, D. B.; Martinez, R.; Hunter, T.; Sudarsanam, S. Science 2002, 298, 1912–1934. (6) Noble, M. E. M.; Endicott, J. A.; Johnson, L. N. Science 2004, 303, 1800– 1805. (7) Cohen, P. Nat. Rev. Drug Discovery 2002, 1, 309–315. (8) Beveridge, M.; Park, Y. W.; Hermes, J.; Marenghi, A.; Brophy, G.; Santos, A. J. Biomol. Screening 2000, 5, 205–212. 10.1021/ac902804h 2010 American Chemical Society Published on Web 03/04/2010
signal readout formats: radioactive detection, fluorescence, and luminescence. The standard method of measuring kinase activity and inhibition is the use of radiolabeled γ-32/33P-ATP; however, the method could pose threats to human health (handling of hazardous radioactive reagents and wastes) and very cost- and time-consuming (up to 1 week per assay).4-11 Generally, compounds must be tested experimentally against many kinases to determine their selectivity because the binding target of most small molecule kinase inhibitors is the ATP-binding site in the corresponding kinase. Therefore, there is increasing interest on developing nonradioactive methods to monitor kinase activity and inhibition, such as immunoassays with fluorophore-labeled antiphosphopeptide antibodies, nanoparticle-based biosensors.12-15 In particular, the combination of optical or mass spectrometric detection, microarray-based high-throughput assays have been used to monitor kinase activity on the surface of peptide chips.3,10,16-23 These assays have enormous benefits to biomedical and kinase-based research since they enable one to detect enzymatic activities and evaluate potential substrates and inhibitors on a large-scale with the lowest reagent consumptions. Currently, microarray techniques mainly rely on the use of fluorescent dye (9) Davies, S. P.; Reddy, H.; Caivano, M.; Cohen, P. Biochem. J. 2000, 351, 95–105. (10) Schutkowski, M.; Reineke, U.; Reimer, U. ChemBioChem. 2005, 6, 513– 521. (11) Simard, J. R.; Gru ¨ tter, C.; Pawar, V.; Aust, B.; Wolf, A.; Rabiller, M.; Wulfert, S.; Robubi, A.; Klu ¨ ter, S.; Ottmann, C.; Rauh, D. J. Am. Chem. Soc. 2009, 131, 18478–18488. (12) Rothman, D. M.; Shults, M. D.; Imperiali, B. Trends Cell Biol. 2005, 15, 502–510. (13) Umezawa, Y. Biosens. Bioelectron. 2005, 20, 2504–2511. (14) Wang, Z. X.; Levy, R.; Fernig, D. G.; Brust, M. J. Am. Chem. Soc. 2006, 128, 2214–2215. (15) Kerman, K.; Kraatz, H. B. Chem. Commun. 2008, 47, 5019–5021. (16) Shults, M. D.; Janes, K. A.; Lauffenburger, D. A.; Imperiali, B. Nat. Methods 2005, 2, 277–284. (17) Johnson, S. A.; Hunter, T. Nat. Methods 2005, 2, 17–25. (18) Rychlewski, L.; Kschischo, M.; Dong, L.; Schutkowski, M.; Reimer, U. J. Mol. Biol. 2004, 336, 307–311. (19) Rupcich, N.; Green, J. R. A.; Brennan, J. Anal. Chem. 2005, 77, 8013– 8019. (20) Maynard, J. A.; Myhre, R.; Roy, B. Curr. Opin. Chem. Biol. 2007, 11, 306– 315. (21) Cretich, M.; Damin, F.; Pirri, G.; Chiari, M. Biomol. Eng. 2006, 2, 77–88. (22) Sun, L. L.; Liu, D. J.; Wang, Z. X. Anal. Chem. 2007, 79, 773–777. (23) Uttamchandani, M.; Lu, C. H. S.; Yao, S. Q. Acc. Chem. Res. 2009, 42, 1183–1192.
Analytical Chemistry, Vol. 82, No. 7, April 1, 2010
3067
labels.16,24-26 However, these fluorescent dyes have several drawbacks (e.g., overlapping emission spectral features and nonuniform fluorophore photobleaching rates) which lead to reduced sensitivity and poor reliability of the microarray. To overcome the inherent demerits of fluorescent dyes, Mirkin and co-workers have developed a series of RLS readout microarray formats with higher sensitivity and selectivity, which enable one to detect a specific biomolecular recognition process in practical samples, such as patient serum.27-30 Using a similar detection principle, we also developed a RLS-based microarray format for studying kinase functionality and inhibition.22,31 This is achieved by replacing the radiolabeled cosubstrate ATP with γ-biotin-ATP which would bind specifically to the gold nanoparticles modified with avidin. Here, we report a microarray-based spectroscopic assay with two readout principles, fluorescence and RLS, which can be applied for kinase inhibitor screening. Compared to our previous reports,22,31 the cosubstrate is ATP, avoiding the use of chemically modified ATP (i.e., γ-biotin-ATP), that enables one to reduce the cost- and laborconsumption. The recognition is achieved by antigen-antibody and biotin-avidin reactions which provide high affinity and selectivity. This new method is simple since the dye modified avidin is commercially available, and the preparation of the peptide modified gold nanoparticles is faster ( 15) and Z′ > 0.6) has been obtained. Comparing the S/N and Z′ values Analytical Chemistry, Vol. 82, No. 7, April 1, 2010
3069
Figure 2. Evaluation of the assay performance. The observed RLS/fluorescence intensities (red dots) are shown in comparison with control measurements (black squares). The mean values and standard deviations extracted from the data were used to calculate the S/N ratio and Z′ factor. The enzymatic reaction is carried out with 50 µM ATP and 100 units of PKA (a and c) or 4.4 µg/mL LCK (b and d), respectively. The same reactions without any added kinase were used as control measurements.
Figure 3. Light scattering (a and b)/fluorescence (c and d) images and corresponding data analysis of peptide microarrays without inhibitor or with 10 µM inhibitors, respectively. The kinase is PKA (a and c) or LCK (b and d), respectively. The signals have been corrected for background noise (positive control) and normalized to the average signal intensity obtained in the absence of inhibitors (negative control), respectively.
of these two readout formats (as shown in Figure 2), we found that the assay performance of RLS format is better than that of the fluorescence format. 3070
Analytical Chemistry, Vol. 82, No. 7, April 1, 2010
Inhibition Assay. To demonstrate the utility of this assay for inhibitor screening, we examined the inhibition of kinase PKA/LCK with the compounds from the commercial EMD Protein Kinase
Figure 4. Effect of various concentrations of inhibitors in the LCK solution on the light scattering images of the peptide microarrays and corresponding IC50 curves for inhibitors. All of the IC50 curves have similar scales as shown in the graph. The signals have been corrected for background noise (positive control) and normalized to the average RLS intensity obtained in the absence of inhibitors (negative control).
Inhibitor Library I. The library consists of 80 compounds which are potent and reversible protein kinase inhibitors. Each slide was divided into 12 independent 5 × 5 subarrays by a PTFE masker, which includes negative-control reactions without added compounds and positive-control reactions without kinase to set high and low boundaries for RLS/fluorescence signals, respectively (as shown in Figures S1 and S2 in the Supporting Information). The inhibition efficiency of a specific inhibitor can be measured by the RLS/fluorescence intensity because more efficient inhibitors lead to lower levels of phosphorylation, which in turn gives a decreased RLS/fluorescence intensity. Different RLS/fluorescence images and correspondent data analysis of the inhibition assays are shown in Figure 3. As expected, in the presence of inhibitors, the RLS/fluorescence intensity is decreased by increasing the efficiency of the inhibitor. This suggests that our method not only has the potential to screen for inhibitor activity qualitatively but can also be developed to yield quantitative data on the efficiencies of different inhibitors. From this screen, we identified 11 primary hits (2 for PKA and 9 for LCK) defined as compounds that reduced the signal for the kinase-catalytic phosphorylated reaction by >50% relative to the negative control reaction. IC50 Value Determination. We determine the IC50 values for the 11 primary kinase inhibitors to further demonstrate the ability of this RLS/fluorescence based microarray format. Figure 4 and (37) Yguerabide, J.; Yguerabide, E. E. J. Cell. Biochem. 2001, 37 (Suppl.), 71– 81.
Figures S3-S5 in the Supporting Information show the resulting array images and corresponding IC50 curves for the 11 inhibitors. In this case, the RLS/fluorescence intensity is decreased by increasing the concentration of inhibitor, indicating the relative levels of substrate phosphorylation and inhibition. The IC50 values of all inhibitors are summarized in Table 1. The IC50 values determined Table 1. IC50 Comparison and Detection Signal Recovery Rate IC50, found recovery rate (%) IC50, inhibitors kinase (RLS/fluorescence)a reported RLS/fluorescencea,b C7 F9 H11 C7 C10 D4 D7 D8 F2 F10 G2 H10
PKA PKA PKA LCK LCK LCK LCK LCK LCK LCK LCK LCK
44 µM/32 µM 429 nM/495 nM 12 nM/16 nM 1.7 µM/3.2 µM 7.1 µM/8.2 µM 4.4 µM/6.0 µM 2.8 µM/2.4 µM 22 nM/26 nM 5.4 nM/7.4 nM 90 nM/122 nM 7.1 µM/6.0 µM 168 nM/142 nM
none 570 nM38 7 nM39 none none none none 50 nM40 4 nM41 88 nM42 none 131 nM44
98/98 84/80 78/72 93/95 96/93 95/94 96/94 73/74 72/73 75/79 94/93 87/84
a Generally, the values of the IC50 and recovery rates obtained by RLS readout format agree with the results from the fluorescent readout format, and these values were comparable with the literature reported values. However, RLS can be excited by a broad spectrum white light source, which would greatly reduce the cost of instrumentation. b The concentration of ATP is 5 × 106 µM; the concentration of the inhibitor is 10 µM.
Analytical Chemistry, Vol. 82, No. 7, April 1, 2010
3071
Figure 5. The effect of various concentrations of ATP in the LCK solution on the light scattering images of the peptide microarrays and corresponding recovery curves for the inhibitors. All of recovery curves have similar scales as shown in the graph. The signals have been corrected for background noise (positive control) and normalized to the average RLS intensity obtained in the absence of inhibitors (negative control).
using the assay with either RLS or fluorescence readout format were comparable with the literature reported values.38-43 These results confirm that this microarray format has the potential to screen kinase inhibitors as well as determine the IC50 values. Most interesting, we found that staurosporine analogues (the molecular structures of these compounds are shown in Figure S6 in the Supporting Information) display a certain degree of inhibition of PKA and following the inhibition efficiency order, C7 > F9 > H11. This result suggests that the 5-oxo-5H-indolo[2,3a]pyrrolo[3,4-c]carbazole scaffold (part A) enables binding in the ATP-binding cleft of PKA and inhibits the enzymatic activity. In addition, the structure of the substituent group (part B) at 13N and 16N of 5-oxo-5H-indolo[2,3-a]pyrrolo[3,4-c]carbazole is also a critical factor of the inhibition efficiency of these compounds; a larger and more hydrophobic substituent group leads to stronger inhibition. This experimental result suggests that the approach could be used to study the binding affinity of the inhibitor with kinase and the design of new inhibitors. ATP Competition Assay. The effect of ATP on the inhibition efficiency of a specific inhibitor was also tested because most of the small molecule kinase inhibitors are ATP-competitors (type I inhibitor).9,44 As expected, the inhibition efficiencies of these inhibitors are decreased by increasing the ATP concentration in reaction mixtures (as shown in Figure 5 and Figures S7-S9 in the Supporting Information). This suggests that all of these inhibitors are bound (38) Fabbro, D.; Ruetz, S.; Bobis, S. Anti-Cancer Drug Des. 2000, 15, 17–28. (39) Couldwell, W. T.; Hinton, D. R.; He, S. FEBS Lett. 1994, 345, 43–46. (40) Burchat, A. F.; Calderwood, D. J.; Hirst, G. C.; Holman, N. J.; Johnston, D. N.; Munschauer, R.; Rafferty, P.; Tometzki, G. B. Bioorg. Med. Chem. 2000, 10, 2171–2174. (41) Traxler, P.; Bold, G.; Frei, J. J. Med. Chem. 1997, 40, 3601–3616. (42) Tian, G.; Cory, M.; Smith, A. A.; Knight, W. B. Biochemistry 2001, 40, 7084–7091. (43) Zhang, Q.; Liu, Y.; Gao, F.; Ding, Q.; Cho, C.; Hur, W.; Jin, Y.; Uno, T.; Joazeiro, C. A. P.; Gray, N. J. Am. Chem. Soc. 2006, 128, 2182–2183. (44) Young, P. R.; McLaughlin, M. M.; Kumar, S.; Kassis, S.; Doyle, M. L.; McNulty, D.; Gallagher, T. F.; Fisher, S.; McDonnell, P. C.; Carr, S. A.; Huddleston, M. J.; Seibel, G.; Porter, T. G.; Livi, G. P.; Adams, J. L.; Lee, J. C. J. Biol. Chem. 1997, 272, 12116–12121.
3072
Analytical Chemistry, Vol. 82, No. 7, April 1, 2010
with the corresponding kinase in a type I binding mode. For the compound with a low inhibiting ability (IC50 > 1 µM), the detection signal intensity of the phosphorylation can be nearly fully recovered (recovery rate > 90%). However, for the compound with a high inhibiting ability (IC50 < 90 nM), clear inhibition (>20%) can still be observed when the ratio of inhibitor and ATP is 1:5 × 104. This implies that these inhibitors bind much more tightly to the kinase to inactivate the phosphorylation reaction than ATP binds to kinase to activate the phosphorylation process. The recoveries of these inhibitors are also summarized in Table 1. The experimental result indicates that the microarray-based assay has great promise to study the inhibition mechanism of inhibitors. CONCLUSIONS A microarray-based spectroscopic assay for kinase functionality and inhibition has been developed. The assay circumvents the need for radioactive labeling usually required for assaying kinase activities and offers double spectroscopic detection procedures in microarray format to detect protein phosphorylation and inhibition. The utility of the assay has been demonstrated by primary (high-throughput screening) and secondary screening (IC50) using the protein kinases, PKA and LCK. This would open up possibilities for the future to develop microarray-based highthroughput technologies in drug discovery. ACKNOWLEDGMENT The authors thank the NSFC (Grant No. 20675080) for financial support. SUPPORTING INFORMATION AVAILABLE Additional information as noted in text. This material is available free of charge via the Internet at http://pubs.acs.org.
Received for review December 9, 2009. Accepted February 15, 2010. AC902804H
