Energy Transfer-Based Multiplexed Assay of ... - ACS Publications

May 6, 2008 - Rapid and sensitive assay of proteases and their inhibition in a high-throughput manner is of great significance in the diagnostic and p...
31 downloads 10 Views 1MB Size
Download PDF
Anal. Chem. 2008, 80, 4634–4641

Energy Transfer-Based Multiplexed Assay of Proteases by Using Gold Nanoparticle and Quantum Dot Conjugates on a Surface Young-Pil Kim,† Young-Hee Oh,† Eunkeu Oh,† Sungho Ko,‡ Min-Kyu Han,† and Hak-Sung Kim*,† Department of Biological Sciences, Korea Advanced Institute of Science and Technology (KAIST), Daejeon 305-701, Republic of Korea, and Korea Food Research Institute (KFRI), Sungnam-Si 463-746, Republic of Korea Rapid and sensitive assay of proteases and their inhibition in a high-throughput manner is of great significance in the diagnostic and pharmaceutical fields. We developed a multiplexed assay system of proteases and their inhibition by measuring the energy transfer between quantum dots (QDs) and gold nanoparticles (AuNPs) on a glass slide. In this system, while the photoluminescence (PL) of donor QDs immobilized on a surface was quenched due to the presence of AuNPs (energy acceptor) in close proximity, the protease activity caused modulation in the efficiency of the energy transfer between the acceptor and donor, thus enabling the protease assay. In comparison to the QD-dye system, the conjugate of the QD-AuNP gave rise to higher energy transfer efficiency, resulting in quantitative assay of proteases with more sensitivity. When matrix metalloproteinase, caspase, and thrombin were tested, a multiplexed assay was successfully achieved since the AuNP could be used as a common energy acceptor in conjunction with QDs having different colors. Our system is anticipated to find applications in the diagnosis of protease-related diseases and screening of potential drugs with high sensitivity in a high-throughput way. Proteases have been of particular interest because they are involved in major human diseases such as cancer, AIDS, inflammation, and neurodegenerative diseases,1–3 and typical examples include matrix metalloproteinases (MMPs), caspases, and thrombin. Thus, a method to assay proteases and their inhibition with high sensitivity in a multiplexed manner is of great significance in diagnosis of protease-relevant diseases and development of potential drugs.1–4 Over the past decade, a number of approaches to assay proteases have been reported mainly based on liquid chromatography or gel electrophoresis,5,6 but most of them have * To whom correspondence should be addressed. Phone: +82-42-869-2616. Fax: +82-42-869-2610. E-mail: [email protected]. † Korea Advanced Institute of Science and Technology (KAIST). ‡ Korea Food Research Institute (KFRI). (1) Overall, C. M.; Kleifeld, O. Nat. Rev. Cancer 2006, 6, 227–239. (2) Concha, N. O.; Abdel-Meguid, S. S. Curr. Med. Chem. 2002, 9, 713–726. (3) Schwienhorst, A. Cell. Mol. Life Sci. 2006, 63, 2773–2791. (4) White, C. M. Am. Heart J. 2005, 149, S54-S60. (5) Bjurlin, M. A.; Bloomer, S.; Nelson, C. J. Biotechnol. Lett. 2002, 24, 191– 195. (6) Zhao, Z.; Raftery, M. J.; Niu, X. M.; Daja, M. M.; Russell, P. J. Electrophoresis 2004, 25, 1142–1148.

4634

Analytical Chemistry, Vol. 80, No. 12, June 15, 2008

a limitation on a multiplexed and high-throughput assay. With the progress of fluorescence-based techniques,7,8 much attention has been paid to Fo¨rster resonance energy transfer (FRET) for in vitro and in vivo assay of proteases.9–12 In the FRET-based system, quantum dots (QDs) have been widely used due to their intrinsic optical properties including high quantum yield, less photobleaching, and size-tunable photoluminescence with broad excitation and narrow emission bandwidths.13,14 In particular, multiple binding of an energy acceptor per QD is expected to increase the overall energy transfer efficiency.15,16 Recently, it was proposed that the use of gold nanoparticles (AuNPs) as an energy acceptor extends the effective energy transfer distance up to 20 nm, resulting in a high energy transfer efficiency.17–21 AuNP is known to have a superior quenching efficiency in a broad range of wavelengths compared to other organic quenchers. We previously reported an inhibition assay of a target analyte in solution by using the energy transfer between QDs and AuNPs.22,23 (7) Harris, J. L.; Backes, B. J.; Leonetti, F.; Mahrus, S.; Ellman, J. A.; Craik, C. S. Proc. Natl. Acad. Sci. U.S.A. 2000, 97, 7754–7759. (8) Marme, N.; Knemeyer, J. P.; Wolfrum, J.; Sauer, M. Angew. Chem., Int. Ed. 2004, 43, 3798–3801. (9) Shi, L.; De Paoli, V.; Rosenzweig, N.; Rosenzweig, Z. J. Am. Chem. Soc. 2006, 128, 10378–10379. (10) Medintz, I. L.; Clapp, A. R.; Brunel, F. M.; Tiefenbrunn, T.; Uyeda, H. T.; Chang, E. L.; Deschamps, J. R.; Dawson, P. E.; Mattoussi, H. Nat. Mater. 2006, 5, 581–589. (11) Chang, E.; Miller, J. S.; Sun, J.; Yu, W. W.; Colvin, V. L.; Drezek, R.; West, J. L. Biochem. Biophys. Res. Commun. 2005, 334, 1317–1321. (12) Stockholm, D.; Bartoli, M.; Sillon, G.; Bourg, N.; Davoust, J.; Richard, I. J. Mol. Biol. 2005, 346, 215–222. (13) Bruchez, M., Jr.; Moronne, M.; Gin, P.; Weiss, S.; Alivisatos, A. P. Science 1998, 281, 2013–2016. (14) Willard, D. M.; Mutschler, T.; Yu, M.; Jung, J.; Van Orden, A. Anal. Bioanal. Chem. 2006, 384, 564–571. (15) Clapp, A. R.; Medintz, I. L.; Mauro, J. M.; Fisher, B. R.; Bawendi, M. G.; Mattoussi, H. J. Am. Chem. Soc. 2004, 126, 301–310. (16) Zhang, C. Y.; Yeh, H. C.; Kuroki, M. T.; Wang, T. H. Nat. Mater. 2005, 4, 826–831. (17) Yun, C. S.; Javier, A.; Jennings, T.; Fisher, M.; Hira, S.; Peterson, S.; Hopkins, B.; Reich, N. O.; Strouse, G. F. J. Am. Chem. Soc. 2005, 127, 3115–3119. (18) Jennings, T. L.; Singh, M. P.; Strouse, G. F. J. Am. Chem. Soc. 2006, 128, 5462–5467. (19) Jennings, T. L.; Schlatterer, J. C.; Singh, M. P.; Greenbaum, N. L.; Strouse, G. F. Nano Lett. 2006, 6, 1318–1324. (20) Dulkeith, E.; Morteani, A. C.; Niedereichholz, T.; Klar, T. A.; Feldmann, J.; Levi, S. A.; van Veggel, F. C.; Reinhoudt, D. N.; Moller, M.; Gittins, D. I. Phys. Rev. Lett. 2002, 89, 203002. (21) Pons, T.; Medintz, I. L.; Sapsford, K. E.; Higashiya, S.; Grimes, A. F.; English, D. S.; Mattoussi, H. Nano Lett. 2007, 7, 3157–3164. (22) Oh, E.; Hong, M.-Y.; Lee, D.; Nam, S.-H.; Yoon, H. C.; Kim, H.-S. J. Am. Chem. Soc. 2005, 127, 3270–3271. 10.1021/ac702416e CCC: $40.75  2008 American Chemical Society Published on Web 05/06/2008

Despite numerous attempts employing the energy transfer, however, demand for an assay system on a surface is very high because it allows a simple and sensitive assay in a high-throughput way compared to a solution-based one. In addition, it enables more reliable analyses with no aggregation of nanoparticles and much smaller reaction volume (a few nanoliters). Here we demonstrate a multiplexed assay of proteases and their inhibition by using AuNP and QD conjugates as nanoprobes on a glass slide. Unlike other fluorescent assays employing a single fluorophore-based intensity, our energy transfer-based approach can effectively reduce photobleaching and background noise. The chemical feature of the QD-AuNP conjugate was also investigated in comparison with the QD-dye conjugates in terms of the energy transfer efficiency, quenching constant, and maximum effective distance. Details are reported herein. EXPERIMENTAL SECTION Materials. Matrix metalloproteinase-7 (MMP-7) was purchased from Calbiochem. Matrix metalloproteinase-2 (MMP-2) and caspase-3 were purchased from Sigma. Thrombin was purchased from Novagen. Monomaleimide-functionalized gold nanoparticles (1.4 nm in diameter) were purchased from Nanoprobe. Strepetavidin-conjugated QDs with different colors (SAQD525, SA-QD605, and SA-QD655) were obtained from Invitrogen. N-Hydroxy-1,3-di-(4-methoxybenzenesulfonyl)-5,5-dimethyl[1,3]-piperazine-2-carboxamide (NHDDPC) was purchased from Calbiochem, and N-CBZ-VAL-ALA-ASP fluorometryl ketone (ZVAD-FMK) and 4-(2-aminoethyl)benzenesulfonyl fluoride hydrochloride (AEBSF) were from Sigma. A hydrogel-type aminereactive glass (Nexterion Slide H) was obtained from Schott Nexterion (Germany). BioPepMMP7 (CRPLALWRSK-biotin) and BioPepTHR (biotin-GKGGLVPRGSGC-NH3), RPLALWRSK-biotin, and CRPLALWRSK were synthesized from Peptron Inc. (Korea). PepCASP (biotin-GRRGDEVDGGRRC-NH3) and Cy5-labeled peptide (Cy5-RPLALWRSK-biotin) were synthesized from AnyGen Inc. (Korea). Silver enhancer kit was purchased from Sigma. Synthesis of Peptide-Conjugated AuNPs. Monomaleimidefunctionalized AuNPs (Nanogold, 1.4 nm in diameter) were dissolved in 100 µL of dimethyl sulfoxide (DMSO) and then diluted to 1 mL with distilled water (final concentration ) 6 nmol mL-1). In the typical coupling reaction of the AuNPs with the respective biotinylated, cysteine-terminated peptide, a solution of AuNPs was added to an equivalent volume of cysteine-terminated peptide dissolved in distilled water. A 10-fold molar excess of peptide to the AuNP was used to obtain sufficient binding of the peptides to the AuNPs. The resulting solution was incubated for 18 h at 4 °C without stirring. After incubation, the solution was spin-filtered by using a microcentricon (YM-10) with an exclusion cutoff of 10 000 Da to separate the unbound peptides from the AuNP conjugates (AuNP clusters typically have a molecular weight of 15 000), followed by subsequently washing with distilled water. The concentrated conjugates were resuspended in distilled water or appropriate buffer solution. The final concentration was determined from the UV-vis spectrum of the conjugate. Since most peptides have no absorbance at 420 nm, (23) (a) Oh, E.; Lee, D.; Kim, Y.-P.; Cha, S. Y.; Oh, D.-B.; Kang, H. A.; Kim, J.; Kim, H.-S. Angew. Chem., Int. Ed. 2006, 45, 7959–7963. (b) Kim, Y.-P.; Oh, Y.-H; Oh, E.; Kim, H.-S. Biochip J. 2007, 1, 228–233.

the molar concentration of the AuNP-peptide conjugates was calculated by dividing the measured absorbance by the extinction coefficient (1.1 × 105 M-1 cm-1) for AuNPs at 420 nm (the manufacturer’s instructions). As a control, free AuNPs were synthesized by the overnight binding of monomaleimide AuNP and 10-fold excess β-mercaptoethanol. Unbound β-mercaptoethanol was extracted by using a microcentricon (YM-10). The resulting free AuNPs were retrieved and quantified by spectrometric analysis. Protease Assay in Solution. For quenching experiments, the SA-QD605 was mixed with varying amounts of either peptideconjugated AuNPs or Cy5 in a 96-well plate for 1 h at room temperature to allow specific association between SA and biotin. The final concentration of the QDs in aqueous solution was typically 10 nM (corresponding to 1 pmol in a 100 µL of reaction volume). After incubation, the fluorescence was measured at an excitation wavelength of 430 nm by using a microplate reader (Infinite M200, TECAN, Austria). The mean and standard deviation of the intensity were calculated from quadruple experiments. The relative quenching efficiency (QE) of an energy acceptor for QDs was determined by measuring the photoluminescence (PL) intensity of QDs in the presence and absence of energy acceptor. The concentrations of SA-QD605 and Cy5-peptide-biotin were determined from the extinction coefficients of QD605 (1.1 × 106 M-1 cm-1 at 488 nm) and Cy5 (2.5 × 105 cm-1 M-1 at 649 nm), respectively. Protease assay was initiated by addition of a 10 µL protease solution to microwells containing nanoprobes composed of QD-AuNP or QD-Cy5. Each enzyme (MMP-7, caspase, and thrombin) was dissolved in 10 mM HEPES buffer (pH 7.4, 150 mM NaCl, 5 mM CaCl2) at a final concentration of 100 ng mL-1. The 96-well plate was incubated at 37 °C for 60-120 min with a tight seal to prevent water evaporation. The equal amount of nonspecific protease was used as a control experiment. Following a protease reaction, PL intensity in each well was measured. The resulting data were analyzed by using Excel (Microsoft 2003) or SigmaPlot (ver. 10.0) to estimate the PL intensity and relative quenching efficiency. Protease Assay on a Glass Slide. SA-QDs with different colors (QD525, QD605, and QD655) were arrayed directly onto NHS-derivatized hydrogel glass slide (Schott Nexterion) in quadruple spot format by using a robotic arrayer (Microsys, Cartesian Technologies, Irvine, CA) equipped with CMP 3 spotting pins (Telechem International, Sunnyvale, CA). The spotting pin repetitively delivered approximately 1 nL of 10 nM QDs solution. Following a spotting, the slide was incubated in a chamber at 60-70% relative humidity and room temperature for 1 h and then immersed in a solution of 2% BSA (50 mM borate buffer pH 8.5) for 1 h to block the remaining NHS groups. The slide was then rinsed with distilled water and subsequently incubated in a solution containing 500 nM peptide-conjugated AuNPs. For protease assay, a multiwell-type chambered silicon coverslip (φ 3 mm × H 1 mm, Sigma) was overlaid onto a glass slide with exposure of the spotting area in each well. A 10 µL volume of enzyme solution (10 mM HEPES buffer pH 7.4, 150 mM NaCl, 5 mM CaCl2) containing the respective proteases (100 ng mL-1 of MMP-7, caspase-3, or thrombin) was added to the wells formed by the chambered silicon coverslip; this surface was placed in a plate and incubated at 37 °C for 60-120 min. Nonspecific protease was used as negative control experiments. Inhibitor assay for Analytical Chemistry, Vol. 80, No. 12, June 15, 2008

4635

Scheme 1. Schematic Principle of Protease Assay by Using the AuNP-QD Conjugates as Nanoprobes on a Glass Slidea

a

(A) Construction of the peptide-conjugated AuNP; (B) changes in the PL intensities of nanoprobes by the protease action in the absence and presence of its inhibitor.

Figure 1. Changes in the PL intensity of donor QD605 by different types of acceptors (quenchers) as a function of the molar ratio between donor and acceptor: (A) Pep-AuNPs, (B) Pep-Cy5, and (C) free AuNPs. The molar ratios of acceptor to donor were varied: donor only, 1:1, 2.5:1, 5:1, 10:1, 20:1, 30:1, 40:1, and 50:1 (from top to bottom). (D) Quenching efficiencies (QE) as a function of acceptor concentration: Pep-AuNPs (closed circle), Pep-Cy5 (open circle), and free AuNPs (closed triangle). The concentration of donor QDs was 10 nM. The QE was calculated as described in the Supporting Information. The error bars indicate the standard deviation in quadruplicate experiments.

respective protease was performed by incubation of protease solution in the presence of a specific inhibitor (NHDDPC, Z-VADFMK, or AEBSF). After incubation, the surface was rinsed three times with distilled water, dried with a stream of N2, and then subjected to fluorescence analysis. 4636

Analytical Chemistry, Vol. 80, No. 12, June 15, 2008

Fluorescence Readout. Fluorescence scanning was carried out by using an ArrayWoRxe slide scanner (Applied Precision, U.S.A.) equipped with a white-light CCD camera. For measurements of the PL intensities from different QDs, the glass slide was scanned using an emission filter (Chroma Tech. Corp., U.S.A.)

Table 1. Typical Features in the Energy Transfer of QD-AuNP and QD-Cy5 in Solution donor

acceptor

expected distance (nm)a

R0 (nm)b

QE (%) (n ) 50)c,d

r (nm)e

SA-QD605

AuNP-CRPLALWRSK-biotin Cy5-RPLALWRSK-biotin

11.3-13.8 10.9-13.4

5.8 6.4

83.2 ± 6.0 60.5 ± 3.2

8.5 ± 0.7 11.4 ± 0.3

a An expected distance is a theoretical maximum distance from the core of the donor to the surface of acceptor that takes into account the full extension of the peptide sequence and the radius of the donor SA-QD. b The Fo ¨rster distance (R0) values of QD-AuNP and QD-Cy5 were calculated by using Fo ¨rster formula (eq 1 in the Supporting Information) and the overlap integral (eq 2 in the Supporting Information). c The quenching efficiency (QE) was calculated according to eq 3 in the Supporting Information. d The binding ratio (n) was estimated from the saturated molar ratio of the donor to the acceptor where the QE did not increase further. e The separation distance (r) was determined according to eq 4 in the Supporting Information.

of 525 (QD525), 605 (QD605), or 655 nm (QD655), respectively, with a common excitation filter of 460 nm (Chroma Tech. Corp., U.S.A.). Multiplexed images were obtained by merging the surface images which were successively scanned using different QD filter sets. After scanning, the spot fluorescence was analyzed using an imaging software (GenePix Pro 4.0, Axon), and the mean value and standard deviation were calculated using the quadruple spots. RESULTS AND DISCUSSION Energy Transfer between QDs and AuNPs. Scheme 1 illustrates our chip-based assay principle. A biotinylated peptide substrate for protease is conjugated to a monomaleimide-functionalized AuNP, and the resulting AuNPs (Pep-AuNPs) are associated with streptavidin (SA)-bound QDs (SA-QDs) that had been deposited on a glass slide to form the AuNPs-QDs conjugates of which the PL is quenched. Addition of protease causes cleavage of the peptide substrate on the AuNP-QD conjugates, resulting in detachment of the AuNPs from the QDs, and the PL of QDs is recovered. Protease inhibitor, however, prevents the recovery of PL of QDs by blocking the action of protease. Thus, our assay relies on modulations in the PL intensities of the AuNP-QD conjugates on a glass surface by the protease action. When considering our assay principle shown in Scheme 1, it is highly desirable for a sensitive protease assay that the AuNP-QD conjugates have a background fluorescence level as low as possible, which is achieved by maximizing the PL quenching of SA-QDs by the Pep-AuNPs. To check this possibility, we first constructed the nanoprobes for assay of MMP-7 and investigated their typical feature in terms of the quenching efficiency (QE) in solution. To accomplish this, the biotinylated peptide substrates for MMP-7 (CRPLALWRSKbiotin) had been conjugated to the monomaleimide-functionalized AuNPs (1.4 nm in diameter, Nanoprobe). Different concentrations of the resulting Pep-AuNPs (AuNP-CRPLALWRSKbiotin) were added to the SA-QDs having a maximum emission at 605 nm, and emission spectra of the constructed nanoprobes were measured. As shown in Figure 1A, the PL intensity decreased with increasing molar ratio of Pep-AuNPs to QDs. In the case of Cy5-labeled peptide substrate (Pep-Cy5; Cy5-CRPLALWRSK-biotin) conjugated to QDs, the PL quenching of QDs was less than that of QD-AuNP over the tested molar ratios, along with the marginal emission increase of Cy5 at 670 nm (Figure 1B). As another control, free AuNPs without peptide substrate were added to the solution containing QDs, and reduction in the PL intensity of QDs was smaller than those

of other acceptors (Figure 1C). We examined the changes in the QE for different types of quenchers as a function of molar ratio between the QD and the respective quencher. As shown in Figure 1D, increase in the QE was most significant by the Pep-AuNPs. Unlike Pep-AuNPs, free AuNPs followed a Stern-Volmer model (see Figure S1 in the Supporting Information). Certainly, a low quenching constant (KD) of free AuNPs reveals the diffusion-driven quenching effect rather than the affinity-induced one. In the case of Pep-Cy5, the theoretical maximum distance between donor and acceptor was expected to range from 10.9 to 13.4 nm (Table 1), which seems to be responsible for the retarded quenching process. This result clearly demonstrates that the energy transfer between QDs and Pep-AuNPs occurred effectively even in a longer distance mainly through the interaction between the donors and acceptors via specific biotin-SA affinity. The use of QDs with different colors also resulted in a similar QE (see Figure S2 in the Supporting Information), which confirms that the AuNPs can be employed as common energy acceptors for different kinds of QDs. The AuNPs might approach the QD energy donor, to a shorter distance than the extended chain distance, due to thermal motions of linker chains.24 To get some insights into this issue, we tested the peptide-conjugated dye as a control set. The QE and separation distance (r) for both QD-AuNP and QD-Cy5 pairs were compared. Although the QD-Cy5 has a longer Fo¨rster distance (R0) than the QD-AuNP (Table 1), the QE of Cy5 was lower than that of AuNP over the tested molar ratios. As a result, the QD-Cy5 has a relatively long separation distance (r ) 11.4 nm), approaching 2R0. If thermal drift is dominant in peptide-conjugated dye, a higher quenching efficiency would be observed than expected; however, the quenching efficiency of peptide-conjugated dye was found to fit well the expected value by a conventional FRET mechanism (Table 1). On the other hand, the energy transfer between the QD and the AuNP led to a shorter separation distance of 8.3 nm than the expected one (11.3-13.8 nm) of core-to-surface between the QD and the AuNP. On the basis of this observation, thermal motions of the peptide linker chains seem not to be significant in our system, even though this possibility could not be completely excluded. A recent report showed that the energy transfer between peptide-linked AuNPs and QDs can effectively occur even beyond the maximum distance in the conventional FRET mechanism.21 In addition, it has been reported elsewhere that the use of AuNPs as an energy acceptor (24) Aguila, A.; Murray, R. W. Langmuir 2000, 16, 5949–5954.

Analytical Chemistry, Vol. 80, No. 12, June 15, 2008

4637

Figure 2. Detection of protease activity by using the QD-AuNP conjugates in solution. (A) Changes in the PL intensity of AuNP-QD conjugates by MMP-7: QD only (black circle), QD-Pep-AuNP (gray triangle), additions of MMP-7 (dark gray square), and MMP-2 (gray diamond). (B) Changes in relative PL of respective AuNP-QD conjugates by caspase and thrombin: SA-QDs only (black), QD-Pep-AuNP (gray), and additions of caspase and thrombin (dark gray). (C) UV-gel electrophoresis images. The SA-QDs were associated with different acceptors, and their migrations were visualized: lane 1, SA-QDs only; lane 2, SA-QDs + RPLALWRSK; lane 3, SA-QDs + RPLALWRSK-biotin; lane 4, SA-QDs + AuNP-CRPLALWRSK; lane 5, SA-QDs + AuNP-CRPLALWRSK-biotin; lane 6, a reaction mixture of MMP-7 with a sample in lane 5 followed by filtration and concentration. A 0.8% agarose gel was used in 1× TAE and run at 100 V for 30 min. The molar ratio of different acceptors per QD was equally maintained at 50.

results in deviation from conventional FRET.17–20 Thus, it is likely that higher quenching efficiency in our energy transfer system might result from the intrinsic property of AuNP rather than thermal motions of the linker chains. The QE between QDs and Pep-AuNPs reached a maximum level of about 83% at the molar ratio of 50 (Table 1). Further increases in the molar ratio did not result in greater enhancement of QE (data not shown). Self-absorbance of metal nanoparticles can be more critical at higher concentrations and might result in the saturation curve as shown in Figure 1D due to reduced quenching ability. However, over a range of 100-500 nM in Figure 1D, free AuNPs showed a relatively linear curve with the increasing concentration, which resulted from diffusion-driven quenching effect as aforementioned. In contrast, biotin-Pep-AuNP and biotin-Pep-Cy5 displayed more saturated curve. We presumed that this discrepancy could be caused by saturation in a binding between the SA-QD and the biotin-Pep-AuNPs, rather than by self-absorbance. The saturated QE at the molar ratio of 50, therefore, may represent the maximum number (∼50) of bound AuNPs per QD. Considering that the number of SA per QD is about 5-10 according to the manufacturer’s manual, total biotin binding sites per QD would be 10-30, since typically 2-3 biotin binding sites are available on a single SA molecule. On the basis of the observation that the estimated number (∼50) from the saturated QE was higher than theoretical binding level (10-30) of AuNPs per QD, we cannot completely exclude self-absorbance effect of AuNPs at high concentration. Protease Assay in Solution Using the Nanoprobes. With the constructed nanoprobes, we first attempted to assay MMP-7 4638

Analytical Chemistry, Vol. 80, No. 12, June 15, 2008

activity in solution. As shown in Figure 2A, a notable increase in the PL of QDs was observed by addition of MMP-7 which specifically cleaves the Ala-Lys bond in the peptide sequence, and approximately 47% of the initial PL of QDs was recovered. As a control, nonspecific protease (MMP-2) was tested, but change in the PL of nanoprobes was negligible. The PL of the nanoprobes was not fully recovered even by addition of MMP7, and this might be attributed to limited accessibility of MMP-7 to the peptide substrate as a result of the steric effect by the peptide substrate itself on the surface of QDs. To check the applicability of this assay system to other proteases, we tested caspase-3 and thrombin. For this, the respective peptide substrates were conjugated to the AuNPs and employed for the construction of nanoprobes for protease assay. As shown in Figure 2B, addition of caspase-3 and thrombin to the respective nanoprobes also gave rise to a significant recovery in the PL of QDs. UV-gel electrophoresis further supports the association between the donor and the acceptor via specific biotin-SA interaction and the recovery in the PL of QDs by the proteolytic activities (Figure 2C). The mobility of negatively charged SA-QDs was retarded only when incubated with the biotinylated peptide (lane 3), compared to the control SA-QDs (lane 1), the SA-QDs conjugated with biotin-free peptides (lane 2), or the biotin-free Pep-AuNPs (lane 4). The biotinylated Pep-AuNPs specifically caused the quenching of the PL intensity of SA-QDs (lane 5), and addition of protease to lane 5 led to a notable recovery in the PL of QDs in a gel (lane 6). Protease Assay Using the Nanoprobes on a Glass Slide. We tested a chip-based assay of proteases and their inhibition using the energy transfer between QDs and AuNPs on a glass

Figure 3. Chip-based assay of MMP-7 and its inhibition by using QD605-AuNP conjugates. (A) SA-QDs only. (B) SA-QDs + Pep-AuNPs. (C) SA-QDs + Pep-AuNPs + MMP-7. (D) SA-QDs + Pep-AuNPs + MMP-7 + inhibitor. Changes in the relative PL intensity and the standard deviation were determined from the quadruple spots.

Figure 4. (A) Fluorescence images from the AuNP-QD conjugates on a glass slide at different concentrations of MMP-7. The calibration curves for (B) MMP-7, (C) caspase-3, and (D) thrombin. The error bars indicate the standard deviation in quadruple spots. The protease activity represents the concentration of cleaved peptides/min (nM min-1).

slide. For this, SA-QDs were deposited on a glass slide in quadruple spots by using a microarrayer. The AuNPs that had been conjugated with the biotinylated peptide substrate for MMP-7 were added to the SA-QDs at the molar ratio of 50 for the construction of nanoprobes, and the change in the PL intensity of the nanoprobes was traced in the presence of MMP-7 and its inhibitor. Association of AuNPs with SA-QDs on a glass slide

was also examined using a silver staining method. As expected, SA-QDs deposited on a glass surface displayed a strong PL intensity as in solution (Figure 3A). However, addition of Pep-AuNPs onto the surface-confined SA-QDs resulted in a drastic decrease in the PL intensity below 20% of the original level (Figure 3B), and a strong signal was observed from a silver staining. This result indicates that the PL quenching was caused Analytical Chemistry, Vol. 80, No. 12, June 15, 2008

4639

Figure 5. Multiplexed assay of proteases by using QDs with different colors on a glass slide. SA-QD525, SA-QD605, and SA-QD655 were used (from left to right). Biotinylated peptide substrates for MMP-7, caspase-3, and thrombin were conjugated to the AuNPs, and then the resulting Pep-AuNPs were associated with SA-QD525, SA-QD605, and SA-QD655, respectively. (A) SA-QDs only. (B) SA-QDs + respective Pep-AuNPs. (C) SA-QDs + Pep-AuNPs + MMP-7. (D) SA-QDs + Pep-AuNPs + caspase-3. (E) SA-QDs + Pep-AuNPs + thrombin. (F) QDs + Pep-AuNPs + mixture of the respective protease and its inhibitor.

by association of the AuNPs with the SA-QDs. When MMP-7 was added and the surface was washed with a buffer solution, the PL of QDs was recovered up to 64% level, and the signal from the silver staining became very weak (Figure 3C). This observation reflects that the AuNPs were detached from the nanoprobes due to cleavage of the peptide substrate by the action of MMP-7. The recovery of PL on a glass slide was found to be somewhat higher than that in a solution-based assay, and this might be attributed to wash-out of the detached AuNPs from the spots and negligible aggregation of the nanoparticles. The presence of inhibitor for MMP-7 gave rise to a negligible change in the PL intensity (Figure 3D), and a similar signal was observed from a silver staining to that in Figure 3B. From these results, it is obvious that our approach by using the nanoprobes could be effectively used for a chip-based assay of MMP-7 and its inhibition. We determined the calibration curve and the detection limit for MMP-7 on a surface. The calibration curve was obtained by plotting the protease activity as a function of protease concentration. The protease activity was calculated by converting the recovered PL intensity before and after protease reaction to the amount of cleaved peptide/min (nM min-1) by using a standard curve (see Figure S3 in the Supporting Information). As a result, the PL intensity of QDs on a glass surface became stronger as the concentration of MMP-7 increased (Figure 4A). The activity of MMP-7 was linearly proportional to the logarithmic concentration of MMP-7, ranging from as low as 10 ng mL-1 to 5 µg mL-1 4640

Analytical Chemistry, Vol. 80, No. 12, June 15, 2008

(Figure 4B), which is within a favorable range, when considering the secretion levels in malignant tissues (