Langmuir 2007, 23, 9513-9521
9513
Surface Immobilization of Catalytic Beacons Based on Ratiometric Fluorescent DNAzyme Sensors: A Systematic Study Daryl P. Wernette,† Carolyn Mead,† Paul W. Bohn,‡ and Yi Lu*,† Department of Chemistry and Beckman Institute for AdVanced Science and Technology, UniVersity of Illinois at Urbana-Champaign, Urbana, Illinois 61801, and Department of Chemical and Biomolecular Engineering and Department of Chemistry, UniVersity of Notre Dame, Notre Dame, Indiana 46556 ReceiVed May 4, 2007. In Final Form: June 21, 2007 DNAzyme-based catalytic beacons have the potential for sensing a large number of relevant analytes. Thus, a systematic investigation of factors affecting their performance when immobilized into gold-coated nanocapillary array membranes (NCAMs) was undertaken. Enzyme immobilization times were varied to determine that as little as 15 min was sufficient for ratiometric detection of Pb2+-specific activity, while immobilization density saturated after 1.5 h. Immobilization of the DNAzymes into NCAMs with 600 nm pore size resulted in higher immobilization efficiency and higher enzymatic activity than that with 200 nm pore size. A poly-T linker length between the tethering thiol and first oligonucleotide, used to extend the DNAzyme above the backfilling mercaptohexanol (MCH) monolayer, had no effect on DNAzyme activity. The backfilling method of immobilization, involving backfilling followed by hybridization, was found most effective for DNAzyme activity compared to immobilization of hybridized DNAzyme complex (a 67% loss of activity) or concurrent enzyme and MCH immobilization (75% loss of activity). The backfilling MCH monolayer provided ∼3.5 times increase in activity compared to DNAzyme assembled without MCH, and was over 5 times more active than shorter and longer backfilling molecules tested. The immobilized DNAzyme retained its optimized performance at 50 mM NaCl. Finally, the generalized immobilization and ratiometric procedure was employed for a uranyl-specific DNAzyme with 25 ( 15 times increase in ratio observed. These findings form a firm basis on which practical applications of catalytic beacons can be realized, including sensors for both Pb2+ and UO22+ ions.
Introduction Molecular beacons are single-stranded oligonucleotides that form a stem-loop structure with a fluorophore and a quencher at each end (Figure 1a).1-9 The fluorescence is quenched when it is free in solution, but the molecular beacon becomes fluorescent when it is hybridized to an oligonucleotide with complementary sequence, because the binding results in physical separation of the fluorophore and the quencher. Since their discovery in the mid-1990s, molecular beacons have been widely used to detect oligonucleotides in a number of applications, such as real-time detection of DNA/RNA hybridization in living cells and clinical assays.10-14 To expand the molecular diagnostic targets beyond DNA or RNA, we reported a catalytic beacon method using DNAzymes * Author to whom correspondence should be addressed: e-mail yi-lu@ uiuc.edu. † University of Illinois at Urbana-Champaign. ‡ University of Notre Dame. (1) Tyagi, S.; Kramer, F. R. Nat. Biotechnol. 1996, 14, 303-308. (2) Du, H.; Disney, M. D.; Miller, B. L.; Krauss, T. D. J. Am. Chem. Soc. 2003, 125, 4012-4013. (3) Du, H.; Strohsahl, C. M.; Camera, J.; Miller, B. L.; Krauss, T. D. J. Am. Chem. Soc. 2005, 127, 7932. (4) Fang, X.; Liu, X.; Schuster, S.; Tan, W. J. Am. Chem. Soc. 1999, 121, 2921-2922. (5) Rajendran, M.; Ellington, A. D. Nucleic Acids Res. 2003, 31, 5700-5713. (6) Nutiu, R.; Mei, S.; Liu, Z.; Li, Y. Pure Appl. Chem. 2004, 76, 1547-1561. (7) Nutiu, R.; Li, Y. Chem.sEur. J. 2004, 10, 1868-1876. (8) Mei, S. H. J.; Liu, Z.; Brennan, J. D.; Li, Y. J. Am. Chem. Soc. 2003, 125, 412-420. (9) Nutiu, R.; Yu, J. M. Y.; Li, Y. ChemBioChem. 2004, 5, 1139-1144. (10) Dirks, R. W.; Molenaar, C.; Tanke, H. J. Methods 2003, 29, 51. (11) Jebbink, J.; Bai, X.; Rogers, B. B.; Dawson, D. B.; Scheuermann, R. H.; Domiati-Saad, R. J. Mol. Diagn. 2003, 5, 15. (12) Landry, M. L.; Garner, R.; Ferguson, D. J. Clin. Microbiol. 2005, 43, 3136. (13) Santangelo, P. J.; Nix, B.; Tsourkas, A.; Bao, G. Nucleic Acids Res. 2004, 32, e57. (14) Yi, J. Z.; Zhang, W. D.; Zhang, D. Y. Nucleic Acids Res. 2006, 34, e81.
Figure 1. Representation of (a) molecular beacon and (b) DNAzyme catalytic beacon constructs for fluorescence based sensing of targets. The target oligonucleotide (a) induces an opening of the molecular beacon structure, or metal ion analyte (b) induces DNAzyme substrate cleavage and release of a fluorescently tagged oligonucleotide fragment.
(also called catalytic DNA or deoxyribozymes) as shown in Figure 1b.15-17 We define catalytic beacons as signaling systems that contain a fluorophore/quencher pair labeled on a catalytic nucleic acid module. DNAzymes are based on single-stranded DNA exhibiting catalytic functions such as RNA cleavage. Although not found in nature yet, DNAzymes that are highly specific for a number of metal ions such as Pb2+, Cu2+, Co2+, and UO22+ have been isolated through in vitro selection.15,18-22 We (15) Li, J.; Lu, Y. J. Am. Chem. Soc. 2000, 122, 10466-10467. (16) Liu, J.; Lu, Y. Anal. Chem. 2003, 75, 6666-6672. (17) Liu, J.; Lu, Y. Methods Mol. Biol. (Totowa, NJ, U.S.) 2006, 335, 275288. (18) Li, Y.; Breaker, R. R. Curr. Opin. Struct. Biol. 1999, 9, 315-323. (19) Carmi, N.; Breaker, R. R. Bioorg. Med. Chem. 2001, 9, 2589-2600. (20) Emilsson, G. M.; Breaker, R. R. Cell Mol. Life Sci. 2002, 59, 596-607. (21) Lu, Y. Chem.sEur. J. 2002, 8, 4588-4596. (22) Liu, J.; Brown, A. K.; Meng, X.; Cropek, D. M.; Istok, J. D.; Watson, D. B.; Lu, Y. Proc. Natl. Acad. Sci. U.S.A. 2007, 104, 2056.
10.1021/la701303k CCC: $37.00 © 2007 American Chemical Society Published on Web 08/04/2007
9514 Langmuir, Vol. 23, No. 18, 2007
subsequently converted these DNAzymes into highly sensitive and selective fluorescent sensors using the catalytic beacon methodology, in which a DNAzyme enzyme strand with a quencher at the 3′-end is hybridized to a substrate strand with a fluorophore at the 5′-end (Figure 1b). In the absence of the metal ion, the system remains minimally fluorescent. In the presence of metal ion, the DNAzyme catalyzes cleavage of the ribonucleotide in the middle of the DNA substrate strand, releasing the DNA fragment containing the fluorophore, resulting in a dramatic increase of fluorescent signal amplitude. For example, a recently reported catalytic beacon for UO22+ has over 1 millionfold selectivity over competing divalent metal ions and a detection limit of 45 pM, better even than inductively coupled plasma mass spectrometry, a standard instrumental analysis method.22 Such a method has been expanded to detect nonmetal ions, such as small organic molecules or complementary DNA, by using allosteric DNA/RNAzymes or aptazymes.23-32 A different catalytic beacon design, placing fluorophore/quencher pairs closer to the cleavage site rather than the ends, has also been demonstrated.6-8,33 Yet another class of catalytic beacons reported is catalytic molecular beacon structures that, by definition, contain a stem-loop module that acts as an antisense conformational trigger with a catalytic nucleic acid.34-37 While the catalytic beacon has been demonstrated in solution, practical application of this method requires advances in a number of areas. A critical issue is nonspecific dehybridization of the enzyme and substrate strand; the catalytic beacon design requires a fine balance between stable hybridization of the enzyme and substrate strands before cleavage and rapid release of the product strands after cleavage. Nonspecific dehybridization in the absence of the analyte can result in high fluorescent background, lower signal-to-noise ratio, and poorer (i.e., higher) detection limits. To address this issue, we immobilized the catalytic beacon on a gold surface, allowing a facile wash of the surface to remove the nonspecifically dehybridized oligonucleotides before detection, resulting in ∼10-fold improvement of detection limit over solution-based systems.38 Furthermore, we introduced a ratiometric fluorescent detection system in which a noncleavable substrate DNA strand was labeled with a different fluorophore from that of cleavable substrate DNA strand; the noncleavable substrate DNA strand had a sequence identical with that of the cleavable substrate except that the scissile ribonucleotide was replaced with a deoxyribonucleotide.39 In this way, the ratiometric system accounted for nonspecific dehybridization as well as differences in DNAzyme surface coverage among different batches. Finally, to improve mass transport of analytes to the surface and increase total fluorescent signals released from the surface, (23) Jose, A. M.; Soukup, G. A.; Breaker, R. R. Nucleic Acids Res. 2001, 29, 1631-1637. (24) Breaker, R. R. Curr. Opin. Biotechnol. 2002, 13, 31-39. (25) Sekella, P. T.; Rueda, D.; Walter, N. G. RNA 2002, 8, 1242. (26) Hartig, J. S.; Najafi-Shoushtari, S. H.; Gruene, I.; Yan, A.; Ellington, A. D.; Famulok, M. Nat. Biotechnol. 2002, 20, 717-722. (27) Achenbach, J. C.; Nutiu, R.; Li, Y. Anal. Chim. Acta 2005, 534, 41-51. (28) Zivarts, M.; Liu, Y.; Breaker, R. R. Nucleic Acids Res. 2005, 33, 622631. (29) Soukup, G. A.; Breaker, R. R. Trends Biotechnol. 1999, 17, 469-476. (30) Soukup, G. A.; Breaker, R. R. Proc. Natl. Acad. Sci. U.S.A. 1999, 96, 3584-3589. (31) Soukup, G. A.; Breaker, R. R. Structure 1999, 7, 783-791. (32) Soukup, G. A.; Breaker, R. R. Curr. Opin. Struct. Biol. 2000, 10, 318325. (33) Chiuman, W.; Li, Y. Nucleic Acids Res. 2007, 35, 401. (34) Porta, H.; Lizardi, P. M. BioTechnology 1995, 13, 161-164. (35) Robertson, M. P.; Ellington, A. D. Nat. Biotechnol. 1999, 17, 62-66. (36) Stojanovic, M. N.; De Prada, P.; Landry, D. W. ChemBioChem 2001, 2, 411-415. (37) Tian, Y.; Mao, C. Talanta 2005, 67, 532.
Wernette et al.
we replaced the planar gold surface with Au-coated nanocapillary array membranes (NCAMs), resulting in over 2-fold increase in immobilized DNAzyme due to the contributed surface area of nanopore walls, and a 4-fold increase in intensity due to the increase in surface area created by electroless deposition of Au (nanofeatured surface roughness).39 Self-assembled monolayers of ssDNA on gold surface have been well characterized.40-45 Factors governing monolayer conformation and molecular beacon construction such as oligonucleotide length, immobilization time, thiol position, and hybridization methods have been explored extensively as they relate to coverage density and hybridization efficiency.46-50 A catalytic beacon, however, differs from a molecular beacon in three important ways. First, while secondary structure is a good indicator of molecular beacon structure (hairpin or duplex), a DNAzyme complex is thought to have a more complicated tertiary structure necessary for activity.51-53 This tertiary structure is more susceptible to perturbations by interactions with neighboring molecules and surfaces due to the weaker forces that dictate the structure compared to interbase hydrogen bonding.53 The second is the requirement that a metal cofactor be incorporated into the DNAzyme complex after hybridization, (in addition to metal ions for folding), in order for activity to be observed.52 Immobilization of a DNAzyme must not prevent mass transport of the metal cofactor to the surface by double-layer screening or other space-charge related factors. Finally, DNAzyme activity is signaled by release of fluorophore into solution, instead of a simple position change of the fluorophore from proximal to distal relative to the surface, as is the case in a molecular beacon. Cleavage and release of signaling fluorophore using a restriction endonuclease has already been demonstrated to improve molecular beacon signaling, whereas the DNAzyme has an analogous self-contained cleavage mechanism.54 Therefore, in contrast to extensive studies of factors affecting nonstructured ssDNA immobilization on surface, no systematic studies on immobilization of active DNAzymes and the environmental and protocol-related effects on the performance of catalytic beacons have been reported. Here we report the effect of immobilization variables, mercaptohexanol backfilling, and NaCl concentration on DNAzyme activity and compare the impact of those factors to their effect on molecular beacon assembly. (38) Swearingen, C. B.; Wernette, D. P.; Cropek, D. M.; Lu, Y.; Sweedler, J. V.; Bohn, P. W. Anal. Chem. 2005, 77, 442-448. (39) Wernette, D. P.; Swearingen, C. B.; Cropek, D. M.; Lu, Y.; Sweedler, J. V.; Bohn, P. W. Analyst 2006, 131, 41-47. (40) Levicky, R.; Herne, T. M.; Tarlov, M. J.; Satija, S. K. J. Am. Chem. Soc. 1998, 120, 9787-9792. (41) Steel, A. B.; Herne, T. M.; Tarlov, M. J. Anal. Chem. 1998, 70, 46704677. (42) Wang, H.; Tang, Z.; Li, Z.; Wang, E. Surf. Sci. 2001, 480, L389-L394. (43) Petrovykh, D. Y.; Kimura-Suda, H.; Tarlov, M. J.; Whitman, L. J. Langmuir 2004, 20, 429-440. (44) Luderer, F.; Walschus, U. Top. Curr. Chem. 2005, 260, 37. (45) Lee, C. Y.; Gong, P.; Harbers, G. M.; Grainger, D. W.; Castner, D. G.; Gamble, L. J. Anal. Chem. 2006, 78, 3316. (46) Herne, T. M.; Tarlov, M. J. J. Am. Chem. Soc. 1997, 119, 8916-8920. (47) Steel, A. B.; Levicky, R. L.; Herne, T. M.; Tarlov, M. J. Biophys. J. 2000, 79, 975-981. (48) Huang, E.; Satjapipat, M.; Han, S. B.; Zhou, F. M. Langmuir 2001, 17, 1215. (49) Wirtz, R.; Walti, C.; Tosch, P.; Pepper, M.; Davies, A. G.; Germishuizen, W. A.; Middelberg, A. P. J. Langmuir 2004, 20, 1527-1530. (50) Gao, Y.; Wolf, L. K.; Georgiadis, R. M. Nucleic Acids Res. 2006, 34, 3370. (51) Liu, J.; Lu, Y. J. Am. Chem. Soc. 2002, 124, 15208-15216. (52) Brown, A. K.; Li, J.; Pavot, C. M. B.; Lu, Y. Biochemistry 2003, 42, 7152-7161. (53) Di Giusto, D. A.; King, G. C. Top. Curr. Chem. 2006, 261, 131. (54) Zuo, X. B.; Yang, X. H.; Wang, K. M.; Tan, W. H.; Li, H. M.; Zhou, L. J.; Wen, H. H.; Zhang, H. Anal. Chim. Acta 2006, 567, 173.
DNAzyme-Based Catalytic Beacons
Langmuir, Vol. 23, No. 18, 2007 9515
Table 1. Sequences of DNAzyme Oligonucleotides Used for Immobilizationa HS-(7)17E Fl-17S(7)-Dy Alexa-17Snc(7) HS-39E Fl-39S Alexa-39Snc a
5′-(C6Thiol)-TTTTTAAAGAGACATCTCTTCTCCGAGCCGGTCGAAATAGTGAGT-3′ 5′-Fluorescein-ACTCACTATrAGGAAGAGATGTCTCTTT-Dabcyl-3′ 5′-Alexa546-ACTCACTATAGGAAGAGATGTCTCTTT-3′ 5′-(C6Thiol)-CACGTCCATCTCTGCAGTCGGGTAGTTAAACCGACCTTCAGACATAGTGAGT-3′ 5′-Fluorescein-ACTCACTATrAGGAAGAGATGGACGTG-3′ 5′-Alexa546-ACTCACTATAGGAAGAGATGGACG TG-3′
rA denotes RNA.
Experimental Section Chemicals, DNA, and Supplies. All reagents were purchased from Sigma-Aldrich, Inc., and used without additional purification. Buffer solutions were prepared with as-received reagents, chelated with Chelex 100 beads for 1 h to remove contaminating divalent metal ions, filtered, and titrated with glacial acetic acid to adjust the pH as desired. Concentrated metal solutions were prepared from acetate salts to a concentration of 10 mM in 10 mM acetic acid to assist in solubility. Polycarbonate membranes were purchased from GE Osmonics Labstore. Oligonucleotides with modification were purchased from Integrated DNA Technologies, Inc., with HPLC purification and used without additional purification. Oligonucleotide sequences and modification are shown in Table 1, while sequence and modification justifications were reported previously.16,22,38 Preparation of Au-Coated NCAM. Electroless deposition of Au on commercial polycarbonate track-etched NCAMs (GE Osmonics Labstore) was performed by modifying previously reported methods.55 Membranes with pore sizes of 200 and 600 nm (internal diameter, i.d.) were cleaned by soaking for 5 min in CH3OH. Membranes were sensitized with Sn2+ by soaking for 45 min in 0.022 M SnCl2 and 0.067M trifluoroacetic acid in a 1:1 mixture of water and methanol. After being rinsed with water, the membranes were activated with Ag+ by immersion in an aqueous solution of 0.035 M AgNO3 for 5 min and subsequent rinsing with water. The final solution used to deposit Au was 0.023 M NaHCO3, 0.118 M Na2SO3, 0.68 M 37% formaldehyde in water, and 0.007 M Oromerse gold solution, part B (Technic, Inc). The membranes were soaked in Au deposition solution at 4 °C for 3 h. The coated membranes were rinsed with water and cleaned in 25% HNO3 overnight, then rinsed with water, and dried with dry N2(g). The Au-coated NCAMs were stored in a dry nitrogen atmosphere until use. Immediately prior to DNAzyme immobilization, the Au-coated NCAMs were cleaned by exposure to O3(g) for 20 min. Preparation of Ratiometric DNAzyme Sensor Surface. Assembly of thiolated-DNA on Au and hybridization of complementary DNA followed previously reported methods with minor modifications40,46,56 and is portrayed in Figure 2. Immobilization of HS(7)17E, or enzyme strand, on Au-NCAM was achieved by soaking cleaned Au (ca. 0.3 × 0.3 cm2) in 1 M potassium phosphate buffer (pH ) 6.9) and 1 µM HS-(7)17E for 90 min. NCAMs were then thoroughly rinsed in deionized (DI) water and immediately soaked in 1 mM mercaptohexanol (MCH) for 5 min. Subsequently, NCAMs were thoroughly rinsed in 50 mM Tris-acetate buffer (pH ) 7.4) and 1 M NaCl. Hybridization of substrate and internal control was accomplished by soaking the immobilized enzyme and MCH mixed monolayer in 50 mM Tris-acetate buffer (pH ) 7.4) and 1 M NaCl containing 1 µM Fl-17S(7)-Dy and 125 nM Alexa546-17Snc(7), an 8:1 ratio, and heating in a 70 °C water bath for 60 min. The bath was then allowed to cool to room temperature over 90 min, resulting in a DNAzyme-coated NCAM. Previously, we reported further cooling to 4 °C during hybridization, but recent experiments showed no increase in hybridization, so the further cooling step was eliminated. Reaction. Prior to use of the prepared DNAzyme-coated NCAM for Pb2+ sensing, it was soaked in a 50 mM Tris-acetate buffer (pH ) 7.7) and 50 mM NaCl solution for 5 min in an effort to remove any physisorbed substrate strands adsorbed in spite of MCH(55) Yu, S. F.; Li, N. C.; Wharton, J.; Martin, C. R. Nano Lett. 2003, 3, 815-818. (56) Aqua, T.; Naaman, R.; Daube, S. S. Langmuir 2003, 19, 10573-10580.
passivation and to rinse away any dissociated or partially dissociated substrate strand at the lower NaCl concentration. Measurements were made by placing the assembled DNAzyme in 600 µL of freshly prepared solution containing Pb2+ (or other divalent metal ion in the case of the specificity experiment) in 50 mM Tris-acetate buffer (pH ) 7.7) and 50 mM NaCl. The DNAzyme NCAM was allowed to react with the Pb2+ solution for 60 min, after which it was removed and rinsed with the reaction solution. For the solution-based specificity experiment, the DNAzyme complex was hybridized in 50 mM Tris (pH 7.7) and 50 mM NaCl at a concentration of 50 nM and reacted with 10 µM divalent metal for a total of 5 min. Fluorescence intensity of the cleaved and uncleaved DNA in solution was determined by use of a 0.5 by 0.5 cm2 quartz cell in a Jobin Yvon Fluoromax-P fluorimeter (λex ) 491 nm, λem ) 518 nm; and λex ) 555 nm, λem ) 571 nm) to measure fluorescein and Alexa546 fluorescence intensity, respectively.
Results and Discussion The molecular beacon and DNAzyme catalytic beacon systems share similarity in terms of immobilization and hybridization of DNA probe. Therefore, factors known to be optimal for monolayer density and hybridization efficiency in the molecular beacon system may directly translate to formation of immobilized DNAzyme complex in the DNAzyme catalytic beacon system.40,45,46,48,50 The primary difference between the two systems, however, is a catalytic event after hybridization for signaling in the DNAzyme catalytic beacon system. That is, in the DNAzyme monolayer, after immobilization and hybridization, an independent reaction creates a signaling event. It is this activity characteristic of the catalytic DNA surfaces that requires additional study for optimal implementation of the DNAzyme-based sensor. Catalytic DNAzyme Activity Characterized by Ratiometric Fluorescence. The ratiometric method used an uncleavable substrate strand (by replacing the substrate RNA base with a DNA base) as a way to monitor the nonspecific release of uncleaved substrate away from immobilized enzyme39 (Figure 3). It was observed that a kinetically slow, osmotically forced release of hybridized substrate caused an increase in detected fluorescence for cleavable substrate in the absence of Pb2+ analyte. The contribution of this nonspecific release was very dependent on the total amount of immobilized DNAzyme on the surface. Introduction of an uncleavable substrate strand with a different fluorophore (Alexa546) provided a way to monitor the amount of nonspecific release of substrate strand in the absence of Pb2+. That is, the noncleavable DNAzyme complex will nonspecifically denature substrate strand at the same rate as the uncleaved Pb2+absent cleavable substrate. The measured ratio of the two fluorophores, fluorescein and Alexa546, then remains constant in the absence of Pb2+. However, after introduction of Pb2+, the cleaved substrate releases an additional component to the detected fluorescein signal, thereby increasing the measured ratio of fluorophores (fluorescein/Alexa546). Furthermore, the amount of nonspecific release also standardized the variability in total amount of DNAzyme immobilized. The use of an uncleavable substrate strand provided a fluorescence signal proportional to the total amount of hybridized substrate. That meant the Alexa546 signal was a measurement of total
9516 Langmuir, Vol. 23, No. 18, 2007
Wernette et al.
Figure 2. Schematic representation of (a) the immobilization process by self-assembly of thiolated enzyme strand, (b) backfilling with short mercaptohexanol self-assembly to force immobilized enzyme into upright conformation and displace nonspecific adsorption, (c) hybridization of fluorophore-labeled substrate strand, and (d) Pb2+ cleavage and subsequent release of fluorescence into solution for detection.
Figure 3. Schematic diagram of the ratiometric internal control method before and after Pb2+ reaction to monitor nonspecific release of substrate from immobilized enzyme. Cleavable substrate is shown with black heads representing fluorescein label; uncleavable substrate, as oligos with white heads representing Alexa546 label; and enzyme, as “bubbled” oligos.
immobilized DNAzyme complex (enzyme + substrate). The bottom line is that constructing a ratio between fluorescein fluorescence, derived from the cleavable substrate, and Alexa546 fluorescence, arising from the uncleavable substrate, provided a measurement of DNAzyme activity independent of total immobilized DNAzyme. Enzyme Immobilization Time Necessary for Maximum DNAzyme Activity. Immobilization time for DNA probes has been studied extensively by many research groups, and total immobilization density has been shown to plateau at 1.5 h of immobilization for shorter oligonucleotides (16-20-mers).46,47 However, the 45-mer enzyme strand may have different adsorption kinetics, and immobilization densities could have a strong effect on DNAzyme activity. To investigate the optimal immobilization time, Au-NCAM surfaces were soaked in 1 µM HS-(7)17E for 15 min (see Table 1 for DNA sequences and modifications) and for 1, 1.5, 2, 4, and 16 h. Surfaces were then continued to be treated through the standard MCH-backfill and hybridization procedures to assemble functional DNAzyme. DNAzyme activity and immobilization were studied by ratiometric fluorescence measurement. The immobilization time for enzyme strand showed little effect on total observed DNAzyme activity, within error, in the absence or presence of Pb2+ (Figure 4a). However, the noncleavable signal showed that total immobilization density plateaus very near 1.5 h, as observed in DNA probe studies (Figure 4b). Again, the ratiometric system standardized total immobilized DNAzyme, resulting in immobilization-time-independent activity measurements. To obtain maximum fluorophore intensities, 1.5 h of immobilization was used in the remainder of the experiments, though it should be noted that as little as 15 min of immobilization time is sufficient to provide an active DNAzyme surface.
Figure 4. (a) Enzyme immobilization time effect on total DNAzyme activity in the absence (O) and presence (b) of 10 µM Pb2+, showing relative time independence. (b) Total immobilized DNAzyme is maximized after 1.5 h of immobilization as monitored by the noncleavable signal.
Membrane Pore Diameter. Previously, the ratiometric internal control method was proven to be effective at calibrating differences in cleaved intensity due to differences in total amount of DNAzyme because of variations in surface area. Calibration was shown by measuring an equal, standardized ratio for 10 µM Pb2+ reactions of two different NCAM tab sizes (2.5 × 2.5 mm2 or 4.5 × 5.0 mm2). Those experiments were conducted with 600 nm pore membranes, with the larger pore diameter chosen to maximize mass transport of DNA and analyte through the pores. A smaller diameter pore membrane may be more desirable in some sensing constructs, such as microfluidic devices. Therefore, an experiment was performed to characterize a 200 nm pore membrane. Comparing a 200-nm to a 600-nm pore-diameter NCAM showed a difference in ratio and Pb2+ activity, despite the ratiometric method’s ability to compensate for differences in surface area between the membranes (Figure 5). The 600 nm pore membrane showed ∼75% higher activity, as indicated by the ratio (solid bar) of fluorescein to Alexa546. The ratiometric method provided some insight into the reasons behind the above observation. First, the 200 nm pore NCAM did indeed have more immobilized DNAzyme, as indicated by an increase in uncleavable substrate signal (Alexa546) used to monitor the total amount of DNAzyme complex (open bar). The
DNAzyme-Based Catalytic Beacons
Figure 5. Pore diameter effects on DNAzyme activity, seen by comparing the activity of a 200 nm pore membrane to that of a 600 nm pore membrane with and without Pb2+. The ratio (solid bars) of cleavable to uncleavable substrate shows higher activity in 600 nm membranes compared to 200 nm membranes, despite more immobilized DNAzyme on the 200 nm pore membranes (open bars). Error bars represent one standard deviation of three independent reactions.
calculated increase in surface area of the 200 nm NCAM was 1.7 times that of the 600 nm NCAM, reflecting an order of magnitude increase in pore density offsetting the decreased pore diameter. The increase in uncleavable substrate signal was only 1.3 times that observed for 600 nm. The slight deficit of observed signal relative to the expected change was attributed to decreased diffusion of enzyme and substrate through the 200 nm pores in the soak-based state used for immobilization and hybridization. Second, despite an observed increase in DNAzyme complex, there was less cleavable fragment fluorescence (fluorescein) observed. Lower cleavable fluorescence must be, in part, attributed to a decrease of activity for the DNAzyme complex within the pores. It is suspected that the pore wall constraints, that is, the curvature of the pore walls forcing DNAzyme into closer proximity to one another, diminish DNAzyme activity by the DNAzyme interacting more with one another or with the pore walls. Finally it is reasonable to consider whether differences in hindered diffusion could, in part, explain the observed differences. However, the Renkin equation applied to small permeants, such as the analytes being considered here, predicts differences in transport ∼1% between 200 and 600 nm pores.57 All experiments described hereafter used the 600 nm pore NCAM to ensure efficient diffusion through the pores and optimal activity of the immobilized DNAzyme. While solute transport effects in 200 nm pores might, at first, seem surprising, such effects have been observed for some time and are usually addressed within the framework of hindered diffusion.58-60 Reductions in solute diffusivity are thought to result from both thermodynamic and hydrodynamic effects. The thermodynamic effect has its origin in a concentration-based reduction in the driving force for diffusion compared to that observed in bulk solution, while the hydrodynamic effect arises from an enhanced viscous drag arising from a proximal pore wall. These have been effectively treated by use of a virial expansion for the concentration dependence of the partition coefficient and intrapore diffusivity.61,62 (57) Renkin, E. M. J. Gen. Physiol. 1954, 38, 225-243. (58) Deen, W. M. AIChE J. 1987, 33, 1409. (59) Nitsche, J. M.; Balgi, G. Ind. Eng. Chem. Res. 1994, 33, 2242. (60) Beerlage, M. A. M.; Peeters, J. M. M.; Nolten, J. A. M.; Mulder, M. H. V.; Strathmann, H. J. Appl. Polym. Sci. 2000, 75, 1180. (61) Shao, J. H.; Baltus, R. E. AIChE J. 2000, 46, 1307. (62) Shao, J. H.; Baltus, R. E. AIChE J. 2000, 46, 1149.
Langmuir, Vol. 23, No. 18, 2007 9517
Figure 6. Effect of poly-T linker length, between 5'-thiol and first complementary base of the enzyme strand, on DNAzyme activity. While no dependence on linker length is observed for 0T, 5T, or 15T, a significant decrease in activity is observed in the 10T system. (Open bars) Uncleavable Alexa546 intensity; (solid bars) Fl/Alexa ratio. Error bars represent one standard deviation of four independent reactions.
Poly-T Linker Length Dependence. Another important variable is the length of the poly-T nucleotide spacer on the enzyme strand. In the present system a 5T linker was employed between the 5′-thiol and the first complementary oligonucleotide in an effort to ensure extension of the complementary region of the DNAzyme complex above the backfilling MCH monolayer. To investigate the linker length dependence on activity, enzyme was immobilized containing no poly-T linker or 5T, 10T, and 15T poly-T linkers and hybridized with the 8:1 ratiometric cleavable/uncleavable solution. The results of reaction in 10 µM Pb2+ solution are shown in Figure 6. There is negligible activity dependence on poly-T linker length, even when no poly-T linker is used. Linker independence is likely because a six-carbon linker, which attached the 5′-thiol to the first oligonucleotide, was sufficient to extend the complementary region above the MCH layer. The observed linker independence was, however, in stark contrast to molecular beacon approaches, which have been shown to have strong dependence on linker length.63 In those systems, longer linkers significantly improved performance of the molecular beacon construct. This observed difference in linker length dependence between molecular beacon and catalytic beacon was not intuitive. In the case of molecular beacons, hybridization to a hairpin loop induces the folding of a fluorophore label away from quencher molecule, resulting in an increase in fluorescence. By using a linker in the MB system, signaling can be improved by more stable MBs on the surface or higher activity of the immobilized MB.63 The exact same conclusion could be made about an immobilized DNAzyme, that is, a spacing linker could make the DNAzyme more stable on the surface or could increase activity of the immobilized DNAzyme. However, the lack of linker dependence for the DNAzyme system suggests the MCH passivation was sufficient for stabilizing the DNAzyme and allowing activity even while DNAzyme complex was in close proximity to the gold surface. The single exception is the 10T linker, which showed a reduced ratio compared to the three other linker systems tested. The linker dependence experiment was repeated with a newly synthesized 10T linker enzyme and the same results were observed. The ratiometric system revealed that the 10T linker released nearly twice the amount of uncleavable substrate, yet slightly less cleavable substrate fluorescence was detected. Mfold, a secondary (63) Yao, G.; Tan, W. H. Anal. Biochem. 2004, 331, 216.
9518 Langmuir, Vol. 23, No. 18, 2007
Figure 7. Hybridization of DNAzyme complex followed by immobilization of dsDNAzyme (right) versus the normal ssDNA immobilize-then-hybridize system (left) shows a 67% loss of activity (ratiometric activity, solid bars) in the prehybridized system. Total immobilized DNAzyme (noncleavable activity, open bars) is slightly increased in the prehybridized system but is not calculated to force intermolecular interactions unless island formation is observed.
structure energy minimization program,64 did not predict any interaction differences between the enzyme and substrate strands. A stock 8:1 hybridization solution was used for all poly-T linker enzymes so the variation was not due to hybridization content. The cause of decreased activity for the 10T linker is unknown but is almost certainly due to an induced tertiary structure that cannot currently be predicted. Immobilization of Hybridized DNAzyme Complex. Despite being able to immobilize enzyme in as little as 15 min, the enzyme surface must still be hybridized with substrate to create the DNAzyme complex, a step that may take up to 2.5 h. In an attempt to decrease surface preparation time, immobilization of hybridized DNAzyme complex (hybridized in solution) was investigated. Most DNA immobilization work has been reported on immobilization of ssDNA probes because most research seeks to use immobilized probes for DNA complementarity sensors or in molecular beacon strategies. After comparison of the normal ssDNA enzyme immobilization procedure to the dsDNAzyme complex immobilization with MCH backfilling, a 67% decrease in activity was observed for the latter (Figure 7, solid bars). It was observed that there is a 26% higher noncleavable signal observed in the dsDNAzyme immobilization system (open bars). The 26% increase in immobilization of prehybridized DNAzyme was still insufficient to force DNAzyme molecules into close proximity, with ∼8-10 nm average distance between molecules.65 Monolayers formed by immobilization of dsDNA can form very dense packed selfassembled monolayers due to the rigidity of the DNA duplex, but that was not observed here, likely because of the structural flexibility of the three-arm structure.66,67 Instead, immobilization of the DNAzyme duplex may lead to interactions with the Au surface that cause partial dehybridization of the DNAzyme complex or result in interactions that are not effectively displaced by MCH backfilling. In either case, the significant reduction in DNAzyme activity means that immobilization of dsDNAzyme complex is not a suitable alternative method for generation of DNAzyme surfaces. Effect of Mercaptohexanol Backfilling Monolayer on DNAzyme Activity. Having observed normal DNAzyme activity with no poly-T linker, we decided to test whether or not the (64) Zuker, M. Nucleic Acids Res. 2003, 31, 3406-3415. (65) Peterson, A. W.; Heaton, R. J.; Georgiadis, R. M. Nucleic Acids Res. 2001, 29, 5163-5168. (66) Yang, M.; Yau, H. C. M.; Chan, H. L. Langmuir 1998, 14, 6121-6129. (67) Sakao, Y.; Ueno, N.; Nakamura, F.; Ito, E.; Hayasi, J.; Hara, M. Mol. Cryst. Liq. Cryst. 2003, 407, 537.
Wernette et al.
Figure 8. Activity of DNAzyme on NCAM in the absence and presence of backfilling MCH monolayer. The MCH monolayer causes ∼3.5 times increase in activity by creating more active DNAzyme complex. (Open bars) Uncleavable Alexa546 intensity; (solid bars) ratio for Fl/Alexa. Error bars represent one standard deviation of four independent reactions.
mercaptohexanol (MCH) backfilling layer was necessary for the DNAzyme activity. Tarlov and co-workers used MCH to displace nonspecific binding of DNA bases to the Au surface while simultaneously ensuring the DNA portion of the monolayer is not so dense as to cause steric or osmotic resistance for complementary oligomers.40,41,43,46,47 The net effect was an overall increase in hybridization efficiency, which has since been well studied.68,69 The NCAMs were assembled with and without MCH to compare the activity of DNAzyme on NCAMs as dependent on the MCH monolayer (Figure 8). It was shown that the presence of MCH provided ∼3.5 times increase in activity for the immobilized DNAzyme. The uncleavable substrate signal was higher without MCH than with it. It has been shown that long, nonspecifically bound oligomers are strongly held by the interacting surface and such oligomers are not easily desorbed from the surface. Therefore, the increased intensity of uncleavable substrate was due to an increase in the total hybridized substrate and not from physisorbed substrate existent due to the lack of an MCH backfill.46,47 However, while more substrate was shown to be hybridized, it is thought to be incompletely hybridized, thus decreasing the amount of fully assembled complex, which caused a lower observed DNAzyme activity. Even though the MCH backfilled surface likely had decreased total DNAzyme due to forced displacement of DNAzyme complex by MCH, the ratiometric internal standard compensated for differences in total immobilized DNAzyme and revealed over 3 times higher activity. While MCH was shown to improve hybridization efficiency in Tarlov’s reports, those studies used shorter oligonucleotides. In the case of shorter oligonucleotides, near-complete adsorption to the gold surface would inhibit hydrogen-bond formation between complementary bases. Statistically, an oligonucleotide of N bases has N1/2 bases chemisorbed to the Au surface.70 However, longer oligonucleotides (higher than 25 bases) would not perfectly lie on the surface, as bases at different locations along the chain can adsorb nearly simultaneously (Figure 9). The result is an oligonucleotide that “snakes” up and down from the surface instead of lying flat along the surface. These raised or bubbled portions of the oligonucleotide are available for base recognition and hybridization.71 While hybridization may not be enough force to remove all nonspecific interactions of the enzyme (68) Park, S.; Brown, K. A.; Hamad-Schifferli, K. Nano Lett. 2004, 4, 1925. (69) Arinaga, K.; Rant, U.; Tornow, M.; Fujita, S.; Abstreiter, G.; Yokoyama, N. Langmuir 2006, 22, 5560. (70) Petrovykh, D. Y.; Kimura-Suda, H.; Whitman, L. J.; Tarlov, M. J. J. Am. Chem. Soc. 2003, 125, 5219-5226.
DNAzyme-Based Catalytic Beacons
Figure 9. Schematic illustration of base availability for hybridization in long versus short oligonucleotides on Au. Long oligonucleotides (>25 bases) have regions that are not adsorbed to the gold surface and can allow partial hybridization to available complementary regions, compared to shorter oligonucleotides that fully adsorb.
strand with the gold surface, it is likely enough to hold the substrate strand in a partially hybridized state.72 Furthermore, the lower activity in the absence of MCH could be due to the assembled DNAzyme complex interacting with the gold surface, rendering it inactive by preventing formation of the active tertiary structure thought necessary for cleavage. Formation of fully hybridized complex could occur, if hybridization was strong enough to displace the nonspecific interactions with gold surface. The resulting fully assembled DNAzyme complex would still be inactive, because the unpassivated gold surface can disrupt tertiary structure. In either case, the ratiometric system showed that the MCH monolayer is necessary for efficient formation of active DNAzyme complex on gold surface. MCH Addition during and after Enzyme Immobilization. MCH was determined to be necessary for DNAzyme activity, but it was unknown whether a backfilling step is necessary. Some reported methods look to immobilize the alkylthiol before and during immobilization of DNA as opposed to the stepwise method proposed by Tarlov.3,56,73,74 To evaluate the effect of backfilling on total DNAzyme immobilization and activity, an experiment was performed to compare Au-NCAMs prepared with a mixed MCH/HS-(7)17E solution to the stepwise backfilling method. If immobilization of MCH and DNA could be achieved simultaneously, it could remove the need for the additional soak and rinse steps in the stepwise methodology, thereby simplifying sensor construction. It was observed that the total immobilized DNAzyme complex, as indicated by the noncleavable substrate signal, was only 33% ( 14% higher than in the stepwise backfilling method standardized by the ratiometric method. However, the total DNAzyme activity, as indicated by the ratiometric signal, was ∼500% higher when the stepwise backfilling method was used (Figure 10). If we return to the hypothesized picture of DNA immobilization, the adsorption of DNA oligos on the gold surface likely occurs simultaneously with MCH immobilization. While the MCH presence does not prevent the thiol tether from forming after adsorption of DNA, the concentration of MCH was insufficient to displace the nonspecific binding of DNA. In the mixed immobilization solution, the MCH cannot form a sufficiently dense monolayer to prevent interaction with the gold surface and disruption of activity. Optimization of the MCH:DNA concentration ratio for immobilization has been studied previously, and it was determined (71) Marie, R.; Jensenius, H.; Thaysen, J.; Christensen, C. B.; Boisen, A. Ultramicroscopy 2002, 91, 29-36. (72) Peterson, A. W.; Wolf, L. K.; Georgiadis, R. M. J. Am. Chem. Soc. 2002, 124, 14601. (73) Erts, D.; Polyakov, B.; Olin, H.; Tuite, E. J. Phys. Chem. B 2003, 107, 3591. (74) Steichen, M.; Buess-Herman, C. Electrochem. Commun. 2005, 7, 416.
Langmuir, Vol. 23, No. 18, 2007 9519
Figure 10. Mixed MCH and enzyme (10:1) solution for immobilization compared to the stepwise MCH backfilling method proposed by Tarlov. Total immobilized DNAzyme complex increases by ∼40% by the stepwise backfilling method and DNAzyme activity increases 300% when compared to the mixed immobilization solution. (Solid bars) Ratiometric DNAzyme activity; (open bars) noncleavable substrate indicating total DNAzyme complex.
that 10:1 provided the ideal ratio for high hybridization efficiency and high surface density.3 Using the 10:1 concentration ratio did not prevent nonspecific interactions with the surface, but increasing MCH concentration in the immobilization solution would decrease immobilized DNAzyme density. It was determined, then, that the stepwise backfilling MCH method was required for DNAzyme activity on gold surfaces. Effect of Backfilling Molecular Length on DNAzyme Activity. By varying the length of the alkylthiol backfilling molecule, it was possible to gain insight into the ideal height and monolayer structure for DNAzyme activity. Shorter carbon chain lengths, such as ethyl and propyl, would provide a very short chain so as not to interfere with DNAzyme tertiary structure but would also assemble into less dense monolayers. Longer chain lengths, >10 carbons, have been shown to assemble into tight monolayers and have been shown to stabilize an immobilized DNA monolayer.75 However, these long chain length molecules may be so long as to inhibit proper DNAzyme folding. To address these hypotheses, backfilling monolayers of mercaptoethanol, mercaptopropanol, butanethiol, mercaptohexanol, and mercaptoundecanol were compared (Figure 11). Mercaptohexanol supports the highest DNAzyme activity, in strong agreement with previous DNA probe reports.40,46 While a recent study shows that mercaptoundecanol provides more stable DNA-alkylthiol monolayers, here we observed that the increased stability occurs at the sacrifice of significant DNAzyme activity (∼90% loss of activity).75 It is proposed that carbon chain lengths shorter than six lead to inefficient displacement of nonspecific DNA adhesion due to low-density monolayers. Au surfaces displaying insufficient passivation permit nonspecific interaction of the DNAzyme with resulting surface lowered DNAzyme activity. In the case of the substantially longer 11mer carbon chain, the overall length perturbed the tertiary structure of active DNAzyme, which resulted in lower measured DNAzyme activity. Again, the ratiometric system is not biased by total immobilized DNAzyme. Therefore, while the total assembled DNAzyme was lower in the mercaptoundecanol monolayer because of increased displacement of immobilized DNA, lower total DNAzyme was not responsible for the observed decrease in measured activity. NaCl Concentration Effects on Assembly and Release. Another important factor for DNAzyme activity is NaCl (75) Lai, R. Y.; Seferos, D. S.; Heeger, A. J.; Bazan, G. C.; Plaxco, K. W. Langmuir 2006, 22, 10796.
9520 Langmuir, Vol. 23, No. 18, 2007
Figure 11. Comparison of the effect of carbon chain length of backfilling molecules on DNAzyme activity. Mercaptohexanol (HS(CH2)6OH) is shown to provide a significantly higher DNAzyme activity compared to higher or lower chain lengths. (O) Activity in the absence of Pb2+; (b) activity in the presence of 10 µM Pb2+. Error bars represent one standard deviation of three independent surfaces.
Figure 12. Effect of NaCl concentration on DNAzyme activity in reaction solution. The system is shown to be optimized at 50 mM NaCl, identical to solution-based reactions, with lower concentrations resulting in unstable complex and higher concentrations inhibiting cleaved fragment release. (Open bars) Uncleavable Alexa546 intensity;(shaded bars) cleavable fluorescein;(solid bars) ratio for Fl/Alexa. Error bars represent one standard deviation of three independent reactions.
concentration due to its effect on complex stability and the release of cleaved fragments. As reported previously, the enzyme/ substrate arms were optimized for a 50 mM NaCl reaction solution during solution activity experiments.38 Here the effect of reaction solution NaCl concentration on immobilized DNAzyme activity was studied after rinsing in the corresponding NaCl solution for 5 min. NaCl concentrations were varied from 0 to 1 M NaCl as shown in Figure 12. The results showed that the system is optimized for 50 mM NaCl, as in solution reactions, and that immobilization of the DNAzyme complex did not alter the optimal NaCl reaction concentration. The independence of assembly conditions relative to the reaction conditions allows a single DNAzyme assembly protocol, while reactions can still be carried out at ideal reaction conditions specific for each immobilized DNAzyme. The ratiometric system allowed for additional analysis to be made, because all DNAzyme NCAMs were hybridized in 1 M NaCl, and the differences in intensity of the substrate strands reflected the effect of NaCl concentration. First, the activity was shown to be reduced at concentrations below 50 mM NaCl (20
Wernette et al.
Figure 13. Immobilization of uranyl-DNAzyme following described general methods and reaction in solution with conditions ideal for the uranyl-DNAzyme shows over 10-fold increase in activity upon reaction of 1 µM UO22+ (solid bar) compared to uranyl-absent samples (open bars).
and 0 mM NaCl) because of decreased complex stability. In the 20 mM NaCl case, the uncleavable signal increased, indicating an unstable complex resulting in fast release of substrate from the enzyme. While the 0 mM NaCl case did not show the increased uncleavable signal, it was likely due to substrate-bound complex already being significantly denatured during the 0 mM NaCl rinse step immediately prior to placement in reaction solution. At concentrations above 50 mM NaCl (200 mM and 1 M NaCl), the complex was stabilized as evidenced by a reduced uncleavable signal. The cleavable signal remained low, however, due to the stabilization of the 9-mer cleaved fragment, which did not release from the enzyme. Shortening of the cleaved fragment was then deemed necessary to allow release after cleavage at higher NaCl concentrations. General Applicability of Immobilization Protocol Using Uranyl-Specific DNAzyme. Having optimized the protocol for surface immobilization using Pb2+-specific DNAzyme, we wished to test how general such a protocol might be when applied to other catalytic beacon systems. To test whether the immobilization procedure is universal, a uranyl-specific DNAzyme, 39E/39S, was incorporated into Au-NCAMs following the protocol established in this study. The notable difference between the uranyl DNAzyme system and the Pb2+ DNAzyme system is that the ideal reaction conditions differ: ideal pH is 5.5 for uranyl, and ideal sodium concentration is 300 mM.22 Indeed, the thiol-modified 39E was immobilized and readily assembled into the 39E/39S DNAzyme upon hybridization of Fl-39S and Alexa-39Snc. Reaction with 1 µM UO22+ provided a (25 ( 15)-fold increase in ratio relative to the uranyl-absent controls (no metals and Pb2+) as shown in Figure 13. The reason for the drastically lower calculated ratios in Figure 13 compared to Pb2+ ratios shown previously is the pH-dependent decrease in quantum yield for fluorescein at pH 5.5. A fluorophore with a higher quantum yield at lower pH would remove such variations. Therefore, the reported, generic immobilization protocol described here is directly applicable to new DNAzyme catalytic beacon systems.
Summary In summary, we have carried out a systematic investigation of factors affecting immobilization of DNAzyme onto goldcoated nanocapillary array membranes and obtained protocols for optimal performance of both Pb2+ and UO22+ catalytic beacons. In the process, we learned that many methods governing ideal immobilization of DNA probes and molecular beacons are also necessary for optimal DNAzyme complex immobilization,
DNAzyme-Based Catalytic Beacons
though the effect on ideality differs between the systems. Specifically, stepwise mercaptohexanol backfilling has little effect on measured hybridization efficiency but does passivate the gold surface so DNAzyme activity is not disrupted. Interaction of the DNAzyme with the gold surface must be prevented for efficient DNAzyme activity. With these observations gained from a detailed study of the optimal conditions for an NCAMimmobilized Pb2+ sensor, it was possible to employ a general immobilization procedure to a uranyl-dependent DNAzyme that retained activity in its ideal reaction solution. These findings
Langmuir, Vol. 23, No. 18, 2007 9521
will greatly facilitate practical applications of immobilized catalytic beacons. Acknowledgment. This material is based upon work supported by the U.S. National Science Foundation through the Science and Technology Center of Advanced Materials for the Purification of Water with Systems (WaterCAMPWS, CTS0120978), by the Strategic Environmental Research and Development Program, and by the U.S. Department of Energy (DEFG02-01ER63179 to Y.L. and DE-FG02-07ER15851 to P.W.B.). LA701303K
