Anal. Chem. 2006, 78, 6484-6489
Signaling Aptamers Created Using Fluorescent Nucleotide Analogues Evaldas Katilius,* Zivile Katiliene, and Neal W. Woodbury
Department of Chemistry and Biochemistry and Center for BioOptical Nanotechnology, Biodesign Institute, Arizona State University, Tempe, Arizona 85287
A new approach to creating fluorescent signaling aptamers using fluorescent nucleotide analogues is presented. The fluorescence quantum yield of nucleotide analogues such as 2-aminopurine strongly depends on base stacking interactions when incorporated into double or single stranded DNA. This property is used to generate a binding-specific fluorescence signal. Aptamers for human r-thombin, immunoglobulin E, and platelet-derived growth factor B were modified with fluorescent nucleotide analogues in positions that undergo conformational changes. The resulting signaling aptamers show a specific, bindinginduced increase in the fluorescence signal of up to 30fold. Conformation-changing positions in these aptamers were identified by screening a set of modified aptamer sequences that each included a fluorescent nucleotide analogue at a different position. The positions for these modifications were estimated by modeling the aptamer secondary structure. It is likely that this approach to producing fluorescent signaling aptamers is of general use for protein-binding aptamers because of their “induced fit” binding mechanism. Aptamers are short, single-stranded nucleic acids (DNA or RNA) isolated from random-sequence libraries using in vitro selection.1-3 Aptamers have been selected to bind to a variety of different targets, ranging from small molecules to proteins.4 The high affinity and selectivity of aptamers have resulted in their increasing popularity for development of novel biosensing applications.5 In vitro selection and synthesis of aptamers make it possible to introduce chemical modifications; therefore, new functionalities can be introduced to complement an aptamer’s propensity to bind its target. One such functionality is signaling of binding through generation of a fluorescence signal. Apparently, aptamer binding via an “induced fit” mechanism6 results in sufficient conformational changes in the aptamer structure during binding that a detectable signal can be produced if the aptamer is labeled with a conformationally sensitive fluorophore.7-10 Initial * To whom correspondence should be addressed. E-mail:
[email protected]. (1) Gold, L.; Polisky, B.; Uhlenbeck, O.; Yarus, M. Annu. Rev. Biochem. 1995, 64, 763-797. (2) Ellington, A. D.; Szostak, J. W. Nature 1990, 346, 818-822. (3) Tuerk, C.; Gold, L. Science 1990, 249, 505-510. (4) Lee, J. F.; Hesselberth, J. R.; Meyers, L. A.; Ellington, A. D. Nucleic Acids Res. 2004, 32 (Database issue), D95-100. (5) Tombelli, S.; Minunni, M.; Mascini, M. Biosens. Bioelectron. 2005, 20, 24242434. (6) Hermann, T.; Patel, D. J. Science 2000, 287, 820-825.
6484 Analytical Chemistry, Vol. 78, No. 18, September 15, 2006
studies have concentrated on external modifications of bases in an ATP binding aptamer sequence using different fluorophores; however, external labeling of bases was shown to be rather insensitive to the small conformational changes induced upon ATP binding (signal increases up to 80% have been demonstrated). In addition, the fluorophore sometimes interfered with binding, greatly decreasing the affinity of the aptamer and, thus, its usefulness.7,8 A more effective signaling (on the order of 2-3fold) was obtained when the ATP binding aptamer sequence was internally modified with bispyrenyl.9 In this case, the signal generation was based on excimer fluorescence due to the dimerization of bispyrenyl upon aptamer binding to the ATP molecules. Most recently, Bodipy fluorophore linked to 2′ ribose was used to convert three different aptamers binding AMP, thyrosinamide, and arginamide into signaling aptamers.10 Sensitivity of Bodipy fluorophores to the local environment when linked to the 2′-ribose position in the aptamer results in significant changes in its fluorescence quantum yield (up to a 3.7-fold increase in fluorescence was observed for the arginamide aptamer). Many other approaches to generating fluorescence signals upon bindinginduced conformational changes in aptamer structure have involved a pair of fluorophores acting as donor and acceptor or as donor and quencher (these approaches have been recently reviewed11-13). In this case, the signals are generated as a result of large conformational changes in the aptamer structure (typically, movements on the order of tens of angstroms), leading to changes in fluorescence energy transfer efficiency between the different dyes. We have developed a new approach for creating signaling aptamers using fluorescent nucleotide analogues as the probes of conformational changes. Nucleotide analogues, such as 2-aminopurine (2AP), 4-amino-6-methylpteridone (6MAP) or 3-methylisoxanthopterin (3MI), were used to modify the sequences of several DNA aptamers. The fluorescence quantum yield of these nucleotide analogues is dependent on the base stacking interactions between the nucleotide analogue and its neighbors in the (7) Jhaveri, S.; Rajendran, M.; Ellington, A. D. Nat. Biotechnol. 2000, 18, 12931297. (8) Jhaveri, S. D.; Kirby, R.; Conrad, R.; Maglott, E. J.; Bowser, M.; Kennedy, R. T.; Glick, G.; Ellington, A. D. J. Am. Chem. Soc. 2000, 122, 2469-2473. (9) Yamana, K.; Ohtani, Y.; Nakano, H.; Saito, I. Bioorg. Med. Chem. Lett. 2003, 13, 3429-3431. (10) Merino, E. J.; Weeks, K. M. J. Am. Chem. Soc. 2005, 127, 12766-12767. (11) Nutiu, R.; Li, Y. Angew. Chem. Int. Ed. 2005, 44, 1061-1065. (12) Nutiu, R.; Li, Y. Methods 2005, 37, 16-25. (13) Rajendran, M.; Ellington, A. D. In Optical Biosensors: Present and Future; Ligler, F. S., Rowe Taitt, C. A., Eds.; Elsevier: New York, 2002, pp 369396. 10.1021/ac060859k CCC: $33.50
© 2006 American Chemical Society Published on Web 08/22/2006
single or double-stranded DNA chain.14,15 Binding of an aptamer to its protein target can dramatically change the stacking of bases and lead to the alteration of a fluorescence signal if fluorescent nucleotide is incorporated at the position that undergoes conformational (base-stacking) changes. In this work, the use of nucleotide analogues to create signaling aptamers was first explored by analyzing the available tertiary structure of a thrombin-binding aptamer bound to its target protein and then incorporating a fluorescent nucleotide analogue in a position that can be inferred to undergo changes in base stacking. Next, a general approach to generate signaling aptamers using fluorescent nucleotide analogues was developed using previously published aptamer sequences for immunoglobulin E (IgE)16 and platelet-derived growth factor B (PDGF),17 for which no tertiary structural information is available. It is shown that simply by screening of a set of modified aptamers, one can pinpoint a position that apparently undergoes base-stacking changes upon aptamer binding to the target molecule. The set of modified sequences for screening can be chosen by computing secondary structure models from the aptamer sequence and choosing the regions most likely to change structure upon binding. EXPERIMENTAL SECTION Oligonucleotides modified with 2AP were purchased from Gene Biotech (Tempe, AZ). Those modified with 6MAP were purchased from Trilink Biotech (San Diego, CA) or Fidelity Systems (Gaithersburg, MD). Oligonucleotides containing 3MI were purchased from Fidelity Systems. Human R-thombin was purchased from Haematologic Technologies (Essex Junction, VT). Human IgE (human myeloma plasma fraction) was purchased from Athen’s Research & Technology (Athens, GA). Plateletderived growth factor B (human recombinant form) was purchased from R&D Systems (Minneapolis, MN). Lyophilized PDGF was dissolved in 4 mM HCl, 0.1% BSA solution according to the manufacturer’s recommendations. The fluorescent nucleotide modified DNA aptamers for R-thrombin (oligonucleotide sequence 5′-GGTTGGXGTGGTTGG-3′; here X shows the modified position) were reversed-phase-HPLC-purified. Stock solutions were prepared in sterile Nanopure water at 100 µM concentration and diluted in TE (10 mM Tris-HCl, pH 8, 1 mM EDTA, pH 8) buffer at working concentrations of 1 µM to 100 nM. Thrombin binding assays were performed in TE buffer, because it has been previously shown that the thrombin-binding aptamer exhibits the same affinity toward thrombin in minimal buffer solutions as in pure water.18-20 Buffer conditions were chosen not to contain any divalent (Mg) or potassium ions, which stabilize the G-tetramer structure of the aptamer in the solution, leading to higher fluorescence intensity of the free aptamer. Fluorescence measurements were performed in a quartz fluorescence microcuvette (Starna Cells, 16.50F-Q-10/ Z15) using a FluoroMax-3 (Jobin Yvon) fluorometer. Fluorescence of 2AP was measured using 315-nm excitation, detecting the spectra from 325 to 500 nm. Fluroescence of 6MAP was measured (14) Hawkins, M. E. Cell Biochem. Biophys. 2001, 34, 257-281. (15) Rist, M. J.; Marino, J. P. Curr. Org. Chem. 2002, 6, 775-793. (16) Wiegand, T. W.; Williams, P. B.; Dreskin, S. C.; Jouvin, M. H.; Kinet, J. P.; Tasset, D. J. Immunol. 1996, 157, 221-230. (17) Green, L. S.; Jellinek, D.; Jenison, R.; Ostman, A.; Heldin, C. H.; Janjic, N. Biochemistry 1996, 35, 14413-14424. (18) Baldrich, E.; O’Sullivan, C. K. Anal. Biochem. 2005, 341, 194-197. (19) Baldrich, E.; Restrepo, A.; O’Sullivan, C. K. Anal. Chem. 2004, 76, 70537063. (20) Ho, H. A.; Leclerc, M. J. Am. Chem. Soc. 2004, 126, 1384-1387.
exciting at 330 nm and collecting emission from 375 to 525 nm. Fluorescence of 3MI was excited at 350 nm and detected from 375 to 525 nm. Initial screening of the 2AP modified IgE and PDGF-binding aptamers (sequences are presented below) was performed on unpurified, desalted DNA oligonucleotides, which were dissolved in sterile Nanopure water at 100 µM concentration and then diluted to 1 µM or 500 nM concentrations in appropriate for binding buffer (10 mM Tris-HCl pH 7.1, 140 mM NaCl, 5 mM KCl, 1 mM MgCl2). Screening for the PDGF signaling aptamer was performed by measuring fluorescence spectra from 500 nM aptamer solutions before and after addition of ∼350 nM PDGF. Screening for the IgE signaling aptamer was performed by measuring fluorescence spectra from 1 µM aptamer solutions before and after addition of 80 nM IgE. In both cases, protein was added directly into the cuvette, and samples were mixed by pipetting and then incubated for ∼2 min before the fluorescence spectra were measured. Spectra were corrected for buffer background fluorescence and Raman scattering before calculating the ratio with and without the protein target. After preliminary screening of modified aptamers, the two aptamers that showed the greatest increase in fluorescence upon binding to PDGF or IgE were purified using polyacrylamide gel electrophoresis before further characterization. To show that the binding-induced fluorescence increases were specific, control experiments were performed by synthesizing mutated versions of the aptamers with the fluorescent nucleotides in the same positions as in the sequences above. The following sequences were synthesized: mutated thrombin-binding aptamer 5′-GAT CGA (2AP)GC GCT AGA-3′, mutated IgE-binding aptamer 5′-GGG GCA CGT TTA TGC AT(2AP) CAT GCT AGT GGC GTG CCC C-3′, and mutated PDGF-binding aptamer 5′-CAC AGG ATA CGG CAC GTA TTC (2AP)CT CAC CAT GAG CCT GTG-3′. For all three oligonucleotides, fluorescence spectra were measured before and after addition of increasing concentrations of thrombin, IgE, or PDGF. The secondary structures of aptamers were modeled using the program mfold.21 Secondary structures were calculated for ambient conditions (23 °C) at 140 mM Na+ and 1 mM Mg2+. The concentration dependence of binding between each signaling aptamer and its target protein was measured by holding the aptamer concentration at either 100 nM (for 6MAP-labeled thrombin aptamer) or 500 nM (for 2AP-modified IgE and PDGF aptamers) and adding increasing concentrations of proteins. After each addition, the solution was mixed directly in a cuvette and incubated for several minutes at room temperature to allow binding to reach equilibrium. Results were corrected for aptamer dilution due to addition of the protein. Measurements were repeated three times. The fluorescence signal detected in the binding assay can be expressed as
F ∝ qAPx + qA(A0 - x)
(1)
Here, F is the fluorescence intensity, qAP is the quantum yield of the aptamer-protein complex, x is the concentration of the aptamer-protein complex, qA is the fluorescence quantum yield of free aptamer in the solution, and A0 is the initial concentration of aptamer in solution. (21) Zuker, M. Nucleic Acids Res. 2003, 31, 3406-3415.
Analytical Chemistry, Vol. 78, No. 18, September 15, 2006
6485
Further, the initial free aptamer fluorescence intensity can be written as
F0 ∝ qAA0.
(2)
The ratio of the fluorescence signals can be reduced to
(
)
qAP F x ) -1 +1 F0 qA A0
(3)
Thus, the ratio of the fluorescence intensities is proportional to the amount of the aptamer-protein complex, which can be calculated at each protein concentration by solving the equation that defines a dissociation constant (assuming a simple binding model)
Kd )
(A0 - x)(P - x) x
(4)
where Kd is the dissociation constant, A0 is the initial free aptamer concentration, and P is the protein concentration. On the basis of these considerations, the fluorescence data from the binding assay was fitted to the following equation.
A0 + P + Kd - x(A0 + P + Kd)2 - 4A0P F -1)R F0 2A0
(5)
Here, R is a proportionality constant that contains the relative fluorescence quantum yield of the bound and unbound forms. Experimental results were fitted by varying R and Kd in the above equation until a minimum χ2 value was achieved. The fit was performed using the nonlinear least-squares fitting function in Origin 7.5 software (OriginLab). RESULTS AND DISCUSSION Thrombin-Binding Aptamer. The thrombin-binding aptamer is a short, 15-nucleotide DNA molecule that has been selected to bind and inhibit thrombin activity.22 The crystal structure of the aptamer-protein complex shows that the aptamer is stabilized in a G-quartet structure (see Figure 1).23,24 G-quartet formation leads to the unstacking of the intervening bases forming the loops. for example, the thymine (T) in position 7 is clearly unstacked relative to its neighboring bases, and at the same time, it does not interact with the protein. Using this structural information, the fluorescent base analogues, 2AP, 6MAP, and 3MI, have been introduced in place of T in position 7. Fluorescence spectra of the aptamer modified with 6MAP before and after addition of thrombin are shown in Figure 2a. Fluorescence of 6MAP is highly quenched in the free aptamer, indicating that the aptamer is in a single-stranded unstructured conformation. Upon addition of near-saturating protein concentration to the sample, an immediate increase in fluorescence is (22) Bock, L. C.; Griffin, L. C.; Latham, J. A.; Vermaas, E. H.; Toole, J. J. Nature 1992, 355, 564-566. (23) Padmanabhan, K.; Padmanabhan, K. P.; Ferrara, J. D.; Sadler, J. E.; Tulinsky, A. J. Biol. Chem. 1993, 268, 17651-17654. (24) Padmanabhan, K.; Tulinsky, A. Acta Crystallogr., Sect. D: Biol. Crystallogr. 1996, 52, 272-282.
6486
Analytical Chemistry, Vol. 78, No. 18, September 15, 2006
Figure 1. Crystal structure of human R-thrombin bound to the thrombin-binding aptamer. This figure was created using the entry 1HAO from the Protein Data Bank. Table 1. Fluorescence Signal Changes in Aptamers Modified with Different Nucleotide Analogs
modified fluorescent aptamer thrombin IgE PDGF
increase in fluorescence signal upon addition of a saturating concentration of target proteina 2AP 3MI 6MAP 10 ( 1 5.6 ( 0.4 5.1 ( 0.4
9.7 ( 0.5 1.5 ( 0.1 2.5 ( 0.2
30 ( 3 1.9 ( 0.2 6.7 ( 0.5
a 2AP, 2-aminopurine; 3MI, 3-methylisoxanthopterin; 6MAP, 4-amino6-methylpteridone.
observed; the relative fluorescence increase is roughly 30-fold when the ratio is calculated from the maximums of the spectra before and after addition of R-thrombin (the ratio is ∼33 if calculated from the integrated areas of the spectra). The 6MAP-modified aptamer fluorescence intensity increases as a function of thrombin concentration, as shown in Figure 2b. The data were fitted to a single-component binding equation. The resulting dissociation constant, Kd, was 12 ( 2 nM, which is very similar to the dissociation constant determined for the unmodified aptamer.22 This shows that modification of the aptamer with the fluorescent nucleotide does not alter its binding affinity appreciably, which is expected, since modification is in a position not involved in binding. As mentioned above, three different fluorescent nucleotide analogues were introduced in place of T at position 7 in the aptamer sequence. The fluorescence from aptamers labeled with 2AP, 3MI, and 6MAP was measured, and the results of the relative fluorescence increases obtained with different fluorescent nucleotides are summarized in Table 1. The best signal enhancement is observed using 6MAP (30-fold increase upon binding), because the fluorescence quantum yield of this nucleotide analogue is most strongly quenched in single-stranded DNA.25 The fluorescence increase for the other two nucleotide analogues is about 10-fold. (25) Hawkins, M. E.; Pfleiderer, W.; Jungmann, O.; Balis, F. M. Anal. Biochem. 2001, 298, 231-240.
Figure 2. (a) Fluorescence spectra of the thrombin-binding aptamer modified with 6MAP in position 7. Spectra of a 1 µM aptamer solution was measured before (solid line) and after (dashed line) addition of 3 µM thrombin. The relative increase in fluorescence between the bound and unbound states of the aptamer is ∼30-fold, as calculated from the ratio of the peak fluorescence intensities. (b) A binding curve showing the fluorescence signal increase as a function of thrombin concentration. Fluorescence spectra of the thrombin-binding aptamer modified with 6MAP were measured after adding increasing concentrations of thrombin to the aptamer solution. The solid squares show experimental data; the line shows a fit of the data to a single component binding equation. The dissociation constant resulting from the fit was 12 ( 2 nM. Notice the break in the x axis.
The relative increase apparently depends on the photophysical characteristics of the nucleotide analogue, that is, the extent to which base stacking interactions affect the fluorescence quantum yield of 2AP, 3MI, and 6MAP. The design of the modified thrombin aptamer was based on the crystal structure of the complex between the aptamer and R-thrombin. However, in most cases, no tertiary (crystal or NMR) structural information is available for aptamer-protein complexes. Thus, to develop a general method to pinpoint positions that might report the binding of an aptamer to its target, we have devised the selection protocol based on the modeling of the aptamer secondary structure and from this, guessing which bases in the aptamer are most likely to be affected by binding. Two different aptamers, previously selected to bind to PDGF and to IgE, were chosen to validate this approach. PDGF-Binding Aptamer. The PDGF-binding aptamer, which was selected to bind and inhibit PDGF,17 has the sequence 5′CAC AGG CTA CGG CAC GTA GAG CAT CAC CAT GAT CCT GTG-3′ Secondary structure calculations for this aptamer using the program mfold yields two different possible structures (see Figure 3a). These secondary structures suggest that the protein binding site is in the center of the three-helix junction (this hypothesis is also supported by protein cross-linking experiments presented in the original aptamer selection study17). At the same time, it can be inferred that the loops in the center of the aptamer should be the most labile regions where bases would undergo major conformational changes due to an “induced fit” binding of the aptamer to the PDGF. Therefore, screening of bases that form loops in the center of the aptamer was performed. Five aptamers with 2AP in the positions 20, 21, 22, 32, and 33 were synthesized and screened for fluorescence signaling upon protein binding. Results of the screening assay (Figure 3b) show that position 22 in the aptamer reports by far the greatest increase in fluorescence due to binding. It should be noted that the initial screening was performed using unpurified aptamers and nonsaturating concentrations of protein; therefore, the relative fluorescence signal change is not the maximum that can be obtained. However, even such a simple approach to screening provides sufficient informa-
tion to select a position that undergoes a conformational change. For further characterization, the aptamer with 2AP at position 22 was chosen. Fluorescence measurements under binding-saturating protein concentration show that this signaling aptamer reports binding to PDGF through an ∼5-fold increase in fluorescence of 2AP. Determination of the dissociation constant for this signaling aptamer from the fluorescence data is, however, complicated because PDGF is a homodimer. As a result, the binding equilibrium is more complicated, and data cannot be fitted to the simple binding model (eq 5) to provide an estimate for the dissociation constant. Nevertheless, the PDGF concentration titration results (see Figure 3c) suggest that the binding affinity of the signaling aptamer is still in nanomolar range, which is similar to the published 100 pM-10 nM range for the Kd of the unmodified aptamer.17 In addition to 2AP, the base analogues 3MI and 6MAP were incorporated in position 22 of the PDGF-binding aptamer to test which of these analogues provided the best signal enhancement (see results in Table 1). The signaling aptamer containing 6MAP as a reporter molecule showed a 6.7-fold increase in fluorescence upon addition of saturating PDGF concentration, whereas 3MI produced a fluorescence increase of ∼2.5 times. IgE-Binding Aptamer. The consensus sequence of the IgEbinding aptamer published in ref 16 is 5′ GGG GCA CGT TTA TCC GTC CCT CCT AGT GGC GTG CCC C 3′. The two most stable secondary structures calculated for this aptamer are presented in Figure 4a. These two structures suggest that the aptamer might undergo a conformational change from a loophairpin structure to a structure with two loops (or vice versa) upon binding. In any case, it can be hypothesized that the protein binding site consists of nucleotides in the loop region. Therefore, screening of the 12-loop positions (positions 14-25) was performed using 2AP to replace the respective bases in the aptamer sequence. As in the above example for the PDGF-binding aptamer, screening of modified aptamers was performed using unpurified oligonucleotides, which might contain trace amounts of uncoupled 2AP, and using substoichiometric amounts of IgE. Screening results (Figure 4b) show that 2AP in position 18 shows ∼15% increase in fluorescence upon addition of IgE protein to the Analytical Chemistry, Vol. 78, No. 18, September 15, 2006
6487
Figure 3. (a) Two secondary structure conformations of the PDGF-binding aptamer calculated using mfold. (b) Five aptamers modified with 2AP were screened for a fluorescence change upon binding. The aptamer modified at the position 22 shows a significant fluorescence change after protein addition. (c) 2AP-modified aptamer fluorescence signal dependence on PDGF concentration. Squares are the experimental data; solid line is a sigmoidal function provided as a guide to the eye.
aptamer solution. The binding curve measured for this aptamer revealed a roughly 5.6-fold increase in fluorescence intensity upon addition of binding-saturating protein concentration (see Figure 4c). The fit of the binding curve data to a single component binding model results in an apparent Kd value of 46 ( 6 nM. This value is severalfold larger than the 10 nM Kd value obtained for the unmodified aptamer sequence.16 This might have been expected, since the cytosine is replaced with a larger 2-aminopurine, and this could interfere with optimal binding to the target protein. However, the binding affinity of the modified aptamer is still comparable to the affinity of unmodified aptamer, whereas the signaling functionality provides a basis to use the aptamer as a biosensor. Fluorescence results obtained using the other two fluorescent nucleotide analogues at position 18 in the IgE-binding aptamer are summarized in Table 1. The ratio of fluorescence in the bound and unbound states using 6MAP or 3MI in this aptamer is somewhat smaller than that observed for 2AP. This may be the result of sequence-dependent quenching of the fluorescence, an effect that has been previously observed for pteridin-based nucleotide analogues.14 The relative fluorescence signal change is also greatly dependent on how much the conformation of the particular base changes upon the aptamer’s binding to the target molecule. To verify that in each case the observed increase in fluorescence was due to specific interactions between the aptamer and its target protein, the fluorescence from each signaling aptamer was also measured upon addition of the other two target proteins; i.e., the thrombin-binding aptamer was tested against IgE and PDGF, the IgE-binding aptamer was tested against thrombin and PDGF, etc. In addition, the fluorescence signal from each of the 6488
Analytical Chemistry, Vol. 78, No. 18, September 15, 2006
signaling aptamers was measured after addition of bovine serum albumin (BSA). In all cases, the fluorescence of the signaling aptamers changed
