Article pubs.acs.org/ac
Mercury/Homocysteine Ligation-Induced ON/OFF-Switching of a T−T Mismatch-Based Oligonucleotide Molecular Beacon Magdalena Stobiecka,† Anthony A. Molinero, Agata Chałupa, and Maria Hepel* Department of Chemistry, State University of New York at Potsdam, Potsdam, New York 13676, United States S Supporting Information *
ABSTRACT: A molecular beacon (MB) with stem-loop (hairpin) DNA structure and with attached fluorophore−quencher pair at the ends of the strand has been applied to study the interactions of Hg2+ ions with a thymine− thymine (T−T) mismatch in Watson−Crick base-pairs and the ligative disassembly of MB·Hg2+ complex by Hg2+ sequestration with small biomolecule ligands. In this work, a five base-pair stem with configuration 5′-GGTGG...CCTCC-3′ for self-hybridization of MB has been utilized. In this configuration, the four GC base-pair binding energy is not sufficient to hybridize fully at intermediate temperatures and to form a hairpin MB conformation. The T−T mismatch built-in into the stem area can effectively bind Hg2+ ions creating a bridge, T−Hg−T. We have found that the T−Hg−T bridge strongly enhances the ability of MB to hybridize, as evidenced by an unusually large MB melting temperature shift observed on bridge formation, ΔTm = +15.1 ± 0.5 °C, for 100 nM MB in MOPS buffer. The observed ΔTm is the largest of the ΔTm found for other MBs and dsDNA structures. By fitting the parameters of the proposed model of reversible MB interactions to the experimental data, we have determined the T−Hg−T bridge formation constant at 25 °C, K1 = 8.92 ± 0.42 × 1017 M−1 from mercury(II) titration data and K1 = 1.04 ± 0.51 × 1018 M−1 from the bridge disassembly data; ΔG° = −24.53 ± 0.13 kcal/ mol. We have found that the biomarker of oxidative stress and cardiovascular disease, homocysteine (Hcys), can sequester Hg2+ ions from the T−Hg−T complex and withdraw Hg2+ ions from MB in the form of stable Hg(Hcys)2H2 complexes. Both the model fitting and independent 1H NMR results on the thymidine−Hg−Hcys system indicate also the high importance of 1:1 complexes. The high value of K1 for T−Hg−T bridge formation enables analytical determinations of low concentrations of Hg2+ (limit of detection LOD = 19 nM or 3.8 ppb, based on 3σ method) and Hcys (LOD = 23 nM, 3σ method). The conditional stability constants for Hg(Hcys)H22+ and Hg(Hcys)2H2 at 52 °C have been determined, β112 = 5.37 ± 0.3 × 1046 M−3, β122 = 3.80 ± 0.6 × 1068 M−4, respectively.
N
ovel designs of DNA sensing platforms1−3 have recently been studied extensively in view of increasing applications of biosensors for disease biomarkers in the field and in point-of-care testing. One of these sensing platforms consists of a hairpin-DNA molecular-beacon (MB) system which offers functionality and unique biorecognition propensity of nucleic acids3−8 and provides a variety of analytical signal transduction options.3,4,8−11 In this work, we have investigated MBs with a single thymine−thymine (T−T) mismatch in the area of MB stem to explore the T−T affinity toward Hg2+ ions and ligative disassembly of T−Hg−T bridge by Hg2+ sequestration with small biomolecule ligand, thioamino acid, homocysteine (Hcys). Due to the extraordinarily high melting temperature shift observed during Hg2+-binding in our 14-mer MB, higher than that reported for duplex dsDNA,12,13 we have anticipated an enhanced T−T affinity toward Hg2+ which would be advantageous for analytical and other applications. The elucidation of interactions of the T−T mismatch with Hg2+ ions is the key element to understanding the nature of artificial T−Hg−T base pair formation12 and its impact on DNA stability and mutations.14 Molecular beacons are single-stranded DNA (ssDNA) structures with complementarity introduced to oligonucleotide © 2012 American Chemical Society
5′ and 3′ ends which can, therefore, self-hybridize forming a characteristic stem-loop (or hairpin) structure. By attaching a fluorophore group (donor) and a quencher group (acceptor) at the ends of an MB sequence, the open and closed states of MB can be readily distinguished on the basis of fluorescence measurements.15−20 In the presence of a target-strand that is complementary to the MB loop sequence, the beacon changes conformation and the stem melts, turning the MB to the open (or ON) state. In the absence of an interacton, the beacon hybridizes to a closed (or OFF) state, and owing to the efficient fluorescence resonance energy transfer (FRET) between the chromophore and the acceptor, the fluorescence emission is strongly quenched. MB can be utilized in identifying molecular markers expressed in tumor cells 11,21,22 enabling early disease detection23−25 and treatment. MB for a specific ssDNAsequence in the loop for gene detection and MB aptamers have also been designed.2,26,27 Recently, an MB attached to a Received: March 4, 2012 Accepted: April 23, 2012 Published: April 23, 2012 4970
dx.doi.org/10.1021/ac300632u | Anal. Chem. 2012, 84, 4970−4978
Analytical Chemistry
Article
Figure 1. Fluorescence switching behavior of a hairpin-oligonucleotide molecular beacon (MB). (A) Fluorescence emission spectra of a 100 nM MB at (1) 22 °C, (2) 52 °C. (B) Change in fluorescence emission upon interactions of MB with Hg2+ ions and Hcys: (1) MB alone, (2) after injection of 100 nM Hg2+, (3) same as 2 but after additional injection of 200 nM Hcys. Conditions: delay time t = 15 min at 52 °C; buffer 10 mM MOPS containing 50 mM NaNO3; all concentrations are final concentrations after mixing.
nanoneedle28 for the detection of intrinsic mRNA in a living cell has been described.29 Different strategies have been applied to translate analyte binding to readily measurable fluorescence change for a fluorophore attached to MB, including mono- and multichromophores.2 A special group of MB with ON/OFF fluorescence switching has been designed for analyzing trace mercury(II) samples30−32 and biomarkers of oxidative stress, glutathione and cysteine4 where Hg2+ binds reversibly to a thymine−thymine (T−T) mismatch in the stem of a molecular beacon4,14,33−35 and is released by withdrawing action of ligands.4 Binding of Hg2+ cations to a T−T mismatch in dsDNA has been first described by Mirkin et al.3,36 and applied in colorimetric analysis of trace mercury(II) and cysteine concentrations using two kinds of functionalized gold nanoparticles. The strong ligating properties of thiols toward mercury(II)37,38 is a prerequisite for the effective competition with mercury-binding propensity of a T−T mismatch. Molecular beacons have also been immobilized on a sensor surface.10,39,40 The reporter group in MB can be replaced with a redox-active indicator providing direct means for an electrochemical detection, e.g., by attaching ferrocene moiety to one end of the MB sequence,9,41 similar as in the layer-by-layer redox-DNA sensors.42 The surface attachment enables us also to utilize an electrochemical quartz crystal nanobalance technique.43−46 In this work, a five base-pair stem with configuration 5′GGTGG...CCTCC-3′ for self-hybridization of MB has been utilized. In this configuration, the four GC base-pair binding energy is not sufficient to hybridize fully at intermediate temperatures and to form a hairpin MB conformation. The T− T mismatch built-in into the stem area can effectively bind Hg2+ ions creating a bridge, T−Hg−T, that enables the stem to fully hybridize. The primary interest of this work was to gain new insights into the nature of T−Hg−T bridge formation and its disassembly and to obtain physicochemical data on bridge formation energy and equilibrium constant. The ligating power of Hcys has been explored. Hcys is a biologically active thioamino acid, considered as the biomarker of oxidative stress which is also involved in cardiovascular disease, kidney failure,
and other diseases.47−53 The MB responses to Hg2+ and Hcys have been analyzed on the basis of the proposed model of reversible MB interactions which enabled the determination of the conditional formation constant KTHgT and the Gibbs free energy for T−Hg−T bridge formation. The model parameters, including the formation constants for Hcys complexes of Hg2+, have been determined from fluorimetric measurements of MB ON/OF switching and results of 1H NMR spectral analysis. In this paper, we present experimental data for a hairpin oligonucleotide MB with metal-ligated fluorescence switching and provide the theoretical treatment of the model MB system, to determine the T−Hg−T bridge energy, and to evaluate the stabilization energy of the bridge. The theoretical background may serve not only for the particular beacon we have investigated but also for other hairpin MB sensing platforms. In particular, it can serve as an example of a model enabling the prediction of the lowest MB concentration at which the sensitivity is still sufficient for the analyte (Hg2+ or Hcys) determination and for which the lowest limit of detection (LOD) can be obtained.
■
MATERIALS AND METHODS A molecular beacon with the sequences 5′-6-FAMCCTCCAAAAGGTGG-DABCYL-3′, where 6-FAM is a fluorescence dye 6-carboxyfluorescein and DABCYL is a quencher 4,4-dimethylamino-azobenzene-4′-carboxylic acid, was synthesized by Eurofins MWG/Operon (Huntsville, AL). Its purity was tested by HPLC, and the melting temperature Tm measurements were made in a 10 mM morpholinepropanesulfonic acid (MOPS) buffer. All other chemicals were of analytical grade purity. All concentrations of added reagents cited in this paper are final concentrations obtained after mixing. Due to the high toxicity of Hg2+, the waste collected after the measurements was disposed appropriately. A spectrometer model LS55 (Perkin-Elmer, Waltham, MA) was used for fluorimetric measurements. For further details, see Supporting Information. 4971
dx.doi.org/10.1021/ac300632u | Anal. Chem. 2012, 84, 4970−4978
Analytical Chemistry
Article
Figure 2. Effect on Hg2+ and Hcys on MB melting curves. (A) Fluorescence spectra for MB·Hg2+ as a function of temperature. (B) The same, after addition of Hcys. (C) Dependence of fluorescence emission intensity IFL on t, forward scan, for (1) MB alone, (2) MB/Hg2+. (D) Dependence of IFL vs t, backward scan, for (1) MB alone, (2) MB + Hg2+ + Hcys, kept for 40 s at each temperature, other conditions: CMB = CHg2+ = 100 nM, CHcys = 200 nM, buffer 10 mM MOPS with 50 mM NaNO3, pH 7.45.
■
the temperature to 52 °C, an enhanced fluorescence emission is observed (Figure 1A, curve 2). The emission intensity at the maximum increased more than twice, from I = 200 to I = 470. This is associated with partial dehybridization of the MB stem and more pronounced separation of the quencher from fluorophore. Due to the improvement of the signal and for the comparison with our previous data, the temperature of 52 °C was used in the following experiments. The influence of Hg2+ ions on fluorescence emission of MB is illustrated in Figure 1B. At 52 °C and in the absence of Hg2+ ions, the MB is open, in the ON state (curve 1). After the addition of 100 nM Hg2+ (final concentration) to 100 nM MB solution (final concentration), the fluorescence emission signal has decreased considerably. The fluorescence was quenched approximately 41% in relation to a pure 100 nM MB solution, from I1 = 467 to I2 = 276. It indicates that Hg2+ ions stabilize the T−T mismatch and bring the quencher and fluorophore close together turning the MB to the “OFF” state.
RESULTS AND DISCUSSION Fluorescence Switching Behavior of a T−T Mismatch Molecular Beacon. The molecular beacon used in this study was a single-stranded oligonucleotide with the sequence: 5′CCTCCAAAAGGTGG-3′ designed for the ligation of Hg2+ by the T−T mismatch in the stem area (5′-CCTCC...GGTGG-3′). A fluorophore 6-FAM was attached to 5′ end of the sequence, and a quencher DABCYL was attached to the 3′ end. Due to the hybridization between complementary bases of nucleotides, the MB forms a stem-loop structure, in which the fluorescent dye is in close proximity to the quencher. In this closed (or OFF state), the molecular beacon emits virtually no light when illuminated by the excitation beam since the FAM dye is quenched by DABCYL54 through the efficient FRET. When the MB is open (e.g., at higher temperature or by action of chemical agents), the fluorescence emission is high due to large separation of the fluorophore−quencher pair. The emission spectrum for MB for λex = 480 nm is presented in Figure 1A, curve 1, and shows a maximum at λem = 520 nm. By increasing 4972
dx.doi.org/10.1021/ac300632u | Anal. Chem. 2012, 84, 4970−4978
Analytical Chemistry
Article
Figure 3. MB fluorescence emission switching characteristics. (A) Fluorescence spectra of a 200 nM MB after addition of different concentrations of Hg2+, C0Hg, from 0 to 220 nM, step 20 nM. (B) Dependence of IFL vs C0Hg: (1) experimental data fit (52 °C), R2 = 0.988, σ = 3.31, K1,52C(MB·Hg2+) = 4.72 × 1016 M−1, β2((MB)2Hg2+) = 1.5 × 1014 M−2, β2,2((MB)2(Hg2+)2) = 1.0 × 1020 M−3, εFL= 2.6 × 109 M−1, F0 = 32 (see modeling section for details); (2) saturation (OFF state), (3) fully open MB (ON state). (C) Equilibrium concentration distribution for various complexes: (1) MBfree, (2) MB·Hg2+, (3) (MB)2Hg2+, (4) (MB)2(Hg2+)2. (D) Fluorescence emission spectra of MB·Hg2+ recorded for different additions of Hcys and heating to 52 °C; C0MB = C0Hg = 100 nM. (E) Dependence of IFL vs C0Hcys: (1) experimental data fit (52 °C), R2 = 0.985, σ = 4.53, K1(MB·Hg2+) = 5.45 × 1016 M−1, β112(Hg(Hcys)H22+) = 5.37 × 1046 M−1, β122(Hg(Hcys)2H2) = 3.80 × 1068 M−2, εFL= 3.15 × 109 M−1, F0 = 192; (2) saturation (OFF state), (3) fully open MB(ON state). (F) Equilibrium concentration distribution for various complexes: (1) Hg(Hcys)H22+, (2) Hg(Hcys)2H2, (3) MB·Hg2+.
from 20 to 80 °C, and beyond, leads to the increase in FAM fluorescence signal. This is associated with the MB denaturation proceeding through the breaking of Watson−Crick hydrogen bonds between nucleic bases in MB stem and the formation of a linear single-stranded structure. The fluorescence spectra of MB·Hg2+ complex recorded for gradually increasing temperature, before and after interactions with Hcys, are presented in Figure 2A,B, respectively. The melting profiles for rising temperature are shown in Figure 2C. It is clearly seen that the melting curve for MB·Hg2+ complex (curve 1) is shifted toward higher temperatures than the curve for MB−Hg−Hcys system in the presence of Hcys (curve 2). This shift is associated with higher stability of MB·Hg2+ complex in comparison to bare MB which is formed in the process of Hg2+ ligation by Hcys. The
The addition of 200 nM homocysteine (final concentration) results in the enhancement of the emission signal of 100 nM MB.Hg2+ (Figure 1B, curve 3). The intensity of the fluorescence emission signal rose from I2 = 276, before Hcys addition, to I3 = 498, after addition. It indicates that most of the Hg2+ ions have been removed from the stem of the MB. Therefore, the addition of Hcys has induced switching of MB to the ON state. The temporal evolution of fluorescence emission of MB shows that, after the addition of Hcys at 52 °C, emission intensity initially increases during the first 10 min and then remains nearly constant (Figure 1A, inset). The consecutive ON and OFF switching is reversible. Effect of Hg2+ and Hcys on Molecular Beacon Melting Curves. Heating MB solutions to increase the temperature 4973
dx.doi.org/10.1021/ac300632u | Anal. Chem. 2012, 84, 4970−4978
Analytical Chemistry
Article
mercury-rich environment, in DMSO. The 1H NMR spectra of 0.1 M thymidine, without and with 0.2 M Hg(NO3)2, in deuterated DMSO-d6, are shown in Figures S1 and S2. It is apparent that the proton signal from 3-NH at 11.28 ppm decreases with the addition of Hg(NO3)2. At the same time, the C-5 methyl group signal splits into two distinct peaks (Figure 4), one at 1.76 ppm for the free thymidine (If), and one at 1.81
melting temperatures determined from Figure 2 are Tm,1 = 60.7 ± 0.5 °C for MB−Hg2+ and Tm,2 = 45.6 ± 0.5 °C for MB− Hg(Hcys)2. Therefore, a pronounced melting temperature shift ΔTm = +15.1 ± 0.5 °C is observed. Therefore, the T−Hg−T binding considerably increases stability of the duplex, as expected. Note that the observed shift ΔTm = +15.1 °C is much larger than that reported by Torigoe et al.13 for single mismatch 25-base-pair DNA duplex, ΔTm = +4.1 °C, and by Miyake et al.12 for single T−T mismatch 21-mer duplex DNA, ΔTm = +10 °C, where the low stacking energy TT and AA nearest-neighbor bases were used. In the latter case, despite the low-energy stacks, the T−Hg−T bridge is most likely more strained than in our MB since the ratio of WC base-pairs per one T−Hg−T bridge is 24:1, in comparison to only 4:1 in our MB, leading to over 3-fold amplification of the melting temperature shift in the case of our MB structure. It is interesting that the MB that was turned ON by heating to 52 °C in the presence of Hcys can be readily turned OFF by adding an excess of Hg2+ ions and cooling down to 20 °C. As shown in Figure 2D, the cooling curves for MB·Hg2+ complex and for MB treated by Hg2+ leaching, followed by the additional Hg2+-loading, are virtually indistinguishable. These results indicate that Hcys added in the leaching process does not protect the MB oligonucleotide against the ingress of Hg2+ and binding to the T−T mismatch. It appears that Hcys acts solely as a withdrawing ligand for Hg2+. MB−Hg Binding Profile and Homocysteine-Mediated T−Hg−T Mercury Bridge Opening. The experimental dependence of fluorescence emission spectra on total Hg(II) concentration C0Hg is presented in Figure 3A. It is seen that a strong fluorescence quenching occurs with increasing concentrations of Hg(II). The relationship between F and C0Hg is presented in Figure 3B. As indicated earlier, the addition of homocysteine to Hgloaded MB results in the enhancement of the emission signal of MB. The leaching of Hg2+ ions from MB−Hg2+ complex due to Hcys ligation is illustrated in Figure 3D. The leaching process is fast and is virtually completed in 15 min. Therefore, all fluorescence emission spectra were obtained after 15 min following each addition of Hcys aliquots. The dependence of F on C0Hcys is presented in Figure 3E. The high sensitivity of MB measurements enables analytical determinations of low concentrations of Hg2+ (limit of detection LOD = 18.9 nM or 3.8 ppb, based on Boltzmann fit, R2 = 0.9993, σ = 3.31, N = 11, 3σ method) and Hcys (LOD = 23.3 nM, Gaussian fit, R2 = 0.9978, σ = 4.55, N = 6, 3σ method). The latter value compares favorably with Hcys levels in homocysteinuria patients which is 12−35 μM. The detailed analysis of dependences of F on C0Hg and C0Hcys based on model considerations is presented in the next sections. The comparison of fluorescence data obtained in different measurement series indicates that the quantum efficiency of the fluorophore in MB decreases over time. (During the period of 10 months, it decreased by a factor of ca. 2.) Despite the slow decrease of quantum efficiency, the MBs could be used since, in each measurement series, the fluorescence emission intensities for the ON and OFF states were determined, and thus, the relative emission intensity scale could be applied. Only a small decrease in sensitivity was usually observed. NMR Titration of Hg−T Complexes. The evidence of Hg binding to the thymine base has also been obtained in NMR measurements. Earlier studies55 have shown that cytidine, adenosine, and guanosine form 1:1 mercury(II) complexes in
Figure 4. (A−E) 1H NMR spectra for the upfield region for 0.1 M thymidine + 0.2 M HgCl2 solutions in DMSO-d6 containing homocysteine with concentration CHcys [M]: (A) 0.0, (B) 0.05, (C) 0.1, (D) 0.15, and (E) 0.2. (F) Dependence of CT,free on CHcys determined from changes of C-5 methyl group peaks at 1.76 and 1.81 ppm in 1H NMR spectra A−E, with line showing theoretical dependence for K1 = 4.3 M−1, β11 = 2 × 109 M−1 (HgHcys), β12 = 1 × 1016 M−2 (Hg(Hcys)2).
ppm for the thymidine-mercury complex (Ib). By setting the integration of the peak for vinyl proton on C-6 to 1, it was determined that the relative ratio of mercury complex to free thymidine was 44:56. Use of the 3-NH proton signal at 11.28 ppm gave virtually identical results. In an effort to determine the effect of Hcys on the thymidine−mercury complex, 1H NMR spectra for increasing concentrations of Hcys were obtained (Figure 4 and, in full view, Figure S2). They clearly indicate that Hcys competes with T for the mercury. As the concentration ratio of CHcys:CHg(II) increases from 0.25:1 to 1:1, the ratio of the bound-to-free signal intensities Ib/If dramatically decreases. The decrease of the bound-T concentration on CHcys, illustrated in Figure 4F, corroborates our findings from MB 4974
dx.doi.org/10.1021/ac300632u | Anal. Chem. 2012, 84, 4970−4978
Analytical Chemistry
Article
Figure 5. (A) Effect of Hg2+ on bound-MB concentration COFF and (B) dependence of F−FOFF on C0Hg calculated from eq 3, for different values of K1 [M−1]: (1) 1.1 × 1015; (2) 3.6 × 1015; (3) 1.8 × 1016; (4) 1.8 × 1017; C0MB = 1 × 10−7 M; εFL = 3.2 × 109 M−1; (5) high affinity limit; (6) OFF state (A) or ON state (B) level. (C) Dependence of θ on C0L calculated from eqs 7−9 for different values of K2 [M−1]: (1) 3.6 × 1013; (2) 3.6 × 1014; (3) 3.6 × 1015; (4) 3.6 × 1016; C0MB = 1 × 10−7 M. (D) Dependence of F−FOFF on C0L calculated for conditions (C) and εFL = 3.2 × 109 M−1, Fmin = 170, Fmax = 490.
experiments concluding that Hcys competes favorably with T− T mismatches for Hg2+. The analysis of complexes involved (Figure 4F) indicates the predominant role of 1:1 Hg−Hcys complex during the initial stage of Hg(II) sequestration followed by the sharp rise of the 1:2 complex concentration at the end of the titration. The value of K1 = 4.3 M−1 is in line with that for adenosine and guanosine55 (7.2 and 5.9 M−1, respectively). Obviously, there are considerable differences between the thymidyne−Hg(II)−Hcys system studied by NMR and the MB−Hg(II)−Hcys system studied using fluorescence spectroscopy. However, besides the differences, there are also similarities for these two systems. The main similarity is in the direct interaction of Hg(II) with thymine at the same location 3-NH. Moreover, the NMR measurements have provided the key evidence of the formation of 1:1 complexes which otherwise could have been overlooked. Model of Reversible Metal-Ligated MB-Switching. Whereas the molecular beacons for various analytical applications have been successfully developed, the analysis of the fluorescence emission as a function of analyte concentration has been largely left for semiempirical treatments. Here, we present a simple approach describing the fluorescence switching-OFF observed upon the ligation of Hg2+ ions by MB and switching-ON in response to Hg2+-sequestration by Hcys. For the theoretical treatment of the dependence of MB fluorescence emission on CHg, we consider the following reaction as the main reaction of the molecular beacon turnOFF process: MB + Hg 2 + = MB· Hg 2 + CON C Hg2 + COFF
K1 =
COFF CONC Hg2 +
(1b)
where COFF = CMB·Hg2+ and CON = CMB stand for the molecular beacon in the OFF and ON state, respectively. The process of Hg2+ extraction with ligand L proceeds according to the main reaction equation: MB· Hg 2 + + 2Lz − = MB + HgL2 2 − 2z COFF CL CON C HgL2
(2)
where L is the ligand. To illustrate the approach to equilibrium problem for a system represented by reaction eq 1, we present the solution for COFF in the form (for full derivation, see Supporting Information): COFF =
0 0 CMB C Hg 0 0 CMB + C Hg
+ ΦMB (3a)
The first term in the right-hand-side of this equation represents a hyperbolic average of C0MB and C0Hg for quick estimate of COFF, and the second term is the correction factor ΦMB defined as follows: ΦMB =
0 2 0 2 0 0 + K1−1C Hg ((CMB ) + (C Hg ) + K1−1CMB ) 0 0 + C Hg 2(CMB ) 0 0 − 0.5(K1−1 + CMB + C Hg )
1− (1a)
0 0 4CMB C Hg 0 0 2 + C Hg (K1−1 + CMB )
(3b)
0
The dependence of F vs C Hg for switching the molecular beacon OFF is presented in Figure 5A,B for several values of K1
with the conditional equilibrium constant K1 4975
dx.doi.org/10.1021/ac300632u | Anal. Chem. 2012, 84, 4970−4978
Analytical Chemistry
Article
equilibrium constant for T−Hg−T bridge formation at 52 °C, K1,52C = 4.72 ± 0.42 × 1016 M−1, and its Gibbs free energy ΔG°52C = −24.8 ± 0.6 kcal/mol. The value of β12 = 10−6.34 M38 (Hg2+ + 2H2O = Hg(OH)2 + 2H+) has been taken into account for establishing the free Hg2+ concentration, CHg. Note that the value of K1 remains unchanged for the ionic strength μ = 0 and for low μ < 0.1 M due to the apparent canceling of activity coefficients. In addition to the simple hairpin conformation, we have also considered three other MB−Hg2+ complexes, linear (open) MB(Hg2+), (MB)2Hg2+, and (MB)2(Hg2+)2, with the conditional formation constants, K1′, β2, β22. The contributions from other MB−Hg2+ complexes are minor, as shown in Figure 3C. The fitting of model parameters to the experimental data for MB·Hg2+ leaching with Hcys leads to the following values at 52 °C: K1,52C = 5.45 ± 0.51 × 1016 M−1 and ΔG°52C = −24.9 ± 0.7 kcal/mol. Because of the conclusions arising from the 1H NMR experiments performed in this work concerning the involvement of Hg(Hcys)H22+, we have included also this complex in the detailed analysis of MB switching. (The formation of higher HgLn complexes, HgL3 and HgL4, is neglected due to the low concentration of Hcys.) The conditional stability constants for Hcys complexes, Hg(Hcys)H22+ and Hg(Hcys)2H20, obtained from our data are as follows: β112 = 5.37 ± 0.32 × 1046 M−3 (log β112 = 46.73; Hg2+ + Hcys2‑ + 2H+ = Hg(Hcys)H22+), β122 = 3.80 ± 0.43 × 1068 M−4 (log β112 = 68.58; Hg2+ + 2Hcys2‑ + 2H+ = Hg(Hcys)2H20), for 52 °C, ionic strength μ = 0.05 M. They are consistent with literature values for relevant Hg(II)− cysteine complexes38 (log β112 = 46.12, log β122 = 62.04) and Hg(II)−glutathione complexes38 (log β112 = 44.78, log β122 = 61.07), all for μ = 0. The enthalpy of T−Hg−T bridge formation was calculated on the basis of ΔG52C determined above and assuming that ΔS° remains the same as for the duplex DNA since only a T−T mismatch and Hg2+ ion dehydration are involved in the bridge formation. So, ΔS° should be insensitive to the detailed differences between the T−Hg−T bridge strain in MB and in duplex DNA. Thus, the value ΔS° = 11.8 ± 0.8 cal mol−1 K−1 has been assumed after Torigoe et al.13 The calculated value of enthalpy, ΔH° ≈ ΔH°52C = ΔG°52C +TΔS° = −24.8 + (273.15 + 52) × 11.8/ 1000 = −20.9 ± 0.6 kcal/mol, indicates a strong bonding since −ΔH° is larger than that for a single isolated hydrogen bond formation (−ΔH° ≈ 1−5 kcal/mol). The value of ΔG° at 25 °C can now be calculated as follows: ΔG° = ΔH° − TΔS° = −20.9 −298.15 × 11.8/1000 = −24.4 ± 0.6 kcal/mol. It is seen that the T−Hg−T bridge formation is mainly driven by enthalpic contribution to ΔG° (ca. 84%). This is corroborated by our 1H NMR measurements showing very weak temperature dependence of the Hg(II)−T complex stability. The translation of K1,52C values to 25 °C leads to K1 = 8.92 ± 0.42 × 1017 M−1 (for bridge formation data) and K1 = 1.04 ± 0.51 × 1018 M−1 (for bridge disassembly data). This confirms the results obtained by Torigoe et al.,13 carried out for a T−Hg−T bridge in a 25-base-pair DNA duplex with a single T−T mismatch. For the sake of comparison with our data, their constant, K1* = 5.61 ± 1.35 × 105 M−1, can be recalculated to account for the presence of Hg(OH)2 in aqueous solutions and the respective pH dependence; thus, K1 = 2.0 × 1014 at pH = 7.45. Hence, the value of K1 for our MB is 3 orders of magnitude higher than that for a duplex dsDNA. The higher K1 value likely results from the reduced strain on the T−Hg−T bridge due to the short length of the MB stem.
to show how K1 affects the response of MB to the additions of Hg(II). The dependence of fluorescence intensity F on C0Hg has been obtained using the translation FMB = F0 + εFCON
(4) −1
with εF = 2.6 × 10 M and F0 = 32, corresponding to the experimental readings. It is seen that the lower the value of K1 becomes, the larger the deviation of the F−C0Hg profile from the limiting line 5 becomes. It can be shown that, only for K1 ≫ 1 × 1017 M−1, the plot remains linear up to the saturation concentration C0Hg ≈ C0MB. The value of K1 has been adjusted by a factor 102pH‑logβ = 3.63 × 108 where log β12 = −6.34, to account for Hg(OH)2(aq) complex formation and consistency with experimental measurements. The high values of K1 are especially advantageous for analytical applications of molecular beacons for determinations of low concentrations of Hg2+. Further model improvements by considering other species involved in MB switching are discussed vide inf ra. The dependence of F on CL is more complex, even for a single complex HgL22−2z involved in MB reactivity, since it involves solving 3rd degree polynomial (SI, eq 10). Below, we present a simple calculation procedure to generate a family of F vs C0L curves (where C0L is the total ligand concentration) for different values of K2 to see how the value of K2 and the ligand complexing power toward Hg2+ affect the characteristics. The equation for conditional equilibrium constant for reaction 2 can be rewritten as follows 9
K2 =
0 (1 − θ )2 CMB
θC L2
(5)
where θ is the fraction of MB that is in the OFF state, as defined by θ=
COFF 0 CMB
(6)
It enables one to calculate the equilibrium ligand concentration CL for preset values of θ C L = (1 − θ )
0 CMB θK 2
(7)
and the total ligand concentration 0 C L0 = C L + 2(1 − θ )CMB
(8)
The fluorescence emission is given by 0 FMB = εF(1 − θ )CMB
(9)
C0L
Hence, the plots of FMB vs can be readily obtained. The effect of Hg2+-withdrawing ligand L on bound-MB concentration θ and fluorescence emission F−FOFF, calculated using eqs 7−9, is illustrated in Figure 5C,D, for different values of K2, for C0MB = 1 × 10−7 M. It is seen that with the decreasing value of K2 the slope of the dependence F−F0 vs CL is decreasing. For the operation of the beacon in assays for a metal-withdrawing ligand, the higher value of K2 increases the sensitivity and linearity. On the other hand, the lower value of K2 increases the dynamic range while decreasing the method sensitivity. Therefore, K2 acts for the ligand in a similar fashion as K1 acts for metal cation determination, discussed earlier. By fitting parameters of the proposed model of reversible MB interactions to the experimental data for MB quenching by Hg2+ ions (Figure 3B), we have determined the conditional 4976
dx.doi.org/10.1021/ac300632u | Anal. Chem. 2012, 84, 4970−4978
Analytical Chemistry The presented theory allows one to predict the MB behavior at different MB concentrations which is not directly deducible from experiments without theoretical modeling. This is especially important when evaluating the limit of detection (LOD), since a lower LOD can be found for lower MB concentration, with a trade-off in the diminished sensitivity of the method. Also, it follows from the presented theory that the dynamic range increases with decreasing MB concentration.
■
REFERENCES
(1) Mirkin, C. A.; Letzinger, R. L.; Mucic, R. C.; Storhoff, J. J. Nature 1996, 382, 607−609. (2) Nutiu, R.; Li, Y. Methods 2005, 37, 16−25. (3) Lee, J.-S.; Ulmann, P. A.; Han, M. S.; Mirkin, C. A. Nano Lett. 2008, 8, 529−533. (4) Xu, H.; Hepel, M. Anal. Chem. 2011, 83, 813−819. (5) Hepel, M.; Stobiecka, M.; Nowicka, A. In Biosensors and the Environment Health; Hunter, R. J., Preedy, V. R., Eds.; CRC Press: London, 2012 (in press); ISBN 978-1-57808-735−8. (6) Obraztsov, I.; Noworyta, K.; Kutner, W.; Gadde, S.; D’Souza, F. Phys. Status Solidi B 2007, 244, 3861−3867. (7) Galandova, J.; Ovadekova, R.; Ferancova, A.; Labuda, J. Anal. Bioanal. Chem. 2009, 394, 855−861. (8) Zhang, J.; Lao, R.; Song, S.; Yan, Z.; Fan, C. Anal. Chem. 2008, 80, 9029−9033. (9) Lubin, A. A.; Lai, R. Y.; Baker, B. R.; Heeger, A. J.; Plaxco, K. W. Anal. Chem. 2006, 78, 5671−5677. (10) Wei, F.; Wang, J.; Liao, W.; Zimmermann, B. O.; Wong, D. T.; Ho, C. M. Nucleic Acids Res. 2008, 36, e65 1−7. (11) Herr, E. A.; Hatch, A. V.; Throckmorton, D. J.; Tran, H. M.; Brennan, J. S.; Giannobile, W. V.; Singh, A. K. Proc. Natl. Acad. Sci. U.S.A. 2007, 104, 5268−5273. (12) Miyake, Y.; Togashi, H.; Tashiro, M.; Yamaguchi, H.; Oda, S.; Kudo, M.; Tanaka, Y.; Kondo, Y.; Sawa, R.; Fujimoto, T.; Machinami, T.; Ono, A. J. Am. Chem. Soc. 2006, 128, 2172−2173. (13) Torigoe, H.; Ono, A.; Kozasa, T. Chem.Eur. J. 2010, 16, 13218−13225. (14) Urata, H.; Yamaguchi, E.; Funai, T.; Matsumura, Y.; Wada, S. Angew. Chem. 2010, 122, 6666−6669. (15) Tan, W.; Fang, X.; Li, J.; Liu, X. Chem.Eur. J. 2000, 6, 1107− 1111. (16) Li, J. J.; Fang, X.; Schuster, S. M.; Tan, W. Angew. Chem. 2000, 112, 1091−1094. (17) Tan, W.; Wang, K.; Drake, T. J. Curr. Opin. Chem. Biol. 2004, 8, 547−553. (18) Goel, G.; Kumar, A.; Puniya, A. K.; Chen, W.; Singh, K. J. Appl. Microbiol. 2005, 99, 435−442. (19) Li, Y.; Zhou, X.; Ye, D. Biochem. Biophys. Res. Commun. 2008, 373, 457−461. (20) Wang, K.; Tang, Z.; Yang, C. J.; Kim, Y.; Fang, X.; Li, W.; Wu, Y.; Medley, C. D.; Cao, Z.; Li, J.; Colon, P.; Lin, H.; Tan, W. Angew. Chem., Int. Ed. 2008, 47, 2−17. (21) Gormally, E.; Vineis, P.; Matullo, G.; Veglia, F.; Carboux, E.; LeRoux, E.; Peluso, M.; Garte, S.; Guarrera, S.; Munnia, A.; et al. Cancer Res. 2006, 66, 6871−6876. (22) Xue, Y.; An, R.; Zhang, D.; Zhao, J.; Wang, X.; Yang, L.; He, D. Eur. J. Obstet. Gynecol. Reprod. Biol. 2011, 159, 204−208. (23) Culha, M.; Stokes, D. L.; Griffin, G. D.; Vo-Dinh, T. Biosens. Bioelectron. 2004, 19, 1007−1012. (24) McKillen, J.; Hjertner, B.; Millar, A.; McNeilly, F.; Belak, S. r.; Adair, B.; Allan, G. J. Virol. Methods 2007, 140, 155−165. (25) Takacs, T.; Jeney, C.; Kovacs, L.; Mozes, J.; Benczik, M.; Sebe, A. J. Virol. Methods 2008, 149, 153−162. (26) Piatek, A. S.; Tyagi, S.; Pol, A. C.; Telenti, A.; Miller, L. P.; Kramer, F. R.; Alland, D. Nat. Biotechnol. 1998, 16, 359−363. (27) Wu, J.; Huang, C.; Cheng, G.; Zhang, F.; He, P.; Fang, Y. Electrochem. Commun. 2009, 11, 177−180. (28) Kihara, T.; Yoshida, N.; Kitagawa, T.; Nakamura, C.; Nakamura, N.; Miyake, J. Biosens. Bioelectron. 2010, 26, 1449−1454. (29) Bratu, D. P.; Cha, B.; Mhlanga, M. M.; Kramer, F. R.; Tyagi, S. Proc. Natl. Acad. Sci. U.S.A. 2003, 100, 13308−13313. (30) Tang, Y.; He, F.; Yu, M.; Feng, F.; An, L.; Sun, H.; Wang, S.; Li, Y.; Zhu, D. Macromol. Rapid Commun. 2006, 27, 389−392.
CONCLUSIONS Small 14-nucleobase molecular beacons with a T−T mismatch in the stem area and a fluorophore−quencher pair (FAM and DABCYL, respectively) attached to the 5′ and 3′ ends of MB sequence have been applied to study binding of low concentration Hg2+ cations and to determine the T−Hg−T bond energy ΔG°. We have found that the T−Hg−T bridge formation constant for MB studied in this work is 3 orders of magnitude higher than that for a duplex DNA. Also, the MB stabilization energy due to the T−Hg−T bridge formation and the melting temperature shift are higher for our MB, indicating the relaxed strain on the T−Hg−T bonding due to the short length of the MB stem used in this work. The fluorescent oligonucleotide beacon structure is a useful platform for determination of bond energies and stability constants of Hg(II)−ligating complexes and for the development of sensitive assays for Hg2+ and biologically relevant thiols. The high negative value of the free energy of T−Hg−T bridge formation, ΔG° = −24.4 ± 0.6 kcal/mol, determined in this work, enables trace analysis of Hg2+ and Hcys solutions. The hairpin oligonucleotide MB we have investigated in this work is not an ideal switch, and the fluorescence emission does not change abruptly between two states, i.e., all-ON and all-OFF. It rather follows the less ideal switching characteristics with a twostate sigmoidal transfer function with a large transition width where the fluorescence is concentration dependent in a linear or pseudolinear manner. In fact, fitting the two-state Boltzmann function to the MB fluorescence data gives the lowest deviation for experimental points than any other function tested. The spring-loaded hairpin structure, as such, is a physical switch. Yet the wide temperature range over which the MB melting takes place shows that it is not an abrupt switch, but one that fulfills dynamic chemical equilibria, such as those we have presented in this paper in a model of reversible metal-ligated MB switching. ASSOCIATED CONTENT
S Supporting Information *
Further experimental details, derivation of equations for MB interactions model, and two 1H NMR figures. This material is available free of charge via the Internet at http://pubs.acs.org.
■
ACKNOWLEDGMENTS
This work was partially supported by the U.S. DoD Grant AS073218.
■
■
■
Article
AUTHOR INFORMATION
Corresponding Author
*Phone: +1-315-267-2264. Fax: +1-315-267-3170. E-mail:
[email protected]. Website: www2.potsdam.edu/hepelmr. Present Address †
Department of Physics, Warsaw University of Life Sciences SGGW, 02776 Warsaw, Poland. Notes
The authors declare no competing financial interest. 4977
dx.doi.org/10.1021/ac300632u | Anal. Chem. 2012, 84, 4970−4978
Analytical Chemistry
Article
(31) Knecht, M. R.; Sethi, M. Anal. Bioanal. Chem. 2009, 394, 33−46. (32) Torabia, S.-F.; Lu, Y. Faraday Discuss. 2011, 149, 125−135. (33) Kiy, M. M.; Jacobi, Z. E.; Liu, J. Chem.Eur. J. 2012, 18, 1202− 1208. (34) Ono, A.; Togashi, H. Angew. Chem., Int. Ed. 2004, 43, 4300− 4302. (35) Chiang, C.-K.; Huang, C.-C.; Liu, C.-W.; Chang, H.-T. Anal. Chem. 2008, 80, 3716−3721. (36) Lee, J.-S.; Han., M. S.; Mirkin, C. A. Angew. Chem., Int. Ed. 2007, 46, 4093−4096. (37) Stricks, W.; Kolthoff, I. M. J. Am. Chem. Soc. 1953, 75, 5673− 5681. (38) Cardiano, P.; Falcone, G.; Foti, C.; Sammartano, S. J. Chem. Eng. Data 2011, 56, 4741−4750. (39) Liu, X.; Sun, C.; Wu, H.; Zhang, Y.; Jiang, J.; Shen, G.; Yu, R. Electroanalysis 2010, 22, 2110−2116. (40) Carasquilla, C.; Li, Y.; Brennan, J. D. Anal. Chem. 2011, 83, 957−965. (41) Xiao, Y.; Lubin, A. A.; Baker, B. R.; Plaxco, K. W.; Heeger, A. J. Proc. Natl. Acad. Sci. U.S.A. 2006, 103, 16677−16680. (42) Nowicka, A. M.; Kowalczyk, A.; Stojek, Z.; Hepel, M. Biophys. Chem. 2010, 146, 42−53. (43) Papadakis, G.; Tsortos, A.; Bender, F.; Ferapontova, E. E.; Gizeli, E. Anal. Chem. 2012, 84, 1854−1861. (44) Skládal, P. J. Braz. Chem. Soc. 2003, 14, 491−502. (45) Hepel, M. In Interfacial Electrochemistry. Theory, Experiment and Applications; Wieckowski, A., Ed.; Marcel Dekker, Inc.: New York, 1999, pp 599−630. (46) Stobiecka, M.; Hepel, M. Biomaterials 2011, 32, 3312−3321. (47) Nygard, O.; Vollset, S. E.; Refsum, H.; Brattstrom, L.; Ueland, P. M. J. Intern. Med. 1999, 246, 425−454. (48) Medina, M. A.; Urdiales, J. L.; Amores-Sanchez, M. I. Eur. J. Biochem. 2001, 268, 3871−3882. (49) Lao, J. I.; Beyer, K.; Ariza, A. Drug Dev. Res. 2004, 62, 221−230. (50) Chamberlain, K. L. J. Am. Acad. Nurse Prac. 2005, 17, 90−95. (51) Stobiecka, M.; Deeb, J.; Hepel, M. Biophys. Chem. 2010, 146, 98−107. (52) Zalups, R. K.; Barfuss, D. W. J. Am. Soc. Nephrol. 1998, 9, 551− 61. (53) Zalups, R. K.; Ahmad, S. J. Am. Soc. Nephrol. 2004, 15, 2023− 2031. (54) Broude, N. E. Trends Biotechnol. 2002, 20, 249−256. (55) Kan, L. S.; Li, N. C. J. Am. Chem. Soc. 1970, 92, 4823−4827.
4978
dx.doi.org/10.1021/ac300632u | Anal. Chem. 2012, 84, 4970−4978
