Detection of guanine–adenine mismatches by surface plasmon resonance s
http://www.100md.com
《核酸研究医学期刊》
1 Department of Synthetic Chemistry and Biological Chemistry, Faculty of Engineering, Kyoto University, Kyoto, Japan and 2 PRESTO, Japan Science and Technology Agency (JST), Kyoto 615-8510, Japan
*To whom correspondence should be addressed. Tel: +81 75 383 2756; Fax: +81 75 383 2759; Email: nakatani@sbchem.kyoto-u.ac.jp
ABSTRACT
We have discovered a new molecule naphthyridine–azaquinolone hybrid (Npt–Azq) that strongly stabilized the guanine-adenine (G-A) mismatch in duplex DNA. In the presence of Npt–Azq, the melting temperature (Tm) of 5'-d(CTA ACG GAA TG)-3'/3'-d(GAT TGA CTT AC)-5' containing a single G-A mismatch increased by 15.4°C, whereas fully matched duplex increased its Tm only by 2.2°C. Npt–Azq was immobilized on the sensor surface for the surface plasmon resonance (SPR) assay to examine SPR detection of duplexes containing a G-A mismatch. Distinct SPR signals were observed when 27mer DNA containing a G-A mismatch was analyzed by the Npt–Azq immobilized sensor surfaces, whereas the signal of the fully matched duplex was 6-fold weaker in intensity. The SPR signals for the G-A mismatch were proportional to the concentration of DNA in a range up to 1 μM, confirming that the SPR signal is in fact due to the binding of the G-A mismatch to Npt–Azq immobilized on the surface. Examination of all 16 G-A mismatches regarding the flanking sequence revealed that the sensor surface reported here is applicable to eight flanking sequences, covering 50% of all possible G-A mismatches.
INTRODUCTION
As a follow-on to the complete sequencing of the human genome, typing of single nucleotide polymorphisms (SNPs) in an array of disease-related genes is expected to be an indispensable technique for realizing personalized medicine. A number of methods have been developed for SNP typing (1–3), but much study is still needed to design new typing methods that are simple in operation, rapid and accurate in analysis, and low in cost. One of the challenges we have focused on is the reduction and eventual redundancy of labeled oligonucleotides for analysis. The expense of fluorescently labeling oligonucleotides and PCR products limits their application in large-scale typing by many currently available methods. We have reported a conceptually novel method of SNP typing that detects mismatch-containing duplexes with a small molecular ligand immobilized on a gold surface (4). Hybridization of two sets of duplex DNAs that differ from each other by a single nucleotide produces a DNA heteroduplex containing a single mismatched site. Mismatch-containing duplexes can be separated from homoduplexes by either gel electrophoresis (5,6), chemical and enzymatic cleavages at the mismatched site (7–9), or selective capture with mismatch-binding proteins (10,11). While these heteroduplex analyses applied to low-throughput screening are essentially free from oligonucleotide labeling, new technologies for high-throughput analyses are yet to be established. A small molecular ligand that selectively binds to a mismatched site could replace mismatch-binding proteins and bring an innovation to heteroduplex analyses (5,12–20). The ligand naphthyridine dimer (Npt–Npt) strongly and selectively binds to guanine–guanine mismatches in duplex DNA (5,12) (Fig. 1). The binding constant to a G-G mismatch in the 5'-CGG-3'/3'-GGC-5' sequence is 1.9 x 107 M–1. Npt–Npt consisting of two 2-amino-1,8-naphthyridine (Npt) chromophores, and a linker connecting the chromophores is designed so that each Npt produces three hydrogen bonds to each one of the guanines in the G-G mismatch, and the resultant naphthyridine–guanine pair is stabilized by stacking with the flanking base pairs. The proposed binding of Npt–Npt to the G-G mismatch has been verified by the 2D- NOESY spectrum of the complex (12). In addition to Npt–Npt, ligands that selectively and strongly bind to G-A, G-T and A-A mismatches are needed to accomplish SNP typing by a mismatch binding ligand. Taking into account the structure of Npt–Npt as a clue for the molecular design of ligands binding to a G-A mismatch, we have discovered naphthyridine–azaquinolone hybrids (Npt–Azq) where one Npt chromophore in Npt–Npt is replaced by a 8-azaquinolone chromophore (Azq) having complementary hydrogen-bonding surfaces to adenine (Fig. 2). Herein, we report the remarkable stabilization of G-A mismatch DNA by Npt–Azq and the first synthesis of a surface plasmon resonance (SPR) sensor detecting a G-A mismatch by immobilization of Npt–Azq on a gold surface.
Figure 1. Formulas of Npt- and Azq-based hybrid molecules.
Figure 2. Hydrogen bonding patterns of Npt-G and Azq-A.
MATERIALS AND METHODS
Chemistry
In order to know the structure–activity relationship for the binding of Npt–Azq hybrid to G-A mismatches, hybrid molecules consisted of chromophores with different hydrogen-bonding groups were synthesized. These chromophores include Npt, Azq, quinoline (Q) and 2-aminoquinoline (AQ). Synthesis of the hybrid molecules consisting of Npt with Azq and other heterocycles is straightforward using N-(tert-butoxycarbonyl)imino-3,3'-bis(pentafluorophenyl propionate) (12), in which two carboxyl groups are activated as a pentafluorophenyl ester. Mono-substitution of the pentafluoroester with 2-amino-7-methylnaphthyridine produces intermediate amide Npt-OC6F5 that subsequently reacts with the heterocyclic amines to give hybrid ligands (Fig. 3). A similar procedure was used for the synthesis of Azq-based hybrid ligands (Fig. 4). Details of the synthetic procedure is described in the Supplementary Material.
Figure 3. Synthetic scheme of Npt-based hybrids.
Figure 4. Synthetic scheme of Azq-based hybrids.
Synthesis of Npt–Azq
To a solution of N-(tert-butoxycarbonyl)imino-3,3'-bis(pentafluorophenyl propionate) (1.5 g, 2.53 mmol) in dry DMF (5 ml) was added 2-amino-7-methyl-1,8-naphthyridine (180 mg, 1.14 mmol) and diisopropylethylamine (163 mg, 1.26 mmol). The reaction mixture was stirred at room temperature for 15 h. The solvent was evaporated to dryness and the residue was purified by column chromatography on silica gel to give Npt-OC6F5 (1.29 g, 90%) as pale yellow solids: 1H-NMR (CDCl3, 400 MHz) = 9.01 (br, 1H), 8.44 (d, 1H, J = 8.8 Hz), 8.12 (d, 1H, J = 8.8 Hz), 7.99 (d, 1H, J = 8.0 Hz), 7.26 (d, 1H, J = 8.0 Hz), 3.66 (m, 4H), 2.90 (m, 2H), 2.74 (m, 2H), 2.73 (s, 3H), 1.42 (s, 9H), 13C-NMR (CDCl3, 100 MHz) = 163.6, 155.3, 154.6, 153.4, 142.6, 140.1, 139.5, 138.5, 136.6, 121.9, 118.8, 114.5, 80.9, 44.9, 44.4, 37.6, 37.2, 36.7, 33.4, 32.7, 28.6, 25.8, FABMS (NBA), m/e 569 , HRMS calc. for C26H26O5N4F5 569.1821, found 569.1827.
To a solution of Npt-OC6F5 (300 mg, 0.53 mmol) in dry DMF (2 ml) was added 7-(aminomethyl)hydro-8-azaquinolin-2-one (93 mg, 0.53 mmol) and diisopropylethylamine (77 mg, 0.6 mmol). The reaction mixture was stirred at room temperature for 15 h. The solvent was evaporated to dryness and the residue was purified by column chromatography on silica gel to give N-Boc-Npt–Azq (227 mg, 77%) as pale yellow solids: 1H-NMR (CDCl3, 400 MHz) = 11.29 (br, 1H), 9.08 (br, 1H), 8.38 (d, 1H, J = 8.8 Hz), 8.06 (d, 1H, J = 8.0 Hz), 7.95 (d, 1H, J = 8.0 Hz), 7.81 (d, 1H, J = 8.0 Hz), 7.64 (d, 1H, J = 9.6 Hz), 7.27 (d, 1H, J = 8.0 Hz), 7.22 (d, 1H, J = 8.0 Hz), 6.64 (d, 1H, J = 9.6 Hz), 4.63 (d, 2H, J = 6.0 Hz), 3.58 (t, 2H, J = 6.8 Hz), 3.57 (t, 2H, J = 6.8 Hz), 2.71 (t, 2H, J = 6.8 Hz), 2.70 (s, 1H), 2.54 (t, 2H, J = 6.8 Hz), 1.36 (s, 9H), 13C-NMR (CDCl3, 100 MHz) = 171.7, 163.9, 163.4, 159.8, 155.7, 154.5, 153.6, 149.3, 139.3, 139.2, 137.4, 136.9, 123.1, 121.9, 118.8, 118.2, 114.6, 114.0, 80.5, 44.8, 41.8, 37.5, 28.6, 25.7, FABMS (NBA), m/e 560 , HRMS calc. for C29H34O5N7 560.2621, found 560.2618.
To a solution of N-Boc-Npt–Azq (62 mg, 0.11 mmol) in CHCl3 (3 ml) was added ethyl acetate containing 4 M HCl (2 ml) at room temperature and the reaction mixture was stirred at room temperature for 0.5 h. The solvent was evaporated to dryness to give hydrochloride of Npt–Azq (quantitative yield) as white solids: 1H-NMR (CDCl3, 400 MHz) = 11.45 (br, 1H), 8.67 (t, 1H, J = 5.6 Hz), 8.31 (d, 1H, J = 8.8 Hz), 7.98 (d, 1H, J = 8.8 Hz), 7.88 (d, 1H, J = 8.0 Hz), 7.65 (d, 1H, J = 7.6 Hz), 7.53 (d, 1H, J = 9.6 Hz), 7.15 (d, 1H, J = 7.6 Hz), 7.14 (t, 1H, J = 8.0 Hz), 6.58 (d, 1H, J = 9.6 Hz), 4.65 (d, 2H, J = 5.6 Hz), 3.08 (t, 2H, J = 6.0 Hz), 3.06 (t, 2H, J = 6.0 Hz), 2.65 (t, 2H, J = 6.0 Hz), 2.58 (t, 2H, J = 6.0 Hz), 2.57 (s, 3H), 13C NMR (CD3OD, 100 MHz) = 171.8, 171,3, 165.0, 160.2, 159.7, 157.0, 148.5, 147.3, 146.4, 140.3, 140.2, 138.8, 122.5, 120.9, 120.2, 117.6, 117.0, 114.8, 44.1, 43.9, 43.1, 32.5, 30.5, 19.7, FABMS (NBA), m/e 460 , HRMS calc. for C24H26O3N7 460.2095, found 460.2097.
Synthesis of Npt–Azq having an aminoalkyl linker for immobilization onto SPR sensor surface
SPR sensors having Npt–Azq on their surface were synthesized to examine SPR detection of the G-A mismatch in a flow system (Fig. 5). Npt–Azq was immobilized on a dextran matrix coated gold surface (CM5 chip, BIAcore) through a bivalent linker of N-Boc-aminoaldehyde. First, Npt–Azq was tethered by a reductive amination to the linker, which was efficiently prepared from Boc-protected 4-aminobutanoic acid and 3-aminopropionaldehyde diethylacetal (Supplementary Material). Deprotection of a Boc group produced a primary amine, which was subsequently immobilized on the sensor surface by a coupling between the amino group and an activated carboxyl group on the CM5 chip using a standard method with 1-(3-dimethylaminopropyl)-3-ethylcarbodiimide hydrochloride (EDCI) and N-hydroxysuccinimide (NHS).
Figure 5. Synthetic scheme of Npt–Azq immobilized sensor surfaces.
Preparation of a sensor chip carrying Npt–Azq on its surface
SPR measurements were performed with a BIAcore 2000 system (BIAcore, Uppsala, Sweden). Immobilization of Npt–Azq on a sensor chip CM5 (carboxymethylated dextran surface, BIAcore) was carried out using amine coupling kit (BIAcore) in a continuous flow of HBS-N buffer (10 mM HEPES, pH 7.4) containing NaCl (150 mM) at a flow rate of 10 μl/min. A solution (70 μl) of NHS (0.05 M) and EDCI (0.2 M) was injected using the QUICKINJECT command to activate the carboxymethylated dextran surface of a CM5 sensor chip. A solution (70 μl) of Npt–Azq having an aminoalkyl linker (2 mM in borate buffer, pH 9.2) was injected using the QUICKINJECT command on the activated surface. Residual activated surface was completely blocked by injection of a solution (20 μl) of ethanolamine hydrochloride (1.0 M, pH 8.5). Non-covalently bound material was removed by washing with 5 μl of 50 mM NaOH to produce the sensor chip carrying Npt–Azq on its surface. The amount of Npt–Azq immobilized on the surface was modulated by the reaction period of the immobilization, and was monitored as an increase of SPR signal after the deactivation of the unreacted NHS-esters and a conditioning of the surface. Three sensor surfaces carrying Npt–Azq for 527, 722 and 951 RU were obtained by the immobilization for 1, 3 and 7 min with a 2 mM solution of Npt–Azq in a borate buffer (pH 9.2).
Measurements of thermal denaturation profiles of mismatch-containing duplexes
Thermal denaturation profiles of the duplexes 5'-d(CTAA vGw AATG)-3'/3'-d(GATT xyz TTAC)-5' where G-y mismatches were flanked by v-x and w-z base pairs (4.5 or 5.0 μM for each strand) were measured in a sodium cacodylate buffer (10 mM, pH 7.0) containing NaCl (100 mM) using a Shimadzu UV-2550 UV-Vis spectrometer linked to a Peltier temperature controller. The absorbance of the sample was monitored at 260 nm from 4 to 70°C with a heating rate of 1°C/min in the absence and presence of a hybrid ligand. The measurements were carried out at the ligand concentration of 50 or 200 μM. The melting temperature of these duplexes was determined as the temperature crossing the melting curve and the median of two straight lines drawn for the single and duplex region in the melting curve.
General procedure for SPR binding experiments using synthetic oligonucleotides
All measurements were carried out at 25°C in a continuous flow of a buffer (10 mM HEPES, pH 7.4) containing NaCl (1 M) at a flow rate of 30 μl/min. A 1 μM solution of 27mer duplexes 5'-d(GTT ACA GAA TCT VGW AAG CCT AAT ACG)-3'/3'-d(CAA TGT CTT AGA XYZ TTC GGA TTA TGC)-5' containing G-Y mismatches flanked by V-X and W-Z base pairs in the buffer were injected for 180 s for analyzing the association to the sensor surface. The buffer was subsequently injected for another 180 s for analyzing the dissociation of the bound oligomer form the surface. After each analysis, all binding materials were removed by washing with 30 μl of NaOH solution (50 mM). Immediately after washing, this system can be used for the next assay.
RESULTS AND DISCUSSION
UV melting temperature analyses
The bindings of Npt–Azq and other hybrid ligands to G-y mismatches were examined by measuring the melting temperature (Tm) of mismatch-containing duplexes (5.0 μM) in the presence of the ligand. The difference of the melting temperature (Tm) in the absence and presence of the ligands is summarized in Table 1. A large Tm of 26.2°C was obtained for the 11mer duplex containing the G-G mismatch flanked by two G-C base pairs (vGw/xyz = cGg/gGc) in the presence of Npt–Npt. In order to distinguish the 11mer duplexes used for the Tm measurements from the 27mer duplexes used for SPR studies, the flanking sequences to the mismatch of 11mer were described with lowercase letters, whereas uppercase letters were used of the flanking sequences of 27mer. Under these conditions, Npt–Npt also increased the Tm of the G-A mismatch (cGg/gAc) by 12.8°C, whereas only a modest Tm of 4.3°C was observed for the matched duplex. In contrast to Npt–Npt, substitution of one Npt chromophore to Azq in Npt–Azq showed a striking difference in the spectrum for mismatch stabilization. Npt–Azq strongly stabilizes cGg/gAc as indicated by the Tm of 15.4°C, exceeding the Tm of Npt–Npt by 2.6°C. Since the non-specific binding to a fully matched duplex is weaker for Npt–Azq (2.2°C) than Npt–Npt (4.3°C), Npt–Azq is currently the strongest ligand described that stabilizes the G-A mismatch.
Table 1. Tm values for the duplexes containing G-y mismatches in the absence and presence of drugsa
To gain insight into a role of Azq chromophore for the binding to the G-A mismatch, the Azq and Npt chromophores in Npt–Azq were replaced by either quinoline (Q) or aminoquinoline (AQ) chromophores. A dramatic decrease of the Tm from 15.4 to 1.6°C was observed by replacing Azq chromophore in Npt–Azq with Q in Npt–Q. The hydrogen-bonding donor of N-H in Azq was replaced by a non-hydrogen bonding group of C-H. Aminoquinoline–azaquinolone (AQ–Azq) obtained by a substitution of Npt in Npt–Azq with AQ completely lost the binding to both G-A and G-G mismatches (Fig. 6). Furthermore, the azaquinolone dimer (Azq–Azq), the single chromophore Npt and Azq, or the 1:1 mixture of Npt and Azq did not stabilize the G-A mismatch. These results clearly indicated that a covalent attachment of Npt and Azq chromophores is essential for the stabilization of the G-A mismatch. These remarkable effects of the substitution of the Npt and Azq chromophores in Npt–Azq on the stabilization of the G-A mismatch suggests the significance of the hydrogen bonding of Azq to adenine and Npt to guanine (Fig. 2). It has been recently shown that Azq is superior to thymine in the recognition of adenine in duplex and triplex structure (21). These observations are well consistent to our results.
Figure 6. Schematic representation of the effect of structural modification of chromophores in Npt–Azq on the increased melting temperature (Tm) of the G-A mismatch.
Evaluation of the binding of Npt–Azq to G-A mismatches
Having found that Npt–Azq strongly stabilizes the G-A mismatch, and that the hydrogen bonding groups in the two chromophores are essential, the binding of the ligand was studied in detail. First, the UV absorption of the Npt–Azq (20 μM) was measured in the presence of different concentrations of the 11mer G-A mismatch duplex (cGg/gAc) (0–50 μM). In the absence of DNA, the ligand showed an absorption maximum at 320 nm and a shoulder at 333 nm. At increasing cGg/gAc concentrations, these absorptions decreased in intensity with a concomitant red shift of the peak from 320 to 324 nm (Fig. 7). An isosbestic point was observed at 340 nm for the spectral change, indicating that UV absorbance of Npt–Azq linearly changes from the free state to the bound state to cGg/gAc. With the data points at 320 nm, the fraction of the total ligand bound against the molar fraction of cGg/gAc ( / ) was plotted (Fig. 8). The bound fraction of Npt–Azq rapidly increased with increasing cGg/gAc concentration and saturated in the presence of approximately one molar equivalent of DNA. In contrast, the bound fraction of Npt–Npt to the G-A mismatch steadily increased with increasing cGg/gAc concentration and reached saturation with two molar equivalents of DNA. This suggests that the binding of Npt–Azq to the G-A mismatch is stronger than that of Npt–Npt. CD spectra of cGg/gAc in the presence of Npt–Azq showed an increase of the ellipticity at 275 nm in addition to strong and weak CD signals induced at 320 and 345 nm, respectively (Fig. 9). Distinct induced CD signals indicated that Npt–Azq was under the influence of the chiral environment of duplex DNA in the complex.
Figure 7. UV absorption spectra of Npt–Azq (20 μM) in the presence of different concentrations of the G-A mismatch duplex cGg/gAc (0–50 μM). The experiments were conducted in a sodium cacodylate buffer (10 mM, pH 7.0) containing NaCl (100 mM) at 15°C. The concentrations of DNA are 0, 1, 2.5, 5, 7.5, 10, 15, 20, 25 and 37.5 μM.
Figure 8. Plot of the fraction of total ligand bound against the molar fraction of the G-A mismatch duplex cGg/gAc to the concentration of Npt–Azq (filled circles) and Npt–Npt (open circles), respectively. The data were obtained from experiments shown in Figure 7.
Figure 9. CD spectra of cGg/gAc (4.5 μM) measured in 10 mM sodium cacodylate buffer (pH 7.0) and 100 mM NaCl at 25°C in the absence (gray) and presence (black) of 20 μM Npt–Azq.
The effect of the concentration of Npt–Azq on the Tm of cGg/gAc was examined by measuring the melting curve at different drug concentrations. Increasing the concentration of Npt–Azq, the Tm of the duplex increased and reached a plateau (Fig. 10). At 50 μM Npt–Azq, UV-melting profiles were obtained for all G-A mismatches with regard to the sequence flanking to the mismatch (vGw/xAz). The Tm of all G-A mismatches except for tGt/aAa in the absence and presence of Npt–Azq are summarized in Figure 11. The Tm of tGt/aAa was too low to measure under the conditions. The largest Tm was recorded for the sequence of tGg/aAc. The sequence cGg/gAc we have examined in detail had the second largest Tm value. While we have anticipated that the binding of Npt–Azq would be favorable for the G-A mismatches flanking to G-C base pairs due to the increased stacking stabilization of the complex, there seems no obvious rationalization for the sequences stabilized by the drug. The melting curves of the sequences showing the top five Tm values are shown in Figure 12. The energy gains of these five G-A mismatches in the presence of 50 μM Npt–Azq were estimated by curve fitting of the melting profiles. The melting profile of the G-A mismatch in the absence and presence of Npt–Azq were fitted to a two-state model with a non-linear least-squares program by using Sigma Plots (version 2001) (22). The energy gains obtained by these simulations are summarized in Table 2. The largest energy gain of –6.0 kcal/mol (277.15 K) was obtained for cGg/gAc, whereas the sequence tGg/aAc, which recorded the highest Tm value, was ranked third. As we reported earlier, the magnitude of Tm is not consistent with the energy gain by the drug binding especially when the duplex showed a low melting temperature (23).
Figure 10. Concentration dependency of Tm of cGg/gAc (4.5 μM). The melting temperature of cGg/gAc (4.5 μM) was measured in the presence of 0, 5, 10, 15, 25, 50, 100 and 200 μM Npt–Azq in 10 mM sodium cacodylate (pH 7.0) and 100 mM NaCl.
Figure 11. Tm values for G-A mismatches (4.5 μM) of different flanking sequence in the presence of Npt–Azq (50 μM) (black) and melting temperatures of G-A mismatches in the absence of Npt–Azq (white).
Figure 12. Melting curves of G-A mismatch duplexes (4.5 μM) in the absence (dots) and presence (filled circles) of Npt–Azq (50 μM) in 10 mM sodium cacodylate (pH 7.0) and 100 mM NaCl. tGg/aAc, green; cGg/gAc, blue; gGa/cAt, red; tGc/aAg, orange; tGa/aAt, black.
Table 2. Estimated energy gains of the G-A mismatches in the presence of 50 μM Npt–Azqa
SPR analyses
Having discovered that Npt–Azq strongly stabilizes the G-A mismatch DNA, the binding of G-containing mismatches to the SPR sensor where Npt–Azq was immobilized on the surfaces was investigated. SPR analyses of a 1 μM solution of 27mer duplexes 5'-d(GTT ACA GAA TCT VGW AAG CCT AAT ACG)-3'/3'-d(CAA TGT CTT AGA XYZ TTC GGA TTA TGC)-5' containing a G-Y mismatch flanked by G-C base pairs (5'-VGW-3'/3'-XYZ-5' = CGG/GYC, Y = A, G or C) were performed with the sensor carrying Npt–Azq for 951 RU on the surface. To suppress a non-specific absorption of DNA to the sensor surface, binding experiments were carried out under high salt conditions (1 M NaCl and 10 mM HEPES buffer, pH 7.4). A sensorgram obtained for the G-A mismatch (CGG/GAC) showed strong SPR signals, of which intensity reached to 193 RU, after 180 s of the association time (Fig. 13). The SPR signal for the G-G mismatch (CGG/GGC) was also distinct, but much lower in intensity (114 RU) than that obtained for CGG/GAC. In marked contrast, only a weak SPR signal (31 RU) was observed for the G-C match DNA (CGG/GCC). These results are consistent with those of UV-melting studies in the presence of Npt–Azq showing a higher Tm for cGg/gAc than for cGg/gGc and almost negligible stabilization for the G-C match DNA. The distinct differences in the intensity of the SPR signal between the G-A mismatch and G-C match DNA clearly indicate a unique character of the Npt–Azq immobilized sensor surface.
Figure 13. Sensorgrams for the binding of duplexes containing G-Y mismatches to the Npt–Azq immobilized SPR sensor surface. Aliquots of 90 μl of the duplexes (1.0 μM in 10 mM HEPES buffer pH 7.4, 1 M NaCl) containing G-A, G-G mismatch and G-C matches were injected for 180 s to measure the association to the sensor surface.
Concentration dependency for the SPR responses
To further evaluate the fidelity of the novel sensor surfaces detecting the G-A mismatches, the concentration dependency of the SPR signal for the binding of the G-A mismatch to the sensor surface was investigated. The duplex CGG/GAC containing a G-A mismatch of different concentrations (0.13–1.0 μM) was analyzed by three sensor surfaces carrying Npt–Azq for 951, 722 and 527 RU. The SPR responses after the association time of 180 s were plotted against the concentration of the G-A mismatch DNA (Fig. 14). The SPR signals produced by each sensor surface clearly showed a linear correlation to the concentration of CGG/GAC, with a correlation coefficient of 0.99. A linear correlation between SPR intensity and DNA concentrations validated that the observed SPR signals were definitely due to a specific interaction between CGG/GAC and Npt–Azq on the sensor surface. Furthermore, it was suggested that the sensor surface is not only effective to detect the G-A mismatch, but also applicable to quantify G-A mismatch DNA at the concentration range up to 1 μM (24). It is worth noting that the SPR responses of three sensor surfaces were not proportional to the amount of Npt–Azq immobilized on the surface. This is most likely due to the different surfaces of the three sensors produced by independent immobilization processes for the three CM5 chips of research grade. Thus, calibration of the sensor surface is necessary for quantitative applications.
Figure 14. Concentration dependency for the SPR responses for 27mer duplex (CGG/GAC) containing a G-A mismatch. CGG/GAC at concentrations of 0.13, 0.25, 0.5 and 1.0 μM were analyzed by three sensors carrying Npt–Azq for 951 (filled circles), 722 (open circles) and 527 RU (squares) on the surface. Responses after 180 s of the association time were plotted against the DNA concentration.
Effect of the flanking sequence to the G-A mismatch on the SPR responses
To know the scope and limitation of the novel G-A mismatch detecting sensor, the effect of the sequence flanking to the mismatch on the binding to the Npt–Azq immobilized sensor surface was examined. In heteroduplex analyses, it is important to differentiate the mismatch-containing duplex from the fully matched duplex. Thus, the SPR signal of a mismatched duplex relative to that of the fully matched duplex was investigated. DNA duplexes containing G-A mismatches in the sequence of 5'-VGW-3'/3'-XAZ-5' were analyzed by SPR using an Npt–Azq immobilized sensor surface. SPR intensities of 16 G-A mismatches were reported by the relative intensity to the highest signal of the fully matched duplex. G-A mismatches are largely divided into three groups with regard to the affinity to the surface. The first group of G-A mismatches showed a strong response to the Npt–Azq immobilized surfaces. Among 16 duplexes, the strongest SPR signal was observed for TGG/AAC (Fig. 15a). The SPR intensity was more than two times stronger than the signal for CGG/GAC, and 12-fold stronger than that observed for the matched duplex. Besides these two, G-A mismatches in GGA/CAT, TGC/AAG, AGG/TAC and AGA/TAT showed SPR signals that are markedly stronger than the signal of a matched duplex by >2-fold. The fitting of the response curve to a 1:1 Langmuir model with BIAevaluation software (version 3) provided an estimate of the association constant (Ka) of each G-A mismatch to the Npt–Azq immobilized surface. The Ka obtained for the G-A mismatches were 1.8 x 106 M–1 for TGG/AAC, 1.0 x 106 M–1 for CGG/GAC, 9.0 x 105 M–1 for GGA/CAT, 9.2 x 105 M–1 for AGG/TAC 7.5 x 105 M–1 for AGA/TAT. The Ka for the TGC/AAG could not be estimated due to a very slow dissociation of DNA from the surface. The second group of G-A mismatches showed reduced relative intensities compared with those in the first group (Fig. 15b). The SPR responses of AGT/TAA and GGG/CAC are clearly distinguished from that of matched duplex. The relative intensity of the SPR signal to the mismatches is 1.5-fold. The third group of mismatches did not show any noticeable differences in SPR intensity from that of fully matched duplex. These analyses showed that the sensor we reported here would be effective to detect G-A mismatches in eight flanking sequences, covering 50% of all G-A mismatches.
Figure 15. Relative SPR intensities of the binding of G-A mismatches to the Npt–Azq immobilized sensor surfaces. G-A mismatches (1 μM) were analyzed in HEPES buffer for 180 s of association to the sensor surface and then dissociation of the bound DNA from the surfaces. SPR intensities were reported as a relative intensity by setting the maximum intensity of the fully matched duplex to be 1.0. (a) G-A mismatches strongly bind to the sensor. TGG/AAC, yellow; CGG/GAC, black; TGC/AAG, green; GGA/CAT, blue; AGG/TAC, pink; AGA/TAT, aqua; matched duplex, red. (b) G-A mismatches with weak binding to the sensor. AGT/TAA, violet; GGG/CAC, aqua; matched duplex, red.
A sequence-dependent binding of Npt–Azq to the G-A mismatches implies that Npt–Azq not only binds to the G-A mismatches, but also interacts to the bases flanking to the mismatch. SPR analyses showed that the binding kinetics for the G-A mismatch in TGC/AAG are markedly different from those of other G-A mismatches. The association of TGC/AAG to the Npt–Azq immobilized surface and the dissociation from the surface was quite slow. These observations are particularly important for the molecular design of the improved version of Npt–Azq that is aiming to detect the rest of the eight G-A mismatches.
Conclusion
The new molecule Npt–Azq was discovered to be the ligand strongly stabilizing the G-A mismatch. The SPR sensor surface on which Npt–Azq was immobilized was the first sensor detecting the G-A mismatch in duplex DNA. The G-A mismatch is produced by a heteroduplex formation from a pair of duplex DNAs containing a G·C to T·A mutation. The G to T transversion is high in frequency because the oxidation of guanine leading to a formation of 8-oxoguanine eventually resulted in the mutation (25). Oxidation of guanine is known to be sensitively affected by the base 3' side to the guanine (26). The 5'-GG-3' is most easily oxidizable and the 5'-GA-3' is second in 5'-GX-3' sequences. The G in the G-A mismatches that are detectable by the SPR surface are mostly flanked by 3' side G and A, suggesting that the oxidation at those Gs is high in frequency. Thus, the Npt–Azq immobilized sensor surface reported here would be a useful and important tool for the discovery of a G to T mutation by detecting the G-A mismatches.
SUPPLEMENTARY MATERIAL
ACKNOWLEDGEMENTS
This work was supported by Grant-in-Aid for Scientific Research on Priority Areas (C) ‘Medical Genome Science’ from the Ministry of Education, Culture, Sports, Science and Technology of Japan.
REFERENCES
Syv?nen,A.-C. (2001) Accessing genetic variation: genotyping single nucleotide polymorphisms. Nature Rev. Genet., 2, 930–942.
Kwok,P.Y. (2001) Methods for genotyping single nucleotide polymorphisms. Annu. Rev. Genom. Hum. Genet., 2, 235–258.
Schafer,A.J. and Hawkins,J.R. (1998) DNA variation and the future of human genetics. Nat. Biotechnol., 16, 33–39.
Nakatani,K., Sando,S. and Saito,I. (2001) Scanning of guanine-guanine mismatches in DNA by synthetic ligands using surface plasmon resonance assay. Nat. Biotechnol., 19, 51–55.
Nataraj,A.J., Olivos-Glander,I., Kusukawa,N. and Highsmith,W.E.,Jr (1999) Single-strand conformation polymorphism and heteroduplex analysis for gel-based mutation detection. Electrophoresis, 20, 1177–1185.
White,M.B., Carvalho,M., Derse,D., O’Brien,S.J. and Dean,M. (1992) Detecting single base substitution as heteroduplex polymorphisms. Genomics, 12, 301–306.
Myers,R.M., Larin,Z. and Maniatis,T., (1985) Detection of single base substitutions by ribonuclease cleavage at mismatches in RNA–DNA duplexes. Science, 230, 1242–1246.
Rowley,G., Saad,S., Giannelli,F. and Green,P.M., (1995) Ultrarapid mutation detection by multiplex, solid-phase chemical cleavage. Genomics, 30, 574–582.
Roberts,E., Deeble,V.J., Woods,C.G. and Taylor,G.R. (1997) Potassium permanganate and tetraethylammonium chloride are a safe and effective substitute for osmium tetroxide in solid-phase fluorescent chemical cleavage of mismatch. Nucleic Acids Res., 25, 3377–3378.
Fazakerley,G.V., Qignard,E., Woisard,A., Guschlbauer,W., van der Marel,G.A., van Boom,J.H., Jones,M. and Radman,M. (1986) Structures of mismatched base pairs in DNA and their recognition by the Escherichia coli mismatch repair system. EMBO J., 5, 3697–3703.
Smith,J. and Modrich,P. (1996) Mutation detection with MutH, MutL and MutS mismatch repair proteins. Proc. Natl Acad. Sci. USA, 93, 4374–4379.
Nakatani,K., Sando,S., Kumasawa,H., Kikuchi,J. and Saito,I. (2001) Recognition of guanine-guanine mismatches by dimeric form of 2-amino-1,8-naphthyridine. J. Am. Chem. Soc., 123, 12650–12657.
Nakatani,K., Sando,S. and Saito,I. (2000) Recognition of a single guanine bulge by 2-acylamino-1,8-naphthyridine. J. Am. Chem. Soc., 122, 2172–2177.
Nakatani,K., Hagihara,S., Sando,S., Sakamoto,S., Yamaguchi,K., Maesawa,C. and Saito,I. (2003) Induction of a remarkable conformational change in a human telomeric sequence by the binding of naphthyridine dimer: inhibition of the elongation of a telomeric repeat by telomerase. J. Am. Chem. Soc., 125, 662–666.
Smith,E.A., Kyo,M., Kumasawa,H., Nakatani,K., Saito,I. and Corn,R.M. (2002) Chemically induced hairpin formation in DNA monolayers. J. Am. Chem. Soc., 124, 6810–6811.
Nakatani,K., Horie,S. and Saito,I. (2003) Affinity labeling of single guanine bulge. J. Am. Chem. Soc., 125, 8972–8973.
Jackson,B.A. and Barton,J.K. (2000) Recognition of base mismatches in DNA by 5,6-chrysenequinone diimine complexes of rhodium(III): a proposed mechanism for preferential binding in destabilized regions of the double helix. Biochemistry, 39, 6176–6182.
Jackson,B.A. and Barton,J.K. (1997) Recognition of DNA base mismatches by a rhodium intercalator. J. Am. Chem. Soc., 119, 12986–12987.
Jackson,B.A., Alekseyev,V.Y. and Barton,J.K. (1999) A versatile mismatch recognition agent: specific cleavage of a plasmid DNA at a single base mispair. Biochemistry, 38, 4655–4662.
Lacy,E.R., Cox,K.K., Wilson,W.D. and Lee.M. (2002) Recognition of T·G mismatched base pairs in DNA by stacked imidazole-containing polyamides: surface plasmon resonance and circular dichroism studies. Nucleic Acids Res., 30, 1834–1841.
Eldrup,A.B., Christensen,C., Haaima,G. and Nielsen,P.E. (2002) Substituted 1,8-naphthyridin-2(1H)-ones are superior to thymine in the recognition of adenine in duplex as well as triplex structures. J. Am. Chem. Soc., 124, 3254–3262.
Marky,L.A. and Breslauer,K.J. (1987) Calculating thermodynamic data for transitions of any molecularity from equilibrium melting curves. Biopolymers, 26, 1601–1620.
Nakatani,K., Horie,S., Murase,T., Hagihara,S. and Saito,I. (2003) Assessment of the sequence dependency for the binding of 2-aminonaphthyridine to the guanine bulge. Bioorg. Med. Chem., 11, 2347–2353.
Zhu,G., Yang,B. and Jennings,R.N. (2000) Quantitation of basic fibroblast growth factor by immunoassay using BIAcore 2000. J. Pharm. Biomed. Anal., 24, 281–290.
Friedberg,E.C., Walker,G.C. and Siede,W. (1995) DNA Repair and Mutagenesis. ASM Press, Washington, DC, pp. 1–58.
Saito,I., Takayama,M., Sugiyama,H., Nakatani,K., Tsuchida,A. and Yamamoto,M. (1995) Photoinduced DNA cleavage via electron transfer: demonstration that guanine residues located 5' to guanine are the most electron-donating sites. J. Am. Chem. Soc., 117, 6406–6407.(Shinya Hagihara1, Hiroyuki Kumasawa1, Yu)
*To whom correspondence should be addressed. Tel: +81 75 383 2756; Fax: +81 75 383 2759; Email: nakatani@sbchem.kyoto-u.ac.jp
ABSTRACT
We have discovered a new molecule naphthyridine–azaquinolone hybrid (Npt–Azq) that strongly stabilized the guanine-adenine (G-A) mismatch in duplex DNA. In the presence of Npt–Azq, the melting temperature (Tm) of 5'-d(CTA ACG GAA TG)-3'/3'-d(GAT TGA CTT AC)-5' containing a single G-A mismatch increased by 15.4°C, whereas fully matched duplex increased its Tm only by 2.2°C. Npt–Azq was immobilized on the sensor surface for the surface plasmon resonance (SPR) assay to examine SPR detection of duplexes containing a G-A mismatch. Distinct SPR signals were observed when 27mer DNA containing a G-A mismatch was analyzed by the Npt–Azq immobilized sensor surfaces, whereas the signal of the fully matched duplex was 6-fold weaker in intensity. The SPR signals for the G-A mismatch were proportional to the concentration of DNA in a range up to 1 μM, confirming that the SPR signal is in fact due to the binding of the G-A mismatch to Npt–Azq immobilized on the surface. Examination of all 16 G-A mismatches regarding the flanking sequence revealed that the sensor surface reported here is applicable to eight flanking sequences, covering 50% of all possible G-A mismatches.
INTRODUCTION
As a follow-on to the complete sequencing of the human genome, typing of single nucleotide polymorphisms (SNPs) in an array of disease-related genes is expected to be an indispensable technique for realizing personalized medicine. A number of methods have been developed for SNP typing (1–3), but much study is still needed to design new typing methods that are simple in operation, rapid and accurate in analysis, and low in cost. One of the challenges we have focused on is the reduction and eventual redundancy of labeled oligonucleotides for analysis. The expense of fluorescently labeling oligonucleotides and PCR products limits their application in large-scale typing by many currently available methods. We have reported a conceptually novel method of SNP typing that detects mismatch-containing duplexes with a small molecular ligand immobilized on a gold surface (4). Hybridization of two sets of duplex DNAs that differ from each other by a single nucleotide produces a DNA heteroduplex containing a single mismatched site. Mismatch-containing duplexes can be separated from homoduplexes by either gel electrophoresis (5,6), chemical and enzymatic cleavages at the mismatched site (7–9), or selective capture with mismatch-binding proteins (10,11). While these heteroduplex analyses applied to low-throughput screening are essentially free from oligonucleotide labeling, new technologies for high-throughput analyses are yet to be established. A small molecular ligand that selectively binds to a mismatched site could replace mismatch-binding proteins and bring an innovation to heteroduplex analyses (5,12–20). The ligand naphthyridine dimer (Npt–Npt) strongly and selectively binds to guanine–guanine mismatches in duplex DNA (5,12) (Fig. 1). The binding constant to a G-G mismatch in the 5'-CGG-3'/3'-GGC-5' sequence is 1.9 x 107 M–1. Npt–Npt consisting of two 2-amino-1,8-naphthyridine (Npt) chromophores, and a linker connecting the chromophores is designed so that each Npt produces three hydrogen bonds to each one of the guanines in the G-G mismatch, and the resultant naphthyridine–guanine pair is stabilized by stacking with the flanking base pairs. The proposed binding of Npt–Npt to the G-G mismatch has been verified by the 2D- NOESY spectrum of the complex (12). In addition to Npt–Npt, ligands that selectively and strongly bind to G-A, G-T and A-A mismatches are needed to accomplish SNP typing by a mismatch binding ligand. Taking into account the structure of Npt–Npt as a clue for the molecular design of ligands binding to a G-A mismatch, we have discovered naphthyridine–azaquinolone hybrids (Npt–Azq) where one Npt chromophore in Npt–Npt is replaced by a 8-azaquinolone chromophore (Azq) having complementary hydrogen-bonding surfaces to adenine (Fig. 2). Herein, we report the remarkable stabilization of G-A mismatch DNA by Npt–Azq and the first synthesis of a surface plasmon resonance (SPR) sensor detecting a G-A mismatch by immobilization of Npt–Azq on a gold surface.
Figure 1. Formulas of Npt- and Azq-based hybrid molecules.
Figure 2. Hydrogen bonding patterns of Npt-G and Azq-A.
MATERIALS AND METHODS
Chemistry
In order to know the structure–activity relationship for the binding of Npt–Azq hybrid to G-A mismatches, hybrid molecules consisted of chromophores with different hydrogen-bonding groups were synthesized. These chromophores include Npt, Azq, quinoline (Q) and 2-aminoquinoline (AQ). Synthesis of the hybrid molecules consisting of Npt with Azq and other heterocycles is straightforward using N-(tert-butoxycarbonyl)imino-3,3'-bis(pentafluorophenyl propionate) (12), in which two carboxyl groups are activated as a pentafluorophenyl ester. Mono-substitution of the pentafluoroester with 2-amino-7-methylnaphthyridine produces intermediate amide Npt-OC6F5 that subsequently reacts with the heterocyclic amines to give hybrid ligands (Fig. 3). A similar procedure was used for the synthesis of Azq-based hybrid ligands (Fig. 4). Details of the synthetic procedure is described in the Supplementary Material.
Figure 3. Synthetic scheme of Npt-based hybrids.
Figure 4. Synthetic scheme of Azq-based hybrids.
Synthesis of Npt–Azq
To a solution of N-(tert-butoxycarbonyl)imino-3,3'-bis(pentafluorophenyl propionate) (1.5 g, 2.53 mmol) in dry DMF (5 ml) was added 2-amino-7-methyl-1,8-naphthyridine (180 mg, 1.14 mmol) and diisopropylethylamine (163 mg, 1.26 mmol). The reaction mixture was stirred at room temperature for 15 h. The solvent was evaporated to dryness and the residue was purified by column chromatography on silica gel to give Npt-OC6F5 (1.29 g, 90%) as pale yellow solids: 1H-NMR (CDCl3, 400 MHz) = 9.01 (br, 1H), 8.44 (d, 1H, J = 8.8 Hz), 8.12 (d, 1H, J = 8.8 Hz), 7.99 (d, 1H, J = 8.0 Hz), 7.26 (d, 1H, J = 8.0 Hz), 3.66 (m, 4H), 2.90 (m, 2H), 2.74 (m, 2H), 2.73 (s, 3H), 1.42 (s, 9H), 13C-NMR (CDCl3, 100 MHz) = 163.6, 155.3, 154.6, 153.4, 142.6, 140.1, 139.5, 138.5, 136.6, 121.9, 118.8, 114.5, 80.9, 44.9, 44.4, 37.6, 37.2, 36.7, 33.4, 32.7, 28.6, 25.8, FABMS (NBA), m/e 569 , HRMS calc. for C26H26O5N4F5 569.1821, found 569.1827.
To a solution of Npt-OC6F5 (300 mg, 0.53 mmol) in dry DMF (2 ml) was added 7-(aminomethyl)hydro-8-azaquinolin-2-one (93 mg, 0.53 mmol) and diisopropylethylamine (77 mg, 0.6 mmol). The reaction mixture was stirred at room temperature for 15 h. The solvent was evaporated to dryness and the residue was purified by column chromatography on silica gel to give N-Boc-Npt–Azq (227 mg, 77%) as pale yellow solids: 1H-NMR (CDCl3, 400 MHz) = 11.29 (br, 1H), 9.08 (br, 1H), 8.38 (d, 1H, J = 8.8 Hz), 8.06 (d, 1H, J = 8.0 Hz), 7.95 (d, 1H, J = 8.0 Hz), 7.81 (d, 1H, J = 8.0 Hz), 7.64 (d, 1H, J = 9.6 Hz), 7.27 (d, 1H, J = 8.0 Hz), 7.22 (d, 1H, J = 8.0 Hz), 6.64 (d, 1H, J = 9.6 Hz), 4.63 (d, 2H, J = 6.0 Hz), 3.58 (t, 2H, J = 6.8 Hz), 3.57 (t, 2H, J = 6.8 Hz), 2.71 (t, 2H, J = 6.8 Hz), 2.70 (s, 1H), 2.54 (t, 2H, J = 6.8 Hz), 1.36 (s, 9H), 13C-NMR (CDCl3, 100 MHz) = 171.7, 163.9, 163.4, 159.8, 155.7, 154.5, 153.6, 149.3, 139.3, 139.2, 137.4, 136.9, 123.1, 121.9, 118.8, 118.2, 114.6, 114.0, 80.5, 44.8, 41.8, 37.5, 28.6, 25.7, FABMS (NBA), m/e 560 , HRMS calc. for C29H34O5N7 560.2621, found 560.2618.
To a solution of N-Boc-Npt–Azq (62 mg, 0.11 mmol) in CHCl3 (3 ml) was added ethyl acetate containing 4 M HCl (2 ml) at room temperature and the reaction mixture was stirred at room temperature for 0.5 h. The solvent was evaporated to dryness to give hydrochloride of Npt–Azq (quantitative yield) as white solids: 1H-NMR (CDCl3, 400 MHz) = 11.45 (br, 1H), 8.67 (t, 1H, J = 5.6 Hz), 8.31 (d, 1H, J = 8.8 Hz), 7.98 (d, 1H, J = 8.8 Hz), 7.88 (d, 1H, J = 8.0 Hz), 7.65 (d, 1H, J = 7.6 Hz), 7.53 (d, 1H, J = 9.6 Hz), 7.15 (d, 1H, J = 7.6 Hz), 7.14 (t, 1H, J = 8.0 Hz), 6.58 (d, 1H, J = 9.6 Hz), 4.65 (d, 2H, J = 5.6 Hz), 3.08 (t, 2H, J = 6.0 Hz), 3.06 (t, 2H, J = 6.0 Hz), 2.65 (t, 2H, J = 6.0 Hz), 2.58 (t, 2H, J = 6.0 Hz), 2.57 (s, 3H), 13C NMR (CD3OD, 100 MHz) = 171.8, 171,3, 165.0, 160.2, 159.7, 157.0, 148.5, 147.3, 146.4, 140.3, 140.2, 138.8, 122.5, 120.9, 120.2, 117.6, 117.0, 114.8, 44.1, 43.9, 43.1, 32.5, 30.5, 19.7, FABMS (NBA), m/e 460 , HRMS calc. for C24H26O3N7 460.2095, found 460.2097.
Synthesis of Npt–Azq having an aminoalkyl linker for immobilization onto SPR sensor surface
SPR sensors having Npt–Azq on their surface were synthesized to examine SPR detection of the G-A mismatch in a flow system (Fig. 5). Npt–Azq was immobilized on a dextran matrix coated gold surface (CM5 chip, BIAcore) through a bivalent linker of N-Boc-aminoaldehyde. First, Npt–Azq was tethered by a reductive amination to the linker, which was efficiently prepared from Boc-protected 4-aminobutanoic acid and 3-aminopropionaldehyde diethylacetal (Supplementary Material). Deprotection of a Boc group produced a primary amine, which was subsequently immobilized on the sensor surface by a coupling between the amino group and an activated carboxyl group on the CM5 chip using a standard method with 1-(3-dimethylaminopropyl)-3-ethylcarbodiimide hydrochloride (EDCI) and N-hydroxysuccinimide (NHS).
Figure 5. Synthetic scheme of Npt–Azq immobilized sensor surfaces.
Preparation of a sensor chip carrying Npt–Azq on its surface
SPR measurements were performed with a BIAcore 2000 system (BIAcore, Uppsala, Sweden). Immobilization of Npt–Azq on a sensor chip CM5 (carboxymethylated dextran surface, BIAcore) was carried out using amine coupling kit (BIAcore) in a continuous flow of HBS-N buffer (10 mM HEPES, pH 7.4) containing NaCl (150 mM) at a flow rate of 10 μl/min. A solution (70 μl) of NHS (0.05 M) and EDCI (0.2 M) was injected using the QUICKINJECT command to activate the carboxymethylated dextran surface of a CM5 sensor chip. A solution (70 μl) of Npt–Azq having an aminoalkyl linker (2 mM in borate buffer, pH 9.2) was injected using the QUICKINJECT command on the activated surface. Residual activated surface was completely blocked by injection of a solution (20 μl) of ethanolamine hydrochloride (1.0 M, pH 8.5). Non-covalently bound material was removed by washing with 5 μl of 50 mM NaOH to produce the sensor chip carrying Npt–Azq on its surface. The amount of Npt–Azq immobilized on the surface was modulated by the reaction period of the immobilization, and was monitored as an increase of SPR signal after the deactivation of the unreacted NHS-esters and a conditioning of the surface. Three sensor surfaces carrying Npt–Azq for 527, 722 and 951 RU were obtained by the immobilization for 1, 3 and 7 min with a 2 mM solution of Npt–Azq in a borate buffer (pH 9.2).
Measurements of thermal denaturation profiles of mismatch-containing duplexes
Thermal denaturation profiles of the duplexes 5'-d(CTAA vGw AATG)-3'/3'-d(GATT xyz TTAC)-5' where G-y mismatches were flanked by v-x and w-z base pairs (4.5 or 5.0 μM for each strand) were measured in a sodium cacodylate buffer (10 mM, pH 7.0) containing NaCl (100 mM) using a Shimadzu UV-2550 UV-Vis spectrometer linked to a Peltier temperature controller. The absorbance of the sample was monitored at 260 nm from 4 to 70°C with a heating rate of 1°C/min in the absence and presence of a hybrid ligand. The measurements were carried out at the ligand concentration of 50 or 200 μM. The melting temperature of these duplexes was determined as the temperature crossing the melting curve and the median of two straight lines drawn for the single and duplex region in the melting curve.
General procedure for SPR binding experiments using synthetic oligonucleotides
All measurements were carried out at 25°C in a continuous flow of a buffer (10 mM HEPES, pH 7.4) containing NaCl (1 M) at a flow rate of 30 μl/min. A 1 μM solution of 27mer duplexes 5'-d(GTT ACA GAA TCT VGW AAG CCT AAT ACG)-3'/3'-d(CAA TGT CTT AGA XYZ TTC GGA TTA TGC)-5' containing G-Y mismatches flanked by V-X and W-Z base pairs in the buffer were injected for 180 s for analyzing the association to the sensor surface. The buffer was subsequently injected for another 180 s for analyzing the dissociation of the bound oligomer form the surface. After each analysis, all binding materials were removed by washing with 30 μl of NaOH solution (50 mM). Immediately after washing, this system can be used for the next assay.
RESULTS AND DISCUSSION
UV melting temperature analyses
The bindings of Npt–Azq and other hybrid ligands to G-y mismatches were examined by measuring the melting temperature (Tm) of mismatch-containing duplexes (5.0 μM) in the presence of the ligand. The difference of the melting temperature (Tm) in the absence and presence of the ligands is summarized in Table 1. A large Tm of 26.2°C was obtained for the 11mer duplex containing the G-G mismatch flanked by two G-C base pairs (vGw/xyz = cGg/gGc) in the presence of Npt–Npt. In order to distinguish the 11mer duplexes used for the Tm measurements from the 27mer duplexes used for SPR studies, the flanking sequences to the mismatch of 11mer were described with lowercase letters, whereas uppercase letters were used of the flanking sequences of 27mer. Under these conditions, Npt–Npt also increased the Tm of the G-A mismatch (cGg/gAc) by 12.8°C, whereas only a modest Tm of 4.3°C was observed for the matched duplex. In contrast to Npt–Npt, substitution of one Npt chromophore to Azq in Npt–Azq showed a striking difference in the spectrum for mismatch stabilization. Npt–Azq strongly stabilizes cGg/gAc as indicated by the Tm of 15.4°C, exceeding the Tm of Npt–Npt by 2.6°C. Since the non-specific binding to a fully matched duplex is weaker for Npt–Azq (2.2°C) than Npt–Npt (4.3°C), Npt–Azq is currently the strongest ligand described that stabilizes the G-A mismatch.
Table 1. Tm values for the duplexes containing G-y mismatches in the absence and presence of drugsa
To gain insight into a role of Azq chromophore for the binding to the G-A mismatch, the Azq and Npt chromophores in Npt–Azq were replaced by either quinoline (Q) or aminoquinoline (AQ) chromophores. A dramatic decrease of the Tm from 15.4 to 1.6°C was observed by replacing Azq chromophore in Npt–Azq with Q in Npt–Q. The hydrogen-bonding donor of N-H in Azq was replaced by a non-hydrogen bonding group of C-H. Aminoquinoline–azaquinolone (AQ–Azq) obtained by a substitution of Npt in Npt–Azq with AQ completely lost the binding to both G-A and G-G mismatches (Fig. 6). Furthermore, the azaquinolone dimer (Azq–Azq), the single chromophore Npt and Azq, or the 1:1 mixture of Npt and Azq did not stabilize the G-A mismatch. These results clearly indicated that a covalent attachment of Npt and Azq chromophores is essential for the stabilization of the G-A mismatch. These remarkable effects of the substitution of the Npt and Azq chromophores in Npt–Azq on the stabilization of the G-A mismatch suggests the significance of the hydrogen bonding of Azq to adenine and Npt to guanine (Fig. 2). It has been recently shown that Azq is superior to thymine in the recognition of adenine in duplex and triplex structure (21). These observations are well consistent to our results.
Figure 6. Schematic representation of the effect of structural modification of chromophores in Npt–Azq on the increased melting temperature (Tm) of the G-A mismatch.
Evaluation of the binding of Npt–Azq to G-A mismatches
Having found that Npt–Azq strongly stabilizes the G-A mismatch, and that the hydrogen bonding groups in the two chromophores are essential, the binding of the ligand was studied in detail. First, the UV absorption of the Npt–Azq (20 μM) was measured in the presence of different concentrations of the 11mer G-A mismatch duplex (cGg/gAc) (0–50 μM). In the absence of DNA, the ligand showed an absorption maximum at 320 nm and a shoulder at 333 nm. At increasing cGg/gAc concentrations, these absorptions decreased in intensity with a concomitant red shift of the peak from 320 to 324 nm (Fig. 7). An isosbestic point was observed at 340 nm for the spectral change, indicating that UV absorbance of Npt–Azq linearly changes from the free state to the bound state to cGg/gAc. With the data points at 320 nm, the fraction of the total ligand bound against the molar fraction of cGg/gAc ( / ) was plotted (Fig. 8). The bound fraction of Npt–Azq rapidly increased with increasing cGg/gAc concentration and saturated in the presence of approximately one molar equivalent of DNA. In contrast, the bound fraction of Npt–Npt to the G-A mismatch steadily increased with increasing cGg/gAc concentration and reached saturation with two molar equivalents of DNA. This suggests that the binding of Npt–Azq to the G-A mismatch is stronger than that of Npt–Npt. CD spectra of cGg/gAc in the presence of Npt–Azq showed an increase of the ellipticity at 275 nm in addition to strong and weak CD signals induced at 320 and 345 nm, respectively (Fig. 9). Distinct induced CD signals indicated that Npt–Azq was under the influence of the chiral environment of duplex DNA in the complex.
Figure 7. UV absorption spectra of Npt–Azq (20 μM) in the presence of different concentrations of the G-A mismatch duplex cGg/gAc (0–50 μM). The experiments were conducted in a sodium cacodylate buffer (10 mM, pH 7.0) containing NaCl (100 mM) at 15°C. The concentrations of DNA are 0, 1, 2.5, 5, 7.5, 10, 15, 20, 25 and 37.5 μM.
Figure 8. Plot of the fraction of total ligand bound against the molar fraction of the G-A mismatch duplex cGg/gAc to the concentration of Npt–Azq (filled circles) and Npt–Npt (open circles), respectively. The data were obtained from experiments shown in Figure 7.
Figure 9. CD spectra of cGg/gAc (4.5 μM) measured in 10 mM sodium cacodylate buffer (pH 7.0) and 100 mM NaCl at 25°C in the absence (gray) and presence (black) of 20 μM Npt–Azq.
The effect of the concentration of Npt–Azq on the Tm of cGg/gAc was examined by measuring the melting curve at different drug concentrations. Increasing the concentration of Npt–Azq, the Tm of the duplex increased and reached a plateau (Fig. 10). At 50 μM Npt–Azq, UV-melting profiles were obtained for all G-A mismatches with regard to the sequence flanking to the mismatch (vGw/xAz). The Tm of all G-A mismatches except for tGt/aAa in the absence and presence of Npt–Azq are summarized in Figure 11. The Tm of tGt/aAa was too low to measure under the conditions. The largest Tm was recorded for the sequence of tGg/aAc. The sequence cGg/gAc we have examined in detail had the second largest Tm value. While we have anticipated that the binding of Npt–Azq would be favorable for the G-A mismatches flanking to G-C base pairs due to the increased stacking stabilization of the complex, there seems no obvious rationalization for the sequences stabilized by the drug. The melting curves of the sequences showing the top five Tm values are shown in Figure 12. The energy gains of these five G-A mismatches in the presence of 50 μM Npt–Azq were estimated by curve fitting of the melting profiles. The melting profile of the G-A mismatch in the absence and presence of Npt–Azq were fitted to a two-state model with a non-linear least-squares program by using Sigma Plots (version 2001) (22). The energy gains obtained by these simulations are summarized in Table 2. The largest energy gain of –6.0 kcal/mol (277.15 K) was obtained for cGg/gAc, whereas the sequence tGg/aAc, which recorded the highest Tm value, was ranked third. As we reported earlier, the magnitude of Tm is not consistent with the energy gain by the drug binding especially when the duplex showed a low melting temperature (23).
Figure 10. Concentration dependency of Tm of cGg/gAc (4.5 μM). The melting temperature of cGg/gAc (4.5 μM) was measured in the presence of 0, 5, 10, 15, 25, 50, 100 and 200 μM Npt–Azq in 10 mM sodium cacodylate (pH 7.0) and 100 mM NaCl.
Figure 11. Tm values for G-A mismatches (4.5 μM) of different flanking sequence in the presence of Npt–Azq (50 μM) (black) and melting temperatures of G-A mismatches in the absence of Npt–Azq (white).
Figure 12. Melting curves of G-A mismatch duplexes (4.5 μM) in the absence (dots) and presence (filled circles) of Npt–Azq (50 μM) in 10 mM sodium cacodylate (pH 7.0) and 100 mM NaCl. tGg/aAc, green; cGg/gAc, blue; gGa/cAt, red; tGc/aAg, orange; tGa/aAt, black.
Table 2. Estimated energy gains of the G-A mismatches in the presence of 50 μM Npt–Azqa
SPR analyses
Having discovered that Npt–Azq strongly stabilizes the G-A mismatch DNA, the binding of G-containing mismatches to the SPR sensor where Npt–Azq was immobilized on the surfaces was investigated. SPR analyses of a 1 μM solution of 27mer duplexes 5'-d(GTT ACA GAA TCT VGW AAG CCT AAT ACG)-3'/3'-d(CAA TGT CTT AGA XYZ TTC GGA TTA TGC)-5' containing a G-Y mismatch flanked by G-C base pairs (5'-VGW-3'/3'-XYZ-5' = CGG/GYC, Y = A, G or C) were performed with the sensor carrying Npt–Azq for 951 RU on the surface. To suppress a non-specific absorption of DNA to the sensor surface, binding experiments were carried out under high salt conditions (1 M NaCl and 10 mM HEPES buffer, pH 7.4). A sensorgram obtained for the G-A mismatch (CGG/GAC) showed strong SPR signals, of which intensity reached to 193 RU, after 180 s of the association time (Fig. 13). The SPR signal for the G-G mismatch (CGG/GGC) was also distinct, but much lower in intensity (114 RU) than that obtained for CGG/GAC. In marked contrast, only a weak SPR signal (31 RU) was observed for the G-C match DNA (CGG/GCC). These results are consistent with those of UV-melting studies in the presence of Npt–Azq showing a higher Tm for cGg/gAc than for cGg/gGc and almost negligible stabilization for the G-C match DNA. The distinct differences in the intensity of the SPR signal between the G-A mismatch and G-C match DNA clearly indicate a unique character of the Npt–Azq immobilized sensor surface.
Figure 13. Sensorgrams for the binding of duplexes containing G-Y mismatches to the Npt–Azq immobilized SPR sensor surface. Aliquots of 90 μl of the duplexes (1.0 μM in 10 mM HEPES buffer pH 7.4, 1 M NaCl) containing G-A, G-G mismatch and G-C matches were injected for 180 s to measure the association to the sensor surface.
Concentration dependency for the SPR responses
To further evaluate the fidelity of the novel sensor surfaces detecting the G-A mismatches, the concentration dependency of the SPR signal for the binding of the G-A mismatch to the sensor surface was investigated. The duplex CGG/GAC containing a G-A mismatch of different concentrations (0.13–1.0 μM) was analyzed by three sensor surfaces carrying Npt–Azq for 951, 722 and 527 RU. The SPR responses after the association time of 180 s were plotted against the concentration of the G-A mismatch DNA (Fig. 14). The SPR signals produced by each sensor surface clearly showed a linear correlation to the concentration of CGG/GAC, with a correlation coefficient of 0.99. A linear correlation between SPR intensity and DNA concentrations validated that the observed SPR signals were definitely due to a specific interaction between CGG/GAC and Npt–Azq on the sensor surface. Furthermore, it was suggested that the sensor surface is not only effective to detect the G-A mismatch, but also applicable to quantify G-A mismatch DNA at the concentration range up to 1 μM (24). It is worth noting that the SPR responses of three sensor surfaces were not proportional to the amount of Npt–Azq immobilized on the surface. This is most likely due to the different surfaces of the three sensors produced by independent immobilization processes for the three CM5 chips of research grade. Thus, calibration of the sensor surface is necessary for quantitative applications.
Figure 14. Concentration dependency for the SPR responses for 27mer duplex (CGG/GAC) containing a G-A mismatch. CGG/GAC at concentrations of 0.13, 0.25, 0.5 and 1.0 μM were analyzed by three sensors carrying Npt–Azq for 951 (filled circles), 722 (open circles) and 527 RU (squares) on the surface. Responses after 180 s of the association time were plotted against the DNA concentration.
Effect of the flanking sequence to the G-A mismatch on the SPR responses
To know the scope and limitation of the novel G-A mismatch detecting sensor, the effect of the sequence flanking to the mismatch on the binding to the Npt–Azq immobilized sensor surface was examined. In heteroduplex analyses, it is important to differentiate the mismatch-containing duplex from the fully matched duplex. Thus, the SPR signal of a mismatched duplex relative to that of the fully matched duplex was investigated. DNA duplexes containing G-A mismatches in the sequence of 5'-VGW-3'/3'-XAZ-5' were analyzed by SPR using an Npt–Azq immobilized sensor surface. SPR intensities of 16 G-A mismatches were reported by the relative intensity to the highest signal of the fully matched duplex. G-A mismatches are largely divided into three groups with regard to the affinity to the surface. The first group of G-A mismatches showed a strong response to the Npt–Azq immobilized surfaces. Among 16 duplexes, the strongest SPR signal was observed for TGG/AAC (Fig. 15a). The SPR intensity was more than two times stronger than the signal for CGG/GAC, and 12-fold stronger than that observed for the matched duplex. Besides these two, G-A mismatches in GGA/CAT, TGC/AAG, AGG/TAC and AGA/TAT showed SPR signals that are markedly stronger than the signal of a matched duplex by >2-fold. The fitting of the response curve to a 1:1 Langmuir model with BIAevaluation software (version 3) provided an estimate of the association constant (Ka) of each G-A mismatch to the Npt–Azq immobilized surface. The Ka obtained for the G-A mismatches were 1.8 x 106 M–1 for TGG/AAC, 1.0 x 106 M–1 for CGG/GAC, 9.0 x 105 M–1 for GGA/CAT, 9.2 x 105 M–1 for AGG/TAC 7.5 x 105 M–1 for AGA/TAT. The Ka for the TGC/AAG could not be estimated due to a very slow dissociation of DNA from the surface. The second group of G-A mismatches showed reduced relative intensities compared with those in the first group (Fig. 15b). The SPR responses of AGT/TAA and GGG/CAC are clearly distinguished from that of matched duplex. The relative intensity of the SPR signal to the mismatches is 1.5-fold. The third group of mismatches did not show any noticeable differences in SPR intensity from that of fully matched duplex. These analyses showed that the sensor we reported here would be effective to detect G-A mismatches in eight flanking sequences, covering 50% of all G-A mismatches.
Figure 15. Relative SPR intensities of the binding of G-A mismatches to the Npt–Azq immobilized sensor surfaces. G-A mismatches (1 μM) were analyzed in HEPES buffer for 180 s of association to the sensor surface and then dissociation of the bound DNA from the surfaces. SPR intensities were reported as a relative intensity by setting the maximum intensity of the fully matched duplex to be 1.0. (a) G-A mismatches strongly bind to the sensor. TGG/AAC, yellow; CGG/GAC, black; TGC/AAG, green; GGA/CAT, blue; AGG/TAC, pink; AGA/TAT, aqua; matched duplex, red. (b) G-A mismatches with weak binding to the sensor. AGT/TAA, violet; GGG/CAC, aqua; matched duplex, red.
A sequence-dependent binding of Npt–Azq to the G-A mismatches implies that Npt–Azq not only binds to the G-A mismatches, but also interacts to the bases flanking to the mismatch. SPR analyses showed that the binding kinetics for the G-A mismatch in TGC/AAG are markedly different from those of other G-A mismatches. The association of TGC/AAG to the Npt–Azq immobilized surface and the dissociation from the surface was quite slow. These observations are particularly important for the molecular design of the improved version of Npt–Azq that is aiming to detect the rest of the eight G-A mismatches.
Conclusion
The new molecule Npt–Azq was discovered to be the ligand strongly stabilizing the G-A mismatch. The SPR sensor surface on which Npt–Azq was immobilized was the first sensor detecting the G-A mismatch in duplex DNA. The G-A mismatch is produced by a heteroduplex formation from a pair of duplex DNAs containing a G·C to T·A mutation. The G to T transversion is high in frequency because the oxidation of guanine leading to a formation of 8-oxoguanine eventually resulted in the mutation (25). Oxidation of guanine is known to be sensitively affected by the base 3' side to the guanine (26). The 5'-GG-3' is most easily oxidizable and the 5'-GA-3' is second in 5'-GX-3' sequences. The G in the G-A mismatches that are detectable by the SPR surface are mostly flanked by 3' side G and A, suggesting that the oxidation at those Gs is high in frequency. Thus, the Npt–Azq immobilized sensor surface reported here would be a useful and important tool for the discovery of a G to T mutation by detecting the G-A mismatches.
SUPPLEMENTARY MATERIAL
ACKNOWLEDGEMENTS
This work was supported by Grant-in-Aid for Scientific Research on Priority Areas (C) ‘Medical Genome Science’ from the Ministry of Education, Culture, Sports, Science and Technology of Japan.
REFERENCES
Syv?nen,A.-C. (2001) Accessing genetic variation: genotyping single nucleotide polymorphisms. Nature Rev. Genet., 2, 930–942.
Kwok,P.Y. (2001) Methods for genotyping single nucleotide polymorphisms. Annu. Rev. Genom. Hum. Genet., 2, 235–258.
Schafer,A.J. and Hawkins,J.R. (1998) DNA variation and the future of human genetics. Nat. Biotechnol., 16, 33–39.
Nakatani,K., Sando,S. and Saito,I. (2001) Scanning of guanine-guanine mismatches in DNA by synthetic ligands using surface plasmon resonance assay. Nat. Biotechnol., 19, 51–55.
Nataraj,A.J., Olivos-Glander,I., Kusukawa,N. and Highsmith,W.E.,Jr (1999) Single-strand conformation polymorphism and heteroduplex analysis for gel-based mutation detection. Electrophoresis, 20, 1177–1185.
White,M.B., Carvalho,M., Derse,D., O’Brien,S.J. and Dean,M. (1992) Detecting single base substitution as heteroduplex polymorphisms. Genomics, 12, 301–306.
Myers,R.M., Larin,Z. and Maniatis,T., (1985) Detection of single base substitutions by ribonuclease cleavage at mismatches in RNA–DNA duplexes. Science, 230, 1242–1246.
Rowley,G., Saad,S., Giannelli,F. and Green,P.M., (1995) Ultrarapid mutation detection by multiplex, solid-phase chemical cleavage. Genomics, 30, 574–582.
Roberts,E., Deeble,V.J., Woods,C.G. and Taylor,G.R. (1997) Potassium permanganate and tetraethylammonium chloride are a safe and effective substitute for osmium tetroxide in solid-phase fluorescent chemical cleavage of mismatch. Nucleic Acids Res., 25, 3377–3378.
Fazakerley,G.V., Qignard,E., Woisard,A., Guschlbauer,W., van der Marel,G.A., van Boom,J.H., Jones,M. and Radman,M. (1986) Structures of mismatched base pairs in DNA and their recognition by the Escherichia coli mismatch repair system. EMBO J., 5, 3697–3703.
Smith,J. and Modrich,P. (1996) Mutation detection with MutH, MutL and MutS mismatch repair proteins. Proc. Natl Acad. Sci. USA, 93, 4374–4379.
Nakatani,K., Sando,S., Kumasawa,H., Kikuchi,J. and Saito,I. (2001) Recognition of guanine-guanine mismatches by dimeric form of 2-amino-1,8-naphthyridine. J. Am. Chem. Soc., 123, 12650–12657.
Nakatani,K., Sando,S. and Saito,I. (2000) Recognition of a single guanine bulge by 2-acylamino-1,8-naphthyridine. J. Am. Chem. Soc., 122, 2172–2177.
Nakatani,K., Hagihara,S., Sando,S., Sakamoto,S., Yamaguchi,K., Maesawa,C. and Saito,I. (2003) Induction of a remarkable conformational change in a human telomeric sequence by the binding of naphthyridine dimer: inhibition of the elongation of a telomeric repeat by telomerase. J. Am. Chem. Soc., 125, 662–666.
Smith,E.A., Kyo,M., Kumasawa,H., Nakatani,K., Saito,I. and Corn,R.M. (2002) Chemically induced hairpin formation in DNA monolayers. J. Am. Chem. Soc., 124, 6810–6811.
Nakatani,K., Horie,S. and Saito,I. (2003) Affinity labeling of single guanine bulge. J. Am. Chem. Soc., 125, 8972–8973.
Jackson,B.A. and Barton,J.K. (2000) Recognition of base mismatches in DNA by 5,6-chrysenequinone diimine complexes of rhodium(III): a proposed mechanism for preferential binding in destabilized regions of the double helix. Biochemistry, 39, 6176–6182.
Jackson,B.A. and Barton,J.K. (1997) Recognition of DNA base mismatches by a rhodium intercalator. J. Am. Chem. Soc., 119, 12986–12987.
Jackson,B.A., Alekseyev,V.Y. and Barton,J.K. (1999) A versatile mismatch recognition agent: specific cleavage of a plasmid DNA at a single base mispair. Biochemistry, 38, 4655–4662.
Lacy,E.R., Cox,K.K., Wilson,W.D. and Lee.M. (2002) Recognition of T·G mismatched base pairs in DNA by stacked imidazole-containing polyamides: surface plasmon resonance and circular dichroism studies. Nucleic Acids Res., 30, 1834–1841.
Eldrup,A.B., Christensen,C., Haaima,G. and Nielsen,P.E. (2002) Substituted 1,8-naphthyridin-2(1H)-ones are superior to thymine in the recognition of adenine in duplex as well as triplex structures. J. Am. Chem. Soc., 124, 3254–3262.
Marky,L.A. and Breslauer,K.J. (1987) Calculating thermodynamic data for transitions of any molecularity from equilibrium melting curves. Biopolymers, 26, 1601–1620.
Nakatani,K., Horie,S., Murase,T., Hagihara,S. and Saito,I. (2003) Assessment of the sequence dependency for the binding of 2-aminonaphthyridine to the guanine bulge. Bioorg. Med. Chem., 11, 2347–2353.
Zhu,G., Yang,B. and Jennings,R.N. (2000) Quantitation of basic fibroblast growth factor by immunoassay using BIAcore 2000. J. Pharm. Biomed. Anal., 24, 281–290.
Friedberg,E.C., Walker,G.C. and Siede,W. (1995) DNA Repair and Mutagenesis. ASM Press, Washington, DC, pp. 1–58.
Saito,I., Takayama,M., Sugiyama,H., Nakatani,K., Tsuchida,A. and Yamamoto,M. (1995) Photoinduced DNA cleavage via electron transfer: demonstration that guanine residues located 5' to guanine are the most electron-donating sites. J. Am. Chem. Soc., 117, 6406–6407.(Shinya Hagihara1, Hiroyuki Kumasawa1, Yu)