Article pubs.acs.org/ac
Graphite-Based Nanocomposite Electrochemical Sensor for Multiplex Detection of Adenine, Guanine, Thymine, and Cytosine: A Biomedical Prospect for Studying DNA Damage Khan Loon Ng†,‡ and Sook Mei Khor*,†,§ †
Department of Chemistry, Faculty of Science, University of Malaya, 50603 Kuala Lumpur, Malaysia Wipro Skin Research and Innovation Centre, No. 7 Persiaran Subang Permai, Taman Perindustrian Subang, 47610 Selangor, Malaysia § University of Malaya Centre for Ionic Liquids (UMCiL), University of Malaya, 50603 Kuala Lumpur, Malaysia ‡
S Supporting Information *
ABSTRACT: Guanine (G), adenine (A), thymine (T), and cytosine (C) are the four basic constituents of DNA. Studies on DNA composition have focused especially on DNA damage and genotoxicity. However, the development of a rapid, simple, and multiplex method for the simultaneous measurement of the four DNA bases remains a challenge. In this study, we describe a graphite-based nanocomposite electrode (AurGO/MWCNT/graphite) that uses a simple electro-co-deposition approach. We successfully applied the developed sensor for multiplex detection of G, A, T, and C, using square-wave voltammetry. The sensor was tested using real animal and plant DNA samples in which the hydrolysis of T and C could be achieved with 8 mol L−1 of acid. The electrochemical sensor exhibited excellent sensitivity (G = 178.8 nA/μg mL−1, A = 92.9 nA/μg mL−1, T = 1.4 nA/μg mL−1, and C = 15.1 9 nA/μg mL−1), low limit of detection (G, A = 0.5 μg mL−1; T, C = 1.0 μg mL−1), and high selectivity in the presence of common interfering factors from biological matrixes. The reliability of the established method was assessed by method validation and comparison with the ultraperformance liquid chromatography technique, and a correlation of 103.7% was achieved.
T
food consumption are worthy of further investigation. Various analytical methodologies have been developed to measure DNA base ratios, including liquid chromatography,4,8 gas chromatography,9 fluorescence,10 capillary electrophoresis,11 and electrochemical methods.12−14 Chromatographic techniques are accurate, precise, and selective; however, they require complex sample preparation steps such as DNA enrichment or derivatization,2,9,15 which are time-consuming. In contrast, electrochemical methods are rapid, simple, and sensitive, properties that make these methods suitable for clinical research and applications. Several electrochemical methods have been developed for DNA analysis. Some are based on the detection of DNA bases (nucleic acid),16−18 and others focus on hybridization and detection of single-stranded DNA.19−21 The former type of electrochemical method is more comprehensive, since it assesses the full DNA genome instead of specific genes (the latter type). During electrochemical analyses of DNA, pyrimidine bases (T and C) are especially difficult to oxidize as they require higher anodic applied potential.22,23 Besides, the electron transfer kinetics of DNA bases are slow,15 thus inhibiting
he human chromosome is a biopolymer that comprises a complex combination of deoxyribonucleic acid (DNA) and has a double-helical structure. It encodes genetic information via unique combinations of the four DNA bases: guanine (G), adenine (A), thymine (T), and cytosine (C). DNA is translated into proteins such as enzymes or antibodies. Therefore, the intact combination of these DNA bases is important for preserving the genetic information and function of the organism. Studies have shown that DNA damage is associated with aging, lifestyle stress,1 exposure to ultraviolet or infrared radiation,2,3 and the consumption of carcinogenic chemicals4 and food preservatives.5 These factors contribute to DNA damage via insertions, deletions, or translocations of the DNA bases.6 Failure in detection may lead to mutations and impair the organism’s health via cell death or apoptosis, deregulation of gene expression, and tumorigenesis.1−3 To analyze these changes in detail, comprehensive studies of the G, A, T, and C components are important. By analyzing individual base composition, the intact base combination in DNA can be measured by following the assumption that the mole percentage of G is equal to that of C and that of A is equal to that of T.7 These ratio values should be equal to unity and a deviation could suggest DNA damage. Such an approach is feasible for clinical studies especially in the field of DNA damage, whereby correlations to factors such as lifestyle and © 2017 American Chemical Society
Received: June 23, 2017 Accepted: August 28, 2017 Published: August 28, 2017 10004
DOI: 10.1021/acs.analchem.7b02432 Anal. Chem. 2017, 89, 10004−10012
Article
Analytical Chemistry
Germany), including calf thymus DNA. Ultrapure water (18.2 MΩ, Merck Millipore) was used for the chemical preparations. The graphite electrode used in this study was fabricated according to the methods described in our previously published work.26,27 The electrochemical measurement was performed using an Autolab potentiostat/galvanostat PGSTAT 202 (Utrecht, The Netherlands). Liquid chromatography was performed using ultra performance liquid chromatography (UPLC). The surface morphology of the modified graphite electrode was characterized using the field-emission scanning electron microscope model SU8220 (Hitachi, Tokyo, Japan), equipped with energy dispersive X-ray (EDX) for element composition analysis. Preparation of Au-rGO/MWCNT/Graphite. MWCNT solution (0.2 mg mL−1, pH 7) was centrifuged at 2000g. The clear portion of the solution was pipetted onto the graphite tip and subsequently evaporated to dryness at 105 °C for 3 h. The modified sensor at this stage was designated as MWCNT/ graphite. Subsequently, the gold and GO (Au-GO) were electrodeposited on the MWCNT/graphite surface using cyclic voltammetry (CV) at an applied potential ranging from 0.9 to −1.3 V, with a scan rate of 0.05 V s−1 for 16 cycles; the deposition solution used was a mixture of 50 μg mL−1 graphene oxide and a 0.1 mmol L−1 solution of tetrachloroauric acid prepared in sodium nitrate (0.1 mol L−1). This electrode was designated as Au-rGO/MWCNT/graphite composite sensor and kept at 25 °C until use. Electrochemical and Morphological Characterization of Au-rGO/MWCNT/Graphite Electrode. A solution of ferri/ferrocyanide (1.0 mmol L−1) was used as the redox species, and the CV experiment was performed as follows: starting potential = 0.0 V, first vertex potential = 1.5 V, and second vertex potential = −0.5 V. Six different scan rates that ranged from 0.1 to 1.0 V s−1 were performed for each CV study using both Au-rGO/MWCNT/graphite and bare graphite electrodes. Phosphate buffer (0.1 mol L−1, pH 7) was used as the supporting electrolyte. The electrochemical impedance spectroscopy (EIS) experiment was performed as follows: an applied potential at open circuit potential (OCP) and frequency range from 1.0 kHz to 0.1 Hz. For scanning electron microscopy (SEM) characterization, the tip of the modified surface graphite electrode was disengaged with a cutter and adhered to the SEM sample platform. A low vacuum mode was used, and the electron power was set at 2 keV. Square-Wave Voltammetry (SWV) Analysis of Nucleotide Bases Using Au-rGO/MWCNT/Graphite. SWV was used for determining G, A, T, and C, with the following parameters: scan potential started at 0.2 V and ended at 1.9 V. The method frequency was set at 25 Hz, with a deposition potential of 0.2 V for 5 s. Phosphate buffer (0.1 mol L−1, pH 7) was used as the supporting electrolyte. The calibration linearity, range, limit of detection (LOD), and limit of quantification (LOQ) of the method were also assessed. The same measuring conditions were used for the pH study of G, A, T and C, with pH adjustment using sodium hydroxide 1.0 mol L−1. Acid Hydrolysis and DNA Composition Analysis in Real Samples. For sample analysis, DNA was extracted using ultrapure water containing 0.5 mg mL−1 sodium dodecyl sulfate and 1.5 mg mL−1 sodium chloride and the solution was blended using a homogenizer. Subsequently, absolute ethanol was used to precipitate the DNA. The DNA was hydrolyzed using 4 mL of sulfuric acid (8 mol L−1). The solution was sonicated at 60 °C for 10 min and heated in boiling water for 5 h. The obtained
their detection. Therefore, an ideal electrochemical sensor should possess a wide anodic potential window and have electrocatalytic properties and high sensitivity. Several electrochemical sensors have been developed for detecting DNA bases, including boron-doped diamond electrodes on silica,14,15 polypyrrole/graphene nanocomposites on glassy carbon electrode (GCE),13 graphene/ionic liquid chitosan on GCE,16 polyaniline/multiwalled-carbon-nanotubes (MWCNT) on silica plate,24 and MWCNT/Fe3O4/polydopamine on GCE.12 However, these attempts have been successfully applied only for measuring G and A, probably because of the poor anodic potential window achieved by the sensor. In fact, Gao et al.13 showed that the polypyrrole graphene in GCE electrodes exhibit electroactivity inhibition upon the application of a high anodic potential at +1.8 V. This was probably related to the electrode substrate or platform used, such as silica and GCE. DNA base detection sensors that are based on graphite surfaces have not been studied before and represent a promising approach for further investigation. However, a drawback of their application is their sensitivity (lower effective surface area) and high overpotential energy.25 Surface modification with nanomaterials may be the best approach to overcome these issues. In this present work, we investigated the feasibility of a graphite substrate as the material for sensor development and performed a simple preparation procedure that utilized an electro-co-deposition technique to intercalate the graphene layer with gold particles on the surface of a MWCNT-graphite sensor, as illustrated in Figure 1. To our knowledge, the
Figure 1. Preparation of Au-rGO/MWNCT/graphite electrochemical sensor by electro-co-deposition using cyclic voltammetry.
application of a nanocomposite graphite sensor for the multiplex detection of G, A, T, and C has not yet been reported. The reliability and applicability of the proposed nanocomposite graphite sensor in DNA analysis were tested in real DNA samples from calf thymus and onion.
■
EXPERIMENTAL SECTION Chemicals and Apparatus. For graphite surface modification, a 3 mg mL−1 solution of carboxyl-functionalized multiwalled carbon nanotubes (MWCNT-COOH) was prepared in Britton-Robinson buffer (0.04 mol L−1, pH 7). The electrodeposition solution used was a mixture of 50 μg mL−1 graphene oxide (GO) and 0.1 mmol L−1 tetrachloroauric acid in sodium nitrate (0.1 mol L−1, pH 7). For DNA analysis, standard stocks of A and G were prepared in sodium hydroxide solution (0.1 mol L−1), while standard C and T stocks were prepared in phosphate buffer (0.1 mol L−1, pH 7). All chemicals were purchased from Sigma-Aldrich (Steinheim, 10005
DOI: 10.1021/acs.analchem.7b02432 Anal. Chem. 2017, 89, 10004−10012
Article
Analytical Chemistry clear solution was neutralized with 8 mL of sodium hydroxide solution (8 mol L−1), and the resulting DNA bases were analyzed using SWV as described in section 2.4 after dilution in the supporting electrolyte (0.1 mol L−1 phosphate buffer, pH 7). Validation of DNA Base Analysis in Real Samples Using UPLC with a Photodiode Array Detector (PDA). The following analytical conditions were used for the analysis of DNA bases using UPLC: an isocratic mode was used with the mobile phase ratio of ultrapure water to phosphate buffer pH 7 (0.1 mol L−1) set at 90:10, flow rate of 0.25 mL min−1, and the chromatogram was extracted at 254 nm. The calibration linearity, range, LOD, and LOQ of the UPLC method were also assessed. To further evaluate the reliability of the method, a recovery study of the spikes of G, A, T, and C in calf thymus DNA at concentrations of 25, 50, and 75 μg mL−1 using both SWV and UPLC methods was also conducted. (Note: The sample solution was diluted 5 times with phosphate buffer (0.1 mol L−1, pH 7), prior to SWV analysis. In other words, the final concentrations of the spiked samples were 5, 10, and 15 μg mL−1). Selectivity and Stability Evaluation of Au-rGO/ MWCNT/Graphite Electrode. The selectivity of the AurGO/MWCNT/graphite in the detection of G, A, T, and C was tested by comparing the current response of DNA base standards with and without the presence of interfering factors. The sensor stability was evaluated by measuring the current response of the DNA standard throughout a 1-month duration. The study was performed under two different conditions, at 25 and 45 °C, measuring the frequency at intervals of 1, 3, 5, 7, 14, 21, and 28 days.
Figure 2. (A) Cyclic voltammogram of the electrode-co-deposition of gold and graphene oxide (GO) on multiwalled carbon nanotubes (MWCNT)/graphite surface at the first cycle (dashed line), eighth cycle (dotted line), and 16th cycle (solid line). (B) Cyclic voltammogram of ferrocyanide using bare graphite (dashed line), MWCNT/graphite (solid line), and Au-rGO/MWCNT/graphite (double line).
cm2), respectively. This suggests that more active sites are available for the oxidation process, which is important for sensitivity. This hypothesis was confirmed by the CV analysis of ferri/ferrocyanide solution (Figure 2B). When Au-rGO/ MWCNT/graphite was used, the current response (Ia = 3.17 × 10−5 A) improved by 2.5- and 1.6-fold compared to bare graphite (Ia = 1.29 × 10−5 A) and MWCNT/graphite (Ia = 2.00 × 10−5 A), respectively. The electron transfer efficiency of the Au-rGO/MWCNT/ graphite was measured by calculating the potential difference (ΔEp) between anodic and cathodic peaks of the ferri/ ferrocyanide redox active solution (39). According to the Nernst equation, ΔEp corresponds to 59.16 mV for a reversible reaction that involves a single electron transfer. The ΔEp values calculated for Au-rGO/MWCNT/graphite, MWCNT/graphite, and bare graphite were 73.5 ± 0.3 mV, 92.0 ± 1.4 mV, and 355.0 ± 2.2 mV, respectively, suggesting that the electron transfer efficiency of Au-rGO/MWCNT/graphite was higher than that of bare graphite (Figure 2B). The in-house fabricated bare graphite electrode that we extracted from a used alkaline AA battery showed high ΔEp (355.0 ± 2.2 mV), suggesting a low electron transfer efficiency, even after optimization via successive polishing steps with emery paper and alumina slurry, followed by sonication in ethanol and ultrapure deionized water, as we previously reported in our initial study.26 To overcome these limitations, electrode surface modification is required for its practical use in analytical applications. This was also apparent by comparing the Ks of the graphite sensor using the Laviron equation. The Ks values decreased in the following order: Au-rGO/MWCNT/graphite (27 s−1) > MWCNT/ graphite (8 s−1) > bare graphite (4 s−1). Thus, the Au-rGO/ MWCNT/graphite possessed fast electron transfer kinetic properties. The anodic peak potential of ferri/ferrocyanide using AurGO/MWCNT/graphite (Ea = 0.2675 V) shifted toward a more negative potential by 75.7 mV and 136.7 mV, compared to that of MWCNT/graphite (Ea = 0.3432 V) and bare graphite
■
RESULTS AND DISCUSSION Electrochemical Characterization of Au-rGO/MWCNT/ Graphite Sensor. Gold and graphene oxide (Au-rGO) were electro-co-deposited on MWCNT/graphite by using CV, and the resulting voltammogram is illustrated in Figure 2A, which shows a series of cathodic and anodic peaks. The peak (I) at potential 0.2 V showed an increasing trend of the cathodic current with subsequent CV scan (deposition), suggesting a reduction (electrodeposition) of the tetrachloroauric acid to its elemental gold on the MWCNT/graphite surface. The other cathodic peaks (II and IV) and anodic peak (III) represent the graphene deposition,28 with peaks II and III possibly associated with a reversible redox process of a graphene-functionalized oxide group (COOH, OH). Similarly, at peak IV, the cathodic peak current showing an increasing trend was associated with the irreversible reduction of rGO on the MWCNT/graphite surface. The gold particle intercalated with rGO when performing a CV scan from positive to negative potential, suggesting that gold was deposited first followed by graphene. Therefore, the gold particles were intercalated between the rGO sheets. This unique structure prevented graphene from being agglomerized. The electrochemical characteristics of Au-rGO/MWCNT/ graphite, MWCNT/graphite, and bare graphite were evaluated, including the effective surface area (Eff A), heterogeneous electron transfer rate (Ks), electrode sensitivity, and electrode overpotential. The Eff A of the Au-rGO/MWCNT/graphite was determined using the Randles-Sevcik equation, showed an improvement up to 0.151 ± 0.001 cm−1 corresponding to 0.5to 4-fold expansion of the Eff A compared to MWCNT/ graphite (0.109 ± 0.001 cm2) and bare graphite (0.037 ± 0.001 10006
DOI: 10.1021/acs.analchem.7b02432 Anal. Chem. 2017, 89, 10004−10012
Article
Analytical Chemistry
T and C were more intense than the currents observed by the CV scan.
(Ea = 0.4042 V), respectively (Figure 2B). This shift of peak potential suggested an improvement in the overpotential, indicating a much easier oxidation of the ferrocyanide. This could be attributed to the electrocatalytic properties of the reduced GO and gold. The electron transfer resistance of the fabricated sensor was evaluated using EIS and the Nyquist plot, as illustrated in Figure S1. Both MWCNT/graphite and Au-rGO/MWCNT/ graphite showed a linear line, suggesting a low electron transfer resistance. In contrast to bare graphite, a semicircle graph was observed, suggesting a resistance in electron transfer between the sensor and the interface. Morphology Evaluation of the Au-rGO/MWCNT/ Graphite. The surface of the Au-rGO/MWCNT/graphite was studied using a field emission (FE)-SEM, and the resulting images are illustrated in Figure S2. The GO sheet and gold nanoparticles are shown in Figure S2A, where the GO is a wrinkled, flexible sheet. The surface of the bare graphite was smooth and with a clean surface, as shown in Figure S2B. From the images shown in Figure S2C,D, the MWCNT appeared as a tubular, long-shaped structure, homogeneously distributed on the surface. The presence of gold nanoparticles on the electrode surface was further confirmed by EDX, and they appeared as irregular-shaped (Figure S2A), with an average particle size between 70 and 130 nm. Electrochemical Characteristics of G, A, T, and C. The redox properties of the DNA bases were determined using CV with Au-rGO/MWCNT/graphite. Figure 3A−D illustrate the
Figure 4. Square wave voltammetry (SWV) of (A) guanine, (B) adenine, (C) cytosine, and (D) thymine from pH 2 to 9 using AurGO/MWCNT/graphite electrode and 20 μg mL−1 of the standards. The experiment was performed using SWV with a start potential of 0.2 V; stop potential of 1.9 V; deposition potential of 0.2 V; deposition time of 5 s; and frequency of 25 Hz.
The DNA base response to SWV at various pH conditions was also studied, and the resulting voltammograms are illustrated in Figure 4A−D. At pH 2, only G and A were detected, while at pH 3 and above, all DNA bases were detectable using the Au-rGO/MWCNT/graphite. Upon further evaluation of the anodic peak current (Ipa), the effective pH conditions for the optimized anodic peak current (Ipa) of the multiplex DNA bases detection ranged from pH 5−9. Figure S3A shows the graph of the anodic peak current response under different pH conditions. Both G and A exhibited the highest current response at pH 3 and 4; however, these are not the optimum pH values because of the weak response of T and C, limiting the detection sensitivity. Analysis of the peak current response also suggested that both G and A were more oxidizable in moderately acidic conditions (pH 3−5), giving higher current response, in contrast to T and C, which were more oxidizable at moderately alkaline conditions (pH 8−9). On the basis of the pH optimization study, we selected pH 7 for the multiplex study of the four DNA bases, because pH 7 is reported to be ideal for analyzing biological samples.7 Figure S3B illustrates a graph of the anodic peak potential against the pH values, where the slope can be used to determine the number of electron transfers. According to the Nernst equation, this value corresponds to 59.16 mV/decade for a one electron to one proton transfer process. The slopes obtained for G, A, T, and C corresponded to −55.31 mV pH−1, −62.18 mV pH−1, −56.00 mV pH−1, and −66.17 mV pH−1, respectively. Therefore, the electrochemical reaction of the four DNA bases on the Au-rGO/MWCNT/graphite surface involves a one electron and one proton transfer process, with a possible oxidation reaction on the Au-rGO/MWCNT/graphite interface as postulated in Figure S4. Multiplex Oxidation of G, A, T, and C Using Au-rGO/ MWCNT/Graphite Nanocomposite Electrode. Figure 5 shows the voltammogram for the multiplex detection of G, A, T, and C, using the Au-rGO/MWCNT/graphite sensor, MWCNT/graphite, and bare graphite electrode. Each base was successfully separated in the following order of oxidation potential: guanine (0.77 V) < adenine (1.05 V) < thymine (1.21 V) < cytosine (1.40 V). In contrast, C was not detectable
Figure 3. Cyclic voltammogram of (A) guan,ine, (B) adenine, (C) thymine, and (D) cytosine. The insets in parts C and D are illustration of the enlarged versions of the thymine and cytosine peaks. The solid line and dotted line correspond to the 20 μg mL−1 standard and blank solution scans, respectively. The experiment was performed using CV at starting potential 0.0 V, first vertex = 1.5 V, second vertex = −0.5, and scan rate of 0.1 V s−1.
CV of G, A, T, and C, respectively. Only a single anodic peak (oxidation) was detected in all the DNA bases, suggesting that the redox process was chemically irreversible, with their anodic peak potentials corresponding to 0.74, 1.01, 1.15, and 1.36 V, respectively. No oxidation peak was detected in the blank scan, except for the reduction peak (cathodic) in the blank and other DNA bases (0.45 V) that could result from reducing gold oxide. The current response of C and T DNA bases was weak, because they are less susceptible to oxidation. To improve detection, SWV was used, and the voltammograms are illustrated in Figure 4A−D, showing that the peak currents of 10007
DOI: 10.1021/acs.analchem.7b02432 Anal. Chem. 2017, 89, 10004−10012
Article
Analytical Chemistry
a concentration-dependent increase in current response. A calibration curve was plotted in Figure 6B,C, which illustrates that the correlation coefficients of G, A, T, and C were greater than 0.991. The dynamic range for both G and A was up to 25 μg mL−1, whereas for T and C, it was up to 100 μg mL−1. From the slope of the calibration curve, the order of sensitivity can be expressed as follows: G (178 nA/μg mL−1) > A (92.9 nA/μg mL−1) > C (15.1 nA/μg mL−1) > T (1.29 nA/μg mL−1), suggesting a sensitive detection of G. The LOD values for G, A, T, and C corresponded to 0.5 μg mL−1, 0.5 μg mL−1, 1 μg mL−1, and 1 μg mL−1, respectively. The LOQ values were calculated at 5 times the standard deviation of the LOD, and the values for G, A, T, and C corresponded to 0.8 μg mL−1, 1.2 μg mL−1, 1.4 μg mL−1, and 1.3 μg mL−1, respectively. To determine the fitness using the AU-rGO/MWCNT/graphite in DNA base analysis, the analytical performance was compared to that of UPLC. Figure S5A illustrates the chromatogram of C, G, T, and A using a C18 column. All the DNA bases were successfully separated, and the corresponding peak responses showed a concentration-dependent increase. The calibration plot of G, A, T, and C is illustrated in Figure S5B, showing a good correlation linearity above 0.999. Table S1 summarizes the analytical performance of both SWV and UPLC methods. SWV was more sensitive toward G and C detection, and UPLC was more sensitive toward A and T detection. The sensitivity factor was calculated based on the normalized response of 12 μg mL−1 standard as both methods had a different response unit. From the dynamic range, SWV showed a wider linearity range in C and T analysis and a narrower range for G and A. UPLC showed a better detection limit, which could be explained by the advantages of automated sampling compared to manual addition in SWV. SWV was more rapid, requiring less than 30 s per analysis, in contrast to 8 min for UPLC. We concluded that SWV is as good as the UPLC method for analyzing DNA bases. The analytical performance of the Au-rGO/MWCNT/ graphite was compared to that of other reported electrochemical sensors used for multiplex DNA bases analysis, and the details are summarized in Table 1. This comparison revealed the ability of Au-rGO/MWCNT/graphite sensors to accomplish the multiplex detection of four DNA bases (G, A, T, and C) within a single test, whereas most of the reported electrochemical sensors can only detect A and G, which limits their application in measuring the DNA base ratios (A/T and C/G) for studying DNA damage. The detection limit of A and G is slightly higher (10−6 mol L−1) than that of other sensors (10−7−10−9 mol L−1); however, this does not affect the efficiency of the developed sensor, which is the measurement of total genomic DNA instead of single genes or trace DNA bases. Moreover, the sample size and dilution factor can be adjusted for the required concentration. Although the simultaneous detection of G, A, T, and C has been successfully demonstrated by M. Pumera and colleagues (2012), the authors were not able to integrate the signal response of T and C peaks (due to very low peak signal obtained for T and C), which meant that for the C/T ratio study even by using ER-GO electrode, only the G/A signal ratio could be determined quantitatively.23 In a much earlier work by Kato et al. (2008), the authors claimed that they could detect G, A, and C simultaneously but that they faced difficulty in detecting T; even by using the ECR nanocarbon film electrode, the oxidation peaks of A and T reportedly overlapped. This issue can only be resolved with the aid of a
Figure 5. Square wave voltammogram (SWV) of simultaneous detection of guanine (G), adenine (A), thymine (T), and cytosine (C) using different surface-modified graphite electrodes.
by either MWCNT/graphite or bare graphite electrodes. T was only detectable by MWCNT/graphite but was fused to the adjacent A peak. Both T and C were oxidized less effectively, because it is more difficult for the pyrimidine structure to donate electrons (become oxidized) due to slow electron transfer rates.22,29 The successful detection of T and C by AurGO/MWCNT/graphite could be attributed to the intercalation of gold and rGO layers that improved the Ks and electrocatalytic properties of the sensor. This is because a higher Ks implies a higher Ks/n value, and increasing values of Ks/n are favorable for the diffusion controlled process of T and C. On the other hand, the Ipa values of A (44.6 μA) and G (71.4 μA) were improved significantly by 2- and 10-fold, compared to those for MWCNT/graphite and bare graphite, respectively. Analytical Performance of Au-rGO/MWCNT/Graphite Electrode in Multiplex Analysis of DNA Bases. The SWV response of A, C, T, and G to different concentrations ranged from 2 to 20 μg mL−1, as illustrated in Figure 6A, which shows
Figure 6. (A) Square wave voltammogram (SWV) of simultaneous detection of guanine (G), adenine (A), thymine (T), and cytosine (C) with concentrations ranging from 2 to 25 μg mL−1 for G and A and 6 to 100 μg mL−1 for T and C. (B) Calibration graph of G and A (n = 3). (C) Calibration graph of C and T (n = 3). (D) Correlation graph of SWV results versus ultraperformance liquid chromatography (UPLC) results. All SWV experiments were performed using AurGO/MWCNT/graphite electrode. The supporting electrolyte and standards (G, A, T, and C) were prepared in phosphate buffer (0.1 mol L−1, pH 7). 10008
DOI: 10.1021/acs.analchem.7b02432 Anal. Chem. 2017, 89, 10004−10012
10009
C G A T C G A T C
G A T
G A G A T G A
A
G
DNA bases
SWV
DPV
SWV
DPV
LSV
DPV
CV
method
× × × × × × ×
10−9 10−8 10−7 10−7 10−6 10−6 10−6
3.3 3.7 7.9 9.0
× × × ×
10−6 10−6 10−6 10−6
data not available
data not available for T and C
125 fmol for G and A
2.0 4.0 1.2 1.4 2.6 1.5 5.7
2.5 × 10−7
7.9 × 10
−8
detection limit (mol L−1)
× × × × × × ×
10−6−1.0 10−6−2.0 10−4−5.0 10−4−5.0 10−4−5.0 10−4−8.0 10−4−1.0 × × × × × × ×
1.7 1.9 8.0 9.0
× × × ×
10−4−3.0 10−4−3.0 10−4−7.5 10−4−9.0 × × × ×
data not available
data not available
1.6 5.0 1.5 1.5 6.0 1.3 1.2
10−6 10−6 10−6 10−6
10−9 10−8 10−7 10−7 10−6 10−6 10−7
7.0 × 10−5−1.5 × 10−6
−7
5.0 × 10 −3.0 × 10
−5
linear range (mol L−1)
all G, A, T, and C can be detected simultaneously within a single analysis
only G/A signal ratio can be determined quantitatively but not T/C ratio
only Ipa peaks for G, A, and C are well-separated and can be detected quantitatively. T is found overlapping with A and can only be detected and measured by subtracting the two voltammograms which led to a clearer discrimination of C → T mutation
only G and A can be detected
only G, A and T can be detected
only G and A can be detected
only G and A can be detected
notes
this study
Toh et al., 201223
Yari and Derki, 201612 Kato et al., 200829
Sun et al., 200818 Ferancová et al., 20107 Ye et al., 201417
ref
a CV, cyclic voltammetry; DPV, differential pulse voltammetry; LSV, linear sweep voltammetry; SWV, square-wave voltammetry. bCarbon ionic liquid electrode. cCadmium sulfur−chitosan glassy carbon electrode. dZinc sulfur coated poly(3,4-ethylenedioxythiophene) reduce the graphene oxide hybrid film. eMultiwalled carbon nanotubes-Fe3O4 incorporated polydopamine silver nanoparticles f Intercalation gold nanoparticles-reduced graphene oxide/multiwall carbon nanotubes/graphite electrode
Au-rGO/MWCNT/ graphitef
ER-GO/GCE
ECR Nanocarbon film
MWCNT-Fe3O4@PDAAge
ZnS-PEDOT-rGO/GCd
CdS-CHIT/GCEc
CILE
b
electrodes
Table 1. Comparison of the Analytical Performance Shown by Different Reported Electrochemical Sensors Used for Multiplex DNA Base Analysisa
Analytical Chemistry Article
DOI: 10.1021/acs.analchem.7b02432 Anal. Chem. 2017, 89, 10004−10012
Article
Analytical Chemistry subtraction step on the two voltammograms obtained from a wild-type oligonucleotide and its single-base mismatch oligonucleotides to discriminate the C → T mutation.29 In other words, this is an indirect method for studying DNA damage where a complicated subtraction step would be necessary if a longer chain of single-stranded DNA or a short single-stranded DNA with more than one single nucleotide polymorphism (SNP) mutation is to be studied. Apart from this, for both the above-mentioned reported studies, the simultaneous detection of G, A, T, and C was only tested on a short single-stranded DNA but has not been studied on a double-stranded DNA sample. In contrast, our proposed electrode has successfully overcome the limitations encountered by other researchers for their respective electrochemical sensors, of not being able to detect four DNA bases (G, A, T, and C) quantitatively, directly, and simultaneously. Thus, this method is important and very useful for studying DNA damage (at the wider scope of DNA genome study instead of singlestrand DNA/single gene study), as the ratios of A to T and G to C are measurable using Au-rGO/MWCNT/graphite. Hydrolysis of DNA Samples by Acid Digestion. DNA contains nucleic acid, ribose sugar, and phosphate. To ensure the liberation of G, A, T, and C, a complete hydrolysis of the DNA bases is required. In this present study, three different concentrations of sulfuric acid, corresponding to 1, 4, and 8 mol L−1, were used to hydrolyze standard calf thymus DNA. At lessacidic concentrations (1 and 4 mol L−1), C and T were undetectable by UPLC and SWV. This could be explained by the incomplete hydrolysis of T and C. In contrast, at 8 mol L−1 sulfuric acid, all the DNA bases were detectable. At this concentration, the hydrolysis time used was further optimized at four different time points (4−7 h), and the percentage recovery of G, A, T, and C is illustrated in Figure S6A. At a 5-h hydrolysis time, the recovery of A, C, T, and G was the most efficient. At more than 6-h hydrolysis, C recovery was lower due to acid degradation. When sulfuric acid was substituted with hydrochloric acid, both C and T were not detectable by SWV, as illustrated in Figure S6B. We concluded that the chloride ions of the acid possibly inhibited the oxidation potential of Au-rGO/MWCNT/graphite at high values (>1.3 V), interfering with T and C detection. Thus, sulfuric acid is recommended for the hydrolysis of DNA samples. Analysis of DNA Bases in Real Samples. The developed SWV method was used to analyze two real DNA samples obtained from calf thymus and onion. Both samples were selected owing to differences in the DNA of the source organisms. Table 2 summarizes the composition of DNA bases in onion and calf thymus samples using SWV. The molar percentage of G, A, T, and C in calf thymus DNA corresponded to 21.5 ± 0.5, 27.9 ± 0.4, 28.5 ± 0.4, and 22.0 ± 0.3 mol %, respectively, while for onion DNA the results corresponded to 29.1 ± 0.2, 20.9 ± 0.2, 21.2 ± 0.3, and 28.8 ± 0.1 mol %, respectively. According to Chargaff’s rules, the mole percentage of G should be equal to C, and that of A should be equal to T. On the basis of the results obtained from both calf thymus and onion DNA, the ratios of C to G (onion = 0.99, calf thymus = 1.02) and T to A (onion = 1.01, calf thymus = 1.02) were close to unity, suggesting intact DNA bases. The ratios of purines (G and A) to pyrimidines (C and T) were calculated using the formula (C + G)/(A + T).13,15,30 According to Chargaff’s rules, this value is specific for different organisms. Higher values suggest a stronger interaction between the double-stranded
Table 2. Simultaneous Determination of Guanine, Adenine, Thymine, and Cytosine Contents in Calf Thymus and Onion DNA Samples Using the Square Wave Voltammogram (SWV) Method with the Au-rGO/MWCNT/Graphite Electrochemical Sensor versus Ultraperformance Liquid Chromatography (UPLC) calf thymus DNA SWVa result analyte guanine adenine thymine cytosine
c
(mol %) 21.5 27.9 28.5 22.0
± ± ± ±
0.5 1.1 0.8 0.3
extracted onion DNA
UPLC−PDAb result
c
(mol %) 21.2 28.1 28.5 22.1
± ± ± ±
0.3 0.2 0.5 0.2
SWVa result
c
(mol %) 29.1 20.9 21.2 28.8
± ± ± ±
0.2 0.2 0.3 0.1
UPLC−PDAb resultc (mol %) 29.4 20.7 20.6 29.2
± ± ± ±
0.2 0.2 0.1 0.6
a
Square wave voltammetry. bUltraperformance liquid chromatography photodiode array detector. c3 replicate (n = 3) analyses with standard error at 95% confidence limit.
(DS) DNA, as CG pairing contains three hydrogen bonds and AT pairing only involves two hydrogen bonds. The ratios obtained for onion and calf thymus DNA were 1.38 and 0.77, respectively, suggesting that onion DNA possessed a stronger DS DNA interaction compared to calf thymus DNA. The ratio for calf thymus DNA was also in agreement with the value of 0.77 reported by Gao et al. and Liu et al.,13,30 which used polypyrrole/graphene and polythionine/gold nanoparticles/MWCNT, respectively. The limitation of both the previous studies was their calculations, which were based on the G to A ratio only. In this present study, the DNA base ratio was calculated based on the simultaneous detection of G, A, T, and C. The result reliability was confirmed by UPLC, and the results are summarized in Table 2. Student’s t test showed that both result sets (SWV and UPLC) were not significantly different at a 95% confidence limit. A spike recovery analysis was performed on the sample matrix at three different concentrations (25, 50, and 75 μg mL−1), and the results are shown in Table 3. On average, SWV showed a recovery percentage for all DNA bases at 100.5 ± 0.5%, close to UPLC at 100.3 ± 0.8%. Figure 5D illustrates a correlation graph of analysis results obtained using SWV and UPLC. A linear coefficient of 0.9968 was obtained, and the slope value was equal to 1.0391, suggesting that both methods correlated by a factor of 103.9%. This suggested that the proposed SWV method using Au-rGO/MWCNT/graphite was reliable and comparable to the established UPLC technique. Selectivity and Stability Study of Au-rGO/MWCNTCOOH/Graphite in DNA Bases Analysis. The selectivity of the Au-rGO/MWCNT/graphite sensor in the analysis of G, A, T, and C was studied by evaluating the percentage recovery of DNA bases in the presence of common interfering factors, including uracil, vitamins (niacinamide and pantothenic acid), amino acids (cysteine, serine, glutamine, and tyrosine), glucose, and salt. These interferences are commonly present in animal and plant samples. The percentage recovery of the DNA bases is summarized in Table S2. All DNA base standards were recovered at 92% even with the presence of interfering factors. Uracil is an RNA base that is known to interact with A. The recovery of G, A, T, and C was still above 99%. This suggested that the proposed Au-rGO/MWCNT/graphite sensor was selective toward the detection of G, A, T, and C. 10010
DOI: 10.1021/acs.analchem.7b02432 Anal. Chem. 2017, 89, 10004−10012
Article
Analytical Chemistry
Table 3. Simultaneous Determination of Guanine, Adenine, Thymine, and Cytosine Contents in Calf Thymus DNA Using the Square Wave Voltammogram (SWV) Method with the Au-rGO/MWCNT/Graphite Electrochemical Sensor versus Ultraperformance Liquid Chromatography (UPLC) SWVa analyte guanine
adenine
thymine
cytosine
spiked (μg mL−1) 25 50 75 25 50 75 25 50 75 25 50 75
UPLC−PDAb
resultc (μg mL−1)
recovery (%)
± ± ± ± ± ± ± ± ± ± ± ±
103.7 ± 1.9 103.3 ± 1.2 100.4 ± 1.7 98.5 ± 2.6 96.7 ± 2.0 99.4 ± 2.8 100.4 ± 2.0 101.3 ± 4.5 100.1 ± 1.3 100.6 ± 2.1 100.8 ± 2.0 101.0 ± 3.0
25.9 51.6 75.3 24.6 48.3 74.6 25.1 50.7 75.1 25.2 50.4 75.8
0.5 0.6 1.3 0.7 1.0 2.1 0.5 2.3 1.0 0.5 1.0 2.6
resultc (μg mL−1)
recovery (%)
± ± ± ± ± ± ± ± ± ± ± ±
96.2 ± 2.4 98.3 ± 0.4 103.4 ± 0.2 96.3 ± 1.0 97.4 ± 1.1 101.9 ± 0.5 102.1 ± 4.6 98.8 ± 1.0 103.0 ± 0.4 102.8 ± 0.6 101.6 ± 1.0 101.6 ± 2.8
24.1 49.1 77.6 24.1 48.7 76.4 25.5 49.4 77.3 25.7 50.8 76.2
0.6 0.2 0.2 0.3 0.5 0.4 1.1 0.5 0.3 0.2 0.5 2.1
a
Square wave voltammetry. bUltraperformance liquid chromatography-photodiode array detector. c3 replicate (n = 3) analyses and with standard error at 95% confidence limit. dThe final concentrations after the 5-fold dilution of spiked samples are 5, 10, and 15 μg mL−1.
The storage stability of Au-rGO/MWCNT/graphite was evaluated especially for laboratory or commercial use. Two stability conditions were tested, at 25 °C and at 45 °C, up to 28 days of storage. The results are illustrated in Figure S7A,B, showing that the sensor was stable at 25 °C up to one month. Under the 45 °C storage condition, the sensor response toward T was 10% lower after 14 days and for the sensor response toward C and T, the signal slowly decreased after 7 days. After 28 days, although the recovery was not 100%, at the level of 80% response, the electrode was still able to detect C. According to the Arrhenius equation on the relationship between reaction rate and temperature, chemical stability decreases exponentially with increased temperature. The fabricated Au-rGO/MWCNT/graphite electrode was stable at 45 °C, implying that the sensor could be stored for a longer time at 25 °C.
■
AUTHOR INFORMATION
Corresponding Author
*E-mail:
[email protected]. Fax: +603-7967 4193. ORCID
Sook Mei Khor: 0000-0001-6050-0195 Notes
The authors declare no competing financial interest. The author Ng Khan Loon is currently an employee of Wipro UNZA Holdings Ltd. Other than the concerned author’s input in the manuscript, Wipro UNZA Holdings Ltd. has no direct role in the study design, data analysis, and data collection.
■
■
CONCLUSIONS We have developed a Au-rGO/MWCNT/graphite sensor and successfully used it to measure G, A, T, and C simultaneously using SWV. Au-rGO/MWCNT/graphite was reliable and could be practically used for the simultaneous analysis of G, A, T, and C in real plant and animal DNA samples. The results obtained correlated to those obtained by UPLC at 103.9%. Besides, the newly developed sensor exhibited high sensitivities and low LODs and LOQs compared to UPLC. The developed method was reliable, with recovery of the prespiked DNA bases in sample matrix above 92%. This suggests that the proposed sensor is as good as UPLC and practically suitable for the intended application in clinical or biomedical studies, especially in studies on DNA damage. The developed sensor was also selective toward G, A, T, and C detection in the presence of common biological interfering factors, and it also showed a stable performance over time in storage.
■
Further details on SEM morphology of Au-rGO/ MWCNT/graphite and related EIS result, postulated DNA bases reaction on the sensor surface, UPLC chromatogram, and sensor stability and selectivity study results (PDF)
ACKNOWLEDGMENTS This work was financially supported by the University of Malaya post graduate grant (PPP) Grant PG177-2014B, UMRG-Programme Grant RP012C-14SUS, and the Fundamental Research Grant Scheme (FRGS) from the Ministry of Higher Education of Malaysia (MOHE) Grant FP041-2016. The author would like to thank Wipro Manufacturing services for the postgraduate Ph.D. scholarship funding and support.
■
REFERENCES
(1) Soares, J. P.; Silva, A. M.; Fonseca, S.; Oliveira, M. M.; Peixoto, F.; Gaivao, I.; Mota, M. P. Exp. Gerontol. 2015, 62, 45−52. (2) El-Yazbi, A. F.; Loppnow, G. R. TrAC, Trends Anal. Chem. 2014, 61, 83−91. (3) Sprung, C. N.; Ivashkevich, A.; Forrester, H. B.; Redon, C. E.; Georgakilas, A.; Martin, O. A. Cancer Lett. 2015, 356 (1), 72−81. (4) de Carvalho, A. M.; Carioca, A. A. F.; Fisberg, R. M.; Qi, L.; Marchioni, D. M. Nutrition 2016, 32 (2), 260−264. (5) Hossain, M. Z.; Gilbert, S. F.; Patel, K.; Ghosh, S.; Bhunia, A. K.; Kern, S. E. Food Chem. Toxicol. 2013, 55, 557−567. (6) Shamsi, M. H.; Kraatz, H. B. J. Inorg. Organomet. Polym. Mater. 2013, 23 (1), 4−23. (7) Ferancová, a.; Rengaraj, S.; Kim, Y.; Labuda, J.; Sillanpäa,̈ M. Biosens. Bioelectron. 2010, 26 (2), 314−320.
ASSOCIATED CONTENT
S Supporting Information *
The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.analchem.7b02432. 10011
DOI: 10.1021/acs.analchem.7b02432 Anal. Chem. 2017, 89, 10004−10012
Article
Analytical Chemistry (8) Del Moral, P. G.; Arin, M. J.; Resines, J. A.; Diez, M. T. J. Chromatogr. B: Anal. Technol. Biomed. Life Sci. 2005, 826 (1−2), 257− 260. (9) Glavin, D. P.; Cleaves, H. J.; Buch, A.; Schubert, M.; Aubrey, A.; Bada, J. L.; Mahaffy, P. R. Planet. Space Sci. 2006, 54 (15), 1584−1591. (10) Pang, S.; Zhang, Y.; Wu, C.; Feng, S. Sens. Actuators, B 2016, 222, 857−863. (11) Yang, F. Q.; Ge, L.; Yong, J. W. H.; Tan, S. N.; Li, S. P. J. Pharm. Biomed. Anal. 2009, 50 (3), 307−314. (12) Yari, A.; Derki, S. Sens. Actuators, B 2016, 227, 456−466. (13) Gao, Y. S.; Xu, J. K.; Lu, L. M.; Wu, L. P.; Zhang, K. X.; Nie, T.; Zhu, X. F.; Wu, Y. Biosens. Bioelectron. 2014, 62, 261−267. (14) Niedziałkowski, P.; Ossowski, T.; Zięba, P.; Cirocka, a.; Rochowski, P.; Pogorzelski, S. J.; Ryl, J.; Sobaszek, M.; Bogdanowicz, R. J. Electroanal. Chem. 2015, 756, 84−93. (15) Švorc, L.; Kalcher, K. Sens. Actuators, B 2014, 194, 332−342. (16) Niu, X.; Yang, W.; Ren, J.; Guo, H.; Long, S.; Chen, J.; Gao, J. Electrochim. Acta 2012, 80, 346−353. (17) Ye, X.; Du, Y.; Duan, K.; Lu, D.; Wang, C.; Shi, X. Sens. Actuators, B 2014, 203, 271−281. (18) Sun, W.; Li, Y.; Duan, Y.; Jiao, K. Biosens. Bioelectron. 2008, 24 (4), 988−993. (19) Jayakumar, K.; Rajesh, R.; Dharuman, V.; Venkatasan, R.; Hahn, J. H.; Pandian, S. K. Biosens. Bioelectron. 2012, 31 (1), 406−412. (20) Hasanzadeh, M.; Shadjou, N.; Eskandani, M.; Soleymani, J.; Jafari, F.; de la Guardia, M. TrAC, Trends Anal. Chem. 2014, 53, 137− 149. (21) Wong, E. L. S.; Erohkin, P.; Gooding, J. J. Electrochem. Commun. 2004, 6 (7), 648−654. (22) Oberacher, H.; Erb, R.; Plattner, S.; Chervet, J.-P. TrAC, Trends Anal. Chem. 2015, 70, 100−111. (23) Toh, R. J.; Bonanni, A.; Pumera, M. Electrochem. Commun. 2012, 22 (1), 207−210. (24) Tran, L. D.; Nguyen, D. T.; Nguyen, B. H.; Do, Q. P.; Le Nguyen, H. Talanta 2011, 85 (3), 1560−1565. (25) Wring, S. a.; Hart, J. P. Analyst 1992, 117 (8), 1215. (26) Ng, K. L.; Lee, S. M.; Khor, S. M.; Tan, G. H. Anal. Sci. 2015, 31 (10), 1075−1081. (27) Ng, K. L.; Tan, G. H.; Khor, S. M. Food Chem. 2017, 237, 912− 920. (28) Chen, L.; Tang, Y.; Wang, K.; Liu, C.; Luo, S. Electrochem. Commun. 2011, 13 (2), 133−137. (29) Kato, D.; Sekioka, N.; Ueda, A.; Kurita, R.; Hirono, S.; Suzuki, K.; Niwa, O. Angew. Chem., Int. Ed. 2008, 47 (35), 6681−6684. (30) Liu, H.; Wang, G.; Chen, D.; Zhang, W.; Li, C.; Fang, B. Sens. Actuators, B 2008, 128 (2), 414−421.
10012
DOI: 10.1021/acs.analchem.7b02432 Anal. Chem. 2017, 89, 10004−10012
