Combining a loop-stem aptamer sequence with methylene blue: a simple assay for thrombin detection by resonance light scattering technique

Abstract

An ingenious sensing strategy for detecting thrombin in human serum has been developed on the basis of a hairpin DNA sequence and resonance light scattering (RLS) technique. A thrombin aptamer sequence was embedded inside the hairpin DNA strand (H-eTBA), which was designed to be the loop-stem structure. Moreover, methylene blue (MB) was utilized as the RLS signal indicator according to its different affinity to single or double stranded DNA. Upon the addition of thrombin, the thrombin aptamer inside H-eTBA interacted specifically with thrombin. Thus the conformation of H-eTBA would change. After the introduction of the DNA strand (CTBA), which was complementary to H-eTBA, the amount of double stranded DNA would decrease as a consequence. Later when MB solution was added, the RLS signal would present various response values based on different amounts of thrombin. The determination of thrombin in human serum could be obtained with a detection limit of 0.32 nM and this specific sensor could be applied to detect thrombin practically. Furthermore, this aptasensor showed quite good selectivity and simplicity toward thrombin. Finally, the proposed sensing method showed its superiority with selectivity and practicability, which could be used as a simple platform for thrombin detection.

---

1. Introduction

Thrombin occupies an essential position in many physiological and pathological processes, such as blood coagulation, thrombosis, inflammation, angiogenesis and metastasis.¹ Thrombin can also be utilized as a therapeutic and a biomarker^2,3 for diagnosis of coagulation abnormalities. Obviously, the development of a new strategy that is selective and simple is necessary and crucial for the detection of thrombin.

Aptamers, which are single-stranded DNA molecules or RNA, could selectively bind to different target molecules^4,5 with high affinity and gained broad appeal because of their simple synthesis. Moreover, the properties of excellent stability, wide applicability and superior sensitivity make aptamers a suitable analytical agent in many medical diagnoses.^6–9 Additionally, a hairpin DNA sequence shows high hybridization specificity because of its loop-stem structure. It can easily discriminate the complementary strand from a single-mutation target or mismatched DNA probe. Therefore, a hairpin DNA probe could be used to get a better selectivity in a DNA based probe.

Methylene blue (MB) is a very popular phenothiazine dye. It has demonstrated its different affinities towards single (ss) and double-stranded (ds) DNA.¹⁰ A positively charged MB molecule could accumulate on the surface of the double helix structure of negatively charged dsDNA through electrostatic attractions. This unique affinity of MB with ss and dsDNA allowed the application of MB as an indicator in DNA-based protein assays.¹¹ Resonance light scattering (RLS) is a kind of elastic light scattering, which is produced while the incident beam is close to its molecular absorption band.¹² Pasternack initially established the RLS method, which was developed for analytical application by Huang et al.¹³ Over the next two decades, RLS was applied widely for detecting nucleic acids,¹⁴ anti-cancer drugs^15,16 and proteins,¹⁷ etc. An increasing number of studies indicated that the RLS technique was becoming a most popular testing method in daily application.

Different methods such as optical,^18,19 electrochemical,^20,21 surface enhanced resonance Raman scattering,²² surface plasmon resonance²³ and so on have been developed to detect thrombin. However, not only the conventional techniques mentioned above but also some limited conditions²⁴ have a negative impact on the detection of thrombin. In this paper, a combination of aptamer with hairpin DNA structure revealed significantly improved analytical performance towards thrombin detection and this combination accompanied with the RLS technique has not yet been reported. Particularly, the addition of thrombin could lead to a structural change of the hairpin sequence, after which different amounts of dsDNA or ssDNA would form when a complementary strand of hairpin sequence was added to interact with hairpin DNA. Subsequently, the addition of MB solution expressed distinct affinity toward dsDNA and ssDNA and the system demonstrated a different RLS signal that could be used to make quantitatively analysis. Therefore, an aptamer sensor could be obtained, which could be used successfully for thrombin detection in human serum samples .

2. Experimental

2.1 Regents and apparatus

Oligonucleotides were synthesized by Sangon Biotechnology Co., Ltd. (Shanghai, China) and their sequences are shown in Table 1. (Bases shown in bold font are aptamer segments of thrombin except the single-base mismatched DNA. Bases shown in italic font are the stem segments of the hairpin DNA. Underlined bases are totally complementary to the DNA sequence added subsequently.)

Table 1 Sequences of oligonucleotides used in this work

Name^a	Sequences
a H-eTBA: hairpin DNA embedded a thrombin binding aptamer; CTBA: complementary strand to H-eTBA; Mis-DNA: single-base mismatched DNA (double underline and in bold); non-specific: totally non-specific DNA sequence.
H-eTBA	5′-GA̲A̲T̲T̲C̲ TTAAA̲ G̲G̲T̲T̲G̲G̲T̲G̲T̲G̲G̲T̲T̲G̲ G̲AATTC-3′
CTBA	5′-CCAACCACACCAACCTTTAAGAATT-3′
Mis-DNA	5′-CCAACCACACAACCTTTAAGAATT-3′
Non-specific	5′-GA̲A̲T̲T̲C̲ T̲T̲T̲T̲T̲T̲C̲C̲C̲C̲C̲C̲T̲T̲T̲T̲T̲T̲ G̲AATTC-3′
Complementary strand towards non-specific	5′-CAAAAAAGGGGGGAAAAAAGAATT-3′

---

Methylene blue (MB) was obtained from Aladdin Chemical Reagent Co., Ltd. (Shanghai, China). Human α-thrombin, bovine serum albumin (BSA), trypsin, α-chymotrypsin were purchased from Bomei Biotechnology Co., Ltd. (Hefei, China). Tris–HCl buffer solution (20 mM) containing 50 mM NaCl and 100 mM MgCl₂ was utilized to prepare all solutions. Fresh human serum samples were obtained from the infirmary of Shantou University. Millipore Milli-Q water (18 ΩM cm) supplied by a Millipore Milli-Q water purification system (Bedford, MA. USA) was used in all experiments. All chemicals used for investigations were of analytical grade purity.

The RLS spectra were measured on an F-7000 fluorescence spectrophotometer (Hitachi, Japan) equipped with a 1 cm × 1 cm quartz cuvette. Absorption spectra were recorded on a Lambda-950 UV-vis spectrophotometer (Perkin-Elmer, USA). All pH measurements were made with a PHS-3CA precision acidity meter (Dapu, China).

2.2 Measurement procedure

First, 898 μL of thrombin solution of various concentrations were respectively added to a 1.5 mL centrifugal tube containing 1 μL H-eTBA solution (10 μM) (denatured at 95 °C for 10 min, then cooled to room temperature) in Tris–HCl buffer solution (pH 7.40). After mixing for 25 min at 37 °C, 1 μL CTBA solution (10 μM) was added and the mixture was reacted for another hour at 37 °C. Then 100 μL MB solution (50 μM) was added quickly into the tube and mixed thoroughly at room temperature. The resulting solution was transferred to a 1 cm quartz cuvette for spectral recording without any incubation time. RLS spectra were obtained by scanning synchronously from 225 to 700 nm with Δλ = 0 nm. The excitation and emission slit widths were kept at 5 nm and 2.5 nm, respectively. The decreased RLS intensity of the reaction system was presented as ΔI_RLS = I_RLS⁰ − I_RLS, where I_RLS and I_RLS⁰ were the RLS intensities of the MB-DNA system with and without thrombin.

2.3 Validation test

Additionally, a validation test was proposed to verify that an MB molecule could induce the RLS signal change when it interacted with dsDNA. The validated steps were shown as follow: 1 μL H-eTBA (10 μM) and 1 μL CTBA solution with different concentrations were mixed with 898 μL buffer solution thoroughly in a 1.5 mL centrifugal tube. After incubating at 37 °C for an hour and subsequent cooling down to room temperature, 100 μL of MB solution (50 μM) was added to the tube. Finally, this resulting solution was transferred to a 1 cm micro quartz cuvette for UV-vis spectra recording.

3. Results and discussion

3.1 Characteristics and comparison of RLS spectra

Comparison experiments were performed to identify and characterize the RLS spectra as indicated in Fig. 1. It can be seen that the spectrogram maintained nearly the same shape from curve a to h. The light scattering peak at 370 nm could be regarded as the RLS peak of the reaction system. As shown in Fig. 1, the RLS intensity of MB + H-eTBA + CTBA (curve a) was strong. On the contrary, the RLS intensities of MB + H-eTBA, MB + CTBA, MB + thrombin, thrombin, MB and H-eTBA + CTBA (inset: curve c to h) were rather weak in the whole scanning wavelength range, which could be ignored in this reaction system.

Fig. 1 The RLS intensity spectra of the MB–DNA reaction systems with and without thrombin. (a) MB + H-eTBA + CTBA; (b) MB + H-eTBA + CTBA + thrombin (12.28 nM); (c) MB + H-eTBA; (d) MB + CTBA; (e) MB + thrombin; (f) thrombin; (g) MB; (h) H-eTBA + CTBA. Conditions: MB: 50 μM; H-eTBA: 1 μL, 10 μM; CTBA: 1 μL, 10 μM; pH = 7.40.

However, when thrombin was introduced (12.28 nM, curve b), a decreased RLS intensity was clearly observed. Based on RLS theory,^13,25 light scattering is caused by the presence of particles with diameter less than 1/20 of the incident light wavelength and the RLS intensity is proportional to the number of particles. As above, when thrombin was added to the system, the amount of MB–dsDNA complexes decreased, which could be easily recognized from curve b in Fig. 1. Moreover, from the inset in Fig. 1, it was noted that the RLS intensity of curve c was a little stronger than curve d. This phenomenon demonstrated that MB had reacted with dsDNA to some extent due to the loop-stem structure of H-eTBA.

3.2 Sensing mechanism

A schematic representation of the mechanism of thrombin sensing is illustrated in Scheme 1. Based on the RLS theory mentioned above, the reaction system showed two quite different RLS intensity results when the target thrombin was present or not.

Scheme 1 Schematic illustration of the RLS aptasensor for selective detection of human thrombin.

At the start H-eTBA was present in loop-stem structure (hairpin). When there was no target thrombin, the complementary sequence towards H-eTBA (CTBA) could be highly complementary to H-eTBA. Therefore, when MB was added this could interact with a large number of dsDNA. Thus large-size MB–dsDNA complexes were formed in the reaction system, which could induce a strong RLS signal.

Alternatively, while thrombin was added, H-eTBA could bind with thrombin and thus reduce the RLS intensity. With the addition of thrombin, H-eTBA could react with thrombin through its thrombin aptamer sequence embedded inside the hairpin DNA. Next, CTBA towards H-eTBA was introduced. Thus, there would be less dsDNA in the reaction system because the target thrombin had interacted with some of the H-eTBA. As MB solution was added subsequently, a weak RLS signal was produced because MB interacted with a small number of dsDNA. By monitoring the change in the RLS intensity, the thrombin target could be detected with selectivity and speed.

We then studied the hypochromicity effect of the reaction system through UV-vis spectra. It would present a much more indicative view of the sensing mechanism. As shown in Fig. 2, the absorbance of MB–dsDNA at 666 nm decreased with the reduction of the concentration of CTBA. The hypochromicity demonstrated the fact that MB molecules could accumulate on the surface of the double helix structure, which was produced because of the combination of H-eTBA and CTBA, by electrostatic attractions.²⁶ Meanwhile, the maximum absorption wavelength did not change and this also strongly indicated that MB molecules combined to dsDNA externally.

Fig. 2 The hypochromicity effect of the reaction system. Conditions: MB: 50 μM; H-eTBA: 1 μL, 10 μM; CTBA: 1 μL, 0.1, 1, 2, 4, 8 and 10 μM; pH = 7.40.

3.3 Optimization of experimental conditions

To optimize the sensing conditions, the experimental concentration of H-eTBA and CTBA should be explored at the beginning. As shown in Fig. S1,† we tested the RLS value upon different concentrations of H-eTBA in the absence of CTBA. It showed that 10 μM H-eTBA was the optimized amount. In Fig. S2,† it shows that an RLS maximum was reached after 10 μM CTBA reacted with H-eTBA. So 1 μL, 10 μM were chosen for the detection of thrombin in this work. In addition, we also tested the incubation time presented in Fig. S3.† We found that the RLS intensity hardly changed within 30 min. In order to shorten the time under the experimental conditions, we considered it possible to present our work without any incubation time.

Certainly, the pH value played a very important role in the interaction between MB molecules and DNA. As shown in Fig. 3, the effect of pH was investigated by comparing the results at different pH conditions. In the range of 5.45–9.50, it was apparent that the RLS intensity reached a maximum at pH 7.40. Therefore, pH 7.40 was selected to be the optimal pH value for the detection system.

Fig. 3 The effect of pH value on the RLS intensity. Conditions: MB: 50 μM; H-eTBA: 1 μL, 10 μM; CTBA: 1 μL, 10 μM; thrombin: 4.91 nM; pH value: 5.45, 6.50, 7.40, 8.40 and 9.50. Error bars were the standard deviation of three repetitive measurements.

The effect of MB concentration was tested by carrying through the H-eTBA (1 μL, 10 μM) and CTBA (1 μL, 10 μM) and thrombin (4.91 nM) at pH 7.40. In Fig. 4, the experimental results indicated that the RLS intensity reached a maximum when the MB concentration was 50 μM. It was obvious that the increasing concentration of MB would result in an increase of the RLS signal because more MB molecules interacted with dsDNA. However, the RLS intensity would decrease along with the increasing concentration of MB solution. The likely reason for this was that MB molecules would aggregate into dimers and thus it was not conducive to the combination of MB with dsDNA. As a result, the RLS intensity would decrease clearly. So 50 μM was chosen to be the optimal concentration of MB in this work.

Fig. 4 The effect of the concentration of MB solution on the RLS intensity. Conditions: H-eTBA: 1 μL, 10 μM; CTBA: 1 μL, 10 μM; thrombin: 4.91 nM; MB: 10, 30, 50, 70 and 90 μM; pH = 7.40. Error bars were the standard deviation of three repetitive measurements.

3.4 Quantitative detection of thrombin

On the basis of the above standard procedures and optimal conditions, various concentrations of thrombin were introduced to evaluate the performance of the sensor. As shown in Fig. 5, a higher concentration of thrombin resulted in decreased RLS intensity (curve a to f). This result was in accordance with the inference as above. Thus, the RLS peak of the reaction system at 370 nm could be used to give a quantitative detection of thrombin.

Fig. 5 The RLS spectra of the reaction system upon the addition of human thrombin at different concentrations. a–f: 0, 4.91, 9.82, 12.28, 14.73 and 17.18 nM. Inset: the RLS peak absorbance change is linear with the thrombin concentration in the range from 0 to 17.18 nM. Conditions: MB: 50 μM; H-eTBA: 1 μL, 10 μM; CTBA: 1 μL, 10 μM; pH 7.40. Error bars were the standard deviation of three repetitive measurements.

Furthermore, the proposed sensing strategy had a linear relationship in the range from 0 to 17.18 nM. The equation for the resulting calibration plot was ΔI_RLS = 261.96C_thrombin + 65.38 (R² = 0.9951). The detection limit was 0.32 nM, which was estimated on the concentration corresponding to the mean blank value plus 3 times the standard deviation of the blank value.²⁷

3.5 Selectivity, stability and reproducibility of the assay

In our proposal, it is crucial for H-eTBA to recognize the other variable form of DNA strands because the system is quantified by the amount of dsDNA. Herein, we used a single-base mutation DNA sequence (Mis-DNA) or non-specific aptamer sequence (non-specific) for comparison. The ability to sensitively discriminate that two kinds of DNA sequence was crucial for diagnosis at early stage. With the purpose of testing the selectivity of the loop-stem aptamer sensor, further investigations were performed using the sequences as mentioned above.

As presented in Fig. 6, in a fixed concentration of thrombin at 4.91 nM, two comparison experiments were performed. In the first comparison, a single-base mismatched strand was adopted to replace CTBA. In the second testing, a total non-thrombin specific DNA sensor (non-specific) was used to substitute H-eTBA. Simultaneously, a new complementary strand towards non-specific was utilized to replace CTBA. The procedure of these two independent experiments was the same as the process of detecting thrombin as above.

Fig. 6 Comparison of the aptasnesor selectivity using single-base mismatched strand (Mis-DNA) and non-specific aptamer sequence (non-specific). Conditions: thrombin: 4.91 nM; H-eTBA: 1 μL, 10 μM; CTBA: 1 μL, 10 μM; Mis-DNA: 1 μL, 10 μM; both non-specific DNA and its complementary strand: 1 μL, 10 μM; MB: 50 μM; pH 7.40. Error bars were the standard deviation of three repetitive measurements.

It was clear that when H-eTBA was utilized, the RLS intensity was much stronger than the signal of Mis-DNA. The weaker RLS signal of Mis-DNA implied that the hairpin sensor could discriminate the mutation sequence effectively. The Mis-DNA could not interact with H-eTBA thoroughly and a much weaker RLS signal was shown because MB reacted with fewer dsDNA strands (Fig. 6). This result demonstrated that even with the presence of a single base mutation complementary strand, H-eTBA could still recognize the difference because of its hairpin structure.

Additionally, shown in Fig. 6, the RLS intensity of non-specific sequence was similar to the signal of H-eTBA and even a little stronger than it. Because there was no thrombin aptamer embedded inside the non-specific hairpin strand, non-specific strand and its own complementary could highly combine with each other, which induced a quite strong RLS signal after MB solution was added.

To find a better illustration of the selectivity, we then used BSA, trypsin and α-chymotrypsin (all at 27.25 nM, except for thrombin at 17.18 nM) for interference measuring. As indicated in Fig. S4,† significant change in RLS intensity was only observed for the target thrombin and not for other nontargeted proteins. The compared result indicated that our featured loop-stem aptamer structure could provide a good selectivity for thrombin detection.

The stability of the presented sensor was examined by detecting the RLS response with time variation. We studied the RLS signal of the reaction system after the solution was stored at 4 °C for 5 and 10 days. We found that the RLS response retained 92.4% of the initial RLS signal for 4.91 nM thrombin, demonstrating the good stability. Additionally, the reproducibility of this sensor was studied by analysis of the same concentration of thrombin (4.91 nM) using 3 sensors under the same experimental conditions. Close RLS intensity was obtained with a relative standard deviation (RSD) of 5.3%. The results revealed that our proposal had an acceptable reproducibility.

3.6 Practical application

The performance of the thrombin probe in human serum sample was further investigated. Human serum samples were diluted 3-fold with Tris–HCl buffer solution at the beginning. As given in Table 2, satisfactory recoveries could be achieved using our method. The recovery of those measurements were in the range from 95.6% to 101.6% under the optimal conditions, indicating that the designed thrombin sensor still worked well and gave a more reliable result in real human serum sample applications.

Table 2 Determination results of thrombin in human serum^a

Sample number	Thrombin added (nM)	Thrombin found (nM)	Recovery (%)	RSD (%, n = 3)
a Human serum samples were diluted 3-fold with buffer solution (pH 7.40) before detection. Each data was given as the average value obtained from three independent experiments.
1	4.08	3.90	95.6	3.8
2	6.81	6.69	98.2	5.2
3	9.54	9.69	101.6	4.2
4	12.26	12.35	100.7	7.3
5	14.98	14.80	98.8	6.1

---

Different methods for detecting thrombin were then compared and the results are presented in Table 3. It can be concluded that with the measurements of SERS (Surface-enhanced Raman Scattering), electrochemical, colorimetric or fluorescence methods, each one had either a narrow linear range or a high detection limit.

Table 3 Comparison of different methods for the detection of thrombin

Method	Sensor	Linear range (nM)	Detection limit (nM)	Ref.
SERS	Au/SH-TBA/TB/TBA/AuNPs/Raman Probe/AgNPs	Not given	0.5	28
SERS	Si substrase/AuNPs/SH-TBA(-FITC)/TB	1 × 10^−4–0.01	0.02	29
Electrochemical	Magnetic bat/eletrode/AuMNP/SH-Apt1/TB/SH-Apt2-CS-AuNPs-HRP	1.0 × 10^−5–0.01	5.4 × 10^−6	30
Electrochemical	Ferrocene–graphene nanosheets/Ru(bpy)3^2+	0.5–25	0.21	31
Electrochemical	Enzyme-free and non-label fluorescent G-quadruplex DNA	0.01–1	0.005	32
Photoelectrochemical	Cationic macromolecules PEI and MPA modified G-CdS nanocomposites	0.002–0.6	0.001	33
Colorimetric	Nanorose/SH-Apt15/TB/SH-Apt29/nanorose	1.6–30.4	1.0	34
Fluorescence	GO-FAM–peptide complex	2.2–10	2.0	35
RLS	Loop-stem thrombin aptamer DNA sequence	0–17.18	0.32	This assay

---

Compared with reported methods, our method presented a comparable linear range and detection limit, which were presented in Table 3. Additionally, our method was simple because it did not require any complicated experimental procedures and conditions.

4. Conclusion

A sensing strategy using well-designed loop-stem structured DNA sequence was presented to demonstrate the feasibility for simple detection of thrombin. We tactfully combined the loop-stem structure with methylene blue and RLS detecting technique. Because of the thrombin aptamer embedded inside the hairpin DNA and the interaction between methylene blue molecule and dsDNA, the target thrombin could be detected effectively using RLS technique. In addition, the limit of detection was as low as 0.32 nM, which was practical for the detection of thrombin. Moreover, complicated and tedious experimental procedures were unnecessary for thrombin detection by this sensing strategy. Finally, because of its favorable selectivity and practicability, the authors believe that our sensing proposal could give a new promising method for clinical application of detecting thrombin.

Acknowledgements

We are grateful for the financial support from the Natural Science Foundation of Guangdong Province (no. S2011010005208 & no. 2014A030313480), the Science & Technology Project of Guangdong Province (no. 2013B030600001) and the Guangdong High Education Fund of Science and Technology Innovation (no. 2013KJCX0078).

References

1. Q. Zhao and J. Gao, Chem. Commun., 2013, 49, 7720–7722.
2. L. G. Licari and J. P. Kovacic, J. Vet. Emerg. Crit. Care, 2009, 19, 11–22.
3. Y. Kitamoto, E. Nakamura, S. Kudo, H. Tokunaga, E. Murakami, K. Noguchi and T. Imamura, Clin. Chim. Acta, 2008, 398, 159–160.
4. R. Y. Lai, K. W. Plaxco and A. J. Heeger, Anal. Chem., 2007, 79, 229–233.
5. G. S. Bang, S. Y. Cho, N. Lee, B. R. Lee, J. H. Kim and B. G. Kim, Biosens. Bioelectron., 2013, 39, 44–50.
6. D. Li, S. Song and C. Fan, Acc. Chem. Res., 2010, 43, 631–641.
7. Z. L. Jiang, Y. Zhang, A. H. Liang, C. Q. Chen, J. N. Tian and T. S. Li, Plasmonics, 2012, 7, 185–190.
8. J. Zhang, L. H. Wang, D. Pan, S. P. Song, F. Y. C. Boey, H. Zhang and C. H. Fan, Small, 2008, 4, 1196–1200.
9. J. Wang, Y. Cao, G. F. Chen and G. X. Li, Biosens. Bioelectron., 2009, 10, 2171–2176.
10. D. Li, Y. Yan, A. Wieckowska and I. Willner, Chem. Commun., 2007, 34, 3544–3546.
11. E. Farjami, L. Clima, K. V. Gothelf and E. E. Ferapontova, Analyst, 2010, 135, 1443–1448.
12. Z. G. Chen, S. H. Qian, X. Chen, J. H. Chen, Y. J. Lin and J. B. Liu, RSC Adv., 2012, 2, 2562–2567.
13. C. Z. Huang, K. A. Li and S. Y. Tong, Anal. Chem., 1997, 69, 514–520.
14. Z. G. Chen, T. H. Song, Y. R. Peng, X. Chen, J. H. Chen, G. M. Zhang and S. H. Qian, Analyst, 2011, 136, 3927–3933.
15. Z. G. Chen, T. H. Song, S. B. Wang, X. Chen, J. H. Chen and Y. Q. Li, Biosens. Bioelectron., 2010, 25, 1947–1952.
16. Z. G. Chen, Y. R. Peng, M. H. Chen, X. Chen, J. H. Chen and G. M. Zhang, Analyst, 2010, 135, 2653–2660.
17. J. Zheng, G. F. Cheng, P. G. He and Y. Z. Fang, Talanta, 2010, 80, 1868–1872.
18. H. Chang, L. Tang, Y. Wang, J. Jiang and J. Li, Anal. Chem., 2010, 82, 2341–2346.
19. D. W. Huang, C. G. Niu, P. Z. Qin, M. Ruan and G. M. Zeng, Talanta, 2010, 83, 185–189.
20. G. S. Bang, S. Y. Cho and B. G. Kim, Biosens. Bioelectron., 2005, 21, 863–870.
21. P. He, L. Shen, Y. Cao and D. Li, Anal. Chem., 2007, 79, 8024–8029.
22. H. Cho, B. R. Baker, S. W. Hogiu, C. V. Pagba, T. A. Laurence, S. M. Lane, L. P. Lee and J. B. H. Tok, Nano Lett., 2008, 8, 4386–4390.
23. V. Ostatna, H. Vaisocherova, J. Homola and T. Hianik, Anal. Bioanal. Chem., 2008, 391, 1861–1869.
24. Y. Peng, L. D. Li, X. J. Mu and L. Guo, Sens. Actuators, B, 2013, 177, 818–825.
25. J. R. W. Master, R. Thomas, A. G. Hall, L. Hogarth, E. C. Matheson, A. R. Cattan and H. Lohrer, Eur. J. Cancer, 1996, 32, 1248–1253.
26. E. C. Long and J. K. Barton, Acc. Chem. Res., 1990, 23, 271–273.
27. G. Sener, L. Uzun and A. Denizli, Anal. Chem., 2014, 86, 514–520.
28. M. A. Shuman and P. W. Majerus, J. Clin. Invest., 1976, 58, 1249–1258.
29. J. Hu, P. C. Zheng, J. H. Jiang, G. L. Shen, R. Q. Yu and G. K. Liu, Anal. Chem., 2009, 81, 87–93.
30. C. W. Chi, Y. H. Lao, Y. S. Li and L. C. Chen, Biosens. Bioelectron., 2011, 26, 3346–3352.
31. B. R. Zhuo, Y. Q. Li, X. Huang, Y. J. Lin, Y. W. Chen and W. H. Gao, Biosens. Bioelectron., 2015, 208, 518–524.
32. Y. Y. Xu, W. J. Zhou, M. Zhou, Y. Xiang, R. Yuan and Y. Q. Chai, Biosens. Bioelectron., 2015, 64, 306–310.
33. L. Shangguan, W. Zhu, Y. C. Xue and S. Q. Liu, Biosens. Bioelectron., 2015, 64, 611–617.
34. G. H. Liang, S. Y. Cai, P. Zhang, Y. Y. Peng, H. Chen, S. Zhang and J. L. Kong, Anal. Chim. Acta, 2011, 689, 243–249.
35. M. Zhang, B. C. Yin, X. F. Wang and B. C. Ye, Chem. Commun., 2011, 47, 2399–2401.

---

Footnotes

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c4ra14729d

‡ Both the authors contributed equally to the paper.

---