Quantitative serine protease assays based on formation of copper(II)–oligopeptide complexes

Abstract

A quantitative protease assay based on the formation of a copper–oligopeptide complex is developed. In this assay, when a tripeptide GGH fragment is cleaved from an oligopeptide chain by serine proteases, the tripeptide quickly forms a pink GGH/Cu²⁺ complex whose concentration can be determined quantitatively by using UV-Vis spectroscopy. Therefore, activities of serine proteases can be determined from the formation rate of the GGH/Cu²⁺ complex. This principle can be used to detect the presence of serine protease in a real-time manner, or measure proteolytic activities of serine protease cleaving different oligopeptide substrates. For example, by using this assay, we demonstrate that trypsin, a model serine protease, is able to cleave two oligopeptides GGGGKGGH (P₁) and GGGGRGGH (P₂). However, the specificity constant (k_cat/K_m) for P₂ is higher than that of P₁ (6.4 × 10³ mM⁻¹ min⁻¹vs. 1.3 × 10³ mM⁻¹ min⁻¹). This result shows that trypsin is more specific toward arginine (R) than lysine (K) in the oligopeptide sequence.

---

Introduction

Proteases are enzymes which can hydrolyze peptide bonds in proteins or peptides. They are very important in many biological processes such as food digestion, blood coagulation, and cell apoptosis.^1–4 Past studies have shown that dysfunction of proteases may cause cardiovascular disease,^5,6 metastatic cancer^7–10 or Alzheimer disease.¹¹ Despite their importance, it is often difficult to detect protease activity in a real-time and quantitative manner. This is because cleavage of proteins or peptides only leads to shorter oligopeptide fragments, and these fragments can only be differentiated based on their molecular weight or mobility. In the past, HPLC,¹² gel electrophoresis^13,14 or mass spectrometry^15–18 was often employed in protease assays to detect oligopeptide fragments. However, these assays are only suitable for measurements of average protease activity because sampling of reaction mixtures cannot be done in real time.

Alternatively, real-time protease assays can be achieved by using oligopeptide substrates labeled with a pair of fluorophore and quencher.¹⁹ After the cleavage of a labeled oligopeptide molecule, the fluorophore/quencher pair is separated from each other and the fluorescence intensity increases. However, this method requires labeling of the oligopeptide substrates, and the activity of the protease may be affected by the fluorophore or quencher. To avoid labeling of oligopeptide substrates, Li et al. developed a label-free fluorescent protease assay by using hemoglobin as a protease substrate. After protease cleavage of hemoglobin, a heme molecule is released to quench the fluorescence, and the protease activity can be correlated with the fluorescence intensity. Recently, gold nanoparticles (AuNPs) and their aggregates are also exploited for the development of protease assays.^20–23 For example, Guarise et al.²⁰ developed a serine protease assay by using cysteine-modified oligopeptides which cross-linked AuNPs. Once the oligopeptides were pretreated with a protease solution, the oligopeptides were cleaved into fragments. As a result, oligopeptides could not cross-link AuNPs and no color change was observed. However, this approach requires a two-stage exposure of oligopeptides to proteases and AuNPs separately. To address this problem, Laromaine et al.²¹ modified the oligopeptide with N-(fluorenyl-9-methoxycarbonyl) (Fmoc) to promote aggregation of AuNPs through π–π interactions. After the addition of a protease solution, Fmoc groups were removed and the aggregates of AuNPs broke apart due to electrostatic repulsion. However, it is difficult to quantify the degree of aggregation of AuNPs. Besides, AuNPs are susceptible to matrix proteins (such as human serum protein) which can induce or prevent AuNPs from forming aggregates.

In the past, binding of oligopeptides to metal ions have been exploited for the detection of metal ions because the binding of oligopeptides to metal ions is very specific. For example, Wu and Liu²⁴ developed a fluorescent chemosensor by using a fluorescent-labeled tetrapeptide Fluor-HGGG to detect Cu²⁺ in aqueous solutions. However, to the best of our knowledge, the formation of an oligopeptide–metal ion complex has not been used in protease assays before. In this study, we used Cu²⁺ and a GGH-containing oligopeptide to develop a serine protease assay for measuring protease activity in real time. The detection principle is based on the formation of a pink GGH/Cu²⁺ complex during the cleavage of the oligopeptide. When the tripeptide GGH is cleaved from the oligopeptide substrate, it forms a pink complex with Cu²⁺ in the solution immediately. This metal complex absorbs visible light at 525 nm and its concentration can be determined quantitatively. Thus, the protease activity can be calculated by using the formation rate of the GGH/Cu²⁺ complex.

Experimental section

Materials

Oligopeptides GGGGKGGH (P₁), GGGGRGGH (P₂), GGGGFGGH (P₃), GGGGWGGH (P₄), and RKRIRRMMPRPS (P₅) (purity >95%) were synthesized by GenicBio (China). Tripeptide glycyl-glycyl-histidine (GGH) was purchased from Bachem (Switzerland). Trypsin (from bovine pancreas, type II, activity, 14 408 U mg⁻¹), chymotrypsin (from bovine pancreas, type II, activity: 40 U mg⁻¹), human serum (from human male AB plasma, U.S.A. origin), sodium carbonate anhydrous (99.5%), sodium sulfide nonahydrate (98%), benzamidine hydrochloride hydrate (99%), and copper(II) sulfate anhydrous (99.9%) were purchased from Sigma-Aldrich (U.S.A). Sodium chloride (99.8%) was purchased from GCE Laboratory Chemicals (Singapore). Calcium chloride dihydrate (99%) was purchased from Merck (Singapore). Tris buffer (1 M, pH 8.0) was purchased from 1^st Base (Singapore) and used as received. Ultrapure water was obtained by using a water purification system from Millipore (U.S.A).

Protease assay

Trypsin was added to 400 μL of Tris buffer (50 mM, pH 8.0) containing 1.0 mM of copper(II) sulfate and 1.0 mM of P₁. The final concentrations of trypsin were between 0 and 400 nM. These solutions were incubated at 37 °C in a thermo-mixer (Eppendorf, Germany). After 2 h, visible spectra of the solutions were recorded by using a UV-Vis spectrometer (Cary 50, Varian) equipped with a temperature controlling system. To measure protease activities, buffer solutions containing 160 nM of trypsin (or 4 μM of chymotrypsin), 1.0 mM of copper(II) sulfate, and different concentrations of oligopeptides were prepared. Subsequently, visible absorption spectra were recorded every 5 min at 37 °C. The reaction rate (v) was determined by using the slope of the GGH concentration versus the reaction time. The Michaelis constant (K_m) and maximum reaction rate (v_max) were determined by using a Lineweaver–Burk plot, and the turnover number (k_cat) was determined by using k_cat = v_max/[E]₀, where [E]₀ is the protease concentration. For protease assay in serum solution, 1 mL of human serum was added into 20 mL of Tris buffer (50 mM, pH 8.0) as a serum solution. Then, 400 μL of the serum solution was spiked with 1.0 mM of copper(II) sulfate, 1.0 mM of P₁ and different concentrations of trypsin (0, 200, 400, and 600 nM). After incubation at 37 °C for 2 h, images were taken by using a digital camera.

Inhibition assay

Tris buffer (50 mM, pH 8.0) containing 480 nM of trypsin was mixed with different amounts of benzamidine hydrochloride (BH). After incubation at 4 °C for 30 min, 120 μL of the above solution was mixed with 120 μL of copper(II) sulfate solution (3.0 mM) and 120 μL of P₁ solution (3.0 mM). The final concentrations of copper(II) sulfate, P₁, and trypsin were 1.0 mM, 1.0 mM, and 160 nM, respectively. The mixed solution was incubated at 37 °C for 2 h. The inhibition efficiency (IE) of BH was determined by using the following equation:where C_1,t and C_2,t are final concentrations of the uncleaved oligopeptides with and without the inhibitor, respectively. C₀ is the initial concentration of oligopeptide P₁. C_1,t and C_2,t can be obtained from the GGH concentration since the molar ratio of cleaved oligopeptide and GGH is 1 : 1.

Results and discussion

Principle of the protease assay

Serine proteases such as trypsin and chymotrypsin were used as model proteases because their activities are not inhibited by Cu²⁺.²⁵ To detect the activity of trypsin, we designed an oligopeptide substrate GGGGKGGH (P₁), which consists of a potential trypsin cleavage site at lysine (K) and a terminal tripeptide residue of GGH. In Tris buffer (pH 8.0), P₁ is able to form a weak complex with Cu²⁺ through imidazole nitrogen and peptide bonds,²⁶ as is evident by a strong absorption peak around 565 nm (Fig. 1A). In this case, the color of the solution is blue. Subsequently, trypsin was added to cleave the oligopeptide. According to the literature,²⁷ this protease is able to cleave oligopeptides at the carboxyl terminal of lysine (K) or arginine (R). When 240 nM of trypsin was added to the solution, the color of the solution gradually turned to pink over 2 h. The color change was accompanied by a blue shift in the absorption spectra from 565 nm to 525 nm (Fig. 1A). The pink color indicates that when trypsin cleaves P₁ at the carboxyl terminal of K, a tripeptide GGH is released. This tripeptide is able to form a highly stable complex with Cu²⁺ (pK_a = 8.05) through a free amino group, an imidazole nitrogen and two deprotonated nitrogen atoms (Fig. 1B).²⁸ Our titration experiment also shows that GGH/Cu²⁺ is more stable than the P₁/Cu²⁺ complex (see ESI†).

Fig. 1 (A) UV-Vis spectra of Tris buffer containing 1.0 mM of Cu²⁺ and 1.0 mM of P₁ without trypsin (solid line) or with trypsin (dash dot line). The inset shows colors of both solutions. (B) Principle of the protease assay. The arrow indicates the cleavage site of the oligopeptide substrate by a protease (X = K, R, F, and W).

Fig. 2A shows effects of trypsin concentration on the colors of the reaction mixtures after 2 h of reaction. Only when the trypsin concentration was higher than 160 nM, the color of the solution changed from blue to pink after 2 h. Fig. 2B shows normalized spectra (against a spectrum without trypsin) of the reaction mixtures. With an increasing trypsin concentration, the absorbance peak at 525 nm gradually increased and the absorbance peak at 624 nm decreased. However, this trend stopped at 240 nM probably because the oligopeptide substrates had been cleaved completely by 240 nM of trypsin. Fig. 2C shows a linear relationship between the normalized absorbance ΔA (A₅₂₅/A₆₂₄ − A⁰₅₂₅/A⁰₆₂₄) and trypsin concentration. In our experiments, A₅₂₅/A₆₂₄ was measured in the presence of trypsin, and A⁰₅₂₅/A⁰₆₂₄ was measured in the absence of trypsin. From Fig. 2C, the limit of detection (LOD) was determined to be 16.7 nM (LOD = 3.3 SD/sensitivity, where SD is the standard deviation of 10 measurements of a blank sample). In Table 1, we compared LODs of several protease assays. The LOD of our protease assay is still higher compared to other assays due to a relatively small extinction coefficient of GGH/Cu²⁺. Further research is needed to lower the LOD of the protease assay reported herein.

Fig. 2 Trypsin assays conducted at different trypsin concentrations. (A) Colors of the reaction mixtures 2 h after the addition of trypsin. (B) Normalized visible spectra of the reaction mixtures (against the spectrum without trypsin). (C) Linear plot of normalized absorbance ΔA (A₅₂₅/A₆₂₄ − A⁰₅₂₅/A⁰₆₂₄) after addition of different concentrations of trypsin.

Table 1 LODs and target proteases of serine protease assays reported in the literature

No.	Method	Target protease	LOD	Ref.
1	AuNPs	Trypsin	5.0 nM	Chen et al.^29
2	AuNPs	Thrombin	5.0 nM	Guarise et al.^20
3	AuNPs	Trypsin	66.7 pM	Xue et al.^30
4	Fluorescent polymer	Trypsin	1.7 nM	Wang et al.^31
5	GGH/Cu^2+	Trypsin	16.7 nM	This work

---

Specificity of the protease assay

To test the specificity of the protease assay, another four oligopeptides, GGGGRGGH (P₂), GGGGFGGH (P₃), GGGGWGGH (P₄), and RKRIRRMMPRPS (P₅) were synthesized. Among these oligopeptides, P₂ has a potential trypsin cleavage site R, while P₃ and P₄ have potential chymotrypsin cleavage sites phenylalanine (F) and tryptophan (W). P₅ is a scrambled oligopeptide with trypsin cleavage sites R and K, but it does not contain any GGH in its sequence. Then, 1.0 mM of oligopeptides P₁ to P₅ were mixed with 1.0 mM of copper(II) sulfate and 160 nM of trypsin (or 4 μM of chymotrypsin). Higher chymotrypsin was used because the activity of chymotrypsin (40 U mg⁻¹) is much smaller than that of trypsin (14 408 U mg⁻¹). Fig. 3 shows that only the solutions of P₁ and P₂ changed from blue to pink in the presence of trypsin. This is because P₁ and P₂ can be cleaved at the carboxyl terminal of K or R by trypsin. As a result, the GGH fragment was released from the oligopeptide and formed a color complex with Cu²⁺. Similarly, in the presence of chymotrypsin, only the solutions of P₃ and P₄ changed color, suggesting that only P₃ and P₄ can be cleaved. When a scrambled oligopeptide P₅ was used, the color of the solution did not change due to the lack of GGH to form a complex with Cu²⁺. Collectively, no cross-response was observed for trypsin and chymotrypsin, suggesting that this assay has good specificity.

Fig. 3 Specificity test of the protease assay with 5 different oligopeptide substrates.

Kinetics of protease reaction

To determine reaction rates of serine proteases, we monitored normalized absorbance (A₅₂₅/A₆₂₄) of different reaction mixtures with different oligopeptide concentrations. Subsequently, the time-dependent GGH concentration was calculated from the normalized absorbance (A₅₂₅/A₆₂₄) and a calibration curve (see ESI†). Fig. 4A shows that when the concentration of P₁ was between 0.33 and 1.0 mM, the concentration of GGH increased linearly in 60 min. From Fig. 4A, the reaction rate of trypsin was determined, and the Lineweaver–Burk plot (Fig. 4B) and rate constants (Table 2) were also obtained. It can be seen that Michaelis constant (K_m) for P₁ is 0.24 mM, and the maximum reaction rate (v_max) is 49.6 μM min⁻¹. This value is close to the protease activity provided by the manufacturer (57.6 μM min⁻¹ for a substrate p-toluenesulfonyl-L-arginine methyl ester). This consistency shows that the Cu²⁺-based protease assay is reliable.

Fig. 4 Reaction kinetics of trypsin. (A and C) are time-dependent concentration of GGH released during cleavage of P₁ and P₂ by trypsin. (B and D) are Lineweaver–Burk plots of trypsin with different substrate concentrations of P₁ and P₂. Concentrations of oligopeptides (P₁ and P₂) are 1.0 mM (squares), 0.67 mM (circles), 0.5 mM (up triangles), 0.4 mM (down triangles), and 0.33 mM (diamonds), respectively. The concentration of trypsin was fixed at 160 nM and the concentration of Cu²⁺ was 1.0 mM for all experiments. The error bar represents a standard deviation of three parallel experiments.

Table 2 Michaelis constants and kinetic parameters for trypsin and chymotrypsin with different oligopeptide substrates

Subs.^a	v max (μM min^−1)	K m (mM)	k cat (min^−1)	k cat/Km (mM^−1 min^−1)
a The oligopeptide substrates of P1 and P2 were cleaved by 160 nM of trypsin, and the substrates of P3 and P4 were cleaved by 4 μM of chymotrypsin, respectively.
P1	49.6	0.24	3.1 × 10^2	1.3 × 10^3
P2	93.5	0.09	5.8 × 10^2	6.4 × 10^3
P3	31.7	0.07	4.0	57
P4	125.2	0.12	16	133

---

Similarly, Fig. 4C shows the reaction rate of trypsin when P₂ was used as a substrate. In Fig. 4C, the concentration of GGH increased linearly in the first 25 min before reaching a plateau. This is because the reaction rate of trypsin is faster when P₂ was used as a substrate. Subsequently, we used the initial reaction rates (between 0 and 25 min) to obtain a Lineweaver–Burk plot (Fig. 4D) and kinetic constants (Table 2). When P₂ was used as a substrate, K_m is 0.09 mM, indicating that P₂ has a higher binding affinity for trypsin. We also notice that when P₂ was used as a substrate, the turn-over number (k_cat) of trypsin is 187% higher than that when P₁ was used. The specificity constant (k_cat/K_m) of trypsin to P₂ is about 4.9-fold higher than P₁, showing that trypsin is more favorable to hydrolyze the R residue rather than K. This is consistent with a previous report showing that trypsin cleaves arginine-vasopressin several times faster than lysine-vasopressin.³²

The reaction rates and Lineweaver–Burk plots of chymotrypsin are shown in Fig. 5, and the kinetic data are shown in Table 2. For chymotrypsin, when P₄ was used as a substrate, the k_cat of chymotrypsin is 4-fold higher than that when P₃ was used. This is probably because chymotrypsin prefers the W residue than F during the hydrolysis of P₃ and P₄. Although the binding affinity (1/K_m) of chymotrypsin for P₄ is 1.7-fold smaller than P₃, k_cat/K_m of chymotrypsin for P₄ is 2.3-fold higher than that for P₃. This is slightly different from the previous report in which k_cat/K_m of chymotrypsin for W is 1.2-fold higher than that for F.³³ More experiments will be needed to verify the discrepancy.

Fig. 5 Reaction kinetics for chymotrypsin. (A and C) are time-dependent concentration of GGH released during cleavage of P₃ and P₄ by chymotrypsin. (B and D) are Lineweaver–Burk plots of chymotrypsin with different substrate concentrations of P₃ and P₄. The concentration of oligopeptides (P₃ and P₄) are 1.0 mM (squares), 0.67 mM (circles), 0.5 mM (up triangles), 0.4 mM (down triangles), and 0.33 mM (diamonds), respectively. The concentration of chymotrypsin was fixed at 4 μM, and the concentration of Cu²⁺ was 1.0 mM for all experiments. The error bar represents a standard deviation of three parallel experiments.

Protease assays in serum solution

Detection of protease activity in complex media is challenging. To understand the performance of this protease assay in complex media, serum solutions containing 1.0 mM of copper(II) sulfate, 1.0 mM of P₁, and different concentrations of trypsin (0, 200, 400, and 600 nM) were prepared. Fig. 6 shows that the solution color still changed from blue to pink when the trypsin concentration was 400 nM or higher, indicating that P₁ was cleaved by trypsin even in the serum solution. To rule out the possibility that the color change was caused by the digestion of serum proteins, a control experiment without P₁ was conducted. Fig. 6 shows that the color of the serum solution without P₁ did not change. These results, when combined, show that this protease assay still functions properly in serum solutions, but the LOD becomes higher. This is probably because some trypsin cleaved serum proteins instead of P₁ when the solution contained serum proteins.

Fig. 6 Trypsin assay performed in serum solutions with or without P₁.

Inhibition of protease

Finally, we evaluated inhibitory potency of a trypsin inhibitor benzamidine hydrochloride (BH) by using this protease assay. It has been demonstrated in the literature that BH is a potent trypsin inhibitor.³⁴Fig. 7A shows that when the BH concentration was 1.6 mM or higher, the color of the solution remained blue. This result shows that the activity of trypsin was inhibited by BH. Fig. 7B shows that the inhibition efficiency (IE) increases with the increasing concentration of BH. From Fig. 7B, the IC₅₀ value of benzamidine hydrochloride for trypsin is 0.4 mM, which is similar to the previously reported value.³⁴

Fig. 7 (A) Inhibition of trypsin activity by using different concentrations of benzamidine hydrochloride. (B) Inhibition efficiency as a function of benzamidine hydrochloride concentration.

Conclusions

A quantitative serine protease assay was successfully developed. Substrates used in this assay were synthetic oligopeptides with a cleavable residue X (X= K, R, F, and W) followed by another 3 residues GGH at their C-terminal ends. The tripeptide GGH can be released and formed a pink complex with Cu²⁺ after the oligopeptides were cleaved by trypsin or chymotrypsin. This assay shows good specificity for different oligopeptides and serine protease as no cross-response was observed. This assay offers many advantages including simplicity, free of labels, and possibility for naked-eye detection. However, the requirement for Cu²⁺ may limit its applications for proteases whose activities are inhibited by metal ions.

Acknowledgements

This work was supported by the National Research Foundation (NRF) in Singapore under grant (NRF 2009 NRF-CRP 001-039).

References

1. H. B. Wang, Q. Zhang, X. Chu, T. T. Chen, J. Ge and R. Q. Yu, Angew. Chem., Int. Ed., 2011, 50, 7065–7069.
2. L. R. Zhang, Q. X. Xu, D. Xing, C. J. Gao and H. W. Xiong, Plant Physiol., 2009, 150, 1773–1783.
3. Y. L. Pan, M. L. Guo, Z. Nie, Y. Huang, Y. Peng, A. F. Liu, M. Qing and S. Z. Yao, Chem. Commun., 2012, 48, 997–999.
4. C. J. Vickers, G. E. Gonzalez-Paez and D. W. Wolan, ACS Chem. Biol., 2013, 8, 1558–1566.
5. F. H. Messerli, Eur. Heart J., 2004, 25, 1475–1476.
6. A. Papazafiropoulou and N. Tentolouris, Hippokratia, 2009, 13, 76–82.
7. P. Chen, R. Selegard, D. Aili and B. Liedberg, Nanoscale, 2013, 5, 8973–8976.
8. T. J. Harris, G. von Maltzahn, A. M. Derfus, E. Ruoslahti and S. N. Bhatia, Angew. Chem., Int. Ed., 2006, 45, 3161–3165.
9. G. von Maltzahn, T. J. Harris, J. H. Park, D. H. Min, A. J. Schmidt, M. J. Sailor and S. N. Bhatia, J. Am. Chem. Soc., 2007, 129, 6064–6065.
10. W. Cheng, Y. L. Chen, F. Yan, L. Ding, S. J. Ding, H. X. Ju and Y. B. Yin, Chem. Commun., 2011, 47, 2877–2879.
11. S. Krizkova, O. Zitka, M. Masarik, V. Adam, M. Stiborova, T. Eckschlager, G. J. Chavis and R. Kizek, TrAC, Trends Anal. Chem., 2011, 30, 1819–1832.
12. C. Q. Zhang, L. Zheng, J. Nurnberg, B. M. Vacari, J. Z. Zhou and Y. Wang, Anal. Biochem., 2014, 445, 14–19.
13. W. T. Zhao, C. L. Yao, X. T. Luo, L. Lin and I. M. Hsing, Electrophoresis, 2012, 33, 1288–1291.
14. D. M. Krizek and M. E. Rick, Electrophoresis, 2001, 22, 946–949.
15. S. D. Lamore, S. X. Qiao, D. Horn and G. T. Wondrak, Photochem. Photobiol., 2010, 86, 1307–1317.
16. Y. P. Kim, B. S. Lee, E. Kim, I. S. Choi, D. W. Moon, T. G. Lee and H. S. Kim, Anal. Chem., 2008, 80, 5094–5102.
17. Y. R. Na, S. Y. Kim, J. T. Gaublomme, A. K. Shalek, M. Jorgolli, H. Park and E. G. Yang, Nano Lett., 2013, 13, 153–158.
18. O. Vosyka, K. R. Vinothkumar, E. V. Wolf, A. J. Brouwer, R. M. J. Liskamp and S. H. L. Verhelst, Proc. Natl. Acad. Sci. U. S. A., 2013, 110, 2472–2477.
19. T. Zauner, R. Berger-Hoffmann, K. Muller, R. Hoffmann and T. Zuchner, Anal. Chem., 2011, 83, 7356–7363.
20. C. Guarise, L. Pasquato, V. De Filippis and P. Scrimin, Proc. Natl. Acad. Sci. U. S. A., 2006, 103, 3978–3982.
21. A. Laromaine, L. L. Koh, M. Murugesan, R. V. Ulijn and M. M. Stevens, J. Am. Chem. Soc., 2007, 129, 4156–4157.
22. G. B. Kim, K. H. Kim, Y. H. Park, S. Ko and Y. P. Kim, Biosens. Bioelectron., 2013, 41, 833–839.
23. X. Ding, D. Ge and K.-L. Yang, Sens. Actuators, B, 2014, 201, 234–239.
24. S. P. Wu and S. R. Liu, Sens. Actuators, B, 2009, 141, 187–191.
25. H. Steinhart, R. Wieninger-Rustemeyer and M. Kirchgessner, Arch. Tierernaehr., 1981, 31, 119–125.
26. G. F. Bryce and F. R. N. Gurd, J. Biol. Chem., 1966, 241, 122–129.
27. M. J. Page and E. Di Cera, Cell. Mol. Life Sci., 2008, 65, 1220–1236.
28. P. Mlynarz, D. Valensin, K. Kociolek, J. Zabrocki, J. Olejnik and H. Kozlowski, New J. Chem., 2002, 26, 264–268.
29. G. C. Chen, Y. S. Xie, H. T. Zhang, P. Wang, H. Y. Cheung, M. S. Yang and H. Y. Sun, RSC Adv., 2014, 4, 6560–6563.
30. W. X. Xue, G. X. Zhang and D. Q. Zhang, Analyst, 2011, 136, 3136–3141.
31. Y. Y. Wang, Y. Zhang and B. Liu, Anal. Chem., 2010, 82, 8604–8610.
32. H. Keilova, Physiol. Chem. Phys., 1969, 1, 100.
33. P. Hudaky, G. Kaslik, I. Venekei and L. Graf, Eur. J. Biochem., 1999, 259, 528–533.
34. W. Y. Li, J. Chen, H. P. Jiao, Q. F. Zhang, H. P. Zhou and C. Yu, Chem. Commun., 2012, 48, 10123–10125.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c4an01731e

---