Application of a DNA-based luminescence switch-on method for the detection of mercury(II) ions in water samples from Hong Kong

Mercury is a highly toxic environmental contaminant that damages the endocrine and central nervous systems. In view of the contamination of Hong Kong territorial waters with anthropogenic pollutants such as trace heavy metals, we have investigated the application of our recently developed DNA-based luminescence methodology for the rapid and sensitive detection of mercury(II) ions in real water samples. The assay was applied to water samples from Shing Mun River, Nam Sang Wai and Lamma Island sea water, representing natural river, wetland and sea water media, respectively. The results showed that the system could function effectively in real water samples under conditions of low turbidity and low metal ion concentrations. However, high turbidity and high metal ion concentrations increased the background signal and reduced the performance of this assay.


Introduction
Mercury is listed as a priority pollutant by many international agencies due to its persistence, bioaccumulation and toxicity in the environment. Mercury is extensively used in agriculture and industry for the manufacture of pesticides, fungicides, electrical goods, paper, batteries and other items, resulting in the release of large amounts of mercury into the environment. Another major source of atmospheric emission of mercury Content from this work may be used under the terms of the Creative Commons Attribution-NonCommercial-ShareAlike 3.0 licence. Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI. 3 These authors contributed equally to this work.
is the combustion of coal in coal-fired power plants (Gibb et al 2000). Consequently, the biogeochemistry of mercury(II) ions (Hg 2+ ) in coastal and estuarial environments has received particular attention (Mason et al 1996, Hines et al 2000, Conaway et al 2003, Schaefer et al 2011. Bioaccumulation and trophic transfer processes in the environment lead to the concentration of mercury(II) in the tissues of high trophic level marine organisms, including edible marine species, posing risks to the local population.
Hong Kong contains a coal-fired station located on Lamma Island operated by the Hong Kong Electric Company (HEC), as well as several large industrial estates which are typically situated near important water systems such as Shing Mun River and Nam Sang Wai. In Hong Kong, mercury(II) was detected in fish at Nam Sang Wai (Kong et al 2005) and Shing Mun River (Zhou and Wong 2000). Feathers of two Ardeid species in Mai Po Marshes next to Nam Sang Wai wetland were measured to contain up to 35.5 µM dry weight of mercury by inductively-coupled plasma mass spectrometry (ICP-MS) (Connell et al 2001). Furthermore, Perna viridis mussels collected at Lamma Island in Hong Kong waters were found to contain 0.3-0.51 µM dry weight of mercury using ICP-MS (Liu and Kueh 2005).
As Hong Kong suffers from mercury pollution, the sensitive monitoring of environmental contaminants is essential to safeguard citizens' health. Mercury(II) can cause damage to the human brain, kidneys and lungs (Clifton 2007, Garrecht and, and can cause diseases such as acrodynia (Shandley and Austin 2011), Hunter-Russell syndrome (Kondo 2000) and Minamata disease (Davidson et al 2004). Therefore, the development of in-field detection methods for mercury(II) ions remains an important challenge (Davidson et al 2004, Holmes et al 2009. Traditional detection methods for mercury(II) ions include atomic absorption spectroscopy (AAS) and ICP-MS. Despite their widespread use in laboratories and industry, their requirement for off-site laboratory and costly instrumentation limits their practical use for the in-field testing of mercury(II) ions. The target-induced structural switching of DNA (Lacroix et al 2011, Tran et al 2010, Mergny 2012 has been widely used in the design of DNA-based sensors (Willner and Zayats 2007, Li et al 2008, Teller et al 2009  In the present study, we have applied this luminescence methodology for the in-field detection of mercury(II) ions in different environmental media including natural river, wetland and sea water samples close to industrial zones.

Materials
The stock solution (1 mM) of platinum(II) complex was prepared in acetonitrile and was kept at −20 • C in the dark before use. Further dilution to designated concentrations was made using Tris buffer (5 mM Tris-HCl, 50 mM NaCl, pH 7.2) or water samples. The mercury(II) chloride stock solution (1 mM) was stored in glass bottles with tight lids, and was further diluted to the designated concentrations by addition into the Tris buffer or water samples. All oligonucleotides were synthesized by Techdragon Inc. (Hong Kong, China). The sequence of the thyminerich single-stranded oligonucleotide T 33 is as follows: 5 -TTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTT-3 .

Water sample collection
Water samples were collected from three locations in Hong Kong representing natural river, wetland and sea water environments, respectively. The sampling area in Shing Mun River was located next to Fo Tan Factory Estate, the sampling area in Nam Sang Wai wetland was located next to Yuen Long Industrial Estate, and the sampling area in Lamma Island was located next to the Lamma Power Station. To obtain the samples, 2 liters of water were collected at a depth of 1 m below the water surface at each sampling station using a stainless steel 45 mm diameter Vertical Bailer Water Sampler.
The water samples collected were split into two 1 liter glass bottles which were immediately capped, chilled on ice and delivered to the laboratory, where they were stored at 4 • C in the dark. Glassware was used throughout the experiments instead of polypropylene in order to prevent mercury(II) ions from binding to the polypropylene surface. All glassware was pre-cleaned according to a reported literature procedure (Polonini et al 2011). Glassware was ultrasonicated for 15 min, followed by rinsing with ethanol and double distilled water and then drying in a forced draft hot air oven at 100 • C overnight.

Sample extraction
The water samples were thoroughly shaken each time before extraction to ensure that the determinant was uniformly distributed within the sample. Samples were extracted at 5 cm under the water level of bottle with an auto-pipette and filtered using a GD/XP 25 mm syringe filter with a 0.45 µm pore size. The blanks were prepared by adding the T 33 (285.7 nM) and [Pt(C ∧ N ∧ N)(4−appt)] + (9.74 µM) to 500 µl of filtered water sample.

Instrumental analysis
The concentrations of mercury(II) ions in untreated water samples were analyzed using inductively-coupled plasma mass spectrometry (ICP-MS) from a commercial testing laboratory (Hong Kong Standards and Testing Centre). Concentrations of mercury(II) ions in the sea water samples were measured by the Hong Kong Standards and Testing Centre (detection limit = 500 nM), while the samples from natural river and wetland were measured by the CMA Testing and Certification Laboratories (detection limit = 50 nM). Quality control for ICP-MS determination was performed by spike-and-recovery analysis.

Characterization of mercury(II) ion detection system in buffer solution
It has been previously reported that single-stranded thyminerich DNA sequences can be induced into a hairpin conformation by mercury(II) ions (Miyake et al 2006) though formation of T-Hg 2+ -T mismatches ( (1, C ∧ N ∧ N = 6-phenyl-2,20-bipyridine; 4-appt = 2-amino-4phenylamino-6-(4-pyridyl)-1,3,5-triazine; figure 2) is weakly emissive in aqueous buffer solution due to complex-solvent interactions that result in the non-radiative decay of its excited state. In the absence of mercury(II) ions, complex 1 is weakly emissive due to its weak interaction with the single-stranded oligonucleotide T 33 . However, in the presence of the mercury(II) ions, complex 1 intercalates into the double-stranded hairpin conformation induced by mercury(II) ions. This protects the metal complex from the aqueous buffer environment and suppressed non-radiative decay, thus enhancing 3 MLCT luminescence. (Chan et al 2009). By measuring the luminescence of the system, the assay was presently determined to achieve a detection limit of 20 nM for  mercury(II) ions in buffer solution, with a maximum intensity fold-change of 5.19 reached at a saturation limit of above 15 µM (table 1).

Characterization of the mercury(II) ion detection system in water samples.
In order to investigate the accuracy and reliability of the luminescent system in different media, the probe was applied to water samples from natural river, wetland and sea water samples. Quality control for the probe was initially planned by comparison with results obtained using inductively-coupled plasma mass spectroscopy (ICP-MS). However, ICP-MS analysis of the three water samples showed that they contained insignificant concentrations of mercury(II) ions. Therefore, the water samples were spiked with mercury(II) ions for the purpose of the experiments. The precision of the system was evaluated by the use of triplicate measurements of the water samples. A calibration curve for the luminescence intensity of the system at a range of mercury(II) concentrations in buffer solution was first constructed (figure 3). The relationship between intensity fold-change and mercury(II) concentration in the buffered solution was found to be linear in the range of 0.5-15 µM (R 2 = 0.9946). The real water samples were then spiked with various concentrations of mercury(II) ions, and the resulting luminescence intensity was compared with the calibration curve to determine the estimated mercury(II) concentration. Spike recovery analysis was performed in triplicate to evaluate the performance of the detection system in water samples compared to buffer solution.
3.2.1. Characterization of the mercury(II) ion detection system in low turbidity and low metal ion water samples. In natural river samples, the turbidity level was determined to be about 28 NTU, compared to 0.77 NTU for the buffer solution. The water sample also contained a relatively low concentration of metal ions compared to that of sea water (Solà and Prat 2006). The luminescence intensity of the system was recorded in triplicate at various concentrations of spiked mercury(II) ions in the natural river water samples ( figure 4). The results showed a linear dynamic range from 0-12 µM (R 2 = 0.9869) for Hg 2+ ions with a maximum emission intensity fold-change of about 4.3, which is comparable to the performance of the assay in buffer solution. At higher concentrations of Hg 2+ ions, the luminescence emission of the system was observed to plateau, presumably due to the saturation of Hg 2+ ion binding sites in the T 33 oligonucleotide. This phenomenon can be found in both aqueous buffer and real water sample, while the luminescence emission of the system in aqueous buffer requires higher concentrations of Hg 2+ ions to attain the plateau compared to that in the real water sample. The system exhibited modest percentage recovery values for spiked mercury(II) ions in Figure 4. Relationship between luminescence intensity fold-change of the system and mercury concentration (nM) in triplicate natural river samples. the natural river water samples as compared to the buffer solution (table 2). The detection limit for mercury(II) ions in the natural river sample was determined to be 100 nM. Significantly, this detection limit for mercury(II) ions is comparable to those offered by testing agencies in Hong Kong (50-500 nM). The higher detection limit for mercury(II) ions in the river sample compared to that for the buffer solution (20 nM) could be due to the shielding of the luminescence signal at low mercury(II) ion concentrations by the moderate turbidity present in the real water sample. Furthermore, the selectivity of our approach for Hg 2+ ions was evaluated by investigating the luminescent response of the system to 100-fold higher concentrations of nine other common metal ions, including Pb 2+ , Mg 2+ , Ca 2+ , Al 3+ , Ti 3+ , Sr 2+ , Zn 2+ , Na + and K + ions in natural river water sample. The results show that only Hg 2+ ions could significantly enhance the luminescence emission of the complex 1/T 33 system (figure 5). The presence of Ca 2+ or Al 3+ ions changed the luminescence response of the system by around 6-10%. These results indicate that the system displays excellent selectivity for Hg 2+ ions over the nine other metal ions tested in the natural river sample.
3.2.2. Characterization of the mercury(II) ion detection in high turbidity. The wetland water samples had an average turbidity measurement of 269.3 NTU (table 3), which was about 350 times higher than that of aqueous buffer solution (0.77 NTU). The high turbidity of the sample, which is likely caused by suspended particles such as organic matters and soil (Blavet et al 2009), resulted in a significant background emission signal (figure 6). The maximum fold-change of the  system of 1.8 was reached at a saturation limit of only 4 µM of mercury(II) ions. The detection limit of the system in wetland samples was determined to be 100 nM, which is comparable to that for the natural river samples (table 4). However, due to the turbidity of the sample matrix, the percentage recovery value for the spiked wetland sample was only about 50% at 4 µM of mercury(II) ions (data not shown).

Characterization of the mercury(II) ion detection in high metal ion water samples.
A high emission background signal was observed in the sea water sample. The four most abundant metal cations in sea water are sodium (469 000 µM), magnesium (53 000 µM), calcium (10 300 µM), and potassium (10 200 µM) (Holmes-Farley 2003). Our previous results have shown that the presence of micromolar metal ions could only induce a minute increase in the background signal of the detection system in the absence of mercury(II) ions. Although the sea water sample had a very low turbidity level similar to that of buffer solution, its detection limit for mercury(II) ions (500 nM) was the highest of all the water samples tested (figure 7). Only a 1.6-fold change in luminescence intensity was recorded at saturating concentrations of mercury(II) ions, and the recovery value

Conclusion
In conclusion, we have successfully applied our recently developed DNA-based luminescent method for the detection of spiked mercury(II) ions in real environmental samples including natural river, wetland and sea water media. The results showed that the luminescent assay functioned effectively under conditions of low turbidity and low heavy metal ion concentrations in the natural river water sample. However, a higher background signal was observed due to the high turbidity of the wetland sample or high concentrations of natural metal ions in the sea water sample. The assay exhibited the best performance in the natural river sample with modest percentage recovery values and a detection limit of 100 nM.