Introduction

Environmental pollution caused by heavy metal ions has become an important worldwide concern [1]. Because of the increasingly common use of silver in macro-usages such as battery, bearing, brazing and soldering production, as well as in the automobile, electronics, photographic, pharmaceutical and imaging industries [2], pollution by silver(I) ions has also received extensive attention [3]. Despite not being a bioaccumulative toxin, Ag+ absorbed by human does inactivate sulfhydryl enzymes, bind to amino, imidazole and carboxy groups of various metabolites, and displace essential metal ions such as Ca2+ and Zn2+ in hydroxyapatite in bone [4, 5], leading to argyria and symptoms such as headache, skin irritation, stomach distress, organ edema, and even death [6]. Therefore, Ag+ has been assigned to the highest toxicity class, surpassed only by Hg2+, and rapid and sensitive detection of it has been of large interest [7].

Among the available conventional techniques, atomic absorption spectroscopy (AAS) [8], electrothermal atomic absorption spectrometry (ETAAS) [9], inductively coupled plasma atomic emission spectrometry (ICP-AES) [10], inductively coupled plasma mass spectroscopy (ICP-MS) [11], voltammetry [12] and potentiometry [13] are the most commonly used for Ag+ detection. Respectable advantages such as high accuracy and excellent sensitivity showed by these techniques, but their wide applications in routine Ag+ detections remain limited due to the disadvantages of needing expensive sophisticated instruments, elaborate sample preparation, well-trained personnel and time consuming procedures [14]. So much focus has been turned to developing new methods which are simple and low cost.

Recent years have witnessed the rapid development of using functional nucleic acids as the molecular recognition elements for the detection of metal ions [15–17]. Generally, the functional nucleic acids includes aptamer, DNAzyme, G-quadruplex binding ligand and specific DNA that can selectively bind to its targets [18]. As for Ag+ detections, the most impressive and widely used functional nucleic acid is silver-specific DNA which riches in cytosine (C) and readily forms a hairpin structure in the C-Ag+-C configuration in the presence of the target Ag+ [19]. Since Ono et al. firstly reported the selective capture of Ag+ by C-C mismatches [20], several C-rich oligonucleotides have been designed as silver-specific DNA and omnifarious silver-specific DNA-based methods have been developed for Ag+ detections by transforming the formation of C-Ag+-C base pairs into colorimetric [21], scattering [19], fluorescent [22–24] and electrochemical [25] signal output. Among these different kinds of signal output, fluorimetry occupies a considerable proportion due to its operational simplicity, high sensitivity and real-time detection [26, 27]. And those fluorescent probes for Ag+ could be divided into two types: “off-on” type, fluorescence enhanced upon Ag+ binding; “on-off” type, fluorescence weakened upon Ag+ binding. It is generally acknowledged that the former type is superior compared to the latter type due to its better overall signal processing in biosystems and specificity of mode of action [28]. However, as the Ag+ belongs to the so-called “silent ions”, the design of “off-on” type probes for it remains challenging and attractive up to now [29]. Therefore, much effort should be devoted to the development of highly selective fluorescent “turn on” probes for Ag+ [30].

In this study, a sensitive Ag+ quantitative method that utilized silver-specific DNA, gold nanoparticles (AuNPs) and Rhodamine B (RhoB) was presented. As shown in Scheme 1, when no target (i.e. Ag+) exists, the silver-specific DNA that named Ag-CC can absorb onto the surface of AuNPs, and the DNA-absorbed AuNPs can maintain dispersion even at a high concentration of NaNO3, greatly quenching the fluorescence of RhoB. However, after the Ag-CC interacts with Ag+ to form C-Ag+-C base pairs, it changes its conformation from a random coil to a hairpin structure, failing to absorb onto the AuNPs. Thus the AuNPs stay bare and aggregate induced by NaNO3, consequently their effect on the fluorescence of RhoB gets weak. As the fluorescence intensity of RhoB was control by the aggregation degree of the AuNPs, which was mediated by the silver-specific DNA, and the DNA amount that plays a part in mediating the AuNPs correlates with the involved Ag+, a fluorescent assay for Ag+ detection in aqueous solution was thus established.

Scheme 1
scheme 1

Schematic representation of the fluorescent detection of Ag+ in aqueous solution with Rhodamine B and silver-specific DNA mediated gold nanoparticles

Full size image

Materials and methods

Chemicals and reagents

The sequence of the silver-specific DNA (i.e. Ag-CC) is referred to previous literature [19], and it was synthesized by Sangon Biotechnology Inc. (Shanghai, China. www.sangon.com). Its sequence is 5′-CTCTCTTCTCTTCATTTTTCAACACAACACAC-3′ (32-mer). HAuCl4 and 3-(N-morpholino) propanesulfonic acid (MOPS) were obtained from Sangon Biotechnology Inc., too (www.sangon.com). Rhodamine B, trisodium citrate, sodium chloride, sodium hydrosulfide, organic compounds used in selectivity study and cationic compounds such as nitrates of Ag+, Na+, K+, Mg2+, Ca2+, Ni2+, Zn2+, Fe3+ and sulfates of Fe2+, Mn2+, Cu2+ were obtained from commercial sources and used without further purification. Standard solution (1 mg · mL−1, 1000 ppm) of Pb2+, Hg2+ and Cd2+ were purchased from Merck Co., Inc. (Germany, www.merck.com) and used after diluted to appropriate concentration with ultrapure water. The 96-well microplate was purchased from Thermo Fisher Scientific Inc. (Nunclon, Denmark, www.thermofisher.com). Ultrapure water that utilized to prepare all aqueous solutions was from a Millipore-MilliQ (Milli-Q plus, Millipore Inc, Bedford, MA, USA, www.millipore.com) system.

Instruments

An analytical transmission electron microscope JEM-2010HT (Hitachi, Japan, www.hitachi.com) was used to observe the images of AuNPs. A Microplate Spectro-photometer M200 Pro (Tecan Group Ltd, Switzerland, www.tecan.com) was employed to determine the aggregation degree of the AuNPs. A steady-state & time-resolved fluorescence spectrofluorometer of the QM/TM/IM model (Photon Technology International, USA, www.pti-nj.com) was applied to detect the fluorescence lifetime of the RhoB that treated differently. A PL-F2X nitrogen laser was used as the excitation source for lifetime measurements, with the resulting fluorescence monitored at right angles through an emission monochromator with a photomultiplier tube and strobe detector. Quartz cells (1*1 cm) were used for all spectral measurements. The obtained fluorescence intensity was normalized by dividing the intensity at each time points with the highest fluorescence intensity, and the time which the normalized intensity 1 corresponded was determined as the start time (0 ns). An F-4500 fluorescence spectrophotometer (Hitachi, Japan, www.hitachi.com) was used to record the fluorescence intensity, with the response time of 0.05 s, PMT voltage of 700 V, scan speed of 1200 nm · min−1, excitation wave length of 520 nm and excitation and emission slits of 10 nm. A Thermostatic incubating device (Eppendorf, China, www.eppendorf.cn) was used to provide the different incubating temperatures at which the final solution that containing all the sensing elements was incubated when studying the influences of the temperature.

Preparation of AuNPs

The AuNPs were synthesized by sodium citrate reduction of HAuCl4 following a literature procedure [31]. Briefly, a sodium citrate solution (1 % (w/v), 10.5 mL) was rapidly added to a boiled solution of HAuCl4 (100 mL, 0.03 % (w/w)) under vigorous stirring. The mixed solution was boiled for 10 min and further stirred for 15 min. The resulting wine-red solution was cooled to room temperature and filtered, then stored in dark glass bottles at 4 °C for further use.

Optimization of the sensing system

The volume of each sensing system was set as 500 μL and the volume of AuNPs added to each sensing system was fixed as 250 μL. The absorbance values at 520 nm (A520) and 650 nm (A650) were measured, and the absorbance ratio A=A650/A520 was recorded to evaluate the aggregation degree of AuNPs. The increased absorbance ratio value ΔA=A-A 0 was calculated to indicate the difference of the aggregation degree of AuNPs, in which A 0 stands for the absorbance ratio of the blank solution and A for the ions treated one. When optimize the NaNO3 concentration, AuNPs in the sensing system was treated with NaNO3 of increasing concentration, and the minimum NaNO3 concentration which can fully aggregate AuNPs was chose. The Ag-CC concentration was optimized by comparing the ΔA of different sensing system in the absence and in the presence of 500 ppb Ag+ which contain Ag-CC of different concentration. The RhoB concentration was optimized by comparing the fluorescence intensity of different sensing system in the absence and in the presence of 800 ppb Ag+ which contain RhoB of different concentration. The fluorescence intensity of the blank was recorded as F 0 and that of the Ag+ treated one was recorded as F, and the relative fluorescence intensity was calculated as F/F 0 , and the RhoB concentration which corresponding to the highest F/F 0 value was chose.

Procedure for Ag+ determination

A typical Ag+ analysis was realized as following steps. Firstly, 7.5 μL of Ag-CC solution (1 μM) was mixed with 10 μL Ag+ solution of appropriate concentration. This solution was allowed to incubate for 10 min at room temperature. Secondly, 250 μL AuNPs was added to the above prepared solution and the solution was diluted with MOPS buffer (10 mM, pH 7.0) to 430 μL, and then the solution was allowed to react for 10 min at room temperature. Finally, 30 μL of 2 M NaNO3 and 40 μL of 100 μM RhoB were added. After incubated for another 10 min under the same condition, the fluorescence intensity of each sample was monitored. Blank experiments were carried out by replacing Ag+ with ultrapure water. The fluorescence intensity of the blank was recorded as F 0 and that of the Ag+ treated one was recorded as F, and the increased fluorescence intensity was calculated as (F−F 0 ). To test the selectivity of the assay, other competing metal ions and a series of common organic compounds instead of Ag+ were added in a similar way.

Results and discussion

Sensing mechanism

The sensing system is composed of silver-specific DNA, AuNPs, NaNO3 and RhoB, and each element is prudently selected for the determination of Ag+. The silver-specific DNA serves as the recognition probe, besides interacts with Ag+ to form C-Ag+-C base pairs; it mediates the aggregation status of AuNPs with the help of NaNO3. The AuNPs were usually used as a colorimetric indicator because of its high extinction coefficients and distinct size-dependent colors, while in this assay it was their excellent quenching properties, which largely depends on their dispersed/aggregated status, that was favored [32]. The section before the addition of RhoB is exactly the constitution of the classical colorimetric assay which based on the aggregation of AuNPs. However, if the concentration of the analyte is low, inconspicuous color change will be obtained, which will affect the sensitivity of the colorimetric assay [33]. Thus RhoB was introduced to convert the colorimetric signal into fluorescent signal for the sake of improving the sensitivity.

Proof-of-concept experiments were performed to demonstrate the mechanism of this assay. As shown in Fig. 1a and b, the pure AuNPs and the silver-specific DNA-coated ones were evenly distributed and maintain disperse, which could effectively weaken the originally high fluorescence of the RhoB (line 1, 2 and 3 in Fig. 1d). The AuNPs aggregated after Ag+ was added (c), accompany with the decrease of their effect on the fluorescence of RhoB, under which the fluorescence of the Rho B recovered to certain extent (line 4 and 5 in Fig. 1d). These results preliminarily verify the feasibility of Ag+ determination of this assay.

Fig. 1
figure 1

The transmission electron microscope (TEM) images of the AuNPs solutions and the fluorescence spectra of the RhoB that treated with different substances. (a) AuNPs; (b) AuNPs + 15 nM Ag-CC + 0.12 M NaNO3; (c) AuNPs + 15 nM Ag-CC + 800 ppb Ag++0.12 M NaNO3; (d) line 1: 8 μM RhoB; line 2: 8 μM RhoB + 250 μL AuNPs; line 3: 8 μM RhoB + 250 μL AuNPs + 15 nM Ag-CC + 0.12 M NaNO3; line 4: 8 μM RhoB + 250 μL AuNPs + 15 nM Ag-CC + 800 ppb Ag++0.12 M NaNO3; line 5: 8 μM RhoB + 250 μL AuNPs + 15 nM Ag-CC + 1500 ppb Ag++0.12 M NaNO3

Full size image

Optimization of RhoB concentration

After the optimization of the NaNO3 and Ag-CC concentration (Electronic Supplementary Material, ESM), the RhoB amount that needed was also investigated. RhoB was chosen because of its water solubility, photostability, high extinction coefficient and quantum yield [34]. RhoB molecules are highly fluorescent, but when they have been adsorbed noncovalently onto the surfaces of AuNPs through electrostatic interactions, fluorescence resonance energy transfer (FRET) and collision occur between RhoB and the AuNPs, result in almost no fluorescence emission [35]. As shown in Fig. 2, under the optimal NaNO3 and DNA concentration, without addition of Ag+, a majority of the 250 μL AuNPs was supposed to maintain disperse status, and the fluorescence of RhoB were completely diminished as the RhoB concentration increases from 0 to 6 μM. When the RhoB concentration exceeds 6 μM, increasing fluorescence of RhoB appeared. When 800 ppb Ag+ was added, fluorescence appeared even the RhoB concentration was just 4 μM. This phenomenon indicates that the introduction of Ag+ induced the aggregation of the AuNPs, and the effect on the fluorescence of RhoB of the aggregated AuNPs was greatly weakened compared to the dispersed ones [33]. The possible mechanism of these results may lie in two routes. One is that the surface-to-volume ratio of the aggregated AuNPs is much lower than that of the dispersed ones, thus the amounts of RhoB absorbed to the surface of aggregated AuNPs is smaller, which leads to weaker FRET between AuNPs and RhoB and decreasing quenching ability of the aggregated AuNPs [33]. The other is the fluorescence inner filter effect of AuNPs [36]. The dispersed AuNPs which has strong absorption at 520 nm can filter a part of excitation light and decrease the available emission light. While in the condition AuNPs is aggregated, the absorbance at 520 nm decreased, which leads to diminished inner filter efficient of AuNPs and more light to excite RhoB, thus more emission light is detected. The inset in the upper left shows the relative fluorescence intensity (F/F 0 ), it reveals that at the RhoB concentration of 8 μM, the F/F 0 obtains its maximum, which decreases as the RhoB concentration increased. In order to acquire prominent fluorescence signal and at the same time avoid high background, the optimal RhoB concentration was thus chose as 8 μM.

Fig. 2
figure 2

The Effect of the RhoB concentration on the fluorescence intensity. The volume of AuNPs was 250 μL, the concentration of NaNO3 was 0.12 M, and the concentration of Ag-CC was 15 nM

Full size image

Study of the fluorescence lifetime

Previous studies claim that both the inner filter effect and the quenching of the dye owing to its adsorption on AuNPs contribute to the observed changes in fluorescence intensity [33, 37]. It is easy to predict the occurrence of inner filter effect as both the strong absorption of the dispersed AuNPs and the excitation wave length of the RhoB are at 520 nm. Therefore it is necessary to carry out fluorescence lifetime study to clarify whether the quenching of fluorescence actually happened. As shown in Fig. 3, the fluorescence decay of the RhoB that treated in different ways was monitored. The normalized fluorescence intensity of the RhoB was plotted against time by fitting to ExpDec1 plot. It could be seen that the curvature of each plot differs from those of others. The obtained fitting equation and relevant parameters were summarized in the inset, where y stands for the normalized intensity, A1 for pre-exponential factor, x for time and t1 for the fluorescence lifetime. It reveals that the lifetime of pure RhoB is 2.0881 ns, and when the RhoB mixed with the disperse AuNPs, its lifetime decreased to 0.6021 ns. However, if the RhoB coexists with the NaNO3-induced aggregate AuNPs, its lifetime mere reduced to 1.5024 ns. When RhoB treated with AuNPs which coated by Ag-CC of appropriate amount and maintained disperse status against NaNO3, its lifetime is 0.9111 ns. After 500 ppb Ag+ was introduced to consume the Ag-CC and induced the aggregation of the AuNPs, the lifetime of the RhoB restores to 2.0508 ns. All these results demonstrate that the lifetime of the RhoB is affected by the aggregation degree of the AuNPs, in which dispersed AuNPs correspond to shorter fluorescence lifetime. As in the proof-of-concept experiments, dispersed AuNPs accompany with weaker fluorescence intensity, it thus could be concluded that the dispersed AuNPs not only decrease the fluorescence intensity of RhoB, but also reduce the fluorescence lifetime of it. These results suggest that besides inner filter effect, the fluorescence decreasing phenomenon caused by the AuNPs was also based on the quenching of fluorescence in which the lifetime is reduced by quencher [38].

Fig. 3
figure 3

Fluorescence decay profile of RhoB that treated in different ways. Sample 1: 8 μM RhoB; sample 2: 8 μM RhoB + 250 μL AuNPs; sample 3: 8 μM RhoB + 250 μL AuNPs + 0.12 M NaNO3; sample 4: 8 μM RhoB + 250 μL AuNPs + 15 nM Ag-CC + 0.12 M NaNO3; sample 5: 8 μM RhoB + 250 μL AuNPs + 15 nM Ag-CC + 500 ppb Ag++0.12 M NaNO3

Full size image

Influences of temperature

In order to further elucidate the mechanism of the assay, studies have been carried out on the influences of the temperature at which the final solution that containing all the sensing elements was incubated. The results shown in Fig. 4 indicate that the increase in fluorescence intensity increased gradually as the temperature elevated from 15 to 35 °C, after which the increase in fluorescence intensity rose obviously until the temperature reached 65 °C, and then the increase in fluorescence intensity tends to flatten. There are mainly two reasons accounting for this phenomenon. Firstly, the Ag-CC that coated on the surface of the AuNPs would detach as the temperature increased, results in more bare AuNPs in the sensing system which would aggregate with the help of NaNO3. As claimed in the “optimization of RhoB concentration” section, aggregated AuNPs corresponding to weaker quenching effect on the fluorescence of RhoB, thus the breaking away of Ag-CC from surface of the AuNPs results from the increasing temperature leads to higher fluorescence intensity. Secondly, as the temperature increased to high, the non-covalent interaction between RhoB and the AuNPs may be weakened, which in turn would reduce the incidence of the FRET and collision between them, bringing about stronger fluorescence. These results further testify the mechanism of the assay. Nevertheless, as our goal is to develop assay which holds the potential for on-site and real-time detection, room temperature (25 °C) was still used in all the following studies.

Fig. 4
figure 4

The influences of the temperature on the increased fluorescence intensity of the final solution. Each sample of the final solution contains 250 μL AuNPs, 0.12 M NaNO3, 15 nM Ag-CC, 8 μM RhoB, 500 ppb Ag+ and MOPS buffer (10 mM, pH 7.0) of appropriate volume

Full size image

Sensitivity of the assay for Ag+ detection

The sensitivity of this assay was explored by measuring the fluorescence signals toward different concentration of Ag+. As shown in Fig. 5a, the fluorescence intensity increases as the Ag+ concentration increases from 0 to 1500 ppb. Figure 5b depicts the increased fluorescence intensity plotted against the concentration of Ag+ by fitting to a DoseResp plot with a correlation coefficient of 0.988. Inset shows the linear responses at low Ag+ concentrations. The slope of the linear regression equation is 0.6284 and the intersection at the y axis is 0.1044, with a correlation coefficient of 0.988. Based on previous reports [39], 3σ/slope was used to determine the limit of detection for Ag+ as 2.73 ppb (~25.31 nM), which is lower than that of the colorimetric assay that based on the aggregation of AuNPs [40, 41]. As the maximum contamination level for Ag+ in drinking water defined by the US Environmental Protection Agency (EPA) and the World Health Organization (WHO) is 50 ppb (460 nM) [19], this assay is fully applicable for Ag+ monitoring in drinking water. Moreover, the dynamic range of this assay (2.73–1500 ppb) could be simply tuned by adjusting the involved amount of Ag-CC, AuNPs and RhoB, which shows great practical utility in practical applications as the desirable concentrations for the same target analyte can be different for various applications [16].

Fig. 5
figure 5

a Fluorescence emission spectra of the sensing system on exposure to 0, 5, 10, 20, 30, 40, 60, 80, 100, 150, 200, 300, 400, 600, 800, 1000, 1250 and 1500 ppb Ag+. b Calibration curve of the sensing system for Ag+. Sensing system: 250 μL AuNPs, 0.12 M NaNO3, 15 nM Ag-CC, 8 μM RhoB, MOPS buffer (10 mM, pH 7.0)

Full size image

Selectivity of the assay for Ag+ detection

The selectivity of the assay was investigated by carrying out control experiments by examining other metal ions, such as Ca2+, Cd2+, Cu2+, Fe2+, Zn2+, Hg2+, K+, Mg2+, Mn2+, Na+, Ni2+, Pb2+ and Fe3+ under identical conditions. And the results in Fig. 6 show that only Ag+ caused remarkable increased fluorescence intensity, other metal ions display no significant interferences. The selectivity of the assay for Ag+ detection over other organic compounds and other commonly existing anions, such as Cl− and HS− were also studied (Fig. S3 and Fig. S4, ESM). It reveals that among the testing organic compounds, except cysteine can bring slight response, other organic compounds cause no interference even at high concentration. As for Cl− and HS−, even the concomitant Cl− achieves the high concentration of 50 ppm, Ag+ could still be identified; In contrast, concomitant HS− of 5 ppm can inhibit the response of the assay, but the inhibiting effect could be screening out by treated with nitric acid of appropriate concentration. More effective approach should be exploited to remove the HS− of high concentration when employed this assay in real sample detection.

Fig. 6
figure 6

Selectivity of the assay for Ag+ detection over other common metal ions. Ions concentrations: 500 ppb; Sensing system: 250 μL AuNPs, 0.12 M NaNO3, 15 nM Ag-CC, 8 μM RhoB, MOPS buffer (10 mM, pH 7.0)

Full size image

Detection of Ag+ in water samples

The application of the present assay was evaluated for the determination of Ag+ in several spiked water samples containing different concentrations of Ag+ and some other metal cations and anions. The results are summarized in Table 1, which display that the mean recovery of samples was in the range of 99.9–102.7 % and the relative standard deviation (RSD) was between 3.49 and 5.21 %. The results reveal the potential application of this assay for Ag+ detection in water samples.

Table 1 Determination of Ag+ in water samples
Full size table

Conclusions

In conclusion, a fluorescent assay for Ag+ detection based on the aggregation of gold nanoparticles caused by silver-specific DNA and its effect on the fluorescence of Rhodamine B was established. The random coil cytosine-rich silver-specific DNA undergoes structural transformation upon capturing Ag+, mediating the dispersion/aggregation status of the gold nanoparticles, which at the same time affects the fluorescence intensity of Rhodamine B. As there was a positive correlation between the fluorescence intensity of the Rhodamine B and the involved amount of Ag+, an “off-on” type fluorescent method was thus developed. By monitoring the fluorescence intensity changes upon addition of Ag+, the Ag+ could be selectively detected from 2.73 to 1500 ppb, with a detection limit of 2.73 ppb. Feasibility in determining spiked water samples was also demonstrated, suggesting this assay holds great practicality for Ag+ detection in real matrices such as tablets, injections or drinking water. Compared with those existing fluorescent methods for Ag+ determination, this assay possesses several superiorities such as requires no DNA labeling, relative wider analytical range, and can in principle be used to detect other metal ions by substituting the silver-specific DNA with other native or artificial bases that selectively bind to other metal ions (Table 2).

Table 2 Figures of merit of fluorescent methods for determination of Ag+
Full size table