Introduction

Uranium, a representative radioactive metal, has been used as the main source of nuclear power generation via nuclear fission1. Uranium can exist in various forms in the environment, such as uranyl fluoride (UO2F2), uranyl tetrafluoride (UF4), uranium dioxide (UO2) and triuranium octoxide (U3O8). The accumulation of these compounds in the human body can lead to severe health problems2. For example, UO2F2 and UF4 cause kidney damage and UO2 and U3O8 can cause certain cancers and mutations by accumulating in the lungs3. Uranium is also found in diverse forms in aqueous conditions. The uranyl ion (UO22+) is the most soluble and common form4,5. Because UO22+ can disturb organ function by accumulating in the skeleton, kidneys, lungs and liver6,7, the detection of UO22+ in natural water is very important.

Traditionally, UO22+ has been detected using various physical and chemical techniques, including inductively coupled plasma-mass spectrometry8, radio spectrometry9, atomic adsorption spectrometry10 and phosphorimetry11. However, these methods present several drawbacks in that they require expensive instruments, involve complicated methods and are time-consuming. Recently, nanomaterial-based UO22+ sensing methods have been developed by employing fluorescence12,13, electrochemistry14, resonance scattering15,16, colorimetry17, magnetoelasticity18 and surface-enhanced Raman scattering (SERS)19,20,21. Among these techniques, SERS offers highly sensitive molecular detection because SERS signals can be obtained from a small number of molecules or even a single molecule located within a sub-10 nm metallic nano-gap (hot spot)22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42. This technique also provides molecular fingerprints and causes less photobleaching43,44,45,46,47,48. Nevertheless, there are only a few reported SERS sensors for UO22+ detection and most of them have not been validated in real environments. Therefore, further improvements in sensitivity, selectivity and reproducibility are still needed for the practical sensing of UO22+ in a real environment using a SERS sensor.

Here, we report an ultrasensitive UO22+ detection method using a plasmonic nanowire (NW) interstice (PNI) sensor combined with a DNAzyme-cleaved reaction. SERS-based sensing methods are often affected by uncontrolled aggregation and size distribution of nanomaterials, which are detrimental to the reproducibility of the sensor49,50,51. A PNI platform composed of a single-crystalline Au NW with an atomically flat surface has shown improved sensitivity and reproducibility31. By combining the DNAzyme cleaving method with a PNI platform, we quantitatively detected UO22+ with an ultralow detection limit of 1 pM and high selectivity. More importantly, the precise detection of UO22+ was demonstrated in several UO22+-contaminated natural water samples, showing the practical applicability of the PNI sensor.

Results and Discussion

DNA-based ligands, including aptamers and DNAzymes, have been widely used to detect various metal ions because they provide several advantages, such as high affinity, selectivity, stability and relative ease of modification52,53. In this experiment, we used a DNAzyme, which has the sequence shown in Fig. 1a, for the detection of UO22+. A DNAzyme is composed of an enzyme strand and a substrate strand. The enzyme strand (green) is a specific sequence that can bind to UO22+. The substrate strand (blue) is hybridized with the enzyme strand. In the middle of the substrate strand, a ribose-adenosine (rA) sequence (red) is present. Ribonucleotides are approximately 100,000-fold more susceptible to hydrolytic cleavage than deoxyribonucleotides13. At the 5′ end of the substrate strand, Cy5 (magenta) is present. Cy5 is a well-known Raman reporter for the 633 nm excitation source27,28,29,30,31,32. When UO22+ was added to DNAzyme, UO22+ bound to the enzyme strand and thus, the rA sequence on the substrate strand was hydrolytically cleaved. The cleaved strand was released from the DNAzyme and then captured by a PNI sensor. The PNI sensor was prepared by laying down a single-crystalline Au NW onto a Au film24 (Fig. 1b). Au NWs were synthesized using a previously reported vapor transport method54 and then modified by the addition of a complementary capture DNA strand to the cleaved strand. The length of capture strand was set as 15 base pairs to increase the hybridization efficiency with cleaved strand55. The PNI sensor with an attached cleaved strand can provide a strong Cy5 SERS signal, enabling the detection of UO22+. The left panel in Fig. 1b shows the result of using a PNI sensor combined with a DNAzyme-cleaved reaction for UO22+ detection. When UO22+ was present in the sample solution at the concentration of 10 nM, 4 major bands at 1185, 1360, 1485 and 1580 cm−1 were clearly detected by the PNI sensor (red spectrum). These bands correspond to the v(C-N)stretch, v(C = C)ring, v(C-C)ring and v(C = N)stretch of Cy5, respectively28. In contrast, when UO22+ was absent from the sample solution, the PNI sensor exhibited a weak SERS signal (blue spectrum). In this way, UO22+ can be detected from the turn-on of the SERS signal. Generally, turn-on type sensors are more suitable than turn-off type sensors for practical applications because they result in fewer false-positive results17.

Figure 1
figure 1

(a) Schematic illustration of UO22+-specific DNAzyme and cleavage of DNAzyme induced by UO22+ (b) Schematic illustration of UO22+ detection by a PNI sensor combined with a DNAzyme-cleaved reaction. The left panel shows the SERS spectra measured from PNI sensors in the absence of (blue spectrum) and the presence of 10 nM UO22+ (red spectrum).

Figure 2a shows the SERS signals measured from the PNI sensors while varying the UO22+ concentration from 1 pM to 100 nM. The samples were prepared by a 1/10 serial dilution and a blank sample solution was prepared as a control. In the blank sample solution, the PNI sensor provided a weak Cy5 SERS signal (black spectrum in Fig. 2a). In the 1 pM sample, the signal was significantly enhanced (red spectrum in Fig. 2a). Figure 2b shows the intensity of the Cy5 1580 cm−1 band plotted as a function of the UO22+ concentration. The SERS signal from the PNI sensor increased as the UO22+ concentration increased, saturating at a concentration of 10 nM. The inset in Fig. 2b shows the dynamic range of the PNI sensor for UO22+ detection. The SERS signal intensity linearly increased within the UO22+ concentration range of 1 pM to 10 nM. This wide dynamic range makes this technique advantageous for the quantitative detection of UO22+. We estimated the detection limit of the PNI sensor to be 1 pM. This detection limit is approximately 1,000-fold lower than the limits of other SERS-based sensors21. Furthermore, the detection limit of this method is comparable to that of the most sensitive detection method, which is based on a resonance scattering spectral technique15.

Figure 2
figure 2

(a) SERS spectra of Cy5 measured from PNI sensors when the UO22+ concentration is varied from 1 pM to 100 nM. (b) Intensity of the Cy5 1580 cm−1 band plotted as a function of the UO22+ concentration. The inset provides the dynamic range of the PNI sensor for UO22+ detection. The data represent the mean plus standard deviation from 7 measurements.

To confirm the selectivity of a PNI sensor combined with a DNAzyme-cleaved reaction, SERS signals were observed from the PNI sensors after mixing DNAzyme with samples containing various metal ions (UO22+, Cd2+, Hg2+, Pb2+, Ca2+, Mg2+, Zn2+, Cu2+, Fe2+, Co2+, Ni2+ and Th4+). The concentration of each metal ion was 10 nM. Figure 3 shows the intensity of the Cy5 1580 cm−1 band measured from the PNI sensors for detecting the various metal ions. Remarkable SERS signal was observed only in the presence of UO22+ (magenta bar in Fig. 3) and no distinct SERS signal was observed in the presence of other metal ions (cyan bars in Fig. 3). These results indicate that the proposed detection method is highly specific for UO22+.

Figure 3
figure 3

Selectivity of a PNI sensor for UO22+ detection.

The tested metal ions are shown on the x axis and their corresponding Cy5 1580 cm−1 band intensities are shown on the y axis. Strong SERS signals were observed only in the presence of UO22+ (magenta bar) and weak SERS signals were observed in the presence of other metal ions (cyan bars). The data represent the mean plus standard deviation from 7 measurements.

Finally, we examined the applicability of a PNI sensor combined with a DNAzyme-cleaved reaction to detect UO22+ in natural water. The natural water samples were obtained from various environments, including a sea, a river, a lake and a tap. Buffer solution was also prepared as a control. Figure 4 shows the intensities of the Cy5 1580 cm−1 band measured from the PNI sensors after mixing DNAzyme with the natural water samples. The magenta bars represent the natural water samples spiked with UO22+ and the cyan bars represent the as-collected natural water samples. Strong SERS signals were observed only after the addition of UO22+ into the natural water samples. Without UO22+, SERS signals were rarely observed. Note that the DNAzyme is sensitive to ionic strength56. The DNAzyme is in the lock-and key mode at the ionic strength of 100 mM or higher56. In addition, the concentration of Mg2+ can affect the DNAzyme activity56. At the lower concentration of 2 mM, the activity is promoted. On the other hand, the activity is inhibited at the higher concentration of 2 mM. In this experiment, the DNAzyme solution and natural water sample were mixed with 9:1 volume ratio before the SERS measurement. The ionic strengths of the mixtures are calculated as 246.43 mM (sea), 131.81 mM (river), 134.72 mM (lake) and 132.07 mM (tap) according to the inductively coupled plasma-optical emission spectroscopy (ICP-OES) and ion chromatography (IC) data of natural water samples (Table S1 and S2 in supplementary information). The concentrations of Mg2+ in the mixtures are 1.31 mM (sea), 0.01 mM (river), 0.02 mM (lake) and 0.01 mM (tap). Since the ionic strengths of mixtures are higher than 100 mM and the concentrations of Mg2+ in the mixtures are below 2 mM, the PNI sensor combined with a DNAzyme-cleaved reaction can detect UO22+ successfully even in natural water samples. The USA’s limit for UO22+ in drinking water is 30 μg/L (approximately 100 nM), however, the long-term consumption and exposure to water that contains UO22+, even at concentrations below this limit, can cause severe toxicity and serious diseases57. We anticipate that the present method can be employed for practical UO22+ detection and aid in the prevention of environmental pollution and human diseases caused by UO22+.

Figure 4
figure 4

Applicability of a PNI sensor for UO22+ detection in natural water.

Natural water samples are shown on the x axis and their corresponding Cy5 1580 cm−1 band intensities are shown on the y axis. The magenta bars represent the natural water samples spiked with UO22+ and the cyan bars represent the as-collected natural water samples. The spiked UO22+ concentration was 10 nM. The strong SERS signals were observed only after the addition of UO22+ into the natural water samples. The data represent the mean plus standard deviation from 7 measurements.

Conclusion

We detected UO22+ by combining an ultrasensitive PNI platform with a DNAzyme-cleaved reaction. The DNAzyme specifically reacted with UO22+ and released a cleaved strand. The PNI sensor sensitively captured the cleaved strand, which enabled the detection of UO22+. We quantitatively detected UO22+ with an ultralow detection limit of 1 pM and high selectivity. Moreover, this method enables the detection of UO22+ in various natural water sources, such as sea, lake, river and tap. A PNI sensor that can precisely detect small quantities of UO22+ in natural water is expected to reveal many environmental pollutants and hence minimize damage to the human body caused by UO22+ exposure.

Methods

Materials

Purified DNA was purchased from Bioneer (Daejeon, Korea). Hg(Ac)2, Mg(Ac)2, Ca(Ac)2, CrCl2, FeCl2, Co(Ac)2, NiCl2, CuCl2, Zn(Ac)2, Cd(Ac)2, Pb(Ac)2, UO2(Ac)2, Au powder (99.99%) and sodium dodecyl sulfate (SDS) were purchased from Sigma–Aldrich. Th(NO3)4 was purchased from Merck. Phosphate buffered saline (PBS) was purchased from Gibco. The natural water samples were collected from the west sea of South Korea, the Gap River and the pond at KAIST. Samples were centrifuged to remove impurities.

Preparation of DNAzyme

The sequence of the substrate strand is 5′-Cy5-TAATACACTCACTAT(rA)GGAAGAGATGGACGTG-3′ and the sequence of the enzyme strand is 5′-CACGTCCATCTCTGCAGTCGGGTAGTTAAACCGACCTTCAGACATAGTGAGT-3′. The sequence of the capture strand is 5′-ATAGTGAGTGTATTA-SH-3′. The substrate and enzyme strands were mixed in a 1× PBS solution (pH 7.4) at the same molar concentration (10 μM each). The hybrid solutions were heated at 95 °C for 10 min and slowly cooled to room temperature.

Synthesis of single-crystalline Au NWs

Single-crystalline Au NWs were synthesized on a c-cut sapphire substrate in a horizontal quartz tube furnace system following the chemical vapor transport method described in a previous report54. An alumina boat containing an Au powder was positioned directly below the heat source. The sapphire substrate was placed a few centimeters downstream from the alumina boat. The heating zone was brought to 1100 °C while the chamber pressure was maintained at 3–5 Torr. Ar gas flowing at 100 sccm was used to transport the Au vapor. Au NWs were grown on the sapphire substrate over a 1 h period.

Preparation of PNI sensors

Au film substrates were prepared on Si substrates by electron beam-assisted deposition of a 10 nm thick film of Cr followed by a 300 nm thick film of Au. The prepared Au films were SERS-inactive by themselves24. The Au films were then cut to 1 cm2 for PNI sensor fabrication. To prepare the capture probe DNA-attached Au NWs, as-synthesized NWs were incubated with 5 mM captured DNA in 1 M KH2PO4 buffer (pH 6.75) at room temperature for 12 h. Next, the Au NWs were rinsed with a 0.2% (w/v) SDS solution for 5 min. The capture probe DNA-attached Au NWs were then transferred onto Au films by a simple attachment and detachment process30. Briefly, the NW-grown c-cut sapphire substrates were inverted onto Au film substrates containing medium. Before both substrates were overlapped, a drop of distilled water was applied as a lubricant. The sapphire substrates were then pushed gently and, after a few seconds, detached. After the attachment and detachment process, the remaining water was dried under flowing N2 gas.

Detection of UO22+ using PNI sensors combined with a DNAzyme-cleaved reaction

DNAzyme in 1× PBS (pH 7.4) with RNase-free water was mixed with sample solutions containing 0.27% (w/v) HNO3 and incubated for 10 min. The pH of the mixed solution was maintained at 5.49 and its ionic strength was 146.43 mM. To stop the enzymatic reaction by shifting the pH of the mixture, 0.1 mM Tris-acetate solution was added. Then, the mixture was dropped onto the PNI sensor and allowed to stand for 2 h. To remove excess DNA, the sensor was rinsed with a 0.2% (w/v) SDS solution for 5 min and then rinsed twice with distilled deionized water.

Instrumentation

SERS spectra were measured using a micro-Raman system on an Olympus BX41 microscope. A 633 nm He/Ne laser (Melles Griot) was used as an excitation source and the laser was focused on samples through a 100× objective (NA = 0.7, Mitutoyo). The laser power directed at the sample was 0.4 mW. The SERS signals were recorded with a thermodynamically cooled electron-multiplying charge-coupled device (Andor) mounted on the spectrometer with a 1200 groove/mm grating. The acquisition time for all SERS spectra was 60 s. A holographic notch filter was used to reject laser light. The concentration of metal ion in natural water samples was analyzed by inductively coupled plasma-optical emission spectroscopy (ICP-OES 720, Agilent) and ion chromatography (881 Compact IC pro, Metrohm Ltd.).

Additional Information

How to cite this article: Gwak, R. et al. Precisely Determining Ultralow level UO22+ in Natural Water with Plasmonic Nanowire Interstice Sensor. Sci. Rep. 6, 19646; doi: 10.1038/srep19646 (2016).