Sequence-Selective Recognition of Nucleic Acids under Extremely

ARTICLE pubs.acs.org/ac

Sequence-Selective Recognition of Nucleic Acids under Extremely Low Salt Conditions Using Nanoparticle Probes Yanbing Zu,*,† Aik Leong Ting,† Guangshun Yi,† and Zhiqiang Gao*,†,‡ † ‡

Institute of Bioengineering and Nanotechnology, 31 Biopolis Way, Singapore 138669 Department of Chemistry, National University of Singapore, 3 Science Drive 3, Singapore 117543

bS Supporting Information ABSTRACT: Extensive secondary structures in nucleic acid targets seriously impede the binding of complementary oligonucleotide probes. We report here a method to conduct the detection under extremely low salt conditions where the secondary structures are less stable and more accessible. A new type of nanoparticle probes prepared by functionalizing gold nanoparticles with nonionic morpholino oligos is employed. Because of the salt-independent hybridization of the probes with nucleic acid targets, nanoparticle assemblies can be formed in 2 mM Tris buffer solutions containing 0
5 mM NaCl, leading to the colorimetric target recognition. The sharp melting transitions of the target
probe hybrids allow discrimination of single-base imperfection, including substitution, deletion, and insertion. The method works effectively in detecting sequences that are likely to form secondary structure. In addition, the study provides direct evidence of the relationship between the aggregate structure and the melting behavior of the DNA-linked nanoparticles.

F

or hybridization-based detection of nucleic acids where oligonucleotide probes are employed, a stringent control over the assay conditions is crucial, especially when the target strand contains extensive secondary structure. On the one hand, although low stringency conditions (i.e., high salt and/or low temperature) favor the formation of target
probe duplexes, secondary structure in the target will also be stabilized, making the sequence which is complementary to the probe largely inaccessible. On the other hand, high stringency conditions (i.e., low salt and/or high temperature) minimize the secondary structure; however, the target
probe duplexes may be rendered less stable, resulting in weak signals. A precise control over the hybridization conditions is not a trivial task, and false negatives and positives are usually unavoidable, which seriously reduces the reliability of the hybridization-based assays. To alleviate this problem, several strategies have been explored.1
5 In a number of studies, nonionic nucleic acid analogs, such as peptide nucleic acid (PNA) and morpholino (MOR), were used as probes in solid-phase assays.3
5 The binding of the nonionic probes and nucleic acid targets is generally insensitive to solution ionic strength; hence, the assay can be carried out in a low-salt solution. Under this condition, the secondary structure is less stable and, therefore, the target sequence is fully accessible. However, a tendency of nonspecific binding has been observed.3 In addition, the surface coverage of the targets hybridized to the substrate could be relatively low due to their repulsive interactions, which decreases the signal intensity. Here, we report a homogeneous method for nucleic acid detection in extremely low salt solutions using MOR/gold nanoparticle (Au NP) conjugates as probes. The nonionic nature of the MOR strand allows the salt-independent binding of the NP probe to the target while the Au r 2011 American Chemical Society

NPs generate optical signals which are sensitive to the hybridization events (Figure 1). In the past decade, a variety of Au NP-based biosensing methods have been proposed.6
10 The analyte-related aggregation of Au NPs shifts the localized surface plasmon resonance (LSPR) absorption peak toward longer wavelength, leading to the colorimetric sensing. Letsinger and Mirkin’s groups pioneered the study of DNA-derivatized Au NPs and developed a homogeneous method for nucleic acid detection based on the formation of target-linked NP networks.6 A striking feature of the approaches is the unusually sharp melting transition of the probe
target hybrid, which enables highly selective sequence recognition. The current study takes advantage of the NP-based detection but functionalizes Au NPs with nonionic MORs instead of negatively charged oligonucleotides. Unlike the assays using DNA/NP probes that require the presence of g50 mM NaCl,11 our method employs the MOR/NP conjugates as probes and is able to recognize nucleic acid targets in the presence of e5 mM NaCl. The extremely low salt conditions greatly simplify the stringency control and allow the detection of the target sequences that are likely to form extensive secondary structure.

’ EXPERIMENTAL SECTION Chemicals. Hydrogen tetrachloroaurate(III) trihydrate (HAuCl4 3 3H2O), trisodium citrate, Zonyl FSN-100 (F(CF2CF2)3
8 Received: January 19, 2011 Accepted: April 25, 2011 Published: April 25, 2011 4090

dx.doi.org/10.1021/ac2001516 | Anal. Chem. 2011, 83, 4090–4094

Analytical Chemistry

ARTICLE

CH2CH2O(CH2CH2O)0
15H), tris(2-carboxyethyl)phosphine (TCEP), and dithiothreitol (DTT) were purchased from SigmaAldrich. All other reagents of certified analytical grade were used as received. Oligonucleotides and MOR oligos were obtained from Sigma (Singapore) and Gene Tools, LLC (Philomath, OR), respectively. Their sequences are listed in Table 1. Au NP Synthesis. Au NPs with an average diameter of 13 nm were synthesized by reduction of HAuCl4 with citrate. All glassware used here was thoroughly washed with freshly prepared aqua regia (HNO3/HCl = 1:3) and then rinsed extensively with deionized water. A solution of 0.01% HAuCl4 (60 mL) was brought to a vigorous boil with stirring, followed by the addition of 1.0 wt % sodium citrate (4.5 mL). The solution was maintained at the boiling point with continuous stirring for about 15 min. After the solution was allowed to cool to room temperature, 0.6 mL of 2 wt % FSN-100 was added. The suspension of FSN-capped Au NPs was stored at 4 °C until further use. Preparation of MOR-Modified Au NPs. The MORs are modified with disulfide amide at the 30 terminals. Each MOR strand contains both a 10-base (T10) spacer and a 15-base specific sequence for hybridization. The MORs were first treated with 0.1 M dithiothreitol (DTT) in 0.2 M phosphate buffer solution (pH 8.0) for 1 h, activating the thiol groups; the MORs were then purified using an NAP-5 column (GE Healthcare). The purified thiolated MORs were stored at 4 °C until use. Before mixing with Au NPs, the purified MORs were incubated with 5 mM tris(2-carboxyethyl)phosphine (pH 7.5) for 10 min to reduce the possible disulfide bond formed between a pair of thiolated MOR strands. The mixture solution containing the thiolated MOR (∼2 μM), the FSN-capped NPs (∼4 nM), sodium dodecyl sulfate (∼0.1 wt %), and phosphate buffer (10 mM, pH 7.5) was allowed to incubate at room temperature for 1
2 h, followed by centrifugation at 10.0 K rpm for 10 min to remove free MORs. The MOR
NP conjugates were resuspended in a 2 mM Tris buffer solution (pH 7.5). The washing step was repeated at least five times. The LSPR absorbance was

Figure 1. Schematic presentation of the colorimetric detection of nucleic acids under extremely low salt conditions.

measured at 523 nm with a standard cuvette, which is related to the concentration of the NPs via Beer’s law (A = εlc). An ε value of 1.5  108 L/(mol 3 cm) at 523 nm was used to determine the concentration of the NP conjugates. Transmission electron microscopy (TEM) images were taken using a Philips microscope (Tecnai 20) operated at an acceleration voltage of 200 kV. Quantification of MOR Loaded on Au NPs. Following the incubation of the NPs with MORs, the mixture solution was centrifuged and the supernatant was collected. The absorbance at 265 nm was measured to determine the concentration of the unreacted MORs in the supernatant. When the concentration difference of the remaining free MORs and the originally added MORs was calculated, the MOR surface density on the NPs was estimated. Assembly of MOR/NP Probes by DNA Targets. Solutions of the specific NP probes and the DNA targets were mixed in 1 mL centrifuge tubes. The tubes were then placed in a dry ice bath, freezing the solutions. After 5 min, the solutions were allowed to thaw at ∼15 °C, followed by melting curve measurements and spot tests. Study of the Melting Behavior. The melting profiles of the DNA-linked aggregates were obtained using a spectrophotometer with a temperature controller (Agilent G1103A). The temperature accuracy is (0.3 °C. The temperature was increased at 1 °C intervals with a holding time of 1 min at each point prior to the measurement. The solution in the cuvette was stirred at 300 rpm during the measurement. Spot Test. To obtain a permanent record of the test and visualize the results conveniently, ∼2 μL of the assay solution was spotted on a C18 reverse-phase thin-layer chromatography (TLC) plate (Merk) and allowed to dry.

’ RESULTS AND DISCUSSION Previous attempts of conjugating PNA oligos with Au NPs were less successful because of the poor stability of the conjugates in aqueous solutions.12 Recently, we reported a method to functionalize Au NPs with another type of nonionic nucleic acid analog, i.e., MOR.13 A MOR strand comprises standard nucleic acid bases but uncharged backbone of morpholine rings connected by phosphorodiamidate groups.14 Compared with PNA, MOR is much more soluble in water. The MOR/NP conjugates were prepared by immobilizing thiolated 25-base MORs on Au NPs (average diameter ∼13 nm). To minimize the nonspecific adsorption of the MOR strands, the Au NPs were capped with nonionic fluorosurfactants (e.g., Zonyl FSN) before mixing with the thiolated MORs. Previous studies indicate that FSN molecules

Table 1. Sequences of the Morpholino Oligos and DNA Strands Used in This Work sequence (50 to 30 )

oligo MOR 1

CGG ACT ATG GAC ACC TTT TTT TTT T-Disulfide amide

MOR 2

AAC CAC ACA ACC TAC TTT TTT TTT T-Disulfide amide

MOR 3

TGG AGG CCC CAG CGA TTT TTT TTT T-Disulfide amide

DNA target 1 (For PM, X = G; for substitutions,

50 /GGT GTC CAT AGT CCG GTA GGT TXT GTG GTT/30

X = T, A, or C; for deletion, X = nil; for insertion, X = GT) DNA target 2 (likely to form stem-loop structure by the pairing of underlined sequences)

(a) 50 /CAC AAC CTA CGG TGT CCA TAG TCC GGT AGG TTG TGT GGT T/30 (b) TCG CTG GGG CCT CCA ATC ATC ATC GTA GGT TGT GTG GTT ATC ATC ATC AAC CAC ACA A (c) GTA GGT TGT GTG GTT ATC ATC ATC TCG CTG GGG CCT CCA ATC ATC ATC TGG AGG CCC C 4091

dx.doi.org/10.1021/ac2001516 |Anal. Chem. 2011, 83, 4090–4094

Analytical Chemistry

Figure 2. Colorimetric detection of nucleic acids in extremely low salt solutions. (a) UV
vis spectra of the mixture of MOR1/NP and MOR2/ NP probes before and after aggregation induced by DNA target 1. (b) The absorbance change as a function of the target concentration.

can be adsorbed at gold surfaces strongly15 and the surfactant layer inhibits the nonspecific adsorption of nucleic acids via the nucleobase
Au interactions while allowing the attachment via the sulfur
Au bonds.16 The resulting MOR/NP conjugates were stably dispersed in a 2 mM Tris buffer solution (Figure 2a). The surface density of MORs on NPs was determined to be ∼180 ( 20 strands/particle. After modification with the nonionic oligos, the zeta potential of the NPs changed from ∼
50 mV to∼
17 mV.17 The conjugate solution was stable for at least 3 months when stored at 5 °C. Two sets of NP conjugates (probes 1 and 2 in Figure 1) were designed so that the DNA targets could act as linkers to align a pair of NP probes through hybridization with the NP-bound MOR sequences. Note that, in our previous report,13 MOR/NP probes were employed to recognize nucleic acid targets based on a noncross-linking mechanism, where the stability change of the probe
target hybrids needed to be measured in the presence of a moderate concentration of salt; while the strategy described herein monitors the cross-linking events under extremely low salt conditions. Solutions of probes 1 and 2 (∼1.5 nM each) containing 2 mM Tris and 0
5 mM NaCl were prepared. After the addition of the perfectly matched DNA target, the solutions were allowed to incubate at room temperature for at least 48 h, but no solution color change was observed, indicating that NP cross-linking did not occur under these experimental conditions. In the low salt solutions, strong electrostatic repulsions between the NPs are expected, which prohibit effective particle collisions; as a result, the inter-NP linkages which depend on hybridization reactions could not be established. However, obvious solution color change was induced by a freezing-thaw process, i.e., freezing the solution in a bath of dry ice (10 min) and then thawing it at room temperature. After this treatment, the solution color changed from red to purple or gray, indicating the formation of target-linked larger particles (Figure 2a). Clearly, the cross-linking of the MOR/NP probes by the DNA targets was remarkably accelerated by the freezing step, similar to that reported for the DNA-functionalized NPs.6 This has been attributed to the high local concentrations of the target strands and the NP probes within pockets in the ice structure.6 Despite the low salt conditions in our study, effective collisions of the NP probes could occur at a certain stage of the freezing process, allowing their cross-linking. In control experiments where no DNA strand was present or the sequence of the added DNA was random, no solution color change was observed after the freezing-thaw process.

ARTICLE

Figure 3. Melting profiles of the DNA(target 1)-linked aggregates of the MOR1/NP and MOR2/NP probes. Concentration of the DNA target, 5 nM. Solution, 2 mM Tris buffer (pH = 7.5) containing different concentrations of NaCl.

Figure 4. TEM images of the DNA-linked aggregates of the MOR1/NP and MOR2/NP obtained in 2 mM Tris buffer (pH = 7.5) containing different concentrations of NaCl. Scale bar, 50 nm.

Figure 2b shows the colorimetric change of the NP solution as a function of target concentration. Even in the absence of NaCl, the 13 nm NP probes responded to the target with concentrations g2 nM. The sensitivity of the NP-based method may be enhanced by employing larger NPs.6c However, our study revealed that, when the NP average diameter is larger than 40 nm, the NP probes become less stable upon the freezingthaw steps. Therefore, larger NPs were not used in the study. The formation of DNA-cross-linked MOR/NP aggregates is reversible. As temperature rises, the dissociation of the DNA
MOR duplexes leads to redispersion of the NP probes, and consequently, the solution color changes back to red. Figure 3 shows the spectral changes as the melting transitions occurred. Interestingly, the full width at half-maximum (fwhm) of the first derivative for the melting profile is salt-concentration dependent. A relatively broad melting curve (fwhm >15 °C) was observed when no NaCl was presented in the 2 mM Tris buffer solution, while the addition of a small amount of salt resulted in sharper spectral changes. In the presence of 5 mM NaCl, the aggregates displayed a melting profile with a fwhm ∼3.3 °C. More concentrated salt (g10 mM NaCl) would make the NP probes irreversibly aggregate upon the freezing-thaw treatment. Sharp melting behavior is highly desirable in the sequencespecific DNA assay, which may greatly improve the detection selectivity. Similar dissociation profiles have also been observed previously for DNA-functionalized Au NPs cross-linked by target DNA strands, where the fwhm value is ∼2.5
4 °C for 13 nm NP 4092

dx.doi.org/10.1021/ac2001516 |Anal. Chem. 2011, 83, 4090–4094

Analytical Chemistry

ARTICLE

Figure 6. Spot test for the highly sequence-selective recognition of DNA target 1 by employing the MOR1/NP and MOR2/NP probes. Figure 5. Melting profiles of DNA(target 1)-linked aggregates of the MOR1/NP and MOR2/NP probes. Concentration of the DNA targets, 5 nM. Solution, 2 mM Tris buffer (pH = 7.5) containing 5 mM NaCl.

Table 2. Values of Melting Temperature and Full Width at Half Maximum of the First Derivative for the Melting Profilesa 1MM

1MM

1MM

1MM

1MM

target

PM

(X = T)

(X = A)

(X = C)

(X = nil)

(X = GT)

Tm (°C)

40.5

28.7

27.3

26.1

26.1

36.5

3.3

4.3

4.3

5.3

3.8

3.2

fwhm (°C) a

Probes, MOR1/NP and MOR2/NP; target, DNA target 1.

probes.6 Two mechanisms, based respectively on the cooperative melting prediction6,11,18 and the phase-transition theory,19 have been proposed to model the melting behavior of the DNA-linked aggregates. Because the DNA-functionalized NPs are negatively charged, the linkages between the NPs could only be established in the presence of a relatively high concentration of salt (>50 mM of NaCl). Our experiments present for the first time examples of the dissociation of the DNA-linked NP aggregates under extremely low salt conditions. The TEM images (see Figure 4) show that the morphologies of the NP aggregates formed in the absence or presence of a small amount of NaCl differ significantly. With no salt added, twodimensional NP assemblies were observed. In the aggregates, each NP interacts with up to 3 neighboring NPs. When 5 mM of NaCl was present in the solution, however, the assemblies are mostly three-dimensional and much more bulky. The different aggregate structures could result from the change of electrostatic interactions between the NPs. The presence of salt greatly reduced the Debye length and screened the electrostatic repulsion, allowing the growth of bulky aggregates. In agreement with the above-mentioned models, the results obtained in this study indicate that the structures of the NP aggregates determine the dissociation behavior. As suggested by the phase-transition mechanism, a sufficiently large size is essential for the cross-linked system to display a sharp transition from a gel-like aggregate to a dilute phase. Similarly, the cooperative melting mechanism requires the establishment of several linkages between a pair of NPs, which would be less possible for the aggregates with two-dimensional structures. Our results provide the direct evidence of the relationship between the aggregate structure and the melting behavior of the DNA-linked NPs. The sharp melting transitions allowed the differentiation of the perfectly matched sequence from the strands with single base imperfections (Figure 5). The Tm and fwhm values for different sequences are summarized in Table 2. The single-base substitutions generally reduce Tm by ∼11.8
14.4 °C. On the basis of the

Figure 7. (a) Assay results for DNA targets likely to form stem-loop structure. The MOR1/NP and MOR2/NP probes were used to recognize target 2a, while the MOR2/NP and MOR3/NP probes were used to recognize targets 2b and 2c. (b) Typical spot test. Assay solution, 2 mM Tris buffer (pH = 7.5) containing 5 mM NaCl.

Tm difference, the order of the base pair stability of the MOR
target duplexes is C:G > C:T > C:A > C:C, which is in good agreement with the destabilizing effects of mismatched base pairs observed for corresponding DNA duplexes.20 One base deletion leads to the lost of one base pair and a decrease of 14.4 °C in Tm. The effect of one base insertion in the target sequence is much less significant (with a decrease of 4 °C in Tm), consistent with the fact that a single base bulge has a smaller influence on the duplex stability.21 The assay could be made simpler by a spot test that also provides a permanent record of the result.6 As shown in Figure 6, a small aliquot (∼2
3 μL) of the NP solution was spotted onto a C18 reverse-phase thin-layer chromatography (TLC) plate. The color of the spot indicates the hybridization state of the NP probes at different temperatures. To examine the ability of the current method in detecting targets with secondary structure, DNA strands likely to form 4093

dx.doi.org/10.1021/ac2001516 |Anal. Chem. 2011, 83, 4090–4094

Analytical Chemistry stem-loop structure (targets 2a
c; in each strand, part of the targeting sequence is designed to locate in the stem moiety) were used to react with specific pairs of the MOR/NP probes. Figure 7 shows the results of the assays. For all the targets, colorimetric responses were observed with a detection limit of ∼2 nM. Clearly, the stability of the secondary structures could be significantly reduced under the low-salt condition (see Supporting Information for the Tm measurement), allowing the successful hybridization of the probes with the DNA targets. In a control study, DNA-functionalized Au NPs were used as the probes. The assay procedure is similar to that described above except that 0.1 M NaCl was added to facilitate the DNA/DNA pairing. It was found that specific pair of the DNA/NP probes worked well in the assay of DNA target 1; however, for the DNA targets 2a
c, no response was observed (see Supporting Information for the details). The study demonstrates the superior performance of MOR/NP probes over their DNA/NP counterparts in recognizing nucleic acid targets bearing secondary structure.

’ CONCLUSION Using a new type of NP probe, MOR/Au NP conjugates, colorimetric detection of nucleic acid targets has been achieved under extremely low salt conditions. The hybridization reaction between the probe and the target is relatively salt-insensitive, which greatly simplifies the stringency control of the assays and enable the analysis of nucleic acid sequences with secondary structures. Like other NP-based nucleic acid detection strategies, sharp melting transitions were observed in this method when a small amount of NaCl was presented. The melting behavior has been found to allow the unambiguous discrimination of the sequences with single-base substitution, deletion, or insertion. The strategy may provide a new tool for the study of nucleic acid variation in a more effective way, especially when the targets are likely to form secondary structure. ’ ASSOCIATED CONTENT

ARTICLE

(6) (a) Elghanian, R.; Storhoff, J. J.; Mucic, R. C.; Letsinger, R. L.; Mirkin, C. A. Science 1997, 277, 1078–1081. (b) Storhoff, J. J.; Elghanian, R.; Mucic, R. C.; Mirkin, C. A.; Letsinger, R. L. J. Am. Chem. Soc. 1998, 120, 1959–1964. (c) Reynolds, R. A.; Mirkin, C. A.; Letsinger, R. L. J. Am. Chem. Soc. 2000, 122, 3795–3796. (7) Sato, K.; Hosokawa, K.; Maeda, M. J. Am. Chem. Soc. 2003, 125, 8102–8103. (8) Li, H. Y.; Rothberg, L. J. J. Am. Chem. Soc. 2004, 126, 10958– 10961. (9) Rosi, N. L.; Mirkin, C. A. Chem. Rev. 2005, 105, 1547–1562. (10) Wilson, R. Chem. Soc. Rev. 2008, 37, 2028–2045. (11) Jin, R. C.; Wu, G. S.; Li, Z.; Mirkin, C. A.; Schatz, G. C. J. Am. Chem. Soc. 2003, 125, 1643–1654. (12) (a) Chakrabarti, R.; Klibanov, A. M. J. Am. Chem. Soc. 2003, 125, 12531–12540. (b) Murphy, D.; Redmond, G.; de la Torre, B. G.; Eritja, R. Helv. Chim. Acta 2004, 87, 2727–2734. (c) Lytton-Jean, A. K. R.; Gibbs-Davis, J. M.; Long, H.; Schatz, G. C.; Mirkin, C. A.; Nguyen, S. T. Adv. Mater. 2009, 21, 706–709. (13) Zu, Y.; Ting, A. L.; Gao, Z. Q. Small 2011, 7, 306–310. (14) Summerton, E. Lett. Pept. Sci. 2003, 10, 215–236. (15) (a) Cha, C.-S.; Zu, Y. Langmuir 1998, 14, 6280–6286. (b) Li, F.; Zu, Y. Anal. Chem. 2004, 76, 1768–1772. (c) Lu, C.; Zu, Y.; Yam, W. W. Anal. Chem. 2007, 79, 666–672. (d) Tang, Y. A.; Yan, J. W.; Zhou, X. S.; Fu, Y. C.; Mao, B. W. Langmuir 2008, 24, 13245–13249. (16) Zu, Y.; Gao, Z. Q. Anal. Chem. 2009, 81, 8523–8528. (17) Although the NPs are densely functionalized with the nonionic oligomers, they are still found to be negatively charged. We believe that the negative zeta potential results from the adsorption of anions, such as Cl
, on the NPs. Note that pH value of the Tris buffer is adjusted by HCl. It has also been found that the MOR/NP conjugates could not be stably dispersed in pure water, which suggests the importance of anion adsorption on the stability of the conjugates. (18) Long, H.; Kudlay, A.; Schatz, G. C. J. Phys. Chem. B 2006, 110, 2918–2926. (19) (a) Lukatsky, D. B.; Frenkel, D. Phys. Rev. Lett. 2004, 92, 068302. (b) Park, Y.; Stroud, D. Phys. Rev. B 2003, 67, 212202. (20) Werntges, H; Steger, G.; Riesner, D.; Fritz, H. J. Nucleic Acids Res. 1986, 14, 3773–3790. (21) (a) Zhu, J.; Wartell, R. M. Biochemistry 1999, 38, 15986–15993. (b) Ke, S. H.; Wartell, R. M. Biochemistry 1995, 34, 4593–4600.

bS

Supporting Information. Additional information as noted in text. This material is available free of charge via the Internet at http://pubs.acs.org.

’ AUTHOR INFORMATION Corresponding Author

*E-mail: [email protected] (Y.Z.); [email protected] (Z.G.).

’ ACKNOWLEDGMENT This work was funded by A*STAR, Republic of Singapore. ’ REFERENCES (1) Nguyen, H. K.; Southern, E. M. Nucleic Acids Res. 2000, 28, 3904–3909. (2) Gamper, H. B.; Gewirtz, A.; Edwards, J.; Hou, Y. M. Biochemistry 2004, 43, 10224–10236. (3) Weiler, J.; Gausepohl, H.; Hauser, N.; Jensen, O. N.; Hoheisel, J. D. Nucleic Acids Res. 1997, 25, 2792–2799. (4) Brandt, O.; Hoheisel, J. D. Trends Biotechnol. 2004, 22, 617–622. (5) Tercero, N.; Wang, K.; Gong, P.; Levicky, R. J. Am. Chem. Soc. 2009, 131, 4953–4961. 4094

dx.doi.org/10.1021/ac2001516 |Anal. Chem. 2011, 83, 4090–4094