Available online at www.sciencedirect.com

www.elsevier.com/locate/yabio

ANALYTICAL

BIOCHEMISTRY

Analytical Biochemistry 375 (2008) 299–305

Detection of single-nucleotide polymorphisms using goldnanoparticles and single-strand-specific nucleases

Yen-Ting Chen a, Chiao-Ling Hsu a, Shao-Yi Hou a,b,*

a Institute of Biotechnology, National Taipei University of Technology, Taipei 106, Taiwanb Department of Chemical Engineering, National Taipei University of Technology, Taipei 106, Taiwan

Received 27 November 2007Available online 8 January 2008

Abstract

The current study reports an assay approach that can detect single-nucleotide polymorphisms (SNPs) and identify the position of thepoint mutation through a single-strand-specific nuclease reaction and a gold nanoparticle assembly. The assay can be implemented viathree steps: a single-strand-specific nuclease reaction that allows the enzyme to truncate the mutant DNA; a purification step that usescapture probe–gold nanoparticles and centrifugation; and a hybridization reaction that induces detector probe–gold nanoparticles, cap-ture probe–gold nanoparticles, and the target DNA to form large DNA-linked three-dimensional aggregates of gold nanoparticles. Athigh temperature (63 �C in the current case), the purple color of the perfect match solution would not change to red, whereas a mis-matched solution becomes red as the assembled gold nanoparticles separate. Using melting analysis, the position of the point mutationcould be identified. This assay provides a convenient colorimetric detection that enables point mutation identification without the needfor expensive mass spectrometry. To our knowledge, this is the first report concerning SNP detection based on a single-strand-specificnuclease reaction and a gold nanoparticle assembly.� 2008 Published by Elsevier Inc.

Keywords: Single-nucleotide polymorphisms; Gold nanoparticles; Single-strand-specific nucleases

Single-nucleotide polymorphisms (SNPs)1 are the mostabundant form of genetic variation [1]. The concept of per-sonalized medicine is rooted in the identification of thesepolymorphisms, providing opportunities in both the diag-nosis and treatment of diseases. Methods for SNP analysisare needed as a technology basis for future personalizeddrug therapies [2]. Up to now, many techniques have beendeveloped for SNP detection. Current SNP assays aredominated by techniques that rely on target hybridizationusing fluorescent or chemiluminescent molecular probesand enzyme activity, for example, oligonucleotide ligationusing DNA ligase [3,4] and enzymatic cleavage using sin-gle-strand-specific nucleases [5–7].

0003-2697/$ - see front matter � 2008 Published by Elsevier Inc.

doi:10.1016/j.ab.2007.12.036

* Corresponding author. Fax: +886 2 2731 7117.E-mail address: f10955@ntut.edu.tw (S.-Y. Hou).

1 Abbreviations used: SNP, single-nucleotide polymorphism; ssDNA,single-stranded DNA; dsDNA, double-stranded DNA; PB, phosphatebuffer; TEM, transmission electron microscopy.

Recently, Mirkin and coworkers invented a DNA anal-ysis method by using two sets of gold nanoparticles withdifferent single-stranded DNA (ssDNA) probes and mixingthem with target DNA [8]. If the target DNA has sequencescomplementary to both of the probes, the nanoparticlescross-link together, and this results in particle aggregationthat changes the color of the solution from red to purple.The color change is due to a red shift in the surface plas-mon resonance of the gold nanoparticles [8]. The goldnanoparticle probes linked with ssDNA have emerged asan attractive alternative to molecular probes for SNPdetection [9–12]. This technique offered several advantagesover conventional fluorophore-based assays with regard toquick and easy readout and no requirement for expensiveinstrumentation [13]. In these methods, ligase was used tocovalently join the two adjacent probes if they matchedthe target DNA at adjacent positions, whereas a mismatchin one of the two adjacent positions failed the ligation.

300 Detection of single-nucleotide polymorphisms / Y.-T. Chen et al. / Anal. Biochem. 375 (2008) 299–305

However, mismatch identification with enzymatic reactionssuch as ligation work well in solution, but the reaction con-ditions typically need to be modified to work well on thegold surfaces [14,15]. Furthermore, these ligation methodscannot detect a single-nucleotide mismatch that is not atadjacent positions of the two ssDNA probes.

The single-strand-specific nucleases catalyze the degra-dation of single-stranded DNA and RNA endonucleolyti-cally to yield 50-phosphoryl-terminated products. Thenucleases prefer ssDNA over double-stranded DNA(dsDNA) by 30,000-fold [16,17] and are used for SNPdetection [5–7]. A single-nucleotide mismatch was detectedwhen the probe and target DNA was not a perfect match.However, these methods used mass spectrometry or gelelectrophoresis for the discrimination of fragments pro-duced by endonuclease cleavage [5–7]. These proceduresare time-consuming and relatively high cost.

In this article, we report a newly developed method todetect SNP based on an ssDNA nuclease reaction and agold nanoparticle assembly. As described below, thismethod has several advantages. Unlike the ligationmethod, which detects a single-nucleotide mismatch onlyat adjacent positions of the two probes, this new methodcan detect a single nucleotide mismatch at different posi-tions of the detector probe. It also identifies the positionof a single-nucleotide mismatch using melting analysis.This method is relatively low cost and fast compared withmass spectrometry and gel electrophoresis methods.

Materials and methods

Reagents and buffers

All oligonucleotides (Table 1) were purchased from Sci-entific Biotech (Taipei, Taiwan). Gold nanoparticles (cat.no. G1652), ethanol (cat. no. 32221), sodium phosphate(cat. no. S9390), and potassium phosphate (cat. no.P5379) all were purchased from Sigma (St. Louis, MO,USA). Mung bean nuclease (cat. no. M4311) and its reac-tion buffer (cat. no. M432A) both were purchased fromPromega (Madison, WI, USA). Deionized water wasobtained through a Direct Q-5 Ultrapure Water System(Millipore, Billerica, MA, USA) and had an electrical resis-tance greater than 18.3 MW.

Table 1Oligonucleotides in this study

Perfect match 5Fluorescein probe 5Capture probe 5Detector probe 5Assay probe 5Anti-assay probe 5M1 5M2 5M3 5M4 5

Note. The bold and underlined bases in M1, M2, M3, and M4 indicate the m

Preparation of oligonucleotide-modified gold nanoparticles

DNA–gold nanoparticle conjugates were prepared usingprocedures described by Thaxton and coworkers [10].Briefly, 20 ll of freshly purified 100 lM thiol-modified oli-gonucleotide was added to 500 ll of gold nanoparticlesolution in an Eppendorf tube. After 16 h, the colloid solu-tion was added to 10 mM phosphate buffer (PB: NaH2PO4/Na2HPO4) solution. Subsequently, the colloids were addedgradually to 0.3 M NaCl by the dropwise addition of 2 MNaCl solution as described previously [10]. The solutionwas centrifuged at 16,100g for 30 min. Following removalof the supernatant, the oily precipitate was washed with0.3 M NaCl (pH 7) and 10 mM PB solution, recentrifuged,and redispersed in 0.3 M NaCl (pH 7) and 10 mM PB solu-tion. After being washed twice, the colloid was resuspendedin 0.01% azide, 0.3 M NaCl (pH 7), and 10 mM PB solu-tion at a final concentration of 10 nM and then stored at4 �C.

Quantitation of alkanethiol–oligonucleotides loaded onnanoparticles

The measurements were described previously [9,18]. Thesurface coverage of probe DNA was estimated using a fluo-rescein probe (Table 1). The fluorescence maximums wereconverted to molar concentrations of the fluorescein–alkanethiol-modified oligonucleotide by interpolation froma standard linear calibration curve. Standard curves wereprepared with known concentrations of fluorophore-labeled oligonucleotides using identical buffer pH and saltconcentrations. The average number of oligonucleotidesper particle was obtained by dividing the measured oligo-nucleotide molar concentration by the original gold nano-particle concentration. The estimated particle surface areawas 1.26 � 10�11 cm2.

Point mutation detection

Fig. 1 is a schematic description of the gold nanoparticleassembly and ssDNA nuclease reaction-based assay in thisstudy. The appropriate oligonucleotide target (40 pmol)was mixed with the assay probe that was complementaryto the target DNA. The conditions of the ssDNA nuclease

0-GCATTATTCGTGTCGCTCAAGGTGGTGTGTCAAGTTCTGG-300-HS-(CH2)12-CCAGAACTTGACACACCACC-FITC-300-HS-(CH2)12-CCAGAACTTGACACACCACC-300-HS-(CH2)12-TTGAGCGACACGAATAATGC-300-CCAGAACTTGACACACCACCTTGAGCGACACGAATAATGC-300-GCATTATTCGTGTCGCTCAAGGTGGTGTGT-300-CCATTATTCGTGTCGCTCAAGGTGGTGTGTCAAGTTCTGG-300-GGATTATTCGTGTCGCTCAAGGTGGTGTGTCAAGTTCTGG-300-GCGTTATTCGTGTCGCTCAAGGTGGTGTGTCAAGTTCTGG-300-GCAATATTCGTGTCGCTCAAGGTGGTGTGTCAAGTTCTGG-30

utant bases.

Fig. 1. Schematic description of the proposed method. The particle aggregation changes the color of the solution from red to purple. The color change isdue to a red shift in the surface plasmon resonance of the gold nanoparticles [8]. (For interpretation of the references to colour in this figure legend, thereader is referred to the web version of this article.)

Detection of single-nucleotide polymorphisms / Y.-T. Chen et al. / Anal. Biochem. 375 (2008) 299–305 301

reaction were described previously [17]. The reaction wasperformed using 0.3 units of mung bean nuclease at 37 �Cfor 30 min in 20 ll of the reaction buffer (0.03 M sodiumacetate, 0.05 M NaCl, 0.001% Triton X-100, and 1 mMZnCl2) from Promega. The solution was heated at 90 �Cfor 10 min to stop the reaction. Subsequently, 100 ll ofcapture probe–gold nanoparticles (10 nM) and 400 pmolof anti-assay probe were added to the solution. The highconcentration of salt (0.3 M NaCl) in the solution alsoinhibited mung bean nuclease activity. After 1 h for hybrid-ization of the nanoparticles with the targets at 40 �C, thesolution was centrifuged at 16,100g for 30 min at 40 �C.Following removal of the supernatant, the oily precipitatewas washed with 0.3 M NaCl (pH 7) and 10 mM PB solu-tion, recentrifuged, and redispersed in 100 ll of 0.3 MNaCl (pH 7) and 10 mM PB solution. The colloid was

mixed with 100 ll of detector probe–gold nanoparticles.After mixing, the solutions were heated to 90 �C for10 min and then were placed at room temperature for 1 hto hybridize the nanoparticles and targets. The solutionswere subjected to colorimetric or UV–Vis analysis(UV-3100, ChromTech, Singapore). The temperature wascontrolled using a water circulating system with a heater.A thermocouple was placed just outside of the curve holderto measure the temperatures of the samples.

Results and discussion

Assay and probe design

A schematic description of this method is shown inFig. 1. The experimental procedures were described in

0.3

0.4

0.5nc

eA

302 Detection of single-nucleotide polymorphisms / Y.-T. Chen et al. / Anal. Biochem. 375 (2008) 299–305

Materials and methods. The target DNA and assay probewere mixed, and the ssDNA nuclease reactions were exe-cuted as described previously [17]. After the reaction, thecapture probe–gold nanoparticle and the anti-assay probewere added to the solution at 40 �C. The capture probe–gold nanoparticle has a 20-base complementarity to thetarget DNA. The anti-assay probe is 10 bases shorter thanthe target. It has a 30-base complementarity to the assayprobe and a 10-base complementarity to the capture probe.The concentration of the anti-assay probe is designed to be10-fold higher than that of the target DNA so as to allowmost of the assay probes to bind to the anti-assay probe,not the target DNA. This also allows most of the targetDNA to bind to the capture probe–gold nanoparticle,not the assay probe. The mixing temperature is designedto be 40 �C to prevent binding between the capture probeand the anti-assay probe that have only a 10-base comple-mentarity to each other. Therefore, the anti-assay probeand the capture probe were designed to have a melting tem-perature lower than 40 �C. The temperature also preventsthe binding between the complex of the assay probe–anti-assay probe and the target DNA that have a 10-base com-plementarity to each other. The target DNA–captureprobe–gold nanoparticle assembly was purified by centrifu-gation at 40 �C. The colloid was mixed with the detector

Fig. 2. TEM image (100,000�) of gold nanoparticles modified withalkanethiol-capped oligonucleotides without (A) or with (B) complemen-tary DNA.

probe–gold nanoparticles, the amount of which was thesame as the capture probe–gold nanoparticles. The solu-tions were subjected to colorimetric or UV–Vis analysisas described previously [9]. The probes were designed toavoid hairpin loops and dimers.

Properties of gold nanoparticles modified with alkanethiol-

capped oligonucleotides

Gold nanoparticles (20 nm) were chemically modifiedwith alkanethiol-capped oligonucleotides. After modifica-tion, a slight shift in the surface plasmon band from 520to 525 nm was observed (data not shown). The opticalproperties were similar to those reported previously [10].The surface coverage of probe DNA was estimated as men-tioned in Materials and methods. The number of oligonu-cleotides attached to an individual nanoparticle was

400 500 600 700 800 9000.0

0.1

0.2Abso

rba

Wavelength (nm)

400 500 600 700 800 9000.0

0.1

0.2

0.3

0.4

0.5

Abso

rban

ce

Wavelength (nm)

B

Fig. 3. UV–Vis spectra of nanoparticle probes without target (solid line),with untreated target (dashed line), and with mung-bean-nuclease-treatedtarget (dotted line) at 60 �C. (A) Perfect match. (B) M4 (point mutation).The sample for the untreated target was taken from the solutioncontaining 5 nM capture probe–gold nanoparticles, 5 nM detectorprobe–gold nanoparticles, and 100 nM perfect match or M4. Theconcentration of perfect match or M4 was 100 nM before the mung beannuclease treatment. The return rate was not calculated.

Detection of single-nucleotide polymorphisms / Y.-T. Chen et al. / Anal. Biochem. 375 (2008) 299–305 303

estimated from the experimental data to be 282. The sur-face coverage of probe DNA was 37 pmol/cm2. This resultis comparable to the literature values of 34 pmol/cm2 [18],50 pmol/cm2 [19], and 21 pmol/cm2 [9].

Hybridization and detection of wild-type or mutant DNA

with gold nanoparticle probes

The gold nanoparticles modified with alkanethiol-capped oligonucleotides were suspended in 0.01% azide,0.3 M NaCl (pH 7), and 10 mM PB solution. After theaddition of perfect match ssDNA in the colloid solution,the formation of large DNA-linked three-dimensionalaggregates of gold nanoparticles was observed by transmis-sion electron microscopy (TEM) (Fig. 2). This process canbe visualized in solution as color changes from red to pur-ple with the naked eye or by UV–Vis spectroscopy(Fig. 3A). The UV–Vis spectra show a red shift in the sur-face plasmon resonance from 525 to 560 nm. The sampleswere treated with mung bean nuclease and processed asdescribed in Fig. 1. It was observed that the kmax remainedat 560 nm for the perfect match sample, whereas the kmax

moved back to 525 nm for the point mutation sample(Fig. 3). The UV–Vis spectra of samples with differentpoint mutation sites are shown in Fig. 4. After the mungbean nuclease cut, the lengths of M2, M3, and M4 were38, 37, and 36 bases, respectively. The spectra of these threesamples are clearly different from the spectrum of the per-fect match sample. This indicates that the point mutationsfor these three samples are detectable. However, the differ-ence between the spectrum of M1 and that of the perfect

Fig. 4. UV–Vis spectra of nanoparticle probes with different mung-bean-nucleprobe–gold nanoparticles, 5 nM detector probe–gold nanoparticles, and targetbean nuclease treatment. The return rate was not calculated.

match is small. This may be due to the fact that there isonly a 1-base difference between M1 and the perfect matchafter the ssDNA nuclease cut.

Melting analysis

To find the point mutation site, melting analysis wasapplied. Fig. 5 shows the UV–Vis spectra of nanoparticleprobes with the mung-bean-nuclease-treated target atdifferent temperatures. At room temperature, the solutionwas purple and the kmax was at 560 nm. The kmax shiftstarted at 64 �C. The detailed melting analysis at the absor-bance ratio of k700/k400 is shown in the Fig. 5. inset. Themelting curve reveals that the melting temperature for theperfect match is 67 �C. The melting temperatures formung-bean-nuclease-treated perfect match, M1, M2, M3,and M4 are 67.0, 65.2, 62.2, 59.4, and 55.6 �C, respectively(Table 2). The 95% confidence interval for mung-bean-nuclease-treated M2’s melting temperature is between61.4 and 63.0 �C. This is not close to the confidence inter-vals for the melting temperatures of M1 and M3, indicatingthat the point mutation of M2 is detectable and that theposition is identified using the proposed method. However,the confidence interval for M1 is very close to that for per-fect match.. The point mutation of M1 might not bedetectable.

Conclusion

It has been demonstrated here, for the first time, thatthe use of a single-strand-specific nuclease and gold

ase-treated DNA samples at 63 �C. The solution contained 5 nM captureDNA. The concentration of the DNA sample was 100 nM before the mung

Fig. 5. UV–Vis spectra of nanoparticle probes with mung-bean-nuclease-treated perfect match at different temperatures. The solution contained 5 nMcapture probe–gold nanoparticles, 5 nM detector probe–gold nanoparticles, and perfect match. The concentration of perfect match was 100 nM before themung bean nuclease treatment. The inset shows the absorbance ratio of k700/k400 at different temperatures.

Table 2Melting temperatures for mung-bean-nuclease-treated target DNA

Target DNA Average meltingtemperaturea (�C)

95% confidenceinterval (�C)

Perfect match 67.0 66.2–67.8M1 65.2 64.3–66.1M2 62.2 61.4–63.0M3 59.4 58.2–60.6M4 55.6 54.8–56.4

a Each value represents the mean of five measurements from fiveseparate experiments.

304 Detection of single-nucleotide polymorphisms / Y.-T. Chen et al. / Anal. Biochem. 375 (2008) 299–305

nanoparticles can detect single nucleotide polymorphismsand identify the position of the point mutation. Mungbean nuclease, a single-strand-specific nuclease, truncatesthe mutant DNA. After purification, detector probe–goldnanoparticles, capture probe–gold nanoparticles, and thetarget DNA form large DNA-linked three-dimensionalaggregates of gold nanoparticles, and the solution ispurple at room temperature. At high temperature, thepurple color of the perfect match solution remains,whereas the mismatched solution changes from purpleto red. Using melting analysis, the position of the pointmutation can be identified based on the different meltingtemperature. The current results show that mutations upto 4 bp from the 50 end of the sequence can be detected.If the mutations are further inside the sequence, the mis-match target will be shorter after mung bean nucleasetreatment. The base pair match between the shorter targetand detector probe will be less. It will result in the melting

temperature being lower. As mentioned in Results anddiscussion, a mismatched solution becomes red as theassembled gold nanoparticles separate at high tempera-ture. The lower melting temperature indicates that themutations further inside the sequence are detectable usingthe proposed method. However, if the melting tempera-ture is too low to be measured, the mutation position willnot be identified. In the current case, the length of theassay probe is 40 bp. It is possible to design a longerprobe. For example, if an assay probe is designed to be80 bp, the detector probe and the capture probe each willbe 40 bp. A target DNA fragment has point mutationthat locates at position 5. After mung bean nuclease treat-ment, the detector probe is 35 and 40 bp match with themismatch target and the perfect match target, respec-tively. The difference of melting temperatures for 35 and40 bp should be detectable. The mutations further insidethe sequence should be detectable. However, the muta-tions for the first several base pairs of the sequence mightnot be detectable. The optimal length of assay probemight be case dependent. This assay provided a conve-nient colorimetric detection that enabled point mutationidentification without the need for expensive mass spec-trometry. In addition, gold nanoparticles possess theadvantage of long-term stability and good biocompatibil-ity with DNA and RNA [20]. This method is a goodscreening test when there are two or more point muta-tions in the target DNA fragment. This assay is generalfor SNP detection and will be useful in clinical diagnosisof gene mutant diseases.

Detection of single-nucleotide polymorphisms / Y.-T. Chen et al. / Anal. Biochem. 375 (2008) 299–305 305

Acknowledgment

This work was supported by the National Science Councilin Taiwan, Republic of China, under grant NSC95-2221-E-027-044.

References

[1] B.W. Kirk, M. Feinsod, R. Favis, R.M. Kliman, F. Barany, Singlenucleotide polymorphism seeking long term association with complexdisease, Nucleic Acids Res. 30 (2002) 3295–3311.

[2] J. Burmeister, V. Bazilyanska, K. Grothe, B. Koehler, I. Dorn,B.D. Warner, E. Diessel, Single nucleotide polymorphism analysis bychip-based hybridization and direct current electrical detection ofgold-labeled DNA, Anal. Bioanal. Chem. 379 (2004) 391–398.

[3] D.A. Nickerson, R. Kaiser, S. Lappin, J. Stewart, L. Hood,U. Landegren, Automated DNA diagnostics using an ELISA-basedoligonucleotide ligation assay, Proc. Natl Acad. Sci. USA 87 (1990)8923–8927.

[4] X. Chen, K.J. Livak, P.Y. Kwok, A homogeneous, ligase-mediatedDNA diagnostic test, Genome Res. 8 (1998) 549–556.

[5] P. Ross, L. Hall, I. Smirnov, L. Haff, High level multiplex genotypingby MALDI–TOF mass spectrometry, Nat. Biotechnol. 16 (1998)1347–1351.

[6] D. Schmalzing, A. Belenky, M.A. Novotny, L. Koutny, O. Salas-Solano, S. El-Difrawy, A. Aram, P. Matsudaira, D. Ehrlich,Microchip electrophoresis: A method for high-speed SNP detection,Nucleic Acids Res. 28 (2000) e43.

[7] B. Ren, J.M. Zhou, M. Komiyama, Straightforward detection ofSNPs in double-stranded DNA by using exonuclease III/nuclease S1/PNA system, Nucleic Acids Res. 32 (2004) e42.

[8] C.A. Mirkin, R.L. Letsinger, R.C. Mucic, J.J. Storhoff, A DNA-based method for rationally assembling nanoparticles into macro-scopic materials, Nature 382 (1996) 607–609.

[9] J. Li, X. Chu, Y. Liu, J.H. Jiang, Z. He, Z. Zhang, G. Shen, R.Q. Yu,A colorimetric method for point mutation detection using high-fidelity DNA ligase, Nucleic Acids Res. 33 (2005) e168.

[10] C.S. Thaxton, D.G. Georganopoulou, C.A. Mirkin, Gold nanopar-ticle probes for the detection of nucleic acid targets, Clin. Chim. Acta363 (2006) 120–126.

[11] W.J. Qin, L.Y.L. Yung, Nanoparticle-based detection and quantifi-cation of DNA with single nucleotide polymorphism (SNP) discrim-ination selectivity, Nucleic Acids Res. 35 (2007) e111.

[12] K. Sato, M. Onoguchi, Y. Sato, K. Hosokawa, M. Maeda, Non-cross-linking gold nanoparticle aggregation for sensitive detection ofsingle-nucleotide polymorphisms: Optimization of the particle diam-eter, Anal. Biochem. 350 (2006) 162–164.

[13] C.H. Liu, Z.P. Li, B.A. Du, X.R. Duan, Y.C. Wang, Silvernanoparticle-based ultrasensitive chemiluminescent detection ofDNA hybridization and single-nucleotide polymorphisms, Anal.Chem. 78 (2006) 3738–3744.

[14] H.J. Lee, A.W. Wark, T.T. Goodrich, S. Fang, R.M. Corn, Surfaceenzyme kinetics for biopolymer microarrays: A combination ofLangmuir and Michaelis–Menten concepts, Langmuir 21 (2005)4050–4057.

[15] S. Fang, H.J. Lee, A.W. Wark, H.M. Kim, R.M. Corn, Determina-tion of ribonuclease H surface enzyme kinetics by surface plasmonresonance imaging and surface plasmon fluorescence spectroscopy,Anal. Chem. 77 (2005) 6528–6534.

[16] W.D. Kroeker, D. Kowalski, M. Laskowski Sr., Mung bean nucleaseI: Physical, chemical, and catalytic properties, Biochemistry 15 (1976)4463–4467.

[17] N. Desai, V. Shankar, Single-strand-specific nucleases, FEMSMicrobiol. Rev. 26 (2003) 457–491.

[18] L.M. Demers, C.A. Mirkin, R.C. Mucic, R.A. Reynolds,R.L. Letsinger, R. Elghanian, G. Viswanadham, A fluorescence-based method for determining the surface coverage and hybridizationefficiency of thiol-capped oligonucleotides bound to gold thin filmsand nanoparticles, Anal. Chem. 72 (2000) 5535–5541.

[19] K. Sato, K. Hosokawa, M. Maeda, Rapid aggregation of goldnanoparticles induced by non-cross-linking DNA hybridization,J. Am. Chem. Soc. 125 (2003) 8102–8103.

[20] S.Y. Hou, H.K. Chen, H.C. Cheng, C.Y. Huang, Development ofzeptomole and attomolar detection sensitivity of biotin–peptide usinga dot–blot gold nanoparticle immunoassay, Anal. Chem. 79 (2007)980–985.